edited by Mapping Ecosystem Services Mapping Ecosystem Services edited by Benjamin Burkhard & Joachim Maes Citation: Burkhard B, Maes J (Eds.) (2017) Mapping Ecosystem Services. Pensoft Publishers, Sofia, 374 pp. First published 2017 ISBN 978-954-642-829-5 (Hardback) ISBN 978-954-642-852-3 (Paperback) ISBN 978-954-642-830-1 (e-book) Pensoft Publishers 12, Prof. Georgi Zlatarski Str. 1111 Sofia, Bulgaria e-mail: info@pensoft.net www.pensoft.net All content is Open Access, distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided that the original author and source are credited. Disclaimer: The opinions and arguments expressed herein belong entirely to the authors. Their views do not necessarily reflect those of the European Commission, Leibniz Universität, the editors or the reviewers. Printed in Bulgaria, March 2017 The content of the book is partially funded by the project ‘En- hancing ecosystem services mapping for policy and decision mak- ing’ (ESMERALDA), which receives funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No. 642007. The publication of the book is funded by the EU and Leibniz Universität Hannover. The editors and several authors of this book are members of the Thematic Working Group on Mapping Ecosystem Services of the Ecosystem Services Partnership (ESP). MAPPING ECOSYSTEM SERVICES Edited by: Benjamin Burkhard, Joachim Maes Ecosystem  Services  Partnership Contents 5 Contents List of contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 Chapter 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 Chapter 2 . Background ecosystem services . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 2 .1 . A short history of the ecosystem services concept . . . . . . . . . . . . . . 29 2 .2 . A natural base for ecosystem services . . . . . . . . . . . . . . . . . . . . . . . . . 33 2 .3 . From nature to society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 2 .4 . Categorisation systems: The classification challenge . . . . . . . . . . . . 42 Chapter 3 . Background mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .47 3 .1 . Basics of cartography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 3 .2 . Mapping techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 3 .3. Map semantics and syntactics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 3 .4 . Tools for mapping ecosystem services . . . . . . . . . . . . . . . . . . . . . . . . . 70 3 .5 . Mapping ecosystem types and conditions . . . . . . . . . . . . . . . . . . . . . 75 3 .6 . Landscape metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 3 .7 . Specific challenges of mapping ecosystem services . . . . . . . . . . . . . 87 Chapter 4 . Ecosystem services quantification . . . . . . . . . . . . . . . . . . . . . . . . . . .91 4 .1 . Biophysical quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 4 .2 . Socio-cultural valuation approaches . . . . . . . . . . . . . . . . . . . . . . . . .102 4 .3 . Economic quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .113 4 .4 . Computer modelling for ecosystem service assessment . . . . . . . .124 4 .5 . Bayesian belief networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136 4 .6 . Applying expert knowledge for ecosystem service quantification . .142 Mapping Ecosystem Services6 Chapter 5 . Ecosystem services mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 5 .1 . What to map? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .149 5 .2 . Where to map? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157 5 .3 . When to map? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .164 5 .4 . Why to map? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .171 5 .5 . Mapping specific ecosystem services . . . . . . . . . . . . . . . . . . . . . . . . . .176 5 .5 .1 . Mapping regulating ecosystem services . . . . . . . . . . . . . . . . . . . . .177 5 .5 .2 . Mapping provisioning ecosystem services . . . . . . . . . . . . . . . . . . .187 5 .5 .3 . Mapping cultural ecosystem services . . . . . . . . . . . . . . . . . . . . . . .197 5 .6 . Integrative approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .210 5 .6 .1 . A tiered approach for ecosystem services mapping . . . . . . . . . . . .211 5 .6 .2 . Participatory GIS approaches for mapping ecosystem services . . . .216 5 .6 .3 . Citizen science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221 5 .6 .4 . Ecosystem services matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .225 5 .7 . Mapping ecosystem services on different scales . . . . . . . . . . . . . . . . . . . .231 5 .7 .1 . Regional ecosystem service mapping approaches. . . . . . . . . . . . . .233 5 .7 .2 . National ecosystem service mapping approaches . . . . . . . . . . . . . .237 5 .7 .3 Global ecosystem services mapping approaches . . . . . . . . . . . . . .244 5 .7 .4 . Mapping marine and coastal ecosystem services . . . . . . . . . . . . . .250 5 .7 .5 . Spatial, temporal and thematic interactions . . . . . . . . . . . . . . . . . .256 Chapter 6 . Uncertainties of ecosystem services mapping . . . . . . . . . . . . . . . 263 6 .1 . Data and quantification issues. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .265 6 .2 . Problematic ecosystem services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .271 Contents 7 6 .3 . Uncertainty measures and maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .281 6 .4 . Map interpretation/end-user issues . . . . . . . . . . . . . . . . . . . . . . . . . . .288 Chapter 7 . Application of ecosystem services maps . . . . . . . . . . . . . . . . . . . . 293 7 .1 . Mapping ecosystem services in national and supra-national policy making . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .295 7 .2 . Application of ecosystem services in spatial planning . . . . . . . . . . .303 7 .3 . Land use sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .308 7 .3 .1 . Mapping urban ecosystem services . . . . . . . . . . . . . . . . . . . . . . . . .310 7 .3 .2 . Ecosystem service maps in agriculture . . . . . . . . . . . . . . . . . . . . . .317 7 .3 .3 . Mapping forest ecosystem services . . . . . . . . . . . . . . . . . . . . . . . .322 7 .3 .4 . Nature protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 7 .4 . Applying ecosystem service mapping in marine areas . . . . . . . . . . .332 7 .5 . Business and industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .339 7 .6 . Mapping health outcomes from ecosystem services . . . . . . . . . . . . . 344 7 .7 . Environmental security: Risk analysis and ecosystem services . . . . .349 7 .8 . Mapping ecosystem services for impact assessment . . . . . . . . . . . . .352 7 .9 . The ecosystem services partnership visualisation tool . . . . . . . . . . .356 Chapte 8 . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365 List of contributors 9 List of Contributors ABDUL MALAK, Dania Arquitecto Francisco Peñalosa Edificio Ada Byron Ampliación Campus de Teatinos 29010 Málaga, Spain Email: daniaabdulmalak@uma.es ADAMESCU, Mihai University of Bucharest Research Center in Systems Ecology and Sustainability Splaiul Independentei, 91-95 050095 Bucharest, Romania Email: adacri@gmail.com AGOSTINI, Vera Science and Consearvation, Caribbean Programme, The Nature Conservancy, 4245 North Fairfax Drive, Suite 100 Arlington, VA 22203-1606, United States Email: vagostini@TNC.ORG ALBERT, Christian Leibniz Universität Hannover Institute of Environmental Planning Herrenhäuser Straße 2 30419 Hannover, Germany Email: albert@umwelt.uni-hannover.de ALKEMADE, Rob PBL Netherlands Environmental Assess- ment Agency PO Box 30314, 2500 GH Den Haag, The Netherlands Email: Rob.Alkemade@pbl.nl ANDERSON, Sharolyn University of South Australia School of Natural and Built Environments GPO Box 2471 Adelaide, South Australia 5001, Australia Email: Sharolyn.Anderson@unisa.edu.au ARKEMA, Katie Stanford Woods Institute 473 Via Ortega, MC: 4205 Stanford, CA 94305, USA Email: karkema@stanford.edu BAGSTAD, Kenneth J . US Geological Survey W 6th Ave Kipling St Lakewood, CO 80225, USA Email: kjbagstad@usgs.gov BANKO, Gebhard Umweltbundesamt Spittelauer Lände 5 1090 Vienna, Austria Email: gebhard.banko@umweltbundesa- mt.at BARTON, David N . Norwegian Institute for Nature Research NINA Gaustadalleen 21 NO-0349 Oslo, Norway Email: David.Barton@nina.no BEAUMONT, Nicola Plymouth Marine Laboratory Prospect Place PL1 3DH The Hoe, Plymouth, United Kingdom Email: nijb@pml.ac.uk BOON, Arjen Department of Ecosystem Dynamics, Marine and Coastal Systems Unit, Del- tares Research Institute Boussinesqweg 1, 2629 HV Delft, Neth- erlands Email: Arjen.Boon@deltares.nl Mapping Ecosystem Services10 BOYANOVA, Kremena Kiel University Institute for Natural Resource Conserva- tion Olshausenstr. 40 24098 Kiel, Germany Email: kbboyanova@gmail.com BRAAT, Leon Alterra Droevendaalsesteeg 6708 PB Wageningen, The Netherlands Email: lcbraat@gmail.com BRANDER, Luke Institute for Environmental Studies VU University Amsterdam De Boelelaan 1087 1081 HV Amsterdam, The Netherlands Email: l.m.brander@vu.nl BROEKX, Steven VITO Boeretang 200 2400 Mol, Belgium Email: steven.broekx@vito.mol BROWN, Claire UNEP World Conservation Monitoring Centre Ecosystem Assessment Programme 219 Huntingdon Road Cambridge, CB3 0DL, United Kingdom Email: Claire.Brown@unep-wcmc.org BURKHARD, Benjamin Leibniz Universität Hannover Institute of Physical Geography and Landscape Ecology Schneiderberg 50 30167 Hannover, Germany Email: burkhard@phygeo.uni-hannover.de CAZACU, Constantin University of Bucharest Research Center in Systems Ecology and Sustainability Splaiul Independentei 91-95 050095 Bucharest, Romania Email: constantin.cazacu@gmail.com CONTI, Michele European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: michele.conti@ec.europa.eu COSTANZA, Robert Crawford School of Public Policy The Australian National University Canberra ACT 2601, Australia Email: rcostanz@gmail.com CROSSMAN, Neville D . University of Adelaide School of Biological Sciences Glen Osmond, SA, 5064, Australia Email: neville.crossman@gmail.com DE GROOT, Rudolf Wageningen University & Research (WUR) Environmental Systems Analysis Group Droevendaalsesteeg 3 6708PB Wageningen, The Netherlands Email: Dolf.deGroot@wur.nl DECLERCK, Fabrice Bioversity International Parc Scientifique Agropolis II 34397 Montpellier Cedex 5, France Email: f.declerck@cgiar.org DENDONCKER, Nicolas University of Namur Department of Geography Rue de Bruxelles 61 5000 Namur, Belgium Email : nicolas.dendoncker@unamur.be DERKZEN, Marthe Vrije Universiteit Amsterdam Environmental Geography Group, De- partment of Earth Sciences List of contributors 11 De Boelelaan 1087 1081 HV Amsterdam, the Netherlands Email: marthe.derkzen@vu.nl DRAKOU, Evangelia G . Institut Universitaire Européen de la Mer Rue Dumont d’Urville, 29280 Plouzané, France Email: Evangelia.Drakou@univ-brest.fr DUNFORD, Robert W . Oxford University Centre for the Environ- ment Environmental Change Institute OX1 3QY South Parks Road, Oxford, UK Email: robert.dunford@ouce.ox.ac.uk EGOH, Benis Council for Scientific and Industrial Research, Natural Resources and The Environment, PO Box 320, Stellenbosch 7599, South Africa School of Agricultural, Earth and Envi- ronmental Sciences, University of Kwa- Zulu-Natal, 27 Private Bag X01, Scotts- ville 3209, South Africa Email: BEgoh@csir.co.za ERHARD, Markus European Environment Agency Kongens Nytorv 6 1050 Copenhagen K, Denmark Email: markus.erhard@eea.europa.eu ESTRADA CARMONA, Natalia Bioversity International Parc Scientifique Agropolis II 34397 Montpellier Cedex 5, France Email: n.e.carmona@cgiar.org FAGERHOLM, Nora University of Copenhagen Department of Geosciences and Natural Resource Management Rolighedsvej 23 1958 Fredriksberg C, Denmark Email: ncf@ign.ku.dk FRANK, Susanne GICON - Großmann Ingenieur Consult GmbH Tiergartenstr. 48 01219 Dresden, Germany Email: S.Frank@gicon.de FÜRST, Christine Martin-Luther University Halle-Witten- berg Institute for Geosciences and Geography Sustainable Landscape Development Von Seckendorff-Platz 4 06120 Halle, Saale, Germany Email: christine.fuerst@geo.uni-halle.de GARCÍA-LLORENTE, Marina Madrid Institute for Rural, Agricultural and Food Research and Development (IMIDRA), Ctra. Madrid-Barcelona (N-II), KM. 38.200, 28802 Alcalá de Henares, Spain Email: marina.garcia.llorente@madrid.org GENELETTI, Davide Department of Civil, Environmental and Mechanical Engineering University of Trento via Mesiano, 77 38123 Trento, Italy Email: davide.geneletti@unitn.it GIORDANO, Alberto Texas State University Department of Geography Evans Liberal Arts, Room 139 San Marcos, TX 78666 Email. ag22@txstate.edu GRÊT-REGAMEY, Adrienne ETH Zurich Swiss Federal Institute of Technology Stefano-Franscini-Platz 5 8093 Zurich, Switzerland Email: gret@nsl.ethz.ch Mapping Ecosystem Services12 GRIFFITHS, Charly The Marine Biological Association of the United Kingdom The Laboratory, Citadel Hill Plymouth, Devon, PL1 2PB, UK GONZALEZ-REDIN, Julen The James Hutton Institute Craigiebuckler Aberdeen AB15 8QH, UK Email: Julen.Gonzalez@hutton.ac.uk GRUNEWALD, Karsten Leibniz Institute of Ecological and Re- gional Development IOER Weberplatz 1 01217 Dresden, Germany Email: k.grunewald@ioer.de GUERRA, Carlos German Centre for Integrative Biodiversity Research iDiv Halle-Jena-Leipzig Martin Luther University Halle-Witten- berg Deutscher Platz 5e 04103 Leipzig, Germany Email: carlos.guerra@idiv.de HAINES-YOUNG, Roy Fabis Consulting Ltd The Paddocks, Chestnut Lane, Barton in Fabis, Nottingham, NG11 0AE, UK Email: Roy.Haines-Young@fabisconsult- ing.com HARRISON, Paula A . Oxford University Centre for the Environ- ment Environmental Change Institute OX1 3QY South Parks Road, Oxford, UK Email: paharriso@aol.com HEISKANEN, Anna-Stiina Finnish Environment Institute (SYKE) Marine Research Centre Mechelininkatu 34a 00260 Helsinki, Finland Email: Anna-Stiina.Heiskanen@ympar- isto.fi HEVIA, Violeta Autonomous University of Madrid Socio-Ecological Laboratory Department of Ecology Office C-201 28049 Madrid, Spain Email: violeta.hevia@uam.es HIEDERER, Roland European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: roland.hiederer@ec.europa.eu HOOPER, Tara Plymouth Marine Laboratory Prospect Place The Hoe, PL1 3DH, Plymouth, United Kingdom Email: tarh@pml.ac.uk HUYNEN, Maud University of Maastricht P.O. Box 616 6200 MD Maastricht, The Netherlands Email: m.huynen@maastrichtuniversity.nl JACOBS, Sander Research Institute for Nature and Forest (INBO) Kliniekstraat 25 1070 Brussels, Belgium Email: sander.jacobs@inbo.be JONES, Sarah Bioversity International Parc Scientifique Agropolis II 34397 Montpellier Cedex 5, France Email: s.jones@cgiar.org KABISCH, Nadja Humboldt University Berlin Geography Department List of contributors 13 Unter den Linden 6 10099 Berlin, Germany Email: nadja.kabisch@geo.hu-berlin.de KELEMEN, Eszter Corvinus University of Budapest Fovam ter 81093 Budapest, Hungary and ESSRG Ltd. Romer Floris u. 38 1024 Budapest, Hungary Email: kelemen.eszter@essrg.hu KEUNE, Hans Research Institute for Nature and Forest (INBO) Kliniekstraat 25 1070 Brussels, Belgium Email: hans.keune@inbo.be KLUG, Hermann Paris-Lodron University Salzburg Schillerstr. 30, Building 15, 3rd Floor 5020 Salzburg, Austria Email: hermann.klug@sbg.ac.at KOPPEROINEN, Leena Finnish Environment Institute (SYKE) Environmental Policy Centre Mechelininkatu 34a 00260 Helsinki, Finland Email: leena.kopperoinen@ymparisto.fi KRUSE, MARION Kiel University Institute for Natural Resource Conserva- tion Olshausenstr. 75 24118 Kiel, Germany Email: mkruse@ecology.uni-kiel.de KUBISZEWSKI, Ida Crawford School of Public Policy The Australian National University Canberra ACT 2601, Australia Email: ida.kubiszewski@anu.edu.au LANDUYT, Dries Ghent University Forest & Nature Lab Geraardsbergsesteenweg 267 9090 Melle-Gontrode, Belgium Email: dries.landuyt@ugent.be LIEKENS, Inge VITO Boeretang 200 2400 Mol, Belgium Email: inge.liekens@vito.be LIQUETE, Camino European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: camino.liquete@ec.europa.eu LUQUE, Sandra National Research Institute of Science and Technology for Environment and Agricul- ture, (IRSTEA), UMR TETIS 500 rue JF Breton, Montpellier 34000 France Email: sandra.luque@irstea.fr MAES, Joachim European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: joachim.maes@ec.europa.eu MANDLE, Lisa Natural Capital Project, Stanford University 371 Serra Mall Stanford, CA 94305 USA Email: lmandle@stanford.edu MARTENS, Pim University of Maastricht P.O. Box 616 6200 MD Maastricht, The Netherlands Email: p.martens@maastrichtuniversity.nl MARTÍN-LÓPEZ, Berta Leuphana University Lüneburg Scharnhorststraße 1 21335 Lüneburg, Germany Email: martinlo@leuphana.de Mapping Ecosystem Services14 MONONEN, Laura Finnish Environment Institute (SYKE) Natural Environment Centre Yliopistokatu 7, 80100 Joensuu Email: laura.mononen@ymparisto.fi MÜLLER, Felix Kiel University Institute for Natural Resource Conserva- tion Olshausenstr. 40 24098 Kiel, Germany Email: fmueller@ecology.uni-kiel.de NEDKOV, Stoyan Bulgarian Academy of Sciences National Institute of Geophysics, Geodesy and Geography Acad. G. Bonchev str., bl. 3 1113 Sofia, Bulgaria Email: snedkov@abv.bg NEALE, Anne US EPA Research Mail Drop: D343-04 109 Alexander Drive Durham, NC 27711, USA Email: neale.anne@epa.gov OOSTERBROEK, Bram University of Maastricht P.O. Box 616 6200 MD Maastricht, The Netherlands Email: bram.oosterbroek@maastrichtuni- versity.nl OTEROS-ROZAS, Elisa Universidad Pablo de Olvide Department of Social Anthropology, Basic Psychology, and Public Health 41013 Seville, Spain Email: elisa.oterosrozas@gmail.com PALOMO, Ignacio Sede Building 1, 1st floor Scientific Campus of the University of the Basque Country 48940 Leioa, Spain Email: ignacio.palomo@bc3research.org PÁRTL, Adam Global Change Research Institute, Czech Academy of Sciences Jirchářích 149/6 Prague 11000, Czechia Email: adam.partl@aonbenfield.com PAYYAPPALIMANA, Unnikrishnan UNU-International Institute for Global Health, Universiti Kebangsaan Malaysia Jalan Yaacob Latiff 56000 Cheras Kuala Lumpur, Malaysia Email: payyappalli@unu.edu PERPIÑA, Carolina European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: carolina.perpina@ec.europa.eu PETZ, Katalin PBL Netherlands Environmental Assess- ment Agency PO Box 30314, 2500 GH Den Haag, The Netherlands Email: katalin.petz@gmail.com POLCE, Chiara European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: chiara.polce@ec.europa.eu POTSCHIN, Marion Fabis Consulting Ltd., The Paddocks, Chestnut Lane, Barton In Fabis, Notting- ham, NG11 0AE, United Kingdom Email: Marion.Potschin@fabisconsulting. com List of contributors 15 PRIESS, Joerg A . Helmholtz Centre for Environmental Research UFZ Computational Landscape Ecology Permoserstraße 15 04318 Leipzig, Germany Email: joerg.priess@ufz.de QIU, Jianxiao University of Florida School of Forest Resources & Conserva- tion Fort Lauderdale Research and Education Center 3205 College Ave, Davie, FL 33314 Email: qiujiangxiao@gmail.com RABE, Sven-Erik ETH Zurich Swiss Federal Institute of Technology Wolfgang-Pauli-Str. 15 8093 Zurich, Switzerland Email: rabes@ethz.ch RIVERO, Inés Marí European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: ines.mari-rivero@ec.europa.eu RUSKULE, Anda Baltic Environmental Forum Antonijas 3-8 Rīga,1010, Latvia Email. Anda.Ruskule@bef.lv SANTOS-MARTÍN, Fernando Autonomous University of Madrid Department of Ecology Office C-201 28049 Madrid, Spain Email: fernando.santos.martin@uam.es SCHNEIDERS, Anik Research Institute for Nature and Forest (INBO) Kliniekstraat 25 1070 Brussels, Belgium Email: anik.schneiders@inbo.be SCHRÖTER, Matthias UFZ - Helmholtz Centre for Environ- mental Research, Department of Ecosys- tem Services Permoserstr. 15, 04318 Leipzig, Germany German Centre for Integrative Biodiversi- ty Research iDiv Halle-Jena-Leipzig Deutscher Platz 5e 04103 Leipzig, Germany Email: matthias.schroeter@idiv.de s SCHULP, Nynke Vrije Universiteit Amsterdam Environmental Geography group Earth Sciences Department De Boelelaan 1087 1081 HV Amsterdam, The Netherlands Email: nynke.schulp@vu.nl SUBRAMANIAN, Suneetha M . UNU-International Institute for Global Health, Universiti Kebangsaan Malaysia Jalan Yaacob Latiff 56000 Cheras Kuala Lumpur, Malaysia Email: subramanian@unu.edu SUTTON, Paul University of Denver Department of Geography 2050 E Iliff Ave Denver, CO 80208 Email. Paul.Sutton@du.edu SYRBE, Ralf-Uwe Leibniz Institute of Ecological and Re- gional Development IOER Weberplatz 1 01217 Dresden, Germany Email: r.syrbe@ioer.de Mapping Ecosystem Services16 TARDIEU, Léa French National Institute of Agricultural Research INRA, AgroParisTech, UMR356 Labora- toire d’Économie Forestière, 54042 Nancy, France Email: lea.tardieu@inra.fr TELLER, Anne European Commission Directorate-General for Environment 1049 Brussels, Belgium Email: anne.teller@ec.europa.eu TENERELLI, Patrizia IRSTEA - National Research Institute of Science and Technology for Environment and Agriculture 500 rue JF BRETON Montpellier 34000 France TENEVA, Lida Stanford University Department of Environmental Earth Systems Science Braun Hall Stanford, CA 94305-2115, USA Email: lteneva@stanford.edu TRAUN, Christoph Paris-Lodron University Salzburg Hellbrunnerstr. 34 5020 Salzburg, Austria Email: christoph.traun@sbg.ac.a VAČKÁŘ, David Charles University Environment Centre José Martího 407/2 162 00 Praha, Czechia Email: david.vackar@czp.cuni.cz VALLECILLO, Sara European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: sara.vallecillo-rodriguez@ec.euro- pa.eu VANDECASTEELE, Ine European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: ine.vandecasteele@ec.europa.eu VEERKAMP, Clara J . PBL Netherlands Environmental Assess- ment Agency PO Box 30314, 2500 GH Den Haag, The Netherlands Email: clara.veerkamp@pbl.nl VERHEYDEN, Wim Research Institute for Nature and Forest (INBO) Kliniekstraat 25 1070 Brussels, Belgium wim.verheyden@inbo.be VIHERVAARA, Petteri Finnish Environment Institute Mechelininkatu 34a, P.O.Box 140, FI- 00251 Helsinki, Finland VIINIKKA, Arto Finnish Environment Institute (SYKE) Environmental Policy Centre Mechelininkatu 34a 00260 Helsinki, Finland Email: arto.viinikka@ymparisto.fi VIITASALO, Markku Finnish Environment Institute SYKE Marine Research Centre/Marine Spatial Planning Mechelininkatu 34a 00260 Helsinki, Finland Email: Markku.Viitasalo@ymparisto.fi List of contributors 17 VIZCAINO, Pilar European Commission – Joint Research Centre Via E. Fermi 2749 21027 Ispra (VA), Italy Email: pilar.vizcaino-martinez@ec.euro- pea.eu WALZ, Ulrich University of Applied Sciences Dresden Chair of Landscape Ecology Friedrich-List-Platz 1 01069 Dresden, Germany Email: ulrich.walz@htw-dresden.de WEIBEL, Bettina ETH Zurich Swiss Federal Institute of Technology Stefano-Franscini-Platz 5 8093 Zurich, Switzerland Email: weibel@nsl.ethz.ch WILLEMEN, Louise University of Twente Department of Natural Resources PO Box 217 7500 AE Enschede, The Netherlands Email. l.l.willemen@utwente.nl ZULIAN, Grazia European Commission - Joint Research Centre Directorate D - Sustainable Resources Unit D3 - Land Resources Via E. Fermi, 2749 - TP 290 21027 Ispra (VA), Italy Email: grazia.zulian@ec.europa.eu Foreword 19 Foreword Mapping Ecosystem Services The world’s economic prosperity and well-being are underpinned by its natural capital, i.e. its biodiversity, including eco- systems that provide essential goods and services for mankind, from fertile soils and multi-functional forests to productive land and seas, from good quality fresh water and clean air to pollination and climate regula- tion and protection against natural disasters. This is the reason why, for example, the first priority objective of the 7th Environment Action Programme (7th EAP) of the Euro- pean Union (EU) is to protect, conserve and enhance the EU natural capital. In order to mainstream biodiversity in our socio-eco- nomic system, the 7th EAP highlights the need to integrate economic indicators with environmental and social indicators, includ- ing by means of natural capital accounting, to measure the changes in the stock of nat- ural capital at a variety of levels, including both continental and national levels. The EU Biodiversity Strategy to 2020 called on Member States to map and assess the state of ecosystems and their services in their national territory by 2014, with the as- sistance of the European Commission. The economic value of such services should also be assessed, and the integration of these val- ues into accounting and reporting systems at EU and national level should be promot- ed by 2020 (see Target 21, Action 5). This specific action aims to provide a knowl- edge base on ecosystems and their services in Europe to underpin the achievement of the six specific biodiversity targets of the strat- egy as well as including a number of other 1 http://ec.europa.eu/environment/nature/biodi versity/strategy/target2/index_en.htm sectoral policies such as agriculture, mari- time affairs and fisheries and cohesion. Mapping ecosystem services is essential to understand how ecosystems contribute to human wellbeing and to support policies which have an impact on natural resourc- es. In 2013, an EU initiative on Mapping and Assessment of Ecosystems and their Services (MAES) was launched and a ded- icated working group was established with Member States, scientific experts and rel- evant stakeholders. The first delivery was the development of a coherent analytical framework2 to be applied by the EU and its Member States in order to ensure consistent approaches. In 2014, a second technical re- port3 was issued which proposes indicators that can be used at European and Member State’s level to map and assess ecosystem ser- vices. The indicators are proposed for the main ecosystems (agro-, forest, freshwater and marine) and the important issue of how the overarching data flow from the reporting of nature directives can be used to assess the condition of ecosystems is also addressed. From the start of MAES, some exploratory work was undertaken in parallel to assess how some of the biophysical indicators could be used for natural capital account- ing. It was also important to ensure that the data flows available at European level and, in particular, those from reporting obligations from Member States would 2 http://ec.europa.eu/environment/nature/knowl edge/ecosystem_assessment/pdf/MAESWork ingPaper2013.pdf 3 http://ec.europa.eu/environment/nature/knowl edge/ecosystem_assessment/pdf/2ndMAESWork ingPaper.pdf Mapping Ecosystem Services20 be used for the mapping and assessment of ecosystems and their condition4. More recently, dedicated work on urban ecosys- tems was initiated with the active contri- bution of many cities and a fourth tech- nical report5 on mapping and assessment of urban ecosystems and their services was published. An overlapping activity on the strengthening of the mapping and assess- ment of soil condition and function in the long-term delivery of ecosystem services is also being developed. In the context of The Economics of Ecosys- tems and Biodiversity (TEEB6), a study of available approaches to assess and value eco- system services in the EU7 was supported by the European Commission to support EU countries in taking forward Action 5 of the EU Biodiversity Strategy. In 2015, a Knowledge Innovation Project on an Integrated System for Natural Capital and Ecosystem Services Accounting (KIP INCA)8 was launched jointly by four Com- mission services (Eurostat, Environment, the Joint Research Centre and Research and Innovation) and the European Environment Agency. This project aims to design and im- plement an integrated accounting system for ecosystems and their services in the EU, to serve a range of information needs and inform decision making of different policy sectors, building on existing work in EU countries. Important ecosystems services provided by nature will therefore be explic- 4 http://ec.europa.eu/environment/nature/knowl edge/ecosystem_assessment/pdf/3rdMAESRe port_Condition.pdf 5 http://ec.europa.eu/environment/nature/knowl edge/ecosystem_assessment/pdf/102.pdf 6 http://teebweb.org/ 7 http://ec.europa.eu/environment/nature/biodi versity/economics/index_en.htm 8 http://ec.europa.eu/environment/nature/capi tal_accounting/index_en.htm itly taken into account and demonstrate, in physical and to the greatest extent possible in monetary terms, the benefits of investing in the sustainable management of ecosys- tems and natural resources. Finally, the European work undertaken un- der Target 2, Action 5, is actively contribut- ing to major ongoing initiatives, such as the global, regional and thematic assessments under the Intergovernmental Platform on Biodiversity and Ecosystem Services (IP- BES9) and the UN guidelines on experi- mental ecosystem accounting from the Sys- tem of Environmental-Economic Accounts (UN SEEA EEA10). At present, with the constructive support of research and innovation projects and ac- tions, such as ESMERALDA11 and with the amount of work already accomplished in the Member States and at EU level, the momen- tum for the next steps is impressive12. The policy developments in Europe, but also in many other countries and at global scale, have spurred the scientific commu- nity to map ecosystem services, to devel- op new methods, to assess uncertainty of maps and to provide practical applications of using maps in various decision-making processes. This book is an excellent sum- mary of the achievements of ecosystem service mapping and provides guidance for scientists, students, practitioners and deci- sion makers who need to map ecosystem services. There are still big challenges ahead of us such as the improvement of the mapping and assessment of the ecosystem condition 9 http://www.ipbes.net/ 10 http://unstats.un.org/unsd/envaccounting/eea_ project/default.asp 11 http://esmeralda-project.eu/ 12 http://biodiversity.europa.eu/maes/maes_countries Foreword 21 and the integration of the assessment of the ecosystem condition with ecosystem services and the construction of the first ecosystem accounts. As highlighted in this book, we are however on a very positive track! Anne Teller European Commission, Directorate-General Environment Introduction 23 Ecosystem services (ES) are the contribu- tions of ecosystem structure and function (in combination with other inputs) to hu- man well-being. This implies that mankind is strongly dependent on well-functioning ecosystems and natural capital that are the basis for a constant flow of ES from nature to society. Therefore, ES have the potential to become a major tool for policy and de- cision making on global, national, regional and local scales. Possible applications are numerous: from sustainable management of natural resources, land use optimisation, environmental protection, nature conserva- tion and restoration, landscape planning, nature-based solutions, climate protection, disaster risk reduction to environmental ed- ucation and research. ES maps constitute a very important tool to bring ES into practical application. Maps can efficiently communicate complex spa- tial information and people generally prefer to look at maps and to explore their content and practical applicability. Thus, ES maps are very useful for raising awareness about areas of ecosystem goods and services supply and demand, environmental education about hu- man dependence on functioning nature and to provide information about interregional ecosystem goods and services flows. Further- more, maps are mandatory instruments for landscape planning, environmental resource management and (spatial) land use opti- misation. To fulfil the requirements of the above-mentioned applications, high quality, robust and consistent data and information on ES supply, flow and demand are needed at different spatial and temporal levels. The interest of policy and decision makers, the business sector and civil society in ES- maps has been steadily increasing in the last years. To bring ES maps into practical ap- plication and to make them useful tools for sustainable decision making is an import- ant step and a responsibility of all parties involved. Maps can be applied to portray trade-offs and synergies for ES as well as spatial congruence or mismatches between supply, flow and demand of different ES. Additionally, flows of services from one eco- system to another and source-sink dynamics can be illustrated. Based on such informa- tion, budgets for ES supply and demand can be calculated on different spatio-tempo- ral scales. Such budgets can help to assess the dependence of a region (or even a whole country) on ES imports or its potential to export certain goods and services. However, in addition to the high application potential of ES maps in sustainable decision-making that would benefit human society, there is also a risk of abusing the maps for further exploitation of natural resources, fostering land conversions or supporting land-grab- bing activities. That is the reason why it is so important to communicate the ES concept properly and to prepare and document all related information carefully and with the best knowledge available. Well-documented maps of ES which are de- veloped following rigorous guidelines and definitions will be of crucial importance for natural capital accounting. Across Europe, as well as elsewhere and at local to global scales, natural capital accounts are being developed with the aim of supporting policies on ag- Chapter 1. Introduction Benjamin Burkhard & Joachim Maes Mapping Ecosystem Services24 riculture, natural resources use or regional development programmes or to support de- cision-making. These accounts are intended to measure and monitor the extent, the con- dition, the services and the benefits of ecosys- tems to support different policies. Regularly updated and high quality geo-referenced data on capacity, use and demand of ES are essen- tial inputs for natural capital accounts. The development of respective ES mapping approaches, models and tools has profited from the increasing popularity of the ES concept in science, especially within the last decade. However, this popularity of ES map- ping studies has, together with the rapid de- velopment of computer-based mapping pro- grammes, also led to an almost inflationary generation of various ES maps. Besides the many very promising and well-derived map- ping products, maps of inferior quality have also, unfortunately, been published. It takes more than just some data and a software package to make a good map that fulfils the criteria of being a geometrically accurate, correctly-scaled and appropriately-explained graphic representation of three-dimensional real space. Cartography, the art and science of graphically representing a geographical area usually on a map, has served humanity since its emergence by providing informa- tion on the environment, resources, risks, paths, connections and barriers. The theory, methods and practical appli- cations of ES mapping are presented in this book, thus bringing together valuable knowledge and techniques from leading experts in the field. The different chapters can be explored to learn what is necessary to make proper and applicable ES maps. This book addresses an audience which is broader than the research community alone. ES are becoming mainstream outside the academic world: national and regional au- thorities are calling for or are involved in large-scale studies to map ES for mapping their natural capital. Cities need ES maps to design, implement or maintain urban green infrastructure. Large businesses start assessing ecosystems and their services on their sites so that they can better understand possible im- pacts of their operations on the environment. Nature managers need to know how parks and reserves contribute to human wellbeing. Whereas, although not all of these stakehold- ers will suddenly start mapping ES, they may rely on consultants, students, ecologists and other researchers to help them with spatial data analysis, to understand problems relat- ed to mapping or to give practical guidance. Full Open Access to this book is provided to better reach this audience. After this introductory chapter, Chapter 2 provides the conceptual ES background, including a short history of the concept, introduces the nature-ecosystem service-hu- man society connections and explains ES categorisation systems. The necessary back- ground of mapping is given in Chapter 3, starting from basic cartography knowledge, methods and tools and ending with the specific challenges of mapping ES. There is no mapping without adequate information or data behind it. Therefore, Chapter 4 is solely dedicated to various ES quantifica- tion approaches. These approaches include biophysical, socio-economic, model-, ex- pert- and citizen-science-based quantifica- tion methods. Chapter 5 on ES mapping is the most extensive of this book. After elab- orating what, where, when and why to map ES, the individual subchapters explain what has to be taken into account when mapping specific or bundles of ES using various (in- cluding integrative) approaches. The chap- ter ends by presenting mapping approach- es on different and interacting scales. Each map represents a more or less complex but generalised model of reality and each model comes with specific uncertainties. Uncer- tainties can be related to data, specific ES Introduction 25 properties or concerning the eventual map interpretation and use. Thus, uncertainties are a highly relevant topic in ES mapping that need to be dealt with properly. The whole of Chapter 6 is therefore solely ded- icated to uncertainties of ES mapping. As mentioned above, there is a broad range of applications for ES maps, which are ex- plained in Chapter 7. Applications include policy making and planning, different land use sectors, human health, risk and impact assessments as well as visualisation. The final Chapter 8 provides some conclusions and synthesises the contents presented in the pre- ceding chapters. Several chapters include practical examples which are meant to facilitate the understand- ing of the sometimes complex and often technical topics. The editors’ and authors’ aim was to present chapters in a profession- al but understandable language in order to facilitate their readability and comprehen- sion. Therefore citations and references were avoided in the text. Instead, footnotes with direct links and suggestions for further read- ing are provided at the end of each chapter. We hope this book is helpful and supports the appropriate mapping of ES! Further reading Burkhard B, Kroll F, Nedkov S, Müller F (2012) Mapping supply, demand and bud- gets of Ecosystem Services. Ecological Indi- cators 21: 17-29. Crossman ND, Burkhard B, Nedkov S, Wil- lemen L, Petz K, Palomo I, Drakou EG, Martín-Lopez B, McPhearson T, Boyano- va K, Alkemade R, Egoh B, Dunbar M, Maes J (2013) A blueprint for mapping and modelling Ecosystem Services, Ecosystem Services 4: 4-14. Egoh B, Drakou EG, Dunbar MB, Maes J, Willemen L (2012) Indicators for map- ping ecosystem services: a review. Report EUR25456EN. Publications Office of the European Union, Luxembourg. Maes J, Crossman ND, Burkhard B (2016) Mapping eocsystem services. In: Potschin M, Haines-Young R, Fish R, Turner RK (Eds) Routledge Handbook of Ecosystem Services. Routledge, London, 188-204. Maes J, Egoh B, Willemen L, Liquete C, Vi- hervaara P, Schägner JP, Grizzetti B, Drakou EG, Notte AL, Zulian G, Bouraoui F, Luisa Paracchini M, Braat L, Bidoglio G (2012) Mapping Ecosystem Services for policy support and decision making in the Euro- pean Union. Ecosystem Services 1: 31-39. Martínez-Harms MJ, Balvanera P (2012) Meth- ods for mapping ecosystem service supply: a review. International Journal of Biodiversity Science, Ecosystem Services & Management 8: 17-25. Pagella TF, Sinclair FL (2014) Development and use of a typology of mapping tools to assess their fitness for supporting man- agement of ecosystem service provision. Landscape Ecology 29: 383-399. Troy A, Wilson MA (2006) Mapping Ecosys- tem Services: Practical challenges and op- portunities in linking GIS and value trans- fer. Ecological Economics 60: 435-449. CHAPTER 2 Background Ecosystem Services Nature has a lot to offer to humans (view from Mount Saana, Finland. Photo: Benjamin Burkhard 2014). Chapter 2 29 Introduction A historic overview of the development of the Ecosystem Services (ES) concept in a few pages is almost impossible and unavoidably biased and, for this chapter, we focused on the main events and publications1. Most authors agree that the term “ecosystem services” was coined in 1981. It was pushed to the background in the 1980s by the sus- tainable development debate but came back strongly in the 1990s with the mainstreaming of ES in professional literature and with an increased attention to their economic value. Over time, the definitions of the concept have evolved with a focus on either the eco- logical basis as ES being the conditions and processes through which natural ecosystems and their species sustain and fulfil human life or at the level of economic importance, where ES are the benefits humans derive, directly or indirectly, from ecosystem func- tions. As a compromise, the TEEB (The Economics of Ecosystems and Biodiversity) study (2008-2010) defined ES as the direct and indirect contributions of ecosystems to human well-being. Despite these differences, all definitions stress the link between (nat- ural) ecosystems and human wellbeing (see Figure 1) and the services are the ‘bridge’ between the human world and the natural world, with only humans being virtually sep- arated from that natural world. 1 Some key publications are listed at the end of this chapter as suggestions for further reading. The ecological roots The term ecosystem function was originally used by ecologists to refer to the set of ecosys- tem processes operating within an ecological system. In the late 1960s and early 1970s, some authors started using the term “func- tions of nature” to describe the ‘work’ done by ecological processes, the space provided and goods delivered to human societies. When describing the flow of ES from nature to society, the need to distinguish ‘functions’ from the fundamental ecological structures and processes was emphasised to highlight that ecosystem functions are the basis for the delivery of a service. Services are actual- ly conceptualisations (‘labels’) of the “useful things” ecosystems “do” for people that pro- vide direct or indirect benefits. Figure 1. Dependence of Human Wellbeing on Natural, Social, Built and Human capital. Source: Costanza et al. 2014. Human Well- Being Natural Capital Human Capital Built Capital Social Capital Ecosystem Services Inter- action 2.1. A short history of the ecosystem services concept Rudolf de Groot, Leon Braat & Robert Costanza Mapping Ecosystem Services30 The socio-cultural roots In the late 1960s and early 1970s, a wave of publications was produced which addressed the notion of the usefulness of nature for society, other than being an object to con- serve based on ethical concerns. Terms such as functions of nature, amenity and spiritual value were used in addition to, but not re- placing, intrinsic values of nature, empha- sising the importance to cultural identity, livelihood and other non-material benefits. This expanding field, recognising the depen- dence of people on nature, finally led to the coining of the term “ecosystem services” in the early 1980s. The economic roots The ways nature provides benefits to humans are discussed throughout economic history from the classical economics period to the consolidation of neo-classical economics and economic sub-disciplines specialised in environmental issues. Some of the classical economists explicitly recognised the contri- bution of nature rendered by ‘natural agents’ or ‘natural forces’. However, although they recognised their value in use, they general- ly denied nature’s services role in exchange value, because they were considered as free, non-appropriable gifts of nature. The phys- iocrat’s belief that land was the primary source of value was followed by the classi- cal economist’s view of labour as the major force behind the production of wealth. Marx considered value to emerge from the combination of labour and nature: “Labour is not the source of all wealth. Nature is just as much the source of use values (and it is surely of such that material wealth consists!) as labour, which itself is only the manifesta- tion of a force of nature”. In the 19th century, industrial growth, tech- nological development and capital accumu- lation led to changes in economic thinking that caused nature to lose importance in eco- nomic analysis. By the second half of the 20th century, land or more generally environmen- tal resources, completely disappeared from the production function and the shift from land and other natural inputs to capital and labour alone and from physical to monetary and more aggregated measures of capital, was completed. In the second half of the 20th century, environmental problems became a topic of interest to some economists who founded the Association for Environmen- tal and Resource Economists in 1979. The undervaluation in public and business deci- sion-making of the contributions by ecosys- tems to welfare was partly explained by the fact that they were not adequately quantified in terms comparable with economic services and manufactured capital. From the perspective of environmental eco- nomics, non-marketed ecosystem services are viewed as positive externalities that, if valued in monetary terms, can be more explicitly in- corporated in economic decision-making. In 1989, the Society for Ecological Economics was founded which conceptualises the eco- nomic system as an open sub-system of the ecosphere exchanging energy, materials and waste flows with the social and ecological systems with which it co-evolves. The focus of neo-classical economists on market-driven efficiency is expanded with issues of equity and scale in relation to biophysical limits and to the physical and social costs involved in economic performance using monetary along with biophysical accounts and other non-monetary valuation languages. Neo-classical and ecological economists dif- fer markedly regarding their approach to the sustainability concept. The so-called “weak sustainability” approach, which assumes the ability to substitute between natural and man- Chapter 2 31 ufactured capital, is typical for neo-classical environmental economists. Ecological econo- mists generally embrace the so-called “strong sustainability” approach, which maintains that natural capital and manufactured capital are in a relation of complementarity rather than of one of substitutability. They also differ with respect to approaches to ES valuation. Monetary valuation, costs versus benefits, of marketed goods and services have been pri- mary in neo-classical approaches, while eco- logical economists tend to show more interest in inclusion of non-monetary and non-mar- ket goods and services approaches. Ecosystem services in policy and practice In the 1970s and 1980s, ecological concerns were framed in economic terms to stress so- cietal dependence on natural ecosystems and raise public interest for biodiversity conser- vation. Already in the 1970s, the concept of ‘natural capital’ was used and shortly there- after several authors started referring to “eco- system (or ecological, or environmental, or natural) services”. The rationale behind the ecosystem service concept was to demon- strate how the disappearance of biodiversity directly affects ecosystem functions that un- derpin critical services for human well-be- ing. The 1997 calculation of the total value of the global natural capital and ES was a milestone in the mainstreaming of ES. The Millennium Ecosystem Assessment (2005)2 constitutes another milestone that firmly placed the ES concept on the policy agenda. The TEEB3 study (2010), building on this initiative, has added a clear economic con- notation. The interest of policy makers has turned to the design of market-based instru- 2 http://www.maweb.org 3 http://www.teebweb.org ments to create economic incentives for con- servation (see Chapter 4.3), e.g. Although one has to be careful that the con- cept is not misused, the benefits of greater awareness of the full spectrum of values of nature outweigh the risk and with the adop- tion of the Aichi-targets (see below) at the CBD convention and the creation of the Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES4 in 2012) as described below the ES-concept has been firmly placed on the political agenda. Espe- cially CBD-Aichi Biodiversity Targets 1 and 2 are relevant: Target 1, “by 2020, at the latest, people are aware of the values of bio- diversity and the steps they can take to con- serve and use it sustainably” and Target 2, “by 2020, at the latest, biodiversity values have been integrated into national and local devel- opment and poverty reduction strategies and planning processes and are being incorporat- ed into national accounting, as appropriate, and reporting systems”. The efforts to achieve these targets, in Europe coordinated by the Mapping and Assessment of Ecosystems and their Services (MAES5) contribute much to greater awareness of the many benefits of na- ture and help to give them more weight in everyday decision-making (see Chapter 7.1). Recently, the business-world is also waking up to the ‘ecosystem services-movement’ and created the Natural Capital Coalition6 to bet- ter account for ES and biodiversity conserva- tion in their business models. Although much has been achieved, even more remains to be done to further develop the ES ‘science’ and embed the concept in ev- eryday policy and practice to enhance nature conservation and sustainable use of ES which is the main objective of the Ecosystem Ser- vices Partnership (ESP), founded in 20087. 4 http://www.ipbes.net 5 http://biodiversity.europa.eu/maes 6 http://www.naturalcapitalcoalition.org 7 http://www.es-partnership.org Mapping Ecosystem Services32 Further reading Braat LC, de Groot RS (2012) The ecosystem services agenda: bridging the worlds of natural science and economics, conserva- tion and development and public and pri- vate policy’. Ecosystem Services 1: 4-15. Costanza R, d’Arge R, de Groot RS, Farber S, Grasso M, Hannon B, Limburg K, Naeem S, O’Neill R, Paruelo J, Raskin RG, Sutton P, van den Belt M (1997) The Value of the World’s Ecosystem Services and Natural Capital. Nature 387: 253-260. Costanza R, de Groot RS, Sutton P, van der Ploeg S, Anderson SJ, Kubiszewski I, Far- ber S, Turner RK (2014) Changes in the global value of ecosystem services. Global Environmental Change 26: 152-158. Daily G (Ed.) (1997) Nature’s Services. Soci- etal Dependence on Natural Ecosystems. Island Press, Washington, D.C., 412 pp. Gómez-Baggethun E, de Groot R, Lomas PL, Montes C (2010) The history of ecosystem services in economic theory and practice: from early notions to markets and pay- ment schemes. Ecological Economics 69: 1209-1218. Potschin M, Haynes-Young R, Fish R, Turner RK (Eds.) (2016) Routledge Handbook of Ecosystem Services. Routledge, T&F Group, 640 pp. Chapter 2 33 2.2. A natural base for ecosystem services Anik Schneiders & Felix Müller Introduction Formally, the natural base for ecosystem ser- vices (ES) arises from the performance of the living and non-living components of an eco- system and the interrelations between them. The respective ecosystems can be character- ised as a result of their structural features, their functional attributes and their organ- isational properties. While the latter items demonstrate the overall schemes of ecological interactions, the self-organising processes and the whole system’s dynamics, the functional viewpoint highlights the flows and pools of energy, water, matter and information. The structural aspect of ecosystems is related to the spatio-temporal characteristics of the biotic and abiotic elements. The focal fea- tures of this viewpoint are the components of biodiversity, which play a significant role for the support of ES. The 2020 targets of the Biodiversity Strategy are focussing on two perspectives: the ‘intrinsic value’ of bio- diversity and the ‘life insurance value’ essen- tial for ES supply (see Chapter 5.1). In the following pages, the second perspective will be discussed by examining the cross-correla- tions between biodiversity, ecological integ- rity, ecosystem functions and ES. Biodiversity within the social- ecological system Ecosystems and society are closely connect- ed within a Social-Ecological-System (SES) (Chapter 2.3). The flow from the ecosystem towards society is generated through the supply of ES. The flow back into the sys- tem is society’s influence on the ecosystem generated by drivers and governance. Each step within the system is related to biodi- versity, which is the total stock or the living part of our natural capital. It determines the self-regulating capacity of the system and the attitudes of biodiversity dynamics, such as resilience or adaptability. Within the system, specific ecological func- tions are essential to support and supply a specific ES: for example, primary produc- tion and pollination for food production, water infiltration capacity for water provi- sion and organic decomposition for soil fer- tility. These specific functions depend upon a specific part of biodiversity and often, increasing biodiversity will optimise these functions. Based on supply and demand, the final ES is generated, e.g. as a yield of food or wood, or a direct use of green infrastructure. Based on the benefits of a service, people will even- tually value the components of biodiversity. This can be an ethical or ‘intrinsic value’, but also a cultural or instrumental value. To complete the circle, the societal impact and the governance flow can be adjusted, which is based upon a biodiversity strategy. Here targets are formulated and adjusted on different scales. In line with these objectives, management plans will be developed and Mapping Ecosystem Services34 implemented and indicators will be chosen to measure the trend and to control the dis- tance to the target. Biodiversity and natural capital Biodiversity as a whole is the ‘living’ part of the natural capital. It is our main capacity to generate ES and to ensure adaptation to environmental changes. Figure 1 shows the essential components of the natural capi- tal and the connection with ES and nature conservation. To characterise biodiversity aspects, each of the four organisation lev- els (gene, species, ecosystem, landscape) should be represented. All levels can be studied from different perspectives: the first perspective is ‘composition’ or the presence or absence of a specific property, such as a specific genetic allele, a rare species or a his- torical landscape. Also for cultural ES, such as ecotourism, the presence or absence of specific or charismatic species or landscapes is crucial. The second perspective is ‘diversi- ty of functions’. This part focuses on indi- cators for specific ecosystem functions such as predation, photosynthesis, carbon flows, or nutrient cycling. This part of biodiversity is important for the supply of many regu- lating ES and for the adaptive capacity to environmental changes and perturbations. The third perspective is ‘structural diversity’: how fragmented is the landscape, how many vegetation layers has a lake or a forest? The landscape patterns or vegetation structures are part of the way people perceive nature and this is closely related to cultural ES such as the maintenance of historical landscapes or unobstructed views. The degree of frag- mentation and connectivity in a landscape are also crucial for the migration capacity of species and their adaptive capacity to climate change. The fourth and last per- spective is ‘stock’, a prerequisite to harvest a provisioning ES, but also to most other ES. To observe the dynamics of these biodiversi- ty components, several indicator approaches are utilised. In most regions there is a dom- inance of ‘composition’ indicators linked with the nature conservation strategy while indicators for diversity of functions, con- nectivity or vegetation structure are rarely developed. Biodiversity, ecosystem functions and services Understanding how key ecosystem func- tions determine ES supply, how it depends on biodiversity and understanding the ef- fects of shortcutting these functions by technological variants is crucial in the search for nature-based solutions. The basic in- terrelations between these components are sketched in Figure 2. In the lower box, basic ecosystem elements and relations are depict- ed. In this work, biodiversity structures are perceived as biotic processors which perform active life processes and which can be distin- guished, e.g. due to their roles in food webs. On the other hand, the abiotic processors, such as features of soil, geomorphology or climate, are creating and degrading concen- tration gradients and determining the living conditions of the biota. Both are linked by genes species ecosystems landsapes Link ES Figure 1. Four complementary perspectives of biodiversity, applicable to four organisation levels (gene, species, ecosystem & landscape). genes species ecosystems landsapes i S Chapter 2 35 ecosystemic process bundles that are the dy- namics or pools and flows of energy, carbon, water and nutrients. All of these elements are operating in complex, self-organised in- teraction schemes. Their characteristics can be aggregated into different groups of functional outcomes. To assess the overall state of these complex schemes, aggregated indicators such as eco- system integrity or ecosystem health are developed. For instance, the indication of ecosystem integrity is based on an accessi- ble number of structural items of biodiver- sity and ecosystem heterogeneity, combined with the functional items representative for the energy balance, the water balance and the matter balance of ecosystems. The aggregation of functional units can also be made to represent specific ES. For exam- ple, photosynthesis leads to the fixation of CO2 which is influenced by the static abiotic site conditions, the dynamics of solar radia- tion, rainfall, evapo-transpiration or air tem- perature, but also by the nutrient and water provision and the state of competition with other plants. The result is an increase in phy- tomass and, on a longer time scale, an input of litter into the soil subsystem, where the carbon can be transferred and sequestered into long-term stable humic compounds. These process sequences are interpreted as a functional subsystem, e.g. as carbon seques- tration. These subsystems are illustrated by the middle box in Figure 2. They connect the system with a potential ES supply (see Chapter 5.1). Normatively it is only rec- ognised as a service delivery if there is a hu- man benefit related to its performance (see Chapter 2.3). In our example, the ability of ecosystems to fix carbon from the atmo- sphere becomes a service because this process can be helpful in mitigating elevated CO2 concentrations in the atmosphere which are responsible for global temperature rises. Therefore the described process sequence is a basic component of the regulating ES global climate regulation. The production of this service emerges from a complex se- quence of interrelated processes, which in turn is influenced by all self-organised eco- system interactions illustrated in Figure 2. Such connections are also responsible for most provisioning ES, because the prima- ry and secondary production functions are strongly linked to the sequestration sequence. Also the regulation of nutrient budgets depends on the cycling and accu- mulating activities of the biotic system com- ponents, as well as the potential of the abi- otic sphere to physically or chemically retain nutrients within the soil matrix. As a result of these process sequences, the seepage wa- ter is filtered and can be used for human purposes, e.g. as drinking water. Finally, cul- tural ES also depend on ecological interac- tions, because resulting ecosystem functions provide the basic preconditions to create and maintain certain structural conditions which human beings perceive as attractive phenomena. As a result, we can observe very complex interrelations between ecosystem functions and ES. Some key functions and structures for 16 ES are listed in Table 1. A steering variable is the direct driver of a service, e.g. primary production for wood production. A supporting variable creates important boundary conditions, e.g. pollination and pest control for crop production. Most eco- system functions serve various ES. But what is the role of biodiversity for each of those functions? Many experimental studies demonstrate that an increase in the variety of genes or species contributes to the optimisation of one of the functions. Sow- ing a grassland ecosystem with more species will, for example, generate a higher biomass. For wood biomass usually a positive diver- Mapping Ecosystem Services36 Supply of ecosystem services Biotic processors Ecosystem process bundles Abiotic processors Regulating services Provisioning services Cultural services Exemplary resulting ecosystem functions Air ltering Secondaryproduction Energy ows & pools Water ows & pools Carbon ows & pools Nutrient ows & pools Groundwater storage Temperature buering Pollination Soil formation Primary production Nutrient retention Carbon sequestration Pest dynamics Nutrient cycling Erosion processes Passive gradient dynamics Abiotic structures Active life- support processes Biodiversity structures Self-organised ecosystem interactions Figure 2. Diagram sketching the relations between ecological structures and processes (self-organised ecosystem interactions), exemplary ecosystem functions and ecosystem services. The interrelations are also described in the following Chapter 2.3. Chapter 2 37 sity-production relationship is found, as a result of synergies between species and a bet- ter utilisation of resources, although some combinations create a negative effect due to competition. The fact that many functions are optimised by a higher biodiversity also means that a loss of diversity will generate a suboptimal function, often compensated by human inputs of energy, materials or technology (Chapter 5.1). It is a reality that technical compensation can lead to a disin- tegration of ES potentials and biodiversity in land use. For example, the correlation be- tween species numbers and productivities is broken by the additional inputs of energy, manpower, fertilisers or pesticides. Thus, to- day, modern agriculture produces the high- est biomass under conditions of (optimally) single-species monocultures. Towards nature-based solutions Each ES can be delivered in a gradient from naturally to technologically based solutions. Table 1. Representation of ecosystem functions and structures steering ( ) or supporting ( ) an ecosystem service or a biodiversity target linked with intrinsic valuation. White fields demonstrate indirect effects. Essential functions or structures for the supply of a service Fo od W oo d pr od uc tio n Pr od uc tio n en er gy c ro ps Ve ni so n W at er p ro du ct io n Po lli na tio n Pe st co nt ro l Pr es er vi ng so il fe rt ili ty Fl oo d co nt ro l C oa sta l P ro te ct io n G lo ba l c lim at e re gu la tio n N ut ie nt re gu la tio n W at er re gu la tio n Re gu la tio n ai r q ua lit y N oi se re m id ia tio n C on tro l e ro sio n ris k G re en sp ac e ou td oo r ac tiv iti es N at ur a 20 00 G re en in fra str uc tu re Provisioning ES Cul- tural ES Nature conser- vation Primary production Animal production Soil formation Nutrient availability / -cycling Decomposition of organic material Carbon storage Conservation carbon stock Storage rain water (infiltration capacity) Ground water retention Storage river water River Drainage Combating soil loss Pollination Pest control Prevent disease Air purification capacity Scattering and absorption sound Buffering coastal storms Regulate population dynamics Regulating ecosystem dynamics, succession Stability ecosystem processes Ecosystem resilience Development of complex ecological networks Develop ecosystem diversity / habitat quality Mapping Ecosystem Services38 Nature-based solutions depend more on biodiversity, generate a lower impact on sur- rounding ecosystems and guarantee a lower impact on other ES and a more sustainable use of the service itself. The use of a service is always a balance between supply and de- mand. In highly populated areas, for most ES the current demand is much higher than the supply. The excessive demand, together with a high drive for more human control, has affected and transformed most natural eco- systems towards the technological side of the gradient, in order to maximise a single service. The supporting and regulating role of biodi- versity is systematically replaced by techno- logical inputs, energy inputs, chemical inputs and management. This is true for nearly all provisioning ES, but also for most regulating and cultural ES. The challenge is to optimise the total supply of a bundle of ES, ensuring ES delivery and maintaining ecosystem func- tioning in the long term. Relying on more nature-based solutions will increase positive and decrease negative interactions. Conclusions • All relationships in social-ecologi- cal-systems are driven by different as- pects of biodiversity. All these interac- tions should be analysed in order to set up biodiversity strategies. • The creation of ES is founded on very complex schemes of ecological interac- tions with very high mutual interde- pendencies. • Understanding how key functions deter- mine ES supply and how they depend on biodiversity and understanding the effect of short-cutting these functions by technological variants, is crucial in the search for nature-based solutions. • Moving towards more nature-based solutions of ES supply, generates posi- tive effects for both biodiversity and the sustainable supply of ES bundles. Further reading Cardinale BJ et al. (2012) Biodiversity Loss and Its Impact on Humanity. Nature 486 (7401): 59-67. Haines-Young R, Potschin MP (2010) The links between biodiversity, ecosystem ser- vices and human well-being. In: Raffaelli D, Frid C (Eds.): Ecosystem Ecology: A New Synthesis. BES Ecological Reviews Series, CUP, Cambridge: 110-139. Kandziora M, Burkhard B, Müller F (2013) Interactions of Ecosystem Properties, Eco- system Integrity and Ecosystem Service Indicators - A Theoretical Matrix Exercise. Ecological Indicators 28 (SI): 54-78. Mace GM, Norris K, Fitter AH (2012) Biodi- versity and Ecosystem Services: A Multi- layered Relationship. Trends in Ecology & Evolution 27 (1): 19-26. Morin X, Fahse L, Scherer-Lorenzen L, Bug- mann H (2011) Tree Species Richness Pro- motes Productivity in Temperate Forests through Strong Complementarity between Species. Ecology Letters 14 (12): 1211-19. Noss R F (1990) Indicators for monitoring biodiversity – a hierarchical approach. Conservation Biology 4: 355-364. Schneiders A, Van Landuyt W, Van Reeth W, Van Daele T (2012) Biodiversity and Ecosystem Services: Complementary Ap- proaches for Ecosystem Management? Ecological Indicators 21: 123-33. Chapter 2 39 2.3. From nature to society Marion Potschin & Roy Haines-Young Although people have always depended on nature, in modern societies it is easy to lose sight of the fact that we still do. Indeed, many have argued that our failure to recog- nise the value of nature and especially the contribution that biodiversity makes to our well-being, explains much of our damaging behaviour towards the environment. It is against this background that the concept of ecosystem services (ES) is so important as it highlights the ways in which people and nature are connected. The links between people and nature are, however, complex and so it is hardly surpris- ing that people have defined ES in different ways. Some think of ES as the benefits that nature provides to people, like security and the basic material we need for a good life. Others view ES as the contributions that the ecosystem makes to such things. These dif- ferences in definition are explored in more detail in Chapter 2.4. For the moment it is sufficient to note that despite differences in the way ES are defined, most commentators agree that there is some kind of ‘pathway’ that goes from ecological structures and pro- cesses at one end through to the well-being of people at the other (Figure 1). This idea can be represented in terms of what we call the ‘cascade model’. It is a way of expand- ing thinking about ecosystems to include Environment Supporting or intermediate services Biophysical structure or process Limit pressures via policy action? Σ Pressures (e.g. woodland habitat or net primary productivity) Benet (e.g. contribution to ascpects of well-being such as health and safety) Value (e.g. willingness to pay for woodland protection or for more woodland, or harvestable products) Service (e.g. ood protection or harvestable products) Function (e.g. slow passage of water, or biomass) Final services CICES The ‘production boundary’ Goods and Benets The Social and Economic System Figure 1. The cascade model. Credit: Haines-Young and Potschin. Linking people and nature: Socio-ecological systems Mapping Ecosystem Services40 people and, as such, it might be described as a ‘socio-ecological system’. Finding out how these socio-ecological systems work and how we can act to sustain them are core issues in the field of ecosystem services. The task not only involves the study of ecology, but also such things as social practices, gov- ernance and institutional structures, tech- nology and, most importantly, the things people value. Note: ‘CICES’ in Figure 1 is the Common International Classification of Ecosystem Ser- vices, it is described in more detail in Chapter 2.4; it is a way of categorising and describing the final services that sit at the interface of na- ture and society. Unpacking the cascade model To understand how socio-ecological systems work, it is useful to ‘unpack’ the cascade model to see the inter-relationships between the elements. Ecosystem services are at the centre of the cascade model which seeks to show how the biophysical elements of the socio-ecological system are connected to the socio-economic ones; ES are at the interface between people and nature. The ‘ecosystem’ is represented by the ecolog- ical structures and processes to the far left of the diagram. Often we simply use some label for a habitat type, such as woodland or grassland (Chapter 3.5), as a catch-all to denote this box, but there is no reason why we cannot also refer to ecological processes, such as ‘primary productivity’ as something that can also occupy this part of the diagram (Chapter 2.2). In either case, given the com- plexity of most ecosystems, when we want to start to understand how they benefit people, then it is helpful to start by identifying those properties and characteristics of the system that are potentially useful to people. This is where the idea of a ‘function’ enters into the discussion. In terms of the cascade model, these are taken to be the ‘subset’ characteris- tics or behaviours that an ecosystem has that determines or ‘underpins’ its capacity to de- liver an ecosystem service. Some people call these underpinning elements ‘supporting’ and ‘intermediate’ services, depending on how closely connected they are to the final service outputs; we believe, however, this terminology deflects attention away from the important characteristics and behaviours of an ecosystem that generate different ser- vices. Thus using our terminology for one of the examples in Figure 1, the primary pro- ductivity of a woodland (i.e. an ecological structure) generates a standing crop of bio- mass (i.e. a functional characteristic of the woodland), parts of which can be harvested (as a ‘provisioning’ service). In the cascade, it is envisaged that services contribute to human well-being through the benefits that they support; for example by improving the health and safety of people or by securing their livelihoods. Services are therefore the various ecosystem stocks and flows (Chapter 5.1) that directly contrib- ute to some kind of benefit through human agency. The difference between a service and a benefit in the cascade model is that bene- fits are the things that people assign value to; they are therefore synonymous with ‘goods’ and ‘products’. The cascade model suggests that it is on the basis of changes in the values of the benefits that people make judgements about the kinds of intervention they might make to protect or enhance the supply of ES; this is indicated by the feedback arrow at the base of the diagram. The importance of ‘values’ is that they can be expressed in many ways; for example, alongside monetary val- ues, people can express the importance they attach to the benefits using moral, aesthetic and spiritual criteria (Chapter 4). Despite the simplicity of the cascade model, it is useful in highlighting a defining char- acteristic of an ecosystem service, namely Chapter 2 41 that they are, in some sense, final outputs from an ecosystem. They are ‘final’, in that they are still connected to the ecological structures and processes that gave rise to them and final in the sense that these links are broken or transformed through some human interaction necessary to realise a benefit. Often this intervention can take the form of some physical action such as harvesting the useful parts of a crop. The interaction might also be non-material and more passive involving, for example, the benefit obtained from the reduction or regulation of some kind of risk (flood risk is the example shown in Figure 1), or the intellectual or spiritual significance of na- ture in a particular cultural context. Thus services are at the point where the ‘pro- duction boundary’ is crossed between the biophysical and the socio-economic parts of the socio-ecological system. Balancing supply and demand Socio-ecological systems are, of course, more complex than Figure 1 suggests. However, this simple diagram does help us to under- stand that all the different elements of the cascade need to be considered if we want to appreciate what an ecosystem service really is and how it connects people and nature1. We need to map and measure indicators across the entire pathway to build up a complete picture. The left hand side of the cascade captures the important elements that deter- mine the capacity of an ecosystem to supply services, while the right hand side identifies the aspects of the demand for them. And un- derstanding the balance between them is at the heart of the contemporary sustainability debate and key to our understanding of the way people and nature are linked. Further reading Potschin M, Haines-Young R (2011) Eco- system Services: Exploring a geographical perspective. Progress in Physical Geogra- phy 35(5): 575-594. Potschin M, Haines-Young R (2016) Defin- ing and measuring ecosystem services. In: Potschin M, Haines-Young R, Fish R, Turner RK (Eds.) Routledge Handbook of Ecosystem Services. Routledge, London and New York: 25-44. 1 see for example: http://www.biodiversity.fi/eco- systemservices/cascade/ Mapping Ecosystem Services42 2.4. Categorisation systems: The classification challenge Roy Haines-Young & Marion Potschin Introduction Categorising and describing ecosystem services (ES) is the basis of any attempt to measure, map or value them. It is the basis of being transparent in what we do, so that we can communicate our findings to others, or test what they conclude. So fundamental is the need to be clear about how we classi- fy ES that it might seem that it is an issue that must already be well and truly resolved. The aim of this chapter is to suggest that this might not, in fact, be the case entirely and that the way we categorise ES is something that still represents a challenge. A number of different typologies, or ways of classifying ES are available, including those used in the Millennium Ecosystem Assess- ment (MA) and The Economics of Ecosys- tems and Biodiversity (TEEB) and a num- ber of national assessments, such as those in the UK, Germany and Spain. The problem with them is that they all approach the clas- sification problem in different ways, involv- ing different scale perspectives and different definitions resulting in the fact that they are not always easy to compare. In order to try to partly overcome this ‘translation problem’, the Common International Classification of Ecosystem Services (CICES) was proposed in 2009 and revised in 2013. A typology trans- lator is available via the OpenNESS-HUGIN website”1. We do not argue that it is better than any other system, but it illustrates the difficulty of 1 http://openness.hugin.com/example/cices designing a classification system that is simple and transparent to use. We will argue that the problem of classification is still worth working on – and it is certainly not something that can be taken for granted. We would encourage ev- eryone to think about it when they embark on any kind of analysis involving ES. The conclusion that we would like to advance is that the ES community probably needs to develop a number of different classifications or typologies that can be used to name and describe all the elements in the cascade that we described in Chapter 2.3, namely: the ecosystem or habitat units that give rise to the ES of interest, the ecological functions that are associated with them, as well as the benefits and beneficiaries whose well-being is dependent on the output of services and, of course, the values that people assign to these benefits. Services can also be classified accord- ing to such criteria as whether they give rise to private or public benefits, whether people can be prevented from accessing the service (‘excludable’ vs ‘non-excludable’), or whether the use of a service by one individual or group affects the use by others (‘rival’ vs ‘non-rival’). The Common International Classification of Ecosystem Services (CICES) CICES was originally developed as part of the work on the System of integrated Environ- Chapter 2 43 mental and Economic Accounting (SEEA) led by the United Nations Statistical Division (UNSD), but it has been used by the wider ecosystem services community to help define indicators of ES, or map them. In designing it, the intention was to provide a way of char- acterising ‘final services’, namely those that interface between ecosystems and society. In this sense, it follows the definition used in TEEB, namely that these final services are the things from which goods and benefits are derived. However, it did try to use as much of the terminology that was already widely em- ployed and so used the categorisation of ‘pro- visioning’, ‘regulating’ and ‘cultural’ services that were made familiar by the MA. Material and energetic outputs from ecosys- tems from which goods and products are de- rived are contained in CICES provisioning services. Regulating services categories refer to all the ways that ecosystems can mediate the environment in which people live or de- pend on in some way and therefore benefit from them in terms of health or security, for example. Finally, the cultural category identified all the non-material characteris- tics of ecosystems that contribute to, or are important for people’s mental or intellectual well-being. CICES is hierarchical in struc- ture, splitting these major ‘sections’ succes- sively into ‘divisions’, ‘groups’ and ‘classes’. Figure 1 illustrates how this works using the example of ‘cereals’. The full version of CICES is available online2. Facing the challenges of categorisation The first challenge that working on CICES showed was how difficult it is to categorise ‘final ecosystem services’. These, according to Boyd and Banzhaf, are the ‘end-products of nature’ who argue that it is important to define them clearly to avoid the problem of ‘double counting’ when we calculate their val- ue; i.e. assessing the importance of a compo- nent of nature more than once generally be- cause it is embedded in, or underpins, a range of different service outputs. More formally these authors suggest final services ‘are compo- nents of nature, directly enjoyed, consumed, or used to yield human well-being’. The prob- lem is that, what constitutes a final service, generally depends on the context in which the assessment or mapping exercise is being made; thus CICES lists potential final services. A second challenge was whether abiotic eco- 2 www.cices.eu Section Division Class Class type Provisioning Cultivated crops Nutrition Non-nutritional biotic materials Group Biomass Water ... ... Cereals Figure 1. The hierarchical structure of CICES illustrated with reference to a provisioning service (cultivat- ed crops - cereals). Credit: Haines-Young and Potschin. Mapping Ecosystem Services44 system outputs like wind or hydropower, or minerals like salt, should be categorised as ‘ecosystem services’. In the end, the argu- ment that the category ‘ecosystem services’ should be restricted to those ecosystem outputs that were dependent on living pro- cesses won the day, because it strengthened arguments about the importance of ‘biodi- versity’ to people; an accompanying provi- sional classification of abiotic services that follows the CICES logic has, however, been developed and is available. It is worth mentioning that the final chal- lenge which we encountered in designing CICES, was the difficulty that people have in distinguishing services and benefits. The distinction is a difficult one to make because it involves deciding where the ‘end-product of nature’ is transformed into a good, prod- uct or benefit, product or benefit as a result of human action of some kind. The distinc- tion we use in CICES is whether the con- nection with the underlying ecological pro- cesses and structures is retained; hence the standing crop of wheat in the field is a final service from an agricultural ecosystem, but the grain in the silo is the good or benefit. The distinction between services and ben- efits is an important one because a single service can give rise to multiple goods and benefits that all need to be identified if ser- vices are to be valued appropriately. In the case of rice for example, in addition to the harvest of the grain, rice straw and husks can be used for animal feed or as raw material for energy. Using CICES – Taking stock In this chapter we have used CICES to ex- plore some of the challenges that we need to face when developing systems for categoris- ing ES. These systems are complex and expe- rience suggests that they will need to be de- veloped in an iterative way, using experience to find out what works where and how nam- ing conventions and definitions can be im- proved. While we have used CICES to illus- trate some of these issues, it is important not to overlook the fact that it is a system that, despite limitations, has been used effectively. For example, CICES forms part of the map- ping framework designed to support the EU’s Biodiversity Strategy to 2020 (the sec- ond report of the Mapping and Assessment of Ecosystem Services (MAES) uses CICES classes to identify a range of indicators that can be used for mapping and assessment purposes3; see also Chapter 7.1). A number of papers have appeared in peer-reviewed scientific literature that have either used CICES or commented upon it as part of their methodological discussion. CICES has, for example, been used as the basis of the German TEEB study as well as the scoping work for a German Nation- al Ecosystem Assessment, NEA-DE. The TEEB report on Agriculture also recom- mends the use of CICES. Elsewhere, CICES has been refined at the most detailed class level to meet the requirements of ecosystem assessment in Belgium. Research in Finland used CICES to develop an indicator frame- work at the national scale. These kinds of applications suggest that the detailed class level in CICES can be useful as building block from broader reporting categories, the advantage being that these broader catego- ries are themselves defined in a transparent way. These types of use illustrate the kinds of application that any good classification system must be able to support. Many more applications can be found – several are listed in the further reading material. 3 see also (accessed 30/01/2016): http://biodiver- sity.europa.eu/maes/#ESTAB Chapter 2 45 Outlook While the applications of CICES suggest that the current framework is appropriate for many uses, it is also clear that we need to think carefully about how such systems can be developed. For example, researchers have suggest that it may need to be adapted to ensure that it is suitable for the assess- ment of marine and coastal ecosystems, or integrated more closely with typologies for describing underlying ecosystem functions. It is probable that marine interests were un- der-represented in the consultations that led to the current CICES version. Thus while the current version of CICES clearly works for many purposes, given the importance of categorising ES in clear and transparent ways, the development of this and other systems needs to be reviewed constantly as our needs and concepts evolve. They are es- sential tools for our mapping and assessment work. It has been suggested, for example, that a classification, such as CICES, might form part of a more general systematic approach or ‘blue print’ for mapping and modelling ecosystem services. Other authors have em- phasised that it is important to develop clas- sification systems, such as CICES, that are ‘geographically and hierarchically consistent’ so that we can make comparisons between re- gions and integrate detailed local studies into a broader geographical understanding. Our concluding point is that, whether CICES has a role to play or not, these kinds of systems will not build themselves. We need to be aware of the challenges that the cate- gorisation of ES still poses and the fact that we have only just started to address them. Note: At the time of writing, version 4.3 is to be used. This version is currently under revi- sion and version 5 is under development. All details are available on the CICES webpage4 . 4 www.cices.eu Further reading Boyd J, Banzhaf S (2007) What are ecosystem services? The need for standardized envi- ronmental accounting units. Ecological Economics 63: 616–626. Haines-Young R, Potschin M (2013) Common International Classification of Ecosystem Services (CICES), version 4.3. Report to the European Environment Agency EEA/ BSS/07/007 (download: www.cices.eu). Potschin M, Haines-Young R (2016a) De- fining and measuring ecosystem services. In: Potschin M, Haines-Young R, Fish R, Turner RK (Eds.) Routledge Handbook of Ecosystem Services. Routledge, London and New York: 25-44. Potschin M, Haines-Young R (2016b) Report on Workshop on “Customising CICES across member states”. Milestone 19 of ESMERALDA (download at: http://www. esmeralda-project.eu/documents/). CHAPTER 3 Background mapping Chapter 3 49 3.1. Basics of cartography Kremena Boyanova & Benjamin Burkhard Introduction Cartography (from Greek χάρτης khartēs, “map”; and γράφειν graphein, “write”) is the art and science of representing geo- graphic data by geographical means. Maps are the main products of cartographic work and are graphic representations of features of an area of the Earth or of any other celes- tial body drawn to scale. Regardless of the map type or the mapping technique applied (Chapter 3.2), every map has a coordinate system, a projection, a scale and includes specific map elements. These attributes usually depend on the size and shape of the mapped geographical area and the graphical design of the map representation that needs to be informative and understandable for the map-user (Chapters 5.4 and 6.4). Geographic Information Systems (GIS) are powerful tools for data Input, Management, Analysis and Presentation (IMAP principle) providing multiple possibilities for a better understanding of the structures and pat- terns of human and natural activities and phenomena (Chapter 3.4). Nevertheless, much of its easy-to-apply default-function- ality can be misleading for an inexperienced map-maker. In the present chapter, we discuss the main characteristics of maps such as coordinate system, geodetic datum, projection, scale and map elements; how to choose them ac- cordingly and what their role is for proper use of a map. The use of GIS has significant- ly simplified mapping and provides a good environment for the visualisation of Ecosys- tem Services (ES). Coordinate systems The coordinate system of a dataset is used to define the positions of the mapped phe- nomena in space. It furthermore acts as a key to combine and integrate different data- sets based on their location. This enables the performance of various integrated analytical operations, such as overlaying or merging data layers from different sources. Coordi- nate systems can be geographic, projected or vertical systems. Geographic coordinate systems A Geographic Coordinate System (GCS) uses a three-dimensional spherical surface to define locations on the Earth, i.e. the Earth is repre- sented as a sphere or a spheroid. A point on that sphere is referenced by its longitude and latitude values. Longitude and latitude are an- gles measured in degrees from the Earth’s cen- tre to a point on its surface. The Prime merid- ian and Equator act as reference for longitude and latitude respectively (Figure 1). 80 60 60 40 40 20 200 Longitude 60° 60°E 55°N 55° Lat. Figure 1. The world as a globe with longitude and latitude values. Mapping Ecosystem Services50 Projected coordinate systems A Projected Coordinate System (PCS) is based on a GCS that is transferred into a flat, two-dimensional surface. For that purpose, a PCS requires a map projection, which is defined by a set of projection pa- rameters that customise the map projection for a particular location. The various map projections are discussed in detail below. Vertical coordinate systems A vertical coordinate system defines the ver- tical position of the dataset from a reference vertical position - usually its elevation (height) or depth from the sea level (Figure 2). While the definition of a geographic or pro- jected coordinate system is obligatory for all datasets, vertical coordinate systems are only needed if the vertical height of data is of rel- evance. Lack of, or wrongly defined, coordi- nate system information leads to problems of spatial data integration. (Figure 3). Therefore it is very important when using digital mapping tools that the used datasets are defined in an eligible coordinate system. Geodetic datum and transfor- mations The geodetic datum defines a) the size and shape of the Earth and b) the orientation and origin of the used coordinate system through a set of constants. The geodetic da- tum can be based on flat, spherical or ellip- soidal Earth models: – Flat Earth models are used over short distances so that the actual Earth curva- ture is insignificant (< 10 km); – Spherical models represent the figure of the Earth as a sphere with a specified radius, leading to deformations in the model which are largest at the poles; used for short range navigation and global distance approximations; and – Ellipsoidal models are the most accurate models of Earth; used for calculations over long distances; the reference ellip- soid is defined by semi-major (equato- rial radius) and flattening (the relation- ship between equatorial and polar radii). The ellipsoidal model can represent the topographical surface of the Earth (actual surface of the land and sea at some moment in time), the sea level (average level of the oceans), the gravity surface of the Earth (gravity model) or the Geoid. The Geoid is the equi-potential surface that the Earth’s oceans would take due to the Earth’s grav- itation and rotation, neglecting all other in- fluences such as winds, currents and tides. The World Geodetic System 1984 (WGS- 84) datum defines geoid heights for the en- tire Earth in a ten by ten degree grid. The +6.3 +5.8 +6.0 mean low water mean sea water Figure 2. Two vertical coordinate systems: mean sea level and mean low water. Shale Water Table Sand and Gravel Sandstone Aquifer River Discharge Point Parks Parcels Districts Streets Landuse Figure 3. Integration of datasets for the same area (inspired by Buckley 1997). Chapter 3 51 Global Positioning System (GPS) is based on the WGS-84. The geodetic datums can be horizontal (lat- itude and longitude), vertical (height) and complete. The transformation between datums requires the application of strict mathematical rules and sets of parameters, depending on the required transformation. Most GIS and mapping platforms support automated transformation between datums and coordinate systems. Map projections Map projections are mathematical repre- sentations of the Earth’s spherical body on a plain surface through mathematical trans- formations from spherical (latitude, longi- tude) to Cartesian (x, y) coordinates. Map projections usually depend for the transfor- mation on a form which can be developed or flattened – a plane, a cone, or a cylin- der - which is attached to the sphere at one point or at one or two standard lines. The respective map projections are referred to as planar, conic and cylindrical (Figure 4). The transformation of a spherical surface into a plane leads to different distortions in the lengths, angles, shapes and areas of the mapped surface. The distortions are usually smallest along the standard lines and close to the attachment point. Depending on the shape and size of the mapped area, appropri- ate projection and standard lines should be selected. Distortions are inevitable and it is impossible to create the “perfectly” projected map that fulfils all map projection properties. The four properties of the map and their re- spective projection types are: – Local shapes of the features on the map are the same as on the Earth’s surface. This conformal projection maintains all angles. – The areas of the features on the Earth are in the same proportions as on the map. Other properties - shape, angle, and distance - are distorted in equal-ar- ea projections. – The scaled distances along the standard lines, or from the attachment point, to all other points on the map are main- tained in equidistant projections. This is not valid along all lines or between any two points on a map. – The directions on the map are correct in the true-direction (azimuthal) projection. It gives the directions (or azimuths) of all points on the map correctly with respect to the centre. Some true-di- rection projections are also conformal, equal-area, or equidistant. For every map, only one or two of those properties can be fulfilled and the cartog- rapher has to make a choice, depending on the purpose and needs of the map (see Chapter 5.4). Flattеnable surfaces Flat maps Figure 4. Developable (flattenable) surfaces (in Monmonier 1996). Mapping Ecosystem Services52 Scale The scale represents the ratio of the distance between two points on the map to the corre- sponding distance on the ground. Thus large scale maps (with a large reciprocal value of the scale, such as 1:5,000) cover small areas with great detail and accuracy, while small scale maps (e.g. 1:1,000,000) cover larger areas in less detail (Figure 5). The map scale also influences generalisation (Chapter 3.4) and symbolisation (Chapter 3.3) of the map. When choosing the map scale, the cartogra- pher should consider: – Purpose of the map - the mapped phe- nomena need to be well-represented in the selected scale; – Map size - the scale need to be adapted to the size of the mapped area and the desired final size (format) of the map; – Detail - the scale need to be adapted to the detail in which the phenomena are mapped. Scale selection Map scales can be expressed as a ratio, a ver- bal statement or as a graphic (bar) scale (Fig- ure 6). On non-analogous (digital) maps, it is essential to use a graphic scale bar (linear bar). A scale bar adjusts to the resolution of the respective display, a parameter which cannot be controlled by the map maker. The variability of map size by using a projector is an example of this problem. Elements of a map Elements of a map are crucial for providing the map-user with critical information about the map content. Making a thematic map is to a large extent a creative act and the choice of map elements depends on the context, au- dience and the preferences of the map-maker. Nevertheless, there are three levels for repre- sentation of the elements of a map, presented here by by their level of relevance (Figure 7): Scale selection Small Mapped earth area Information detail Symbolisation 1:5 000 1:50 000 1:100 000 1:200 000 1:500 000 1:1000 0001:25 000 Large More More generalised Less generalised Less Ratio scales 1 : 10 000 One centimetre (on the map) represents 10 000 centimetres (in reality) (or 100 metres) 10.000 5.000 0 10.000 Miles One centimetre (on the map) represents 25 000 centimetres (in reality) (or 250 metres) One centimetre (on the map) represents 100 000 centimetres (in reality) (or 1 kilometre) One centimetre (on the map) represents 1 000 000 centimetres (in reality) (or 10 kilometres ) 1 : 25 000 1 : 100 000 1 : 1 000 000 Verbal scales Graphic scales Alternating scale bar Hollow scale bar Double alternating scale bar 10.000 5.000 0 10.000 Miles 10.000 5.000 0 10.000 Miles Figure 5. Interaction between map content and scale selection. Figure 6. Examples of ratio, verbal and graphic scales. Chapter 3 53 – Elements that make the proper reading of the map possible and it is recom- mended to add them to all maps: • Scale information; • Map direction – a symbol, usually an arrow, that indicates the true north (the direction to the North Pole); if a coordinate grid (graticule) is added to the map or on small-scale (e.g. conti- nental) maps, a north arrow is not re- quired; • Legend – the legend lists all sym- bols, their sizes, patterns and colours used in the map and the features they depict (see Chapter 3.3); they should appear in the legend exactly as they are found in the body of the map; – Elements that provide context: • Title – should provide a short and clear statement about the map content, usually stating the name of the mapped area and the map theme (in ES maps - the mapped ES) along with the depict- ed year in thematic maps; it should be considered that this information can be included in the map legend title also; • Projection – provides information about the projection and possible dis- tortions in the area, distance, direction and shape of the mapped features; • Cartographer’s name and/or the au- thority responsible for the composition of the map; • Date of production; • Data sources used to create the map. – Elements used selectively to assist effec- tive communication (optional): • Neatlines (clipping lines) – used to frame the map and indicate the exact area of the map; • Locator maps – to place the body of the map within a larger geographical context; • Inset map – a “zoomed in” map of small areas from the map with high rel- evance, where information is too clus- tered for the scale of the map body; • Index maps – when labels or other information cannot be placed effective- ly in the body of the map, they can be input separately to increase readability. Author: Kremena Boyanova Bulgarian Academy of Sciences, 2014 Data source: SWAT model outputs Coordinate System: WGS 1984 UTM Zone 34N Projection: Transverse Mercator Datum: WGS 1984 Units: Meter UPPER OGOSTA WATERSHED Freshwater Supply (average 2000-2005) Basin 0 - no relevant supply 1 - low relevant supply 2 - relevant demand 3 - medium relevant supply 4 - high relevant supply 5 - very high relevant demand Figure 7. Example map and its elements: actual map, scale, north arrow, legend, title, coordinate system and projection, cartographer’s name and institution, date of production, data source and neatline. Author: Kremena Boyanova Bulgarian Academy of Sciences, 2014 Data source: SWAT model outputs Coordinate System: WGS 1984 UTM Zone 34N Projection: Transverse Mercator Datum: WGS 1984 Units: Meter Mapping Ecosystem Services54 Conclusions Cartography is based on a long tradition and comprehensive knowledge of map-creation and map-use. ES map-makers still need to be aware of the general principles, techniques (Chapter 3.2) and logics (Chapter 3.3) of cartography, although with today’s software programmes, it seems all too easy to create lots of maps rather quickly. Digital maps are the main means of map representation now- adays and the main tool for geographic data interpretation, visualisation and communi- cation. They provide multiple opportunities but also ‘traps’ for the map-maker. There- fore, instead of producing large quantities of badly-compiled and misleading maps, ES map producers should harness the available knowledge and techniques in order to sup- port the proper application of ES and ES mapping in science, decision making and society (Chapter 7). Further reading Bugayevskiy LM, Snyder JP (1995) Map Pro- jections - A Reference Manual. Taylor & Francis, Great Britain. Fenna D (2007) Cartographic Science: A Compendium of Map Projections, with Derivations. CRC Press, Boca Raton, Florida. International Hydrographic Bureau (2003) User’s Handbook on Datum Transfor- mations Involving WGS 84. 3rd Edition (Last correction August 2008). Special Publication No. 60. Monaco. Maling DH (1992) Coordinate Systems and Map Projections, 2nd Ed. Pergamon Press. Oxford. Monmonier M (1996) How to lie with maps. 2nd ed. The University of Chicago Press. Pearson F (1990) Map Projection: Theory and Applications. CRC Press, Boca Raton, Florida. Snyder JP (1987) Map Projections - A Work- ing Manual. U.S. Geological Survey Pro- fessional Paper 1395. U.S. Government Printing Office. Washington, D.C. Snyder JP (1993) Flattening the Earth: Two Thousand Years of Map Projections. Uni- versity of Chicago Press. Chicago, Illinois. Online resources ArcGIS (ESRI Desktop Help): http://re- sources.arcgis.com/en/help/ Buckley DJ (1997) The GIS Primer. Pacific Meridian Resources Inc.: http://planet.bot- any.uwc.ac.za/nisl/GIS/GIS_primer/index. htm Further: http://geokov.com/education/map-projec- tion.aspx http://www.progonos.com/furuti http://www.colorado.edu/geography/gcraft/ notes/mapproj/mapproj_f.html http://www.colorado.edu/geography/gcraft/ notes/cartocom/cartocom_ftoc.html http://www.colorado.edu/geography/gcraft/ notes/datum/datum_f.html http://www.librry.arizona.edu/help/how/ find/maps/scale http://awsm-tools.com/geo/convert-datum http://gitta.info/LayoutDesign/en/html/in- dex.html Chapter 3 55 3.2. Mapping techniques Christoph Traun, Hermann Klug & Benjamin Burkhard Introduction Mapping is about the graphical represen- tation of spatio-temporal phenomena. Il- lustrating our complex environment by symbols and graphics requires important decisions: Does the chosen map type prop- erly reflect the Ecosystem Service(s) (ES) to be portrayed? Are more intuitive design choices available to visualise and explain a particular dataset? What happens if the map type does not fit the data? This chapter aims to investigate popular map types like dot maps, choropleth maps, proportional sym- bol maps, isarithmic maps and marker maps. We relate those types to inherent spatial and statistical characteristics of certain ES phe- nomena and give advice on advantages and possible pitfalls related to their usage. Every ES map, whether paper or digital, is a graphical representation of ES in their geo- graphic context. In most cases, such maps are built to facilitate understanding of ES in their spatial (Chapter 5.2) and/or tem- poral (Chapter 5.3) dimension. What kind of ES data should be presented to whom (e.g. general public, scientific community, ES-practitioners) greatly determine the map- ping process: a process of abstraction from geographic reality to the final map. Scientif- ic cartography developed an extensive body of theory and derived practical guidelines to accomplish this process. A major goal there- of is the provision of maps that can be in- tuitively read and correctly understood and used by the intended end user (Chapter 6.4). Matching data and map type Data are the result of measurements (Chap- ter 4.1), modelling (Chapter 4.4) or other quantifications (Chapter 4) of geographic phenomena. Air temperature data, for ex- ample, is typically gathered by taking mea- surements at several point locations. Data on tree diameters might look similar, since it uses the same geometry (points) and is measured on a metric level. However, the represented phenomenon (trees) is entirely different in nature, since trees only exist at discrete locations in space, while atmospher- ic conditions are continuously distributed and can be measured everywhere. Different data models can be used to store, analyse and present spatial data, for example in Geographic Information Systems (GIS): Vector data models represent discrete or continuous spatial phenomena by using points, lines and polygons. Vector data have high accuracy for displaying features with distinct boundaries; vector map data files usually use less memory capacity. Raster data represent the world in a regular grid of cells (pixels). Raster models are often used for continuously varying phenomena or they are the result of remote sensing. It is possible to convert vector to raster data and vice versa. However, based on the differ- ent data model concepts, such conversions normally lead to loss of information and/or data accuracy. Mapping Ecosystem Services56 When defining maps as graphic represen- tations with the aim of facilitating the un- derstanding of spatial phenomena, mapping techniques that properly reflect their main spatial characteristics should be chosen. But what does properly reflect mean? According to the congruence principle from cognitive design, the structure and content of visu- alisations should correspond to the desired structure and content of mental represen- tations. The basic mapping concept of scal- ing geographic space is appropriate in this respect, since distances and directions be- tween entities are adequately represented by the scaled distances and directions of their corresponding map symbols (except when mapping on continental scale and projec- tion distortion is apparent). Thus it facili- tates the development of mental models on the respective spatial configuration. Howev- er, it makes a difference whether a spatially continuous geographic phenomenon like the air is represented as a set of discrete dots or by alternative graphic means correspond- ing better to its spatial continuity. Spatial phenomena can be categorised based on spatial continuity and spatial (in)depen- dence. For each possible combination, Fig- ure 1 suggests a specific mapping technique, as discussed in the following section. While such a scheme can assist in selecting an appropriate thematic mapping technique for quantitative data, there are further corre- sponding considerations: – What is the intended usage of the ES map (Chapter 5.4)? Does it merely act as an interface with the ES relevant en- tities, should it provide an overview on general spatial patterns or is it intended to allow for local comparisons? – Is the data related to individual locations or is it aggregated to enumeration units? – Is the data standardised (e.g. rates) or not (raw counts)? The following section describes important thematic mapping techniques while ad- dressing such considerations. Mapping techniques Common thematic mapping techniques include dot (density) maps, marker maps, choropleth maps, proportional symbol maps and isarithmic maps. Dot (density) maps In their simplest form of one-to-one feature correspondence, dot maps (also known as dot distribution maps) follow a very easy concept: at each location of the mapped en- tity, there is a corresponding small symbol in the map. Although this one-dot-per-fea- ture approach is increasingly popular even in small scales and with very large numbers of features1, dots quickly coalesce to a shad- ing of variable intensity, which might be un- 1 http://demographics.coopercenter.org/DotMap/ Figure 1. Models of geographic phenomena and suggested symbolisation methods. Simplified after MacEachren (1992). Spatially discrete (Phenomenon only occurs at distinct, separate locations) Abrupt changes (Measured properties change abruptly over space) Example: Number of people working in National Parks (continental scale) Example: Strenght of environmental protection laws Consider using a Proportional Symbol map Consider using a Choropleth map Example: Number of salamander sightings in dierent regionsof a national park Example: Change of mean annual temperature from 1950-2015 Consider using a Dot (density) map Consider using a Isarithmic map Smooth changes (Measured properties change smoothly over space) Spatially contiuous (Phenomenon is dened everywhere) Chapter 3 57 favourable for certain applications. In that case, a one-to-many approach is favourable, were each dot represents a fixed number of entities (e.g.: 1 dot = 100 people). The choice of the number of entities per dot is related to the chosen dot size, the scale and the density of feature locations. As a rule of thumb, points should start to coalesce in the map areas of maximum density. Dot maps are especially suited to focus on the distribution patterns of entities or on differences in local densities. When using the dot density approach for polygonal ag- gregated data (e.g. number of people per district), the according number of points is placed within each polygon. To determine the position of each point within its poly- gon, several options apply: – Random point distribution is straight- forward and often used, although it might be misleading in cases with a very uneven distribution (e.g. randomly dis- tributing points representing the pop- ulation of Egypt on the country area). – Adjust the point positioning within a polygon by using information on den- sities in neighbouring polygons. – Use of ancillary information (e.g. settle- ment information from remote sensing data) for more precise point allocation. Dot density maps which are based on aggre- gated data require absolute counts as a ba- sis (e.g. number of persons per county). In addition, the use of an area-preserving map projection (see Chapter 3.1) is essential, since the density impression results from the number of dots per area unit on the map. Heat maps are frequently seen derivatives of dot maps. Instead of showing the actual dots, they use areal colouring to represent their density. Dense areas get more reddish colours (therefore “heat”) while areas with sparse data are normally coloured in blue. Although heat maps are quite popular, it is somewhat difficult to derive actual point feature numbers for a certain area. Marker maps Marker maps are a special form of dot maps that emerged with the advent of web map- ping applications such as Google maps. Ly- ing on top of a topographic base map, every marker or “pushpin” symbolises a feature of interest in its geographic location. With each marker being hyperlinked, the user can obtain additional object information or trigger certain actions, like booking a hotel room. The map itself acts foremost as an interface to data which is structured by its spatial location. Paper maps showing the location of entities often use different symbols for different ob- ject types referenced in a legend. Thus the selection of the currently relevant object is performed visually by the user. Contrary to this, a web map allows the user to query the objects of interest within a database first and then show the query result in the map. Con- sequently, no further graphical differentiation of markers is necessary (but still possible). Point markers are used to depict any type of feature geometry in the map, be it points, lines or areas. The main reason refraining from clickable areal symbols is explained by interaction challenges with other objects lying within the same area. Marker maps are often used to encode qualitative infor- mation. They mainly inform the user about individual locations and the spatial distri- bution pattern of the entities of interest. To prevent markers from coalescing in small scales, different mechanisms for grouping and/or selection can be applied. Mapping Ecosystem Services58 Choropleth maps Choropleth maps are preferably used to map data collected for areal units, such as states, census areas or eco-regions. Their main pur- pose is to provide an overview of quantita- tive spatial patterns across the area of inter- est. To construct a choropleth map, the data for each unit is aggregated into one value. According to their values, the areal units are typically grouped into classes and a colour is assigned to each class. This requires the use of meaningful colour-schemes2 (Chapter 3.3), representing the sequential or diverg- ing nature of the mapped phenomenon. Although choropleth maps are very com- mon, several pitfalls are inherently associat- ed with them: Variation within units is ignored, although the mapped phenomenon might vary con- siderably within (especially larger) units. The boundaries between units often do not align with discontinuities in the mapped phenomenon. Especially the historically defined boundaries of administrative units often poorly align with spatial discontinu- ities of current social or natural processes (Chapter 5.2). Both problems, namely the variation within units and the definition of spatial boundaries apply for many ES and belong to the so called Modifiable Areal Unit Problem (MAUP; see Chapter 6.1). Choropleth maps are only suitable for mapping standardised (“normalised”) data like rates (yield per ha per year) or densi- ties (persons per km²). Mapping absolute values (e.g. counts of persons per unit) is wrong since size differences of individual units will greatly affect the result: large units will tend to have higher values, small units lower ones. Even for experienced map users, it is impossible to mentally disentangle the 2 http://colorbrewer2.org resulting relationship between unit-size and colour for correcting the wrong impression of spatial distributions (compare Figure 2). However, in most cases, standardised values can be easily derived from raw counts. In summary, choropleth maps are a good choice to demonstrate standardised data ag- gregated to areal units, especially if there is little variation within units and the bound- aries of the units are meaningful for the mapped phenomenon. Proportional symbol maps Based on our assumption that ‘larger’ means ‘more’, proportional symbol maps use vari- ation in symbol size to depict quantities. While the size of point symbols can be used to denote quantitative attributes of point features (e.g. spring symbols scaled to wa- ter outputs), scaled point symbols are also used to represent data aggregated to areas, as discussed for choropleth maps. Contrary to the latter, not only is the colour of the areal units modified based on an attribute, but a point symbol is positioned within each area and the size of this symbol is scaled accord- ing to the desired attribute. Since comparing sizes is much easier than comparing shades, proportional symbol maps are especially ef- fective for comparison tasks. According to the scheme in Figure 1, proportional symbol maps best connote spatially discrete entities with spatially unrelated attributes. In con- trast to choropleth maps, they are capable of handling absolute data like raw object counts within differently sized areas. This is possible due to the fact that larger symbols can be related to larger areas quite intuitive- ly (Figure 3). In their basic form, the area of a symbol is scaled proportionally to the magnitude of Chapter 3 59 Figure 2. Only standardised data (rates etc.) should be mapped with choropleth maps. Inspired by Slocum (2009). Mapping relative (area standardised) values: number of objects per km². (Regularly dispersed) distribution of underlying objects. Mapping the absolute number of objects per unit leads to a wrong impression of the spatial distribution. 1km 1km 1 object/km² 1 4 16 objects areal units of dierent size object Figure 3. Symbol size relates well to the size of areal units, making proportional symbol maps capable of mapping non-standardised, absolute values (see Figure 2 for the underlying object distribution). Inspired by Slocum (2009). 16 objects 4 1 0 Mapping Ecosystem Services60 the mapped attribute. However, several vari- ants apply: – Although the subject is controversial, perceptual scaling tries to adjust the symbol size to compensate the empiri- cally tested tendency for underestimat- ing the area of large symbols. – The use of 3D-symbols like spheres or cubes allows scaling proportional- ly to symbol volume instead of area. Although volumes are estimated even more badly, this might be useful when large spans of data values have to be ac- commodated in the map. – For data of extremely large or very small value ranges, data values might be classed and classes are assigned a set of ‘graduated’ symbols. While symbol sizes still represent the order of classes, sym- bols are not proportional to the magni- tude of values any more. Thus additional information (e.g. in a legend) pointing to that fact is crucial for interpretation. – At times, data is composed of sever- al subgroups (e.g. total population by gender or age groups). To show this further subdivision, scaled diagrams can be used instead of plain symbols. Pie charts are often chosen due to their compactness. Often, proportional or graduated symbols will overlap. While overall downscaling might be a solution, a small amount of over- lap is acceptable. Using half-transparent, simple symbols like circles is a good strate- gy to cope with overlap as well. Web maps sometimes use cross-breeds of markers and proportional symbols: instead of permitting marker-overlap in small scales, nearby mark- ers are aggregated into one symbol scaled to the number of markers it contains. Isarithmic maps Many ecosystem processes like climate reg- ulation or air quality regulation take place in a spatially continuous manner. As a con- sequence, the related ES are also gradually varying over space. Isarithmic maps connect points of the same value (at certain intervals) by a line (=isoline) and are especially useful to map such smoothly changing ‘continuous field’ data. The most prominent examples of isolines are contour lines in topographic maps, connecting points of the same eleva- tion. This concept can be used for all types of continuous fields. Isarithmic maps can be combined with areal colouring using con- tinuous colour ramps. Alternatively, the ar- eas between the isolines can be filled with a sequence of classed colours. A combination of isolines with analytical hill-shading inten- sifies the ‘surface’-character of the mapped phenomenon. The construction of isarithmic maps re- quires surface data, commonly modelled as point grid or Triangulated Irregular Net- work (TIN). Grounded on a base value and an interval, isolines are constructed from the field model using spatial interpolation. Using, for example, a base value of 50 and an interval of 100 to display a surface with values ranging between 54 and 320, iso- lines of the value 150 and 250 will be the result. Since isarithmic maps emphasise the continuous, smoothly varying character of a phenomenon, it is advisable to use them for such phenomena even though the data is being provided as discrete samples. As an example, data on ecological vulnerability based on districts could be considered: while each district might have assigned a value indicating its vulnerability, local vulnera- bility might smoothly change over space, independently of sharp district borders. De- pending on the intended message (‘objective representation of risk’ versus ‘hey governor, Chapter 3 61 you are responsible for this highly vulner- able district, act!’) it might make sense to create a continuous vulnerability surface from polygonal data and utilise an isarith- mic map for its communication. When fol- lowing such an approach, it is important to use only standardised (relative) values from enumeration units for surface generation. Methods like pycnophylactic interpolation or area-to-point kriging, guarantee that the overall volume remains constant while the surface is smoothed. Apart from the basic thematic mapping concepts described so far, there are nu- merous other techniques: Cartograms3, dasymetric maps, flow maps, animated maps4 or perspective views are just some examples for techniques meeting more specialised purposes. Choosing an appropriate base map A typical ES map consists of a topograph- ic base map and one or more superimposed thematic layers showing the desired ES data. The base map provides the geographic refer- ence to the ES data, informing the user on location while simultaneously providing a sense of the actual map scale. Depending on the used mapping framework5, there is often a choice between various base maps6. Some base maps can also be edited by the user to highlight or subdue certain object classes. When choosing a base map, several aspects must be considered: – Thematic support: The base map should support the thematic ES information; 3 http://www.worldmapper.org/ 4 http://hint.fm/wind/ 5 http://tools.geofabrik.de/mc/ 6 http://maps.stamen.com therefore it depends on the mapped ES topic, what kind of geographic features should be part of the base map. While some base maps focus on the street net- work7, others emphasise the terrain or highlight administrative boundaries. Users should carefully think about what kind of information is required to sup- port the mapped ES topic. – Visual prominence: Base maps provide ancillary information, thus their place is in the visual background. In a digi- tal context there are two common con- cepts to accomplish this: A dark base map with bright and saturated thematic information on top or a light and un- saturated base map overlaid by darker and more saturated thematic layers. – Visual density: At each scale level, the base map should have approximate- ly the same visual density (number of shown features per area). If the themat- ic ES layers are rather complex, a base map with a rather low visual density (e.g. only coastline and country bound- aries) should be chosen. Generalisation Due to scale limitations it is not possible to show all spatial objects with all their detail in the limited map space. Generalisation aims to represent the ES-information in a level of detail appropriate for a given scale, user group and use context. It is necessary in cases where the visual density in maps is increasing rapidly, symbols overlap or to- pological conflicts become evident due to graphical scaling. Figure 4 shows typical operations applied in the generalisation pro- cess. Although the application of some of those operators can be automated, it is the 7 https://www.openstreetmap.org Mapping Ecosystem Services62 responsibility of the map maker to decide on the relevance of specific ES information. Conclusions Map-makers can harness the broad knowl- edge base, experience and techniques avail- able from cartography. ES-maps display highly complex human-environmental systems, consisting of discrete and contin- uous features. This complexity should also be respectively reflected in the maps, which need to be logical, clear, understandable and well-designed. Further reading Brewer CA (1999) Colour Use Guidelines for Data Representation. Paper presented at the Proceedings of the Section on Statisti- cal Graphics, Alexandria. Krygier J, Wood D (2011) Making Maps, Second Edition: A Visual Guide to Map Design for GIS: The Guilford Press. MacEachren AM (1992) Visualising uncertain information. Review of. Cartographic Per- spectives (13):10-9. MacEachren AM (2004) How Maps Work - Representation, Visualisation and Design. 2 ed. New York, London: The Guilford Press. Muehlenhaus I (2013) Web cartography: map design for interactive and mobile devices: CRC Press. Slocum TA, McMaster RB, Kessler F, Howard HH (2009) Thematic Cartography and Geovisualisation. Clarke KC (Ed.) 3rd ed, Prentice-Hall Series in Geographic Infor- mation Science. Upper Saddle River, NJ: Pearson Prentice Hall. Tversky B (2005) Prolegomenon to scientific visualisations. In Visualisation in science education, Springer, 29-42. Figure 4. Typical operators for generalisation. Modified after Phillipe Thibault (in: Slocum, 2009). Original Map Simplification Smoothing Aggregation Amalgamation Collapse Merging Exaggeration Enhancement Displacement Selection Generalised Map Inlet Road Stream Bay Bay Stream Road Station Salzburg School Airport Salzburg School Airport Station Inlet Chapter 3 63 3.3. Map semantics and syntactics Benjamin Burkhard & Marion Kruse Introduction Map-making and cartography combines sci- ence, arts, aesthetics and techniques that fol- low map-specific logic. Thus, cartography is strongly based on semiotics, the theory and study of signs and symbols. Map symbol- isation is a key attribute of each map that determines the map elements (Chapter 3.1) and their applicability for communication (Chapter 6.4) and other uses (Chapter 7). Knowledge about basic semiotic principles is needed to produce proper ecosystem ser- vices (ES) maps that are fit for purpose. Semiotics comprises semantics, syntactics and pragmatics: • Semantics is the study of the relation- ships between signs and symbols and what they represent, • Syntactics deal with the formal prop- erties of languages and systems of sym- bols and • Pragmatics analyse the relationships between signs and their users. This chapter introduces map semantics and syntactics, which are the basis for the proper use of symbols, patterns and colours for dif- ferent mapping purposes and scales. Chap- ter 6.4 deals with map pragmatics. Graphic variables Map features can be points, lines or areas (polygons). They are positioned on a map relating to their location in reality, the map scale and map projection (Chapter 3.1). Ad- ditional information is communicated by the choice of the map symbols’ shapes, sizes, colour hues, colour values, colour intensi- ties and textures. According to the different graphical variables’ semantics, they are best used to show qualitative or/and quantitative differences. Most ES maps and ES model outputs (Chapter 4.4) are choropleth maps (Chapter 3.2) displaying areas of ES supply or demand. Some ES and landscape features are displayed as point or line features. Figure 1 gives an overview of the six key graphical vari- ables and how they can be used for mapping. Shape The map symbols’ shapes are used to repre- sent qualitative differences in thematic maps. In many cases, the shape is a logical connec- tion to the feature that it represents (e.g. a petrol pump representing a gas station or a bed indicating a hotel). Text or respective let- ters/abbreviations are also often used (e.g. ‘P’ for car parking or ‘H’ for hotel). Shapes or adaptations of shapes are most often used for spatially discrete point features (see Chapter 3.2). They are rarely used for line features but can be applied in the form of cartograms (or anamorphic maps) for area features. In anamorphic maps, mapped areas are resized based on particular indicator values.1 1 See for examples: http://www.worldmapper.org/ Mapping Ecosystem Services64 Another relevant graphic feature that can be related to shape is orientation, which can be, for example, used to indicate directions of ES flows, movements or directional ES con- nections (Chapter 5.2). Topic-specific maps (e.g. geological map or weather maps) con- tain highly complex and specialised sym- bols. These maps often use their own logic of semantics and non-specialists can have difficulties interpreting them. Size Size is mainly applied in graphics to express quantitative differences, i.e. variations in amount or count (such as ‘the more the larger’ and, vice versa, ‘the less the smaller’). Size can also be used to suppress less important fea- tures. The size of point and line features can be chosen accordingly, but following a rule of thumb to choose the difference in size accord- ing to quantitative differences in the features (e.g. double size for a double amount; see ‘Classification of data’ below). For graphical reasons, some smaller linear or point features (e.g. streams) are often enlarged, although the proportional size to other symbols might not represent reality (see ‘generalisation’ Chapter 3.2). The meaning of the different sizes (their semantics) should be explained in the map legend by providing the quantitative numbers that are behind the symbols (Chapter 3.1). Size variations of area features should refer to anamorphic maps or cartograms. Combinations of different visual variables are possible, such as dot maps (see Chapter 3.2), illustrating distributions and densities Figure 1. Key graphic variables and their application in mapping (inspired by understandinggraphics). possible, but too weird to show LinesPoints Shape Colour Hue Colour Value Colour Intensity Texture Size Areas Best to show cartogram cartogram qualitativedifferences qualitative differences qualitative differences qualitative & quantitative differences quantitative differences quantitative differences Chapter 3 65 by the symbols’ shapes and quantities by their sizes. Depending on the scale of the map and the complexity of the landscape to display, shape and size may not offer suffi- cient detail or visibility for small symbols. Colour hue, value and intensity Colour hue is arguably the most powerful of the graphical variables. It can be applied to point, linear and area map elements. Differ- ent colour hues can relatively easily illustrate qualitative differences such as different land cover types in area maps. Variations in co- lour value or intensity are commonly used to portray quantitative differences in both dot and choropleth maps (see Chapter 3.2). When using colour in maps, the map-maker needs to be aware that the different colours have specific meanings for many people and cause different psychological effects when viewed. Figure 2 shows some examples, noting that there are many different interpretations on colours based on the subject area and cul- ture. In many cultures green stands for positive developments whereas red is often related to negative things such as intense heat or danger. Due to the omnipresent nature of map- ping products in all kinds of media, the map-maker needs to be aware that many colours are connected with particular geo- graphic phenomena (e.g. green for forests, blue for water bodies, red or black for urban areas; see for example the European CO- RINE land cover data set)2. Applying such commonly used colour schemes is essential for an easy and correct communication of the map content (see Chapter 6.4). Texture Texture can efficiently illustrate qualitative differences, for example different soil types, land uses or hydrological units. Similar to colour hue, texture can be applied for point, linear and area map features (see Figure 1). In combination with varying colour hues, dif- ferent thematic layers (topics) can be shown in one map. Quantitative differences can be portrayed in choropleth maps by applying increasing or decreasing texture densities (Chapter 3.2). However, mixing too many different texture types or changing direction of linear patterns can result in an over-com- plicated map design and should be avoided. Classification of data The normal map-user has a limited capacity to differentiate between a large number of colour (or grey) values or intensities. There- fore it is often necessary to classify (group) quantitative data that are to be portrayed in the thematic map. A small amount of graph- ic variations then appears in the map based on the reduced number of pre-defined data classes. Aggregating of map features into appropriately-defined classes increases the 2 http://www.eea.europa.eu/data-and-maps/fig- ures/ corine-land-cover-types-2006 Figure 2. Possible psychological associations of different colours for viewers (based on: http://gui- ty-novin.blogspot.de/2014/07/chapter-70-history- of-color-color-wheel.html). power sophistication mystery death intellect friendliness warmth caution cowardice innovation creativity thinking ideas love passion romance danger energy life growth nature money freshness peace sincerity confidence integrity tranquilty royalty luxury wisdom dignity authority maturity security stability hope simplicity cleanliness goodness purity Mapping Ecosystem Services66 readability and the usefulness of the map. Distribution patterns in the landscape can be identified easier. The choice of the ap- propriate data classification method and the number of classes has a significant bearing on the quality of the final map. Data classi- fication should be carried out carefully and with consideration to the data distribution and the purpose of the map. The data distri- bution can be checked by using histograms (Figure 3). The most common data classification meth- ods are: • Equal intervals, • Quantiles, • Natural breaks (Jenks), • Geometric intervals and • Standard deviations. GIS or cartography software programmes (see Chapter 3.4) normally offer algorithms and standardised procedures for classifica- tion of data (Figure 3). Equal intervals The data are divided into equal-sized inter- vals (such as an interval of 2, resulting in the classes 0-2; 2-4; 4-6; etc.)3. Equally-dis- tributed data (showing a rectangular shape in the histogram) would result in equal number of values in each class. However, data is usually normally-distributed with fewer values in the extreme (minimum and maximum) classes. This may lead to unequal representation of values in each class. Nev- ertheless, equal interval data classifications are recommended for many quantitative data and natural phenomena. In combina- tion with equally-spaced colour values or saturations from one class to another, the classified map can normally be understood faster (e.g. the 4th class represents a double quantity compared to the 2nd class; see also Chapter 5.6.4). Quantiles When using quantiles, all available data values are divided into unequal-sized intervals so that the number of values is the same in each class. Different from the equal interval method, each class (including the extremes) have the same number of values. This often leads to maps with more classes portraying the middle value ranges. The map-user has to be aware of the classification method and carefully check the map legend when reading the map. Natural breaks (Jenks) The natural breaks classification method is applied by checking the data distribution (for example in a histogram or in a graph) and placing class breaks around data 3 To avoid double-representation of data, each subsequent class must start with the next higher value than the one before ended (i.e. 0-2; 2.1-4; 4.1-6 etc.) (cp. Figure 4). Figure 3. Example of ArcGIS™ data classification interface with histogram. Chapter 3 67 clusters. This avoids large value variations within one class and highlights differences between different classes. As with quantile- based classified maps, the map-creator and reader must be aware of the effects of this (sometimes subjective) data classification on the resulting map. Figure 4 shows examples of the different classification methods applied to the same dataset. In the example, quantiles produce the most heterogeneous map but there are only minor differences for equal intervals and jenks. The classification method must be carefully selected based on the data set and the desired map product. Common mistakes An inappropriate selection and application of graphic variables or the wrong data clas- sification method can lead to map misinter- pretation (Chapter 6.4), confusion or pro- duction of poor map products. Bad choice of colours, the most powerful graphic vari- able, can render a map useless. A common mistake, which has been heavily stimulated by the seemingly easy map-cre- ation with various GIS, cartography and presentation software programmes, is the choice of too vibrant colour ramps that combine contrasting hues and go across for example red-blue-orange-green-yellow- brown colour schemes ignoring effects that different colours have on the map-user. Most map users may not be able to dif- ferentiate more than six or eight different colours within one map, depending on the map’s complexity and the size of the de- picted symbols. This number may be lower for people with limited colour vision. An- other important consideration when creat- ing colour maps is that colours are not rec- ognisable if the map is reproduced in black and white or greyscale. Even if printed in colour, mismatches can occur between the printed version and the computer screen if different colour models are used. Using bad map symbols that do not follow the logics of map semantics, syntactics and pragmatics often leads to noises in the map- maker/map-user communication (map cod- Figure 4. Effects of different data classification methods on resulting maps. Mapping Ecosystem Services68 ing; Chapter 6.4). Different cultural, societal or educational backgrounds may lead to dif- ferent interpretations of symbols. A cartog- rapher or a trained map-user will interpret a map differently from a novice map user. Another common mapping mistake relates to the Modifiable Areal Unit Problem (MAUP; see Chapters 3.2 and 6.1). MAUP becomes especially relevant when using different co- lour values or intensities in choropleth maps. Additionally, the use and combination of too many different colours, patterns and symbols hamper easy and appropriate map compre- hension by giving the map a ‘nervous’ look. Maps that are overcharged with information might run the risk of being ignored. The map-maker needs to be aware of the fi- nite capacity of the map-user to differentiate between the various graphic variables, espe- cially in complex maps covering large spatial scales (Chapter 5.7). The appropriate classifi- cation of quantitative data is therefore a very important step in thematic map compilation. Solutions When choosing from the different graph- ic variables shape, size, colour hue, value, intensity and texture presented above, the map-maker needs to be aware of the seman- tics and syntactics relevant to the choices. The semantic and psychological effects of different colours should therefore be care- fully considered and maps should not be overloaded with too many different colours. When illustrating quantitative differences, the colour ramp should only have one or two colour hues (Figure 5) and the colour intensities should be adapted according to the quantitative data. It is also advisable to check the visibility of the selected colours after printing in black and white. Texture may be a better choice than colour hue or intensity to portray dif- ferent classes in black and white maps. Regional peculiarities need to be taken into account when compiling ES maps because of the trans-disciplinary and complex nature of ES. Involving stakeholders and harness- ing their local or subject-specific knowledge can help to avoid cultural traps or misin- terpretations. A stepwise process that seeks feedback from stakeholders and map-users can help to improve maps and reduce the number of map misinterpretations. A proper map legend showing all symbols, symbol sizes, colour hues, values, intensities and orientations that appear in the map is mandatory in each map in order to read the map accordingly (see Chapter 3.1). Text in maps, where useful and appropriate, can support the information given by the map symbols. It must be legible, easy to compre- hend and in the language of the map-user. Specifics of ecosystem service maps ES science involves several scientific disci- plines and links multiple topics and quan- tification methods (Chapter 4) at various spatial and temporal scales (Chapter 5.7). Therefore ES mapping includes several chal- Figure 5. Examples for single hue colour ramps. Chapter 3 69 lenges (see Chapter 3.7), that can be related to map semantics and syntactics. The six key graphical variables described above are also applied in ES maps, depend- ing on the ES to be displayed, what has to be mapped (Chapter 5.1), where (Chapter 5.2), when (Chapter 5.3) and why (Chap- ter 5.4). Many regulating ES (Chapter 5.5.1) can, for example, be related to nat- ural phenomena which are often indicated by the choice of texture or orientation (e.g. for flows). Different intensities are depicted with appropriate colour hues. Provisioning ES maps (Chapter 5.5.2) often display ser- vice providing areas (Chapter 5.2), which can be point units (graphical shape) or area units (mostly displayed in choropleth maps; see Chapter 3.2). Quantities of ES supply can be portrayed by size variations (point sources, linear flows) and gradational colour values, intensities or textures. Cultural ES (Chapter 5.5.3) can be related to spatially discrete point features (e.g. iconic land- marks or religious sites displayed by map symbol shape variations) or more continu- ous area features (aesthetic experience based on viewsheds or landscape setting displayed by area features). Conclusions The map-makers have to take responsibility for their products as it is easy to impress or mislead map-users with colourful and at- tractive maps. ES maps are of high political, societal and economic relevance (Chapter 7). Therefore their compilation should close- ly follow the logics and the well-founded knowledge from graphic semiology. Based on the diversity of ES map-makers, map-us- ers, the complex topics to be displayed and their high societal relevance, ES maps need to be designed with care. Well-constructed maps can properly communicate and ex- plain complex ES phenomena. Further reading Bertin J (1967) Sémiologie Graphique. Les diagrammes, les réseaux, les cartes. With Marc Barbut [et al.]. Paris: Gauthier-Vil- lars. (Translation 1983. Semiology of Graphics by William J. Berg). Dent B, Torguson J, Hodler T (2008) Cartog- raphy: Thematic Map Design. 6th edition. McGraw-Hill Science/Engineering/Math. Monmonier M (1996) How to lie with maps. 2nd ed. The University of Chicago Press. Muehrcke PC (2005) Map Use: Reading, Analysis, and Interpretation. 5th ed. J P Pubns. Wood D (1992) The Power of Maps. The Guilford Press. Mapping Ecosystem Services70 3.4. Tools for mapping ecosystem services Ignacio Palomo, Kenneth J. Bagstad, Stoyan Nedkov, Hermann Klug, Mihai Adamescu & Constantin Cazacu Background Mapping tools have evolved impressively in recent decades. From early computerised mapping techniques to current cloud-based mapping approaches, we have witnessed a technological evolution that has facilitated the democratisation of Geographic Infor- mation Systems (GIS). These advances have impacted multiple disciplines including ecosystem service (ES) mapping. The infor- mation that feeds different mapping tools is also increasingly accessible and complex. In this chapter, we review the evolution of mapping tools that are shaping the field of ES mapping together with the different sources of information that exist at this point. We discuss briefly the suitability of these approaches for mapping different ES types and for different scientific and policy aims. Finally, we elaborate on the integra- tion of multiple tools (from desktop ap- plications to sensor, web-based, or mobile devices) and on the future developments of these methods and the possibilities they may open for ES mapping. Introduction ES mapping has achieved rapid progress in a very short time frame. To our knowledge, the first peer-reviewed ecosystem service maps were published in 1996 and, since then, a large number of ad hoc mapping studies have been conducted and a variety of tools have been developed to systematise ES mapping. The progress we have witnessed corresponds to advances in computing pow- er, modelling and GIS, the recognition of a plurality of ES approaches (i.e., participato- ry mapping (Chapter 5.6.2) and biophysi- cal modelling (Chapters 4.1 and 4.4), and the consensus that ES maps provide a direct connection between ES and the landscape and therefore with policy (Chapter 7.1). Description of main mapping software, tools and databases Computing power and data availability that support GIS analysis have evolved substan- tially in recent years. Several freeware GIS platforms have been developed, such as QGIS (Quantum GIS), GRASS GIS (Geo- graphic Resources Analysis Support System GIS), SAGA (System for Automated Geo- scientific Analyses), and gvSIG (Generalitat Valenciana Sistema de Información Geográ- fica) that provide similar functionality to the popular commercial ArcGIS software from ESRI (a list of GIS software is avail- able here1). Specific modelling approaches for mapping ES have been developed by different institu- 1 https://en.wikipedia.org/wiki/List_of_geograph- ic_information_systems_software Chapter 3 71 tions worldwide, resulting in a wide variety of possibilities for ES analysts’ use (Table 1, also see chapter 4.4). Most of these tools are openly available to the public and are con- stantly evolving. Training for the potential users of these tools is of importance for their accessibility and use for decision support. The operational time necessary for their ap- plication to case studies ranges from hours (simple spreadsheet-based tools) to several months (advanced software tools). The use of GIS in ES mapping can take three general approaches: (1) analysis tools built into GIS software packages; (2) disciplinary biophysical models applied for ES assessment (e.g., hydrological models such as the Soil and Water Assessment Tool, SWAT or Vari- able Infiltration Capacity model, VIC for wa- ter-related ES); and (3) integrated modelling tools designed specifically for ES assessment (e.g., InVEST, ARIES). The first approach is applicable for simple land cover-based anal- yses and indicator-based ES mapping (see Chapter 5.6.4) that have been used for exam- ple in Mapping and Assessment of Ecosys- tems and their Services (MAES). The second approach is appropriate for more complex model-based analyses of services that inte- grate expertise from specific disciplines (e.g., ecology for crop pollination or hydrology Tool Platform Scale2 Source Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST) ArcGIS/Stand- alone Municipal to provincial http://www.naturalcapitalproject. org/invest/ Artificial Intelligence for Ecosystem Services (ARIES) Graphical User Interface (GUI)/ Web-based Municipal to provincial http://aries.integratedmodelling. org/ Multiscale Integrated Models of Ecosystem Services (MIMES) Simile software Village/farm to global http://www.afordablefutures.com/ orientation-to-what-we-do/services/ mimes Social Values for Ecosystem Services (Solves) ArcGIS Municipal to provincial http://solves.cr.usgs.gov/ Land Utilisation Capability Indicator (LUCI) ArcGIS Village/farm to provincial http://www.lucitools.org/ Integrated Model to Assess the Global Environment (IMAGE) Set of models Global http://themasites.pbl.nl/models/ image/index.php/Welcome_to_ IMAGE_3.0_Documentation Co$ting Nature Web-based, Google Earth Municipal to provincial http://www.policysupport.org/ costingnature Ecosystem Valuation Toolkit Web-based Municipal to provincial http://esvaluation.org/ ESM-App Android Smartphone app Municipal to provincial http://www.ufz.de/index. php?en=33303 Table 1. List of the most common ES mapping tools. 2 Malinga et al. (2015) define scales as follows: village/farm < 60 km2; municipal 60-8,709 km2; provincial 8,709-83,000 km2; national 83,000-1,220,000 km2; continental > 1,220,000 km2. Mapping Ecosystem Services72 for flood regulation mapping). The third ap- proach extends the second one by utilising modelling tools that can assess trade-offs and scenarios for multiple services. Several ecosystem service valuation data- bases have been developed as well, such as The Economics of Ecosystems and Biodiversity (TEEB) Valuation Database and the Ecosystem Valuation Toolkit and these might be used to create ES maps. The Ecosystem Services Partnership (ESP) Vi- sualisation Tool is a database consisting of ES maps prepared by different researchers intended to promote synthesis of mapping studies (see chapter 7.9). Applicability of mapping tools In-depth assessment of the different map- ping tools is necessary to understand which one will best fit the user´s ES mapping con- text: time and data availability, mapping skills, types of services to map, accuracy re- quired, expected impact in decision-making and overall study aims. This means that no tool fits all criteria perfectly. Some highly complex models can provide policy support in regions with considerable time, data and personnel resources. Other approaches exist that allow ES to be mapped with more lim- ited budgets and shorter time frames. The intended use of the maps (i.e., for raising awareness or direct use in policy-making) will also influence the decision on which tools to use (see Chapter 5.6.1). In many cases, the type of ES under assess- ment will determine the mapping approach or tools to use. Services such as water regu- lation usually require modelling approaches that integrate meteorological databases, veg- etation, soils and topographic data (Chapter 5.5.1), while others such as cultural identity might require a participatory mapping ap- proach (Chapters 5.5.3 and 5.6.2). Oth- er services such as food production might use complex agricultural models or indi- cator-based approaches (Chapter 5.5.2). However, the complex nature of ES and the inter-linkages between provisioning, regulat- ing and cultural services have led to the use of different tools for each ecosystem service. It is also important to consider how different mapping tools account for accuracy, reliabil- ity and uncertainty. Accuracy is established through successful calibration, reliability through successful application in different contexts and uncertainty through methods that estimate and transparently commu- nicate uncertainty. These aspects have not been adequately covered in the past and still need to be developed for several tools. Greater transparency in the presentation of results and associated uncertainties (Chap- ter 6) is needed so that informed decisions can be made about the extent to which ES maps can be used for different purposes and which tools are best applied in different con- texts and locations. Future developments Several challenges lie ahead for mapping ES. These are related to the progress that is cur- rently underway in research and monitor- ing, remote sensing, sensor networks, data storage, data and knowledge integration, data harmonisation and sharing, database and tool maintenance and crowdsourcing, among others. On the technical side, the accumulation of a growing quantity of data raises the chal- lenge of effective storage and analysis of large amounts of data and is leading to an increased emphasis on machine learning, pattern recognition (in complex data or re- mote sensing products), and data mining. Initially high data storage requirements were Chapter 3 73 addressed by large data storage and super- computer facilities, but falling costs of dis- tributed solutions have pushed computing towards scalable clusters of computers, grids and cloud computing, all aimed at increas- ing demand-driven computational power. Some ES modelling approaches using grids include: Tropical Ecology Assessment and Monitoring (TEAM) Network, Web-based Data Access and Analysis Environments for Ecosystem Services, ARIES, enviroGRIDS and biodiversity virtual e-laboratory (Bio- Vel). The advantage of grids/clouds is that they are on-demand, self-service approach- es, so the user can unilaterally obtain the necessary computing capabilities, such as server time and network storage, without having to interact with each service’s pro- vider. Cloud-based modelling tools and in- terfaces (e.g., OpenMI) will enable the joint development of and access to modelling and visualisation tools. The ongoing development and maintenance of ES mapping tools (including free open- source software) require adequate funding. Further integration of ES mapping tools with policy will contribute to ongoing de- velopments in the field and a tailored ap- proach towards decision-making aims. Disclaimer Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. or any other Government or by the authors of this article. Further reading Bagstad KJ, Semmens DJ, Waage S, Winthrop R (2013) A comparative assessment of de- cision-support tools for ecosystem services quantification and valuation. Ecosystem Services 5: 27-39. Bagstad KJ, Reed JM, Semmens DJ, Sher- rouse BC, Troy A (2015) Linking biophys- ical models and public preferences for eco- system service assessments: a case study for the Southern Rocky Mountains. Regional Environmental Change: 1-14. Bateman IJ, Jones AP, Lovett AA, Lake IR, Day BH (2002) Applying Geographical Infor- mation Systems (GIS) to environmental and resource economics. Environmental & Resource Economics 22: 219-269. Crossman ND, Burkhard B, Nedkov S, Wil- lemen L, Petz K, Palomo I, Drakou E, Martín-López B, McPhearson T, Boyano- va K, Alkemade R, Egoh B, Dunbar MB, Maes J (2013) A blueprint for mapping and modelling ecosystem services. Ecosys- tem Services 4: 4-14. Drakou EG, Crossman ND, Willemen L, Burkhard B, Palomo I, Maes J, Peedell S (2015) A visualization and data-sharing tool for ecosystem service maps: Lessons learnt, challenges and the way forward. Ecosystem Services 13: 134-140. Eade JDO, Moran D (1996) Spatial econom- ic valuation: Benefits transfer using geo- graphical information systems. Journal of Environmental Management 48: 97-110. Mapping Ecosystem Services74 Klug H, Kmoch A (2015) Operationalizing environmental indicators for real time multi-purpose decision making and action support. Ecological Modelling 295: 66-74. Malinga A, Gordon L, Jewitt G, Lindborg R (2015) Mapping ecosystem services across scales and continents – a review. Ecosys- tem Services 13: 57-63. Nelson E, Mendoza G, Regetz J, Polasky S, Tallis H, Cameron RD, Chan KMA, Dai- ly GC, Goldstein J, Kareiva PM, Lonsdorf E, Naidoo R, Ricketts TH, Shaw RM (2009) Modeling multiple ecosystem ser- vices, biodiversity conservation, commod- ity production and tradeoffs at landscape scales. Frontiers in Ecology and the Envi- ronment 7(1): 4-11. Schröter M, Remme RP, Sumarga E, Barton DN, Hein L (2015) Lessons learned for spatial modelling of ecosystem services in support of ecosystem accounting. Ecosys- tem Services 13: 64-69. Stoll S, Frenzel M, Burkhard B, Adamescu M, Augustaitis A, Baeßler C et al. (2015) As- sessment of ecosystem integrity and service gradients across Europe using the LTER Europe network. Ecological Modelling 295: 75-87. Chapter 3 75 3.5. Mapping ecosystem types and conditions Markus Erhard, Gebhard Banko, Dania Abdul Malak & Fernando Santos Martin Introduction Ecosystems are defined by the UN Con- vention on Biological Diversity (CBD) as ‘a dynamic complex of plant, animal and micro-organism communities and their non-living environment interacting as a functional unit’. Ecosystems are therefore, by definition, multi-functional. Each eco- system provides a series of services for hu- man well-being either directly, for example, as food and fibre, or more indirectly by, for example, providing clean air and water, pre- venting floods or providing recreational or spiritual benefits. Ecosystems contain a multitude of living organisms that have adapted to survive and reproduce in a particular physical and chemical environment, i.e. their natural condition. Anything that causes a change in the physical or chemical characteristics of the environment has the potential to change an ecosystem’s condition, its biodiversity and functionality and, consequently, its ca- pacity to provide services. Up to the pres- ent, ecosystem service (ES) assessments have been based on ecosystem extent and spatial distribution as basic input parameters. The inclusion of condition assessment would add value in terms of ecosystem quality. The provision of timber, for example, not only depends on the availability of forests, but also on the species composition and age class distributions of the forests. Pollination services might be highest in grass- and crop- lands but are also highly influenced by plant species diversity of these ecosystems which again are also triggered by nutrient content and management. The concept of ecosystem mapping and conditions assessment can be applied at all spatial and temporal scales. Spatial explic- itness is important to characterise ecosys- tems in terms of their natural conditions determined by climate, geology, soil prop- erties, elevation etc. and, in terms of their physical and chemical conditions, how they are influenced by anthropogenic pressures. Local or regional assessments require more detailed information for adequate decision support. Usually national and continental mapping is less detailed but provides im- portant information at the strategic level. In any case, active stakeholder involvement is recommended to design and adapt the assessments for successful implementation into the decision process. Mapping ecosystem types Should no map of ecosystems or habitats be available, a feasible proxy has to be developed as shown in Figure 1. The basic geometry and main classes in appropriate spatial resolution can be derived directly from satellite images1 or from existing land cover / land use maps. 1 See e.g. http://www.earthobservations.org/geoss.php Mapping Ecosystem Services76 For policy-relevant information, the map should be re-classified using an ecosystem typology which represents the most im- portant types of their human management to make best use of their services, e.g. by agriculture, forestry, fisheries, water man- agement, nature protection or territorial planning. These management lines are also usually implemented in the respective legis- lations which are important cornerstones in the decision-making process. In case further geometric refinements are required, these can be performed by, for example, integrat- ing more detailed information about rivers and lakes, green linear elements such as hedgerows or detailed maps of urban areas or protected areas. If needed, such a basic map can be further refined thematically by providing more de- tailed information about the natural char- acteristics of the ecosystems and their bio- diversity. GIS–based, so-called envelope- or niche-modelling as developed for habitat or climate change impact studies, allows the combination of non-spatially referenced species or habitat information with a set of environmental parameters such as elevation, soil, geology, climate, phenology, potential natural vegetation etc. to delineate the most likely areas of ecosystem presence. This probability mapping of ecosystem presence depends on the accuracy of the descriptors of its natural boundaries (e.g. alpine meadows or calcareous broadleaf for- ests) and the availability and quality of the respective data to delineate and map these boundaries. Further enhancement can be performed by attributing statistical informa- tion e.g. crop yields or forest inventory data to the respective ecosystem classes. Corine Land Cover Bathymetry, EUSeaMap Ecosystem type map basic version Ecosystem condition maps Ecosystem Services Ecosystem map enhanced version Geometric renement HRL forest, agriculture, wetlands, water bodies, sealing riparian areas, green linear features Thematic renement Elevation, soil, geology, climate, phenology, potential,natural vegetation EUNIS data Species and Habitat data Condition indicators HD, BD, WFD, MSFD Cross-walk land cover - ecosystem type, marine classes Mapping ecosystem extent Mapping ecosystem condition Pressure maps EU Habitat Directive (HD), Bird Directive (BD), Water Framework Directive (WFD), Marine Strategy Framework Directive (MSFD) Figure 1. Work flow for ecosystem mapping and condition assessment. Chapter 3 77 Mapping ecosystem conditions Mapping of ecosystem types provides infor- mation on the natural conditions. To assess the current capability of ecosystems to pro- vide services for human well-being requires information about their current conditions which are induced by human activities. For decision support, the most compre- hensive and informative approach for the assessment of ecosystem conditions should include direct mapping and assessments in combination with information about the direct and indirect pressures which induce these conditions. This approach provides information on both the current environ- mental state and expected changes due to constant, increasing or decreasing pressures. Additionally, important information for risk assessments can be derived. Time lags between pressures and changes in ecosystem conditions are often triggered by buffering processes which indicate the resilience of species and ecosystems to the different types of stress factors affecting their condition. For better understanding of the different pro- cesses affecting ecosystem condition and the link to human activities, the DPSIR (Driv- ers, Pressures, State or Condition, Impact, Response) approach is often used (Figure 2). Drivers to cover our demand for ES and oth- er natural resources induce pressures which affect ecosystem conditions. The impacts should create (policy) responses which should again change the drivers and the way we manage our environment to cope with neg- ative impacts. The DPSIR approach should be considered not as absolute but relative to the ecosystem processes under consideration. The nutrient conditions of agro-ecosystems, for example, are the pressures for freshwater ecosystems and both conditions are pressures for marine ecosystems. Pressures affect ecosystem conditions ei- ther by concentration (e.g. ozone) or by accumulation (e.g. nitrogen and pollution load). The Millennium Ecosystem Assess- ment 2005 identified five different anthro- pogenic main pressures affecting ecosystem conditions: habitat change, climate change, invasive species, land management and pol- lution/nutrient enrichment. Human pressures are either direct, i.e. mainly from land use, or indirect, i.e. by air pollution or anthropogenic climate change. Important for ecosystem condi- tions are the strength of the pressure signal, its persistence if cumulative and its change over time. Time-series of observed chang- es in pressures are, therefore, important to analyse the causal connectivities between pressures and current condition for each ecosystem type and each spatial unit. The trend in pressures also provides a first in- sight into the expected changes in the near future. Decreasing observed trends may Drivers Population, economic growth, technology Response Policy measures to reduce impacts (protection, pollution reduction, land management Pressures Habitat change, climate change, over-exploitation, invasive species, pollution State/condition Habitat quality, species abundance and diversity, water quality etc. Impacts Change in ecosystem state (habitat loss or degradation, change in species abundance etc.) Figure 2. DPSIR framework for assessing ecosystem condition. Mapping Ecosystem Services78 For the implementation of the EU Biodiversity Strategy to 2020, Member Countries1 and European institutions perform ecosystem assessments in their territories. The European map is based on a proxy following the scheme as outlined in Figure 1 based on Corine land cover (CLC)2. Geometry of the basic map was further refined using the High Resolution Layers of Copernicus land services2 and re-classified to eight aggregated ecosystem types: urban, cropland, grassland, forests and woodland, heathland and shrubs, sparsely vegetated land, wetlands and rivers and lakes. For Europe’s seas, only a very simplified classification mainly based on EUSeaMap sea-floor mapping3 and bathymetry data is currently implemented. The basic version was thematically enhanced using the non-spatially referenced habitat information of the European Nature Information System (EUNIS) database4 in combination with a set of environmental parameters to delineate the most likely areas of ecosystem presence. 1 http://biodiversity.europa.eu/maes/maes_countries 2 http://land.copernicus.eu/pan-european 3 http://www.emodnet.eu/seabed-habitats 4 http://eunis.eea.europa.eu/ Box 1 . European ecosystem map Ecosystem map (aggregated) Marine waters Inland vegetation and habitats Inland unvegetated or sparsely vegetated habitats Human made constructions and habitats Non classied areas Marine seabed and coastal habitats Inland surface waters Open waters Arctic, alpine and subalpine scrub and grassland Mediterranean-mountain scrub and brushes Heathland scrub Grasslands and land dominated by forbs Broad leaved deciduous and evergreen woodland Mixed deciduous and coniferous woodland Coniferous and broad leaved evergreen woodland Wetlands - mires, bogs and fens Regularly or recently cultivated agricultural, horticultural and domestic habitats Tundra Screes, inland clis Snow or ice-dominated habitats Constructed, industrial and other articial habitats Unclassied areas Outside area of interest European regional seas Sublittoral sediment Marine habitats Coastal habitats Inland waters and shores Infralittoral and circalittoral rock and other hard substrata Figure 3. Ecosystem map of Europe Version 2.1 (higher resolution map can be downloaded at: http://www.eea.europa.eu/data-and-maps/data/ecosystem-types-of-europe). Chapter 3 79 indicate further improvement of ecosystem conditions and vice versa, i.e. important information for decision-making about measures to mitigate and adapt to positive or negative effects. In practice, information on ecosystem con- ditions is often insufficient for appropriate mapping and assessments. Another problem is the mapping and assessment of the com- bined effects of pressures on the ecosystem condition. Usually spatially explicit maps of the different pressures and their gradients across the area under investigation can be produced but knowledge about the com- bined effects on biodiversity and ecosystem structure and function is still insufficient. So in many cases, proxy indicators have to be used to indicate the current ecosystem con- dition as illustrated in Box 2. The way forward Ecosystem type mapping and condition as- sessments have to be further improved mak- ing use of new information and data flows from research, reporting and other sources. A major issue is the lack of detailed informa- tion on how the ecosystem condition affects ecosystem service delivery. The delivery of ecosystem services depends on the biologi- cal, physical and chemical processes and the biodiversity involved (Chapters 2.2 and 2.3) but there are few quantitative data to model and assess how these processes and function- al traits are affected by pressures such as pol- lution, management or climate change and their combined effects. Further research is needed to fill these gaps and improve our knowledge about the relationships between pressures – ecosystem conditions, related biodiversity and ecosystem service capacity. Figure 4 shows an example how the ecosystem condition can be derived by combining ecosystem mapping, report- ed data of the European Habitat Directive and statistical data. The combination of information in different units often requires re-scaling from absolute to relative values, e.g. from ‘low’ to ‘high’. Box 2 . Mapping ecosystem condition Figure 4. Map of European cropland conditions. Aggregated indicator for management intensity pressure on cropland as combination of land management and crop yield Very low Low Medium High Very high Non-cropland No data Outside coverage Mapping Ecosystem Services80 Maes J, Teller A, Erhard M et al. (2013) Map- ping and Assessment of Ecosystems and their Services. An analytical framework for ecosystem assessments under action 5 of the EU biodiversity strategy to 2020. Publications office of the European Union, Luxembourg http://ec.europa.eu/environ- ment/nature/knowledge/ecosystem_as- sessment/pdf/MAESWorkingPaper2013. pdf accessed 12 December 2015. MAES information platform: Mapping and Assessment of Ecosystems and their Ser- vices (MAES): http://biodiversity.europa. eu/maes accessed 31 May 2016. Millennium Ecosystem Assessment (2005) Ecosystems and human well-being: Syn- thesis, Millennium Ecosystem Assessment, Island Press, Washington, DC, USA: http://www.maweb.org/en/index.aspx ac- cessed 31 May 2016. Potschin M, Haines-Young R, Fish R, Turner RK (Eds.) (2016) Routledge Handbook of Ecosystem Services. Routledge London and New York 629 pp. Further reading De Groot RS, Alkemade R, Braat L, Hein L, Willemen L (2010) Challenges in integrat- ing the concept of ecosystem services and values in landscape planning, management and decision making. Ecological Com- plexity 7: 260-272. EEA (2016) Mapping and assessing the condi- tion of Europe’s ecosystems: progress and challenges. EEA report 03/2016 http:// www.eea.europa.eu/publications/map- ping-europes-ecosystems accessed 31 May 2016. Harrison PA, Berry PM, Simpson G, Haslett JR, Blicharska M, Bucur M, Dunford R, Egoh B, Garcia-Llorente M, Geam N, Geertsema W, Lommelen E, Meiresonne L, Turkelboom F (2014) Linkages between biodiversity attributes and ecosystem ser- vices: A systematic review. Ecosystem Ser- vices 9: 191-203. Pressures Indicators for ecosystem condition assessment Habitat change Land cover change, land take / sealing, fragmentation, land abandonment, river regulation, dams Climate change Changes in temperature, precipitation, humidity, seasonality, extreme events, fires, droughts, frost, floods, storms, average river flows, sea (surface) temperature, sea level rise Invasive alien species Introduction or expansion of invasive alien species, diseases Land/sea use or exploitation Intensification, irrigation, degradation / desertification, erosion, (over-) harvesting, deforestation, water extraction, (over-) fishing, aquaculture, mining Pollution and nutrient enrichment Fertiliser and pesticides application, air pollution, acid and nitrogen deposition, soil contamination, water quality Table 1. Pressures and indicators for ecosystem condition assessment. Chapter 3 81 3.6. Landscape metrics Susanne Frank & Ulrich Walz Introduction Landscape metrics have been used to derive indicators in landscape ecology and related disciplines for decades. More than one hun- dred metrics have been developed for the purpose of describing processes and land- scape functions in the form of mathemati- cal terms. After a very enthusiastic time, the focus at present is on meaningful, simpler measures that can be applied in practice. Meanwhile, landscape metrics play a crucial role not only in science, but also in practical issues, such as spatial planning or biodiver- sity monitoring. Most frequently applied metrics are used to discover biodiversity or landscape fragmentation. Although great advances have already been made, new met- rics continue to be developed. Regularly used metrics are further tested and updat- ed regarding their interpretation. This sub- ject is not without controversy. A question which is often raised in research circles is: Which role can landscape metrics play within the set of indicators for ES map- ping and assessment? The following sections address this question. Regarding the ES cascade model (Chapter 2.3) and following on from it, landscape structures support biodiversity and ecosys- tem functions that are the basis for the fi- nal provision of ecosystem services (ES) to humans. The crucial question is whether landscape metrics applied to land use / land cover maps can provide direct or indirect in- dications on the provision of ES. So far, landscape metrics have been applied to indicate cultural ES (e.g. recreation, landscape aesthetics) and regulating ES (e.g. soil erosion, biological pest control). However they are predominantly applied to measure ecological functioning (biodiversi- ty, connectivity, soil quality) and land use processes (land consumption, fragmenta- tion, urban sprawl). Within this chapter, we review the know- ledge of pattern-related challenges in ES mapping, using the examples of habitat con- nectivity and scenic attraction. We contrib- ute to a better understanding of the reasons for challenges in mapping structure-depen- dent ES and we demonstrate some methods for addressing them. Landscape metrics as method for ES mapping? Landscape metrics are tools which can be used to bridge the methodological gap be- tween landscape structure and ES provision. They take the visible spatial manifestation of land use patterns into account. Compo- sition and configuration of patches (homo- geneous units of one property, e.g. land use type) are key features of maps. Hence, land- scape metrics and mapping are inherently interrelated. Table 1 provides an overview of selected landscape metrics which are appli- cable for mapping and assessment of ES. Landscape metrics quantify physical land- scape structures which themselves deter- mine processes and functions. Although some landscape structures can be measured and related to the provision of specific ES, Mapping Ecosystem Services82 Structure/landscape metric Process/function Mapping target Dimension of Biodiversity Shannon’s diversity index, Patch density Pattern heterogeneity and variety Landscape diversity Shape index Natural conditions Species diversity Proximity index, Nearest neighbour index Isolation, Habitat connectivity Species diversity Effective mesh size Fragmentation Species diversity Provisioning service Total patch area (of arable land) Food and fodder production Food and fodder Total patch area (of forested/arable land) Biomass production Biomass Total patch area of lakes Food (fish) Regulating service No. / length of landscape elements (hedges, tree lines) Soil erosion due to water runoff Mass flow Edge length (of hedges, forests and other ecotones) Habitat provision forpollinators (fringe structures) Pollination Shannon’s diversity index / Heterogeneity of agricultural areas Population development Pest control Cultural service Total patch area (of water), Edge length of waters Attraction, Complexity Landscape aesthetics Shape index Hemeroby index Complexity and Natural conditions Landscape aesthetics No. of landscape elements Legibility, mystery Landscape aesthetics Table 1. Examples for suitable landscape metrics indicating biodiversity and ES (provisioning, regulating, cultural; following CICES (2013)), without claim to completeness. Chapter 3 83 direct functional interpretation of single metrics regarding ES remains limited. For the assessment and interpretation of land- scape metrics, for example, a normative as- sessment basis is required which relates the current situation of landscape structure to a reference or target situation of ES provision. However, landscape structure is important information in a more complex evaluation of ES. Landscape metrics have therefore to be considered as meaningful parameters to- gether with others in ES mapping and eval- uation. A sound application of landscape metrics is possible considering two dimen- sions of biodiversity, species and landscape diversity. Many species and species commu- nities rely on specific landscape structures or landscape elements and their interrelations. Provisioning services strongly depend on the extent of managed land and the land use intensity. However, for quantification of productivity or food provision, further information, for example on soil quality or soil management, is essential to derive a valid estimation on food provision. Re- garding regulating services, landscape met- rics also comprise the potential to provide supplementary information. Indices like edge length or the number of landscape ele- ments can quantify some preconditions for functions and services. Although modelling and/or measurement of species abundance or mass flows (qualitative data; Chapter 4.1) cannot be replaced by structural indicators, they have to be considered as one important part of the required information. Strong interrelations between indices of biodiversity and landscape aesthetics can be identified as potential of landscape met- rics application in mapping ES. Cultural services such as the potential of landscapes for human recreation are interrelated with structural aspects. However, landscape aes- thetics is just one of many spiritual, expe- riential and educational services. Landscape metrics can therefore serve as a complemen- tary mapping and assessment method for cultural ES. Application of landscape metrics for ES mapping Landscape metrics have been applied in several ES mapping and assessment stud- ies. Two examples illustrate how they can be related to ecological integrity, consid- ered as the basis for any ES provision (Box 1) and scenic attraction, as an example for cultural ES (Box 2). In a similar manner, spatial structures strongly determine reg- ulating ES. The regulation of soil erosion, for example, can be estimated using the number and the spatial arrangement of landscape elements, such as hedgerows. These elements reduce slope lengths which is one driving factor for soil erosion. How- ever, suitable landscape metrics, such as the number of patches or edge length, have not been frequently used for assessment and mapping of regulating ES. Furthermore, small–scale landscape ele- ments such as ecotones at forest borders, single-trees, hedgerows including field mar- gins are important for the regulating of ES pollination. Mapping Ecosystem Services84 Box 1 . Example for the application of landscape metrics at the regional scale: evaluation of the landscape structures’ impact on biodiversity One approach, how landscape metrics can contribute to ES mapping and assessment, is illustrated in Figure 1. Six metrics were applied to an administrative planning region (3,434 km²) in middle Saxony, Germany to evaluate ecological integrity as the precondition for biodiversity and the cultural service land- scape aesthetics. They were implemented into the land use change simulation software GISCAME as sup- plementary indicators. In this software, the basic evaluation of ES is based on land use types. An additional landscape structure add-on makes the impact of composition and configuration visible and assessable. The example focuses on habitat connectivity which is a landscape function related to biodiversity (see Table 1). Using the “moving window” method, which is independent from any administrative or geo- graphical zoning, combined with a cost-distance analysis, local landscape pattern were examined across space and interpreted. The size of the moving window is determined by the action radius of a target species. On the basis of the degree of hemeroby of land use types, near-to-nature areas were identified, as well as core habitat areas and functionally connected areas. The latter were defined as potential habi- tat areas which are too small and not compact enough to provide habitat core areas. Nevertheless, they are close enough to another core area to be appropriate habitats for species moving through a landscape. Such functionally connected areas were also considered as part of the habitat network. (Semi-)natural areas were considered as isolated and therefore not contributing to the habitat network if they were separated by roads, urban areas and similar land use types acting as barriers. The map in has been classified according to the functional interpretation of a land use map. It can serve scientists as well as spatial planners to identify i) the share of land which contributes to a habitat network and ii) its spatial distribution. This information allows spatially explicit conclusions on pri- ority areas for enhancement of the connectivity and on the overall state of habitat connectivity as one influencing factor of biodiversity. 20 km (Semi-)natural area without connection Functionally connected (semi-)natural area (Semi-)natural area: core area Wood production Food and Fodder Ecological integrity Drought risk regulation Soil erosion protect C- sequestrationRecreation Connected (semi-)natural area 31.9% Effective Mesh Size of unfragmanted areas: 4.29/km2 Shannon’s Diversity Index: 2.66 Core Area Index of (semi-)natural areas: 19:94 Shape Index of (semi-)natural areas: 1.50 Patch Dencity: 0.32/km2 1 Further information: www.giscame.com Chapter 3 85 Box 2 . Example for the application of landscape metrics at the national scale: application of ES to estimate the cultural ES scenic attraction Based on the natural amenities and features, a model for assessing the scenic attraction of landscapes is presented. It is a suitability analysis of an area for nature-based recreation, assuming that certain features of the landscape have a positive or negative impact on the attraction of the landscape and recreation. In this model, landscape metrics are used for several parameters. The relief diversity, the proportion of open space, the hemeroby Index, the density of forest-dominated ecotones, the density of water edges (without coasts), the coastlines and the proportion of unfragmented open space greater 50 km² were selected. The relief diversity (ratio 3D / 2D) reflects not only the maximum height difference (relief energy), but also the cumulative differences in altitude. A low proportion of open space indicates urban or densely built-up areas which can decrease the natural attraction of the landscape by the strong influence of tech- nical artefacts. In congruence with the hemeroby index, the natural condition is an important factor for the attraction of landscapes. With the density of ecotones dominated by trees and shrubs and the densi- ty of water edges, landscape diversity and structure are taken into account. This parameter characterises mainly the variety and edge effects. Since the coasts play a very important role in terms of attraction and recreation, they are represented by their own parameters - coastlines. Finally, the disturbing effect of fragmentation by the transport network is considered with the parameter ‘proportion of unfragmented open spaces greater than 50 km²’. All data used were based on the official land use data of the state and federal German survey authorities (ATKIS Basis DLM or land cover model LBM-DE ) in vector format collected in 2010. The indicator of the scenic attraction was calculated based on a 5-km grid (standardised according to EU INSPIRE directive). To determine the five classes of scenic attraction, the standard deviation from the nationwide average was used. The reason behind this approach is mainly to use no fixed scale, but starting from the average val- ues of the scenic attraction, to be able to make state- ments as to whether an area is rather less, or rather more scenically attractive. Landscapes which are significantly affected by anthropogenic impacts and thus often are particularly fragmented, intensively farmed or settled, can be found in the class “less at- tractive”. Average attractive landscapes already meet recreational functions in a regional context, while very or particularly attractive landscapes represent targets for nature-related tourism and are mostly well-known nationwide. Monitoring the development of scenic attraction using this aggregated indicator would provide deci- sion-makers with indications as to where the scenic attraction is particularly reduced or has improved. The information derived from the aggregated, landscape metrics-based indicator reveals that individual changes affect the landscape values in their sum. Furthermore, spatial information on the scenic attraction can be used to avoid encroachments in scenic highly attractive areas and thus to achieve better management. Mapping Ecosystem Services86 Conclusions Landscape metrics and ES mapping are inherently related topics since landscape metrics quantify spatial characteristics of landscape patterns. Therefore, we recom- mend the application of landscape metrics in the context of ES mapping and also ES assessment. These indices have the power to support the identification and monitoring of spatial characteristics of landscapes which have implications on the perfor mance of biodiversity and several ES. Some dimensions of biodiversity and cultur- al ES can be comprehensively indicated by landscape metrics. The validity and verifiability of landscape metrics, however, is limited. They quantify and illustrate processes and/or functions, which can serve as surrogates for specific ES. Due to such indirect links to ES, landscape metrics should only be used as supplementa- ry indicators for ES assessments. In the case of landscape aesthetics and recreation, they can be more directly linked to the ES pro- vision. However, landscape metrics describe structural aspects of ES (which are import- ant and should not be forgotten), but usually additional information (e.g. data on quality of land use) is necessary. Still, they have a great significance in terms of mapping. The spatial interpretation of land use maps with the help of landscape metrics serves as a valuable method for communicating ES-related issues. With regard to the current application for ES mapping and assessment in science and practice, we foresee a large capacity for future application of landscape metrics, especially in practice. The benefit of using landscape metrics for ES mapping is currently below its estimated potential. Further reading Burkhard B, Kandziora M, Hou Y, Müller F (2014) Ecosystem Service Potentials, Flows and Demands - Concepts for Spa- tial Localisation, Indication and Quantifi- cation. Landscape Online 34: 1-32. Botequilha Leitao A, Ahern J, McGarigal K (2006) Measuring Landscapes. A Planner’s Handbook. Island Press; 2nd edition. Frank S, Fürst C, Koschke L, Makeschin F (2012) A contribution towards a transfer of the ecosystem service concept to land- scape planning using landscape metrics. Ecological Indicators 21: 30-38. Haines-Young R, Potschin M (2013) Com- mon International Classification of Eco- systemServices (CICES), version 4.3. Report to the European Environment Agency EEA/BSS/07/007 (download: www.cices.eu). McGarigal (2015) FRAGSTATS HELP. V. 4.2, 21 April 2015, available online: http://www.umass.edu/landeco/research/ fragstats/documents/fragstats.help.4.2.pdf Walz U, Stein C (2014) Indicators of hem- eroby for the monitoring of landscapes in Germany. Journal for Nature Conserva- tion 22 (3): 279-289. Walz U (2015) Indicators to monitor the structural diversity of landscapes. Ecolog- ical Modelling 295: 88-106. Chapter 3 87 3.7. Specific challenges of mapping ecosystem services Joachim Maes Ecosystems are spatially explicit and so too are their conditions and their capacity to provide ecosystem services (ES). The differ- ent biomes and ecosystems that cover the earth’s surface deliver various ES bundles at different quantities and qualities. These ser- vices are often consumed or used at other places. Mapping ES thus makes good sense, in particular to quantify and sum stocks and flows (Chapter 5.1) of services at different spatial scales (Chapter 5.7). Furthermore, maps are very powerful tools for communicating and organising data. It is little wonder that geography is a major subject at school. Most people are familiar with maps to navigate or to find places for holidays or recreation. Maps are used to present data and compare the performance of countries and regions across the world for virtually all possible indicators. Many of us have still paper maps in our cars or dig- ital maps on our cell phones, as well as the popular Google Maps which are an essential tool and benefit to our lives. It follows that there is a strong basis in our society for maps and mapping and thus for mapping ecosystem services as well. In par- ticular, there is a demand from policy-mak- ers to map ES (see Chapter 7.1) and to build natural capital accounts which should be based on the reliable geo-referenced data of ecosystems. Despite the popularity of maps, they are pitfalls as well. Some claim that “maps have an air of authority”. Which means that maps and their content are often taken for granted. Yet, ES mapping is challenging for a number of reasons. These are listed here while referring to the next chapters which present and discuss solutions for addressing these challenges. An often heard challenge is that not all ES can be mapped. Review articles typically found that regulating and provisioning ES are most frequently mapped but cultural ES less so. As for regulating ES, most efforts have gone to mapping climate regulation while for provisioning ES, the focus is on food, water and timber. Evidently, these mapping studies have largely profited from knowledge stemming from environmental sciences and agricultural and forestry research. Howev- er, substantial progress in mapping ES has been made in the recent decade (see chap- ters 5.5.1, 5.5.2 and 5.5.3) and solutions have been found to map services which were previously thought impossible to map (see chapter 6.2). Particular advancements have been realised to map certain cultural ES or to map regulating ES which involve service providing areas (Chapter 5.2) that operate at very small spatial scales (such as pollina- tion or biological control). A specific challenge is related to the trans-dis- ciplinary nature of ecosystem services. ES research has become a major academic field, drawing on various academic disci- plines, perspectives and research approach- es. The multifaceted ES concept includes, in addition, a normative component. This exposes ES maps (and the researchers who created them) to the general critique of not being sufficiently inclusive and to the spe- Mapping Ecosystem Services88 cific critique from disciplinary specialists of oversimplifying detailed ecological pro- cesses that are underpinning ES. To both challenges the ES mapping community has responded well. Chapter 5.6 demonstrates how different views expressed by different stakeholders and researchers can be accom- modated in the ES framework. Mapping ES nowadays is not restricted to natural sciences but includes social and economic sciences as well. Furthermore, recent studies promote the adoption of a tiered mapping approach which allows increasing levels of spatial and ecological details to be incorporated in map- ping studies (chapter 5.6.1). Besides these thematic challenges, there are significant technical challenges to map ES. A question which often arises relates to what ES maps should express: ES potentials, flows or demand (Chapter 5.1)? ES are realised when humans benefit from them. At this point, supply meets demand and ES “flow” from where they are generated to where they are received (Chapter 5.2). These flows are dynamic over time and therefore difficult to capture on maps; stocks exhibit less dynam- ics and are therefore easier to map. A typical example is climate regulation; this service is often mapped by the carbon stock in soil or above-ground vegetation assuming that the stock is related to the capacity to provide a flow of service. Carbon capture as such is less mapped. The notion of stocks and flows is crucial for accounting purposes. The size of the stock is not necessarily related to the magnitude of ES flows, so this challenge needs to be addressed when ES maps are ap- plied in decision-making contexts. The selection of an appropriate spatial scale and an appropriate mapping unit is another important issue and remains a challenge for ES mapping studies (Chapter 5.7). Ecolog- ical processes occur at different spatial and temporal scales. Pollination by insects is, for example, a very local ES which takes place in a specific period of the year when tem- perature allows bees and other pollinators to be active. Groundwater recharge, in con- trast, is a large-scale process which usually is measured in decades. ES related to water, climate and atmosphere demonstrate entire- ly different behaviour from services related to soil. They require different quantification approaches and are measured for different spatial units. This results in maps which vary across scale and spatial unit. Bringing them together in a series of consistent and har- monised ES maps for spatial planning and policy support requires application of spatial operations (such as upscaling, downscaling, spatial statistics) which, in turn, may intro- duce uncertainties (Chapter 6). Using scal- able indicators (e.g. indicators which can be measured at different spatial scales such as the density of trees) could overcome errors that arise when local data are upscaled or when global data are downscaled. But such indicators are not always available. In partic- ular for water, air and soil, related ES mea- surements are mostly local and not scalable to larger spatial scales. ES mapping could thus be substantially ad- vanced by a more systematic development of cross-case comparisons and methods. Several chapters of this book touch on these challenges related to spatial scale and pro- vide solutions for dealing with uncertain- ties arising from spatial data handling (dif- ferent sections under 5.7). As more efforts and research are focused on these areas, it seems likely that datasets generated at dif- ferent spatial and temporal scales and, us- ing different types of data, will complement one another to provide a coherent message regarding the health of global ecosystems, biodiversity and the benefits they confer upon society. The different thematic and methodological challenges are sources of uncertainties that Chapter 3 89 should be considered when using ES maps. ES map-makers should try to detect sourc- es of uncertainty and give guidance on how to deal with them (Chapter 6.3). Of equal importance is transparency. The map-mak- er should be clear about how the maps are generated. A helpful tool is provided by the Blueprint for mapping and modelling ES (see further reading and Chapter 7.9). The prima- ry purpose of this blueprint is to provide a template and checklist of information need- ed for those carrying out an ES modelling and mapping study. A second purpose is to reduce uncertainties associated with quanti- fying and mapping of ES and thereby help to close the gap between theory and practice. Further reading Abson DJ, von Wehrden H, Baumgärtner S, Fischer J, Hanspach J, Härdtle W, Hein- richs H, Klein AM, Lang DJ, Martens P, Walmsley D (2014) Ecosystem services as a boundary object for sustainability. Eco- logical Economics 103: 29-37. Crossman ND, Burkhard B Nedkov S, Wil- lemen L, Petz K, Palomo I, Drakou EG, Martín-Lopez B, McPhearson T, Boyano- va K, Alkemade R, Egoh B, Dunbar MB, Maes J (2013) A blueprint for mapping and modelling ecosystem services. Ecosys- tem Services 4: 4-14. Dick J, Maes J, Smith RI, Paracchini ML, Zulian G (2014) Cross-scale analysis of ecosystem services identified and assessed at local and European level. Ecological In- dicators 38: 20-30. Hauck J, Görg C, Varjopuro R, Ratamäki O, Maes J, Wittmer H, Jax K (2013) Maps have an air of authority: Potential benefits and challenges of ecosystem service maps at different levels of decision making. Eco- system Services 4: 25-32. Martnez-Harms MJ, Balvanera P (2012) Methods for mapping ecosystem service supply: A review. International Journal of Biodiversity Science, Ecosystems Services and Management 8: 17-25. Chapter 4 91 Chapter 4 Ecosystem services quantification Mapping Ecosystem Services92 The economic value of the best place on earth has been quantified (Photo: Benjamin Burkhard 2008). Chapter 4 93 4.1. Biophysical quantification Petteri Vihervaara, Laura Mononen, Fernando Santos, Mihai Adamescu, Constantin Cazacu, Sandra Luque, Davide Geneletti & Joachim Maes Introduction Ecosystem services (ES) arise when eco- logical structures and ecological processes directly or indirectly contribute to human well-being and meet a certain demand from people. This flow of ES from ecosystems to society is well represented by the ES cascade concept (see Chapter 2.3). Ecosystems pro- vide the necessary structure and processes that underpin ecosystem functions which are defined as the capacity or potential to deliver services. ES are derived from eco- system functions and represent the realised flow of services in relation to the benefits and values of people. This model is useful for quantifying ES. Consider the follow- ing example: wetlands (an ecosystem or a structure) provide habitat for bacteria which break down excess nitrogen (denitrification, a process). This results in the removal of ni- trogen from the water (a service) resulting in better water quality (a benefit). People can value increased water quality in multiple ways (e.g., by expressing their willingness to pay for clean water). Each of these different steps can be quantified using biophysical, economic or social valuation methods. This chapter focuses on biophysical quanti- fication which is the measurement of ES in biophysical units. Biophysical units are used to express, for example, quantities of wa- ter abstracted from a lake, area of forest or stocks of carbon in the soil. Looking at the ES cascade, it seems evident that biophysi- cal quantification focuses, in particular, on the measurement of ecosystem structures, processes, functions and service flows (also known as the left side or the supply side of the cascade). Benefits and values (also known as the right side or demand side of the cascade) are more often measured using social (see Chapter 4.2) or economic units (see Chapter 4.3). Nonetheless, benefits and values can sometimes be expressed in bio- physical units as well. Consider again the above example of water purification in wet- lands. The benefit from this ecosystem ser- vice is clean water and this can be expressed as the concentration of pollutant substances. To quantify ES along the different compo- nents of the ES cascade, we need to address two questions: what do we measure and how do we measure (Figure 1)? For the purpose of this chapter, we assume that the question as to why we measure (e.g., policy questions, scope of an ecosystem assessment) has been answered. The first question is addressed in the scien- tific literature by developing and proposing indicators. Ecosystem service indicators are used to monitor the state or trends of ecosys- tems and ecosystem service delivery within a determined time interval. In recent years a substantial indicator base has been devel- oped world wide to assess or measure ES. Once an indicator is proposed or selected for inclusion in an ecosystem assessment, the second question becomes important: how can we measure the service or the indi- Mapping Ecosystem Services94 cator in biophysical terms or units? Which methods or procedures should be applied to come to an reasonable estimate of the quan- tity of service provided? What to measure: Ecosystem service indicators ES indicators are information that efficiently communicates the characteristics and trends of ES, making it possible for policy-makers to understand the condition, trends and rate of change in ES. Different indicators can be used to measure or indicate a single ecosystem service. The choice for an indicator depends on many factors in- cluding the purpose, the audience, its position on the ES cascade, the spatial and temporal scale considered and the availability of data. Purpose and target audience are important criteria for selecting or designing indicators for ES. It makes a difference if indicators are used to inform policy makers, journalists, conservation and land managers, scientists or students. Not everybody has an equal understanding of the flow of ES which is indeed a relatively complex concept. There- fore, indicators are sometimes expressed in relative terms by setting a reference value equal to, for instance, 100 and by calculat- ing other values relative to this reference. This facilitates interpretation for some user groups. Of equal importance is the purpose of an indicator. Why is it used? Many ES indicators are proposed to report the state and trends of ES under different biodiver- sity policies from global to local scale. But such indicators are not necessarily useful for application by spatial planners or for sci- entific support to river basin management. Consider pollination, a regulating ecosys- Figure 1. Biophysical quantification of ecosystem services (Icons by Freepik). Direct measurement Field observations Field experiments Surveys and questionaires Indirect measurement Remote sensing and earth observation (NDVI, land cover, surface tempera- ture, …) Socio-economic data Proxy indicators Purpose of the assessment Target audience Position on the ES cascade Spatial and temporal scale Availability of data Indirect measurement Expert based, statistical and process based models of ecosystems and ecosys- tem services What to measure? 1 2 3 How to measure? Select an appropriate indicator Select an appropriate method Biophysical quantification of ecosystem services Chapter 4 95 tem service. A scientist could be interested in the diversity and density of different bee and bumblebee populations; a farmer may wish to know how far he can rely on wild pollination to help pollinate his fruit trees; a biodiversity policy officer may need to know if, at national scale, pollination services are declining or increasing. Clearly, these stake- holders have different information requests which require different indicators with dif- ferent biophysical units although pollina- tion is the common denominator. The above example also illustrates the im- portance of spatial and temporal scales. The issue of scale is frequently presented in all textbooks on ecology as biodiversity and the ecological processes it supports (and thus also the delivery of ES) are heavily dependent on time and space. Processes are influenced by different time cycles (day-night, seasons) and take place at different rates (see also Chapter 5.3). The self-purifying capacity of water is, for instance, highly dependent on the veloc- ity at which water flows. Water purification services, for example, which can be measured by the amount of pollutant removed, differ between fast running streams and stagnant lakes with the latter ecosystems having, in general, a higher capacity (more time) to re- move nitrogen but a lower capacity to clean organic pollution. Also spatial scale matters. Bees and bumblebees deliver their polli- nation services within a distance of a few hundred metres whereas the storage of car- bon in trees operates at almost global scale. Indicators and, in particular, their units of measurement have to consider the scale at which ES are relevant. Sometimes indicators are designed to be scale independent. This means they can be upscaled or downscaled, a very useful technique for mapping. An important question often raised in litera- ture on ES is: should indicators measure the stock and the flow? A service flow refers to the actual use of the actual benefits people receive from ecosystems. A stock refers to the capacity of ecosystems to deliver those benefits. Flows are always expressed per unit of time. Timber production serves as a good example to illustrate the difference between an indicator which measures the stock and an indicator which measures the flow. Tim- ber production is often measured by quan- tifying the harvest (how much timber is cut, usually expressed in a volume of wood per unit area and per unit of time, for example, m3/ha/year). Sometimes timber production can also be indicated by the available timber stock which can be harvested. This difference is subtle for the case of timber. If the stock is harvested, stock becomes flow. However, for other services, the difference between stock and flow is important because indicators for stock and flow cannot always be expressed in the same units. Wetlands have a certain capacity to clean water but it is not always straightforward to express this capacity in terms of pollutant removal (e.g., amount of nitrogen removed or immobilised in the sediment in kg/ha/year). Often the size of the wetland (in ha) is used as proxy to indi- cate this capacity. The rationale is that larger wetlands have more capacity to purify water than smaller wetlands. In this context, the concept of ecosystem condition is import- ant as well (see Chapter 3.5). Not only the quantity (spatial extent) of an ecosystem is important to assess the physical values of ES capacity, ecosystem quality or ecosystem condition is also an important determinant of ecosystem delivery. Changes in ecosys- tems through degradation can thus alter the flows of ES and should thus be measured as well by indicators. A final remark on indicators relates to com- posite indicators or indices which aggregate different sorts of information into a single number. Usually such indicators are made for specific purposes or to inform on partic- ular challenges with a single value. In a sim- ilar context for ES, such indicators exist but Mapping Ecosystem Services96 usually they are composed of normalised ver- sions of indicators for single services which are summed or aggregated. They cannot be quan- tified directly but depend on separate quanti- fication of their individual components. This chapter does not provide a list with indicators for ES for the simple reason that there are hundreds of indicators available. Many countries and regions have developed ES indicator sets; the setting of global or re- gional biodiversity targets has also spurred the development of indicators. Further- more, the application of the ES concept for planning, natural resources management and conservation has created additional in- dicators. Therefore we list in Table 1 some important initiatives where readers can find a selection of indicators, organised from global to sectorial initiatives. In summary, ES indicators express what to measure when quantifying ES in a biophys- ical manner. Good ES indicators come with information on their place on the ES cas- cade, on the available data, on the targeted audience and the objective and on whether they assess a stock or a flow. Scale Location Publication Global Measuring Nature’s Benefits: A Preliminary Roadmap for Improving Ecosystem Service Indicators (http://pdf.wri.org/measuring_natures_ benefits.pdf ) http://www.bipindicators.net/ (report ISBN 92-9225-376-X) Measuring ecosystem services: Guidance on developing ecosystem service indicators (ISBN: 978-92-807-4919-5) http://es-partnership.org/community/workings-groups/thematic-work- ing-groups/twg-3-es-indicators/ A Global System for Monitoring Ecosystem Service Change (doi: 10.1525/bio.2012.62.11.7) Sub-global European Union website: http://biodiversity.europa.eu/maes/mapping-ecosystems article: doi:10.1016/j.ecoser.2015.10.023 National Finland website: http://www.biodiversity.fi/ecosystemservices/home article: doi:10.1016/j.ecolind.2015.03.041 Canada Website: https://www.ec.gc.ca/indicateurs-indicators/ Switzerland Website: http://www.bafu.admin.ch/publikationen/publikation/01587/ index.html?lang=en Germany article: Towards a national set of ecosystem service indicators: Insights from Germany (doi:10.1016/j.ecolind.2015.08.050) Spain Website: http://www.ecomilenio.es/informe-de-resultados-eme/1760 Article: doi:10.1371/journal.pone.0073249 Table 1. Examples of sources, websites and key publications for ecosystem service indicators. Chapter 4 97 How to measure? Indicators must be measured but how is this done for ES? Some of the above given examples already provide the answer. The number of bees on a farmland, the timber harvest from a forest or the denitrification in a wetland can all be monitored or mea- sured with different methods or devices. Yet measuring stocks or flows of ES is less ev- ident than it seems. Here we present three approaches which can be considered to quantify biophysical stocks and flows of ES: direct measurements, indirect measurement and (numerical) modelling. Direct measurements of ecosystem services Direct measurements of an ecosystem ser- vice indicator is the actual measurement of a state, a quantity or a process from observa- tions, monitoring, surveys or questionnaires which cover the entire study area in a repre- sentative manner. Direct measurements of ES deliver a biophysical value of ES in physical units which correspond to the units of the indicator. Direct measurements quantify or measure a stock or a flow value. Direct mea- surements are also referred to as primary data. Examples of direct measurements of ES (see also Table 2) are counting the number of visitors visiting a national park (nature based recreation); measuring the total vol- ume of timber in a forest stand (timber pro- duction); monitoring the release of nitrous oxides of a reed bed or deposition of sulphur dioxide on leaves (water and air filtration); recording the crop yield of a farm (crops); measuring the volumetric capacity of a flood plain (flood control); monitoring over time the improvement of water quality (water purification); measuring the abstraction of water from ground water layers (water pro- vision) or asking citizens how many times they visit a forest to pick berries, mushrooms or chestnuts (wild food products). When the spatial extent or relative surface area of ecosystems is used to approximate ES, also botanical and forest inventories, permanent plots or any other direct observation on the terrain can be used as proxy. In certain cases remote sensing can be considered also as di- rect measurement. These examples of direct measurement share a number of characteristics. They are time and resource consuming and thus costly, mostly suitable for carrying out at site level or local scale and they measure tangible flows of ES, in particular for provisioning ES. Direct measurements are also feasible in case of a clearly defined service providing species (or areas) such as pollination, bird watching or biological control. As many of these indicators are effectively measured for other reasons, it is not always needed to set up expensive measurement schemes. Most provisioning ES including crops, fish, timber and water are recorded by national and regional governments. Fur- thermore, certain species groups and taxa are monitored to assess trends in biodiversity. TESSA1 is a toolkit for rapid assessment of ES at site level which provides many proce- dures and suggestions for on-site measure- ment of ES. Direct measurements and the use of primary data are the most accurate way to quantify ES but they become impractical and expen- sive beyond the site level or they are simply not available for all ES. Therefore the next step to consider for bio- physical quantification is indirect measure- ments. 1 http://tessa.tools/ Mapping Ecosystem Services98 Ecosystem services What to measure How to measure (method) (CICES class) Indicator Direct Indirect Model Cultivated crops Crop yield (tonne/ha/year) Crop statistics (obtained through official reporting) Remote sensing of crop biomass using NDVI and aerial photo analysis for long temporal changes Coupling structural observations with remote sensing information Crop production models Reared animals and their outputs Livestock (heads/ ha) Livestock statistics (head counts obtained by reporting) Wild plants, algae and their outputs Wild berry yield (tonne/ha/year) Field observations and surveys of people harvesting wild fruits Species distribution models; ecological production model Animals from in- situ aquaculture Fish yield (tonne/ ha/year) Aquaculture statistics (obtained through official reporting) Fish production models Water (Nutrition) Water abstracted (m3/year) Water statistics (obtained through official reporting) Remote sensing of water bodies and soil moisture Water balance models Biomass (Materials) Timber growing stock (m3/ha) and timber harvest (m3/ha/year) Forest stand measurements and forest statistics Remote sensing of forest biomass using NDVI Timber production models (Mediation of waste, toxics and other nuisances) Area occupied by riparian forests (ha) Site observations Earth observation land cover data Nitrogen and Sulphur removal in the atmosphere or in water bodies (kg/ha/year) Measurement of deposition of NO2 and SO2; field measurement of denitrification in water bodies Remote sensing of canopy structure (leaf area index) Transport and fate models for N and S Mass stabilisation and control of erosion rates Soil erosion risk (tonne/ha/year) Field measurements of soil erosion Soil erosion models (RUSLE) Flood protection Area of floodplain and wetlands (ha) Site observations Elevation models and data; aerial photo analysis; remote sensing of land cover Modelling water transport Table 2. Examples of different methods to measure ecosystem service indicators Chapter 4 99 Indirect measurements of ES Indirect measurements of ES deliver a bio- physical value in physical units but this value needs further interpretation, certain assump- tions or data processing, or it needs to be combined in a model with other sources of environmental information before it can be used to measure an ecosystem service. Indi- rect measurements of ES deliver a biophysical value of ES in physical units which are differ- ent from the units of the selected indicator. In many cases, variables that are collected through remote sensing qualify as indirect measurement. Examples for terrestrial eco- systems are land surface temperature, NDVI (Normalised Difference Vegetation Index), land cover, water layers, leaf area index and primary production. Examples for marine ecosystems include sea surface temperature, chlorophyll A concentration and suspended solids. Many of these data products do not measure stocks or flows of ES but they are highly useful to quantify global climate reg- ulation as well as all those ES which depend directly on the vegetation biomass of ecosys- tems to regulate or mediate the environment. Soil protection and water regulation, for ex- ample, are strongly driven by the presence of vegetation which can be inferred from earth observation datasets. Local climate regulation can be inferred from spatially and temporally explicit patterns of surface temperature. Air filtration by trees and forest is directly related to the canopy structure which, in turn, can be measured by the leaf area index. In addition, micro-climate regulation in cities (tempera- ture reduction during heat waves through evapotranspiration and provision of shade) can be approximated by measuring the total surface area of urban forest. A specific role is reserved for land cover and land use data which are used for both direct and indirect quantification of ES. Detailed and accurate information on the extent of Ecosystem services What to measure How to measure (method) Pollination and seed dispersal Pollination potential; number and abundance of pollinator species (number/m2) Field sampling of pollinator species; counts of bee hives Species distribution models; ecological modelling of habitat suitability Decomposition and fixing processes Area of nitrogen fixing crops (ha) Field surveys; crop statistics (obtained through official reporting) Crop production models Global climate regulation by reduction of greenhouse gas concentrations Carbon storage (in soil or aboveground biomass) (tonne/ ha); carbon sequestration (tonne/ha/year) On-site measurements of carbon stock and carbon fluxes Remote sensing of vegetation Carbon cycle models Physical and experiential interactions Visitor statistics (number/year) Visitor data and questionnaires of visitors Monitoring parking lots, mapping trails or camping sites Modelling potential use of nature reserves by people Ecosystem services What to measure How to measure (method) (CICES class) Indicator Direct Indirect Model Cultivated crops Crop yield (tonne/ha/year) Crop statistics (obtained through official reporting) Remote sensing of crop biomass using NDVI and aerial photo analysis for long temporal changes Coupling structural observations with remote sensing information Crop production models Reared animals and their outputs Livestock (heads/ ha) Livestock statistics (head counts obtained by reporting) Wild plants, algae and their outputs Wild berry yield (tonne/ha/year) Field observations and surveys of people harvesting wild fruits Species distribution models; ecological production model Animals from in- situ aquaculture Fish yield (tonne/ ha/year) Aquaculture statistics (obtained through official reporting) Fish production models Water (Nutrition) Water abstracted (m3/year) Water statistics (obtained through official reporting) Remote sensing of water bodies and soil moisture Water balance models Biomass (Materials) Timber growing stock (m3/ha) and timber harvest (m3/ha/year) Forest stand measurements and forest statistics Remote sensing of forest biomass using NDVI Timber production models (Mediation of waste, toxics and other nuisances) Area occupied by riparian forests (ha) Site observations Earth observation land cover data Nitrogen and Sulphur removal in the atmosphere or in water bodies (kg/ha/year) Measurement of deposition of NO2 and SO2; field measurement of denitrification in water bodies Remote sensing of canopy structure (leaf area index) Transport and fate models for N and S Mass stabilisation and control of erosion rates Soil erosion risk (tonne/ha/year) Field measurements of soil erosion Soil erosion models (RUSLE) Flood protection Area of floodplain and wetlands (ha) Site observations Elevation models and data; aerial photo analysis; remote sensing of land cover Modelling water transport Table 2. Examples of different methods to measure ecosystem service indicators Mapping Ecosystem Services100 ecosystems or of ecosystem service provid- ing units, constitute an essential data basis for all ecosystem assessments. Importantly, land data can also be used to quantify de- mand for ES. Not all indirect measurements are provided by earth observation. The density of trails and camping sites may provide an indirect esti- mate of recreation and tourism (Table 2). Indirect measurements, in particular earth observation, offer substantial advantages. They provide consistent sources of infor- mation often with global coverage and they are regularly updated which makes them suitable for natural capital accounting and monitoring trends. Modelling as alternative to quantify ES ES modelling can be used to quantify ES if no direct or indirect measurements are avail- able. This is virtually always the case in any ecosystem assessment. With ES modelling, we understand the simulation of supply, use and demand of ES based on ecological and socio-economic input data or knowledge. Models can vary from simple expert based scoring systems to complex ecological mod- els which simulate the planetary cycles of carbon, nitrogen and water. More details are also available in Chapter 4.4 In the context of biophysical quantification, models can be used for spatial and temporal gap filling of direct and indirect measure- ments, extrapolation of direct and indirect measurements, modelling ES for which there are no measurements available or for scenario analysis. For regulating services, modelling is some- times the only option in order to quantify actual ecosystem service flows. This is partic- ularly evident when ecosystems are regulating or mediating stocks and flows of soil, carbon, nitrogen, water or pollutants. Consider soil protection - also termed as erosion regula- tion or erosion control – which is the role ecosystems and vegetation plays in retaining soil or avoiding soil being eroded as a result of wind or run-off water. Soil erosion can be measured directly on sites which are prone to erosion, usually cropland on slopes. Howev- er, estimating the quantity of soil that is not eroded due to the protective cover of vegeta- tion cannot be measured. It can however be modelled by comparing the amount of soil erosion with a model which simulates the presence of vegetation with a model where the protective vegetation cover is deliberately set to zero or to parameters which correspond to parameters for cropland or bare soil. The difference between these two models results in an estimate of avoided soil erosion and can represent the realised service flow. A similar rationale applies to water purification, air quality regulation or other services which ex- ert control on the fate and transport of abiot- ic and organic material. Implementing biophysical methods for decision-making Ecosystem service assessments have increas- ingly been used to support environmental management policies, mainly based on bio- physical and economic indicators. There- fore ES assessments have to integrate data and information on biophysical ecosystem components, including biodiversity, with socio-economic system components and the societal and policy contexts in which they are embedded. Quantification of ES using biophysical methods have been used for a number of perspectives and for a variety of purposes, Chapter 4 101 including landscape management, natural capital accounting, awareness raising, prior- ity setting of projects or policies and policy instrument design. However, transferring the outcomes of the biophysical assess- ments to policy is not straightforward and some additional work is required to ensure a minimum degree of consistency and avoid over-simplistic conclusions. Different methods are relevant at different policy levels (ranging from international, EU, national, regional and local scales). Existing literature frequently acknowledg- es that, in these cases, the interrelationship between different scales must be taken into consideration, which can pose significant challenges. Broad framings for these meth- ods include the work done globally of the Inter-governmental Platform on Biodiver- sity and Ecosystem Services (IPBES) and the Mapping and Assessment of Ecosystems and their Services (MAES) in the context of the EU Biodiversity Strategy. The initial methodological work on biophysical meth- ods will be the basis for the assessment of the economic value of ES and promote the integration of these values into accounting and reporting systems. Conclusions “You can’t manage what you don’t measure”. This well-known expression is also valid for ES which is, in essence, a concept to guide and support the management of natural resources, ecosystems and socio-ecological systems. ES represent the flows of materi- al, energy and information from ecosystems to society. Accurate measurement of these flows as well as the extent and the condition of ecosystems which support these flows is therefore key to base decisions, to monitor progress to biodiversity targets and to create a sound knowledge base for natural capital. Further reading Boerema A, Rebelo AJ, Bodi MB, Esler KJ, Meire P (2016) Are ecosystem services adequately quantified? Journal of Ap- plied Ecology. DOI: 10.1111/1365- 2664.12696. De Araujo Barbosa CC, Atkinson PM, Dear- ing JA (2015) Remote sensing of ecosys- tem services: a synthetic review. Ecological Indicators 52: 430-443. Kareiva P, Tallis H, Ricketts TH, Daily GC, Polasky S (2011) Natural Capital: Theory and Practice of Mapping Ecosystem Ser- vices. Oxford University Press, Oxford. Mononen L, Auvinen AP, Ahokumpu AL, Rönkä M, Aarras N, Tolvanen H, Kamp- pinen M, Viirret E, Kumpula T, Viher- vaara P (2016) National ecosystem service indicators: Measures of social-ecological sustainability. Ecological Indicators 61: 27-37. Peh KS-H et al. (2013) TESSA: A toolkit for rapid assessment of ecosystem services at sites of biodiversity conservation impor- tance. Ecosystem Services 5: e51-e57. Pettorelli N, Owen HJF, Duncan C (2016) How do we want satellite remote sensing to support biodiversity conservation glob- ally? Methods in Ecology and Evolution 7: 656-665. Mapping Ecosystem Services102 4.2. Socio-cultural valuation approaches Fernando Santos-Martín, Eszter Kelemen, Marina García-Llorente, Sander Jacobs, Elisa Oteros-Rozas, David N. Barton, Ignacio Palomo, Violeta Hevia & Berta Martín-López Introduction Any evaluation of ES requires an integrat- ed analysis, taking into account the supply and demand of ES and their biophysical, socio-cultural and economic value dimen- sions (see Chapters 4.1, 4.2 and 4.3, respec- tively). Recent literature has acknowledged that many of the contributions on ES valu- ation still use the term ‘value’ exclusively in a monetary sense, ignoring the broader con- tributions of ecosystems and biodiversity to society in terms of cultural, therapeutic, ar- tistic, inspirational, educational, spiritual or aesthetic values. To fill this scientific gap, literature on so- cio-cultural valuation approaches has grown in the last ten years, mostly related to cultur- al ES (Figure 1). The recent increase in the number of scientific papers on socio-cultural valuation of ES coincides with the creation of the Intergovernmental Platform of Biodi- versity and Ecosystem Services (IPBES) in 2012. Some of the challenges addressed by IPBES are related with socio-cultural valua- tion of ES, such as the inclusion of different knowledge-systems or the recognition of value pluralism. Despite the increase in the number of publi- cations, socio-cultural valuation approaches have not yet formalised a common meth- odological framework. Designing a meth- odological framework, able to explore ways of representing cognitive, emotional and ethical responses to nature, alongside ways of expressing preferences, needs and the desires of people in relation to ES, is very much needed. In this context, the present chapter aims to contribute to this challenge through the review of socio-cultural valu- ation methods that have been frequently applied in ES literature. Socio-cultural valuation is defined in this chapter as an umbrella term for those meth- ods that aim to analyse human preferences towards ES in non-monetary units. Under this umbrella,  terms such as ‘psycho-cultural valuation’, ‘social valuation’, ‘deliberative val- uation’, ‘qualitative valuation’ and ‘subjective assessment’ represent valuation approaches that aim to uncover individual and collective values and perceptions of ES without relying on market logic and monetary metrics. A comprehensive review There are multiple approaches to uncover so- cio-cultural values of ES depending on data availability and the purpose of the valuation. In this chapter, we will focus on seven meth- ods that are frequently used in literature. Chapter 4 103 Figure 1. Trends in the scientific literature exploring socio-cultural valuation approaches for cultural ES.1 Preference assessment is a direct consulta- tive method that assesses the individual and social importance of ES by analysing moti- vations, perceptions, knowledge and associ- ated values of ES. Data is collected through free-listing exercises, ecosystem service rank- ing, rating, or other selection mechanisms. Techniques for weighting the preferences related to impacts on the ecosystem service of different management alternatives such as multi-criteria analysis are examples of inte- grated preference assessment valuation. In the same manner, but aiming at a more quantitative indicator of socio-cultural val- ues of ES, the time use method creates hy- pothetical scenarios for willingness to give up time (WTT). This method estimates the value of ES by asking people how much time they are willing to dedicate for a change in the quantity or quality of a given ecosystem service. This method is not only a non-mon- etary metric, but also a way of measuring the willingness to actively contribute to na- ture conservation through practical actions. Photo-elicitation surveys seek to uncover the socio-cultural value of ES by translat- ing people’s visual experiences, perceptions and preferences of landscapes into ecosys- tem service values. The use of photo-elic- itation surveys has proven to be a useful technique for eliciting socio-cultural values of ES as it uses a communication channel (i.e. photographs) which is easily under- stood by multiple social actors (for instance see Chapter 7.3.3). Narrative methods differ from the pre- vious three as they are mainly used to col- lect qualitative data. By using narrative methods (e.g. structured, semi-structured and unstructured interviews, focus groups, participant observation, content analysis, voice and video recording of events, artistic expression, etc.), participants can articulate the plural and heterogeneous values of ES through their own stories and direct actions (both verbally and visually). Three other approaches, frequently used in socio-cultural valuation, focus on the inte- gration of knowledge systems, disciplines and diverse data. Participatory mapping of ES (or sometimes referred to as partici- patory geographical information systems or review and standarized PGIS, see Box 1) as- sesses the spatial distribution of ES accord- ing to the perceptions and knowledge of stakeholders via workshops and/or surveys. PGIS facilitates the participation of various stakeholders (e.g. community members, en- vironmental professionals, NGO represen- tatives, decision-makers, etc.) integrating their perceptions, knowledge and values in maps of ES (see Chapter 5.6.2). 1 Note: this illustration is not representing the total number of published papers on cultural services valuation, but the timeline of publications of the most relevant papers which focus on six cultural ES: non-extractive recreation and tourism (e.g. outdoor recreation, ecotourism), (2) extractive recreation and tourism (e.g. sport fishing, recre- ational hunting), (3) local ecological knowledge, (4) scientific knowledge and environmental edu- cation, (5) spiritual interactions with nature and (6) aesthetic experience. Mapping Ecosystem Services104 Scenario planning combines various tools and techniques (e.g. interviews, brainstorm- ing or visioning exercises in workshops, often complemented with modelling) to develop plausible and internally consistent descriptions of alternative futures, where values of ES can be elicited. Assumptions about future events or trends are questioned and uncertainties are made explicit to estab- lish transparent links between changes in ES and human well-being. Deliberative methods comprise various tools and techniques to engage and empower non-scientific participants. These methods (e.g. valuation workshops, citizens’ juries, photo-voice, etc.) invite stakeholders and cit- izens to form their preferences for ES togeth- er through an open dialogue. Deliberative methods can address ethical beliefs, moral commitments and social norms and are often used in combination with other approaches (e.g. mapping or monetary valuation). Scrutiny of specific socio- cultural valuation methods The diversity of socio-cultural methods described above is determined by different methodological requirements (Table 1) and the ability of the different methods to pro- vide different outputs and to uncover dif- ferent types of values (Table 2). Regarding methodological requirements, socio-cul- tural methods can be clustered into three different groups: (1) methods that require multiple observations as they are quantita- tive methods and are usually developed in collaboration with scholars from the same field (i.e. preference assessment, time-use and photo-elicitation), (2) methods based on qualitative data that are usually applied in collaboration with non-academic stake- holders (i.e. narratives), (3) methods that are able to gather qualitative and quantita- tive data by collaborating with scholars from other fields and non-academic stakeholders (for instance PGIS, participatory scenario planning and deliberative valuation), also called integrated approaches (Table 1). This third group of methods has been applied to uncover ES values at national scales (and international in the case of scenarios) while the first two groups are not usually applied at such broad scales. Further, the third type of methods can contribute to social learning and knowledge co-production as it fosters discussion between different stakeholder groups regarding the importance of differ- ent ES (deliberative valuation), their spatial distribution (PGIS) and the future trends of ES and their implications for human well- being (participatory scenario planning). PGIS is also the most suitable method to pro- vide spatial outputs, although preference as- sessment, time use and photo-elicitation may also contribute with spatially explicit results by estimating representative values for differ- ent geographical areas. PGIS is particularly suited to identify ecosystem service benefit- ing areas, i.e. places where use or demand of ES converge (see Chapters 5.2 and 5.6.2). Despite all developments regarding socio-cul- tural valuation of ES, the question of how so- cio-cultural valuation methods can elicit the broad range of values associated with nature is still relatively unexplored. Following the conceptual definitions provided for value cat- egories in the Total Economic Value (TEV), the Economics of Ecosystems and Biodiver- sity (TEEB) and the IPBES, an integrative approach to socio-cultural valuation meth- ods has the capacity to uncover most of the different value categories (Table 2). Broadly speaking, Table 1 shows that some methods are more specific towards certain value types (e.g. narrative methods), while other meth- ods are generally able to capture multiple values, but not specifically designed for any value type in particular (e.g. participatory Chapter 4 105 scenario planning or deliberative valuation). All value types are appropriately covered by one or more methods, but all methods have blind spots, which imply bias and condition- al application. Consequently, using multiple methods is necessary to cover all values types. The resulting analyses reflect the extent to which diverse valuation methods capture specific value types or have integrative po- tential, as well as which set of complemen- tary methods can be applied to capture mul- tiple values. SOCIO-CULTURAL METHODS SPATIAL SCALE DATA COLLABORATION RESOURCES Lo ca l R eg io na l N at io na l A m ou nt o f d at a Q ua lit at iv e Q ua nt it at iv e R es ea rc he rs ’ o w n fie ld R es ea rc he rs ’ o th er fie ld N on -a ca de m ic st ak eh ol de rs T im e Ec on om ic Preference assessment ● ● ● ● ● ● ● ● Time use ● ● ● ● ● ● ● Photo-elicitation surveys ● ● ● ● ● ● ● ● Narratives ● ● ● ● ● ● Participatory GIS (PGIS) ● ● ● ● ● ● ● Scenario planning ● ● ● ● ● ● ● Deliberative valuation ● ● ● ● ● ● ● Table 1. Methodological requirements of socio-cultural methods for valuing ES. Methods are evaluated according to their suitability to value ES at different spatial scales and to uncover quantitative or qualitative data - (●) high, (●) moderate, (●) low - and according to the level of requirements in terms of data, collabo- ration, time and economic resources - (●) high, ( ) medium, ( ) low - Source: Kelemen et al. (2015). Mapping Ecosystem Services106 Internal variability of socio- cultural valuation methods A key similarity amongst socio-cultur- al methods is the assumption that values of ES are rooted in individuals and, at the same time, shaped by individuals’ social and cultural context. In fact, socio-cultural ap- proaches have the capacity to elicit collective and shared values of ES through participato- ry and deliberative techniques that go beyond the aggregation of individual preferences. So- cio-cultural valuation methods aim at valu- ing ES in a considered way by discovering the psychological, historical, cultural, social, ecological and political contexts and condi- tions, as well as social perceptions that shape individually held or commonly shared values. Variability among methods makes socio- cultural valuation capable of flexible Table 2. Main socio-cultural methods are presented in relation to their capacity to integrate different types of values - (●) high, (●) moderate, (●) low, (○) not appropriate - and according to their capacity to integrate values - (●) high, ( ) medium, ( ) low - Source: Kelemen et al. (2015). SOCIO- CULTURAL METHODS IPBES values TEEB values Total Economic Value In te gr at iv e Po te nt ia l In tr in si c R el at io na l In st ru m en ta l Ec ol og ic al So ci o- cu lt ur al M on et ar y D ir ec t u se v al ue s In di re ct u se v al ue s Ex is te nc e va lu es B eq ue st v al ue s O pt io n va lu es Preference assessment ● ● ● ● ● ○ ● ● ● ● ● Time use ● ● ● ● ● ● ● ● ● ● ● Photo- elicitation surveys ● ● ● ● ● ● ● ● ● ● ● Narratives ● ● ● ● ● ○ ○ ○ ● ● ● Participatory GIS (PGIS) ● ● ● ● ● ● ● ● ● ● ● ● Scenario planning ● ● ● ● ● ● ● ● ● ● ● ● Deliberative valuation ● ● ● ● ● ● ● ● ● ● ● ● Degree of values captured by all methods ● Chapter 4 107 SOCIO- CULTURAL METHODS IPBES values TEEB values Total Economic Value In te gr at iv e Po te nt ia l In tr in si c R el at io na l In st ru m en ta l Ec ol og ic al So ci o- cu lt ur al M on et ar y D ir ec t u se v al ue s In di re ct u se v al ue s Ex is te nc e va lu es B eq ue st v al ue s O pt io n va lu es Preference assessment ● ● ● ● ● ○ ● ● ● ● ● Time use ● ● ● ● ● ● ● ● ● ● ● Photo- elicitation surveys ● ● ● ● ● ● ● ● ● ● ● Narratives ● ● ● ● ● ○ ○ ○ ● ● ● Participatory GIS (PGIS) ● ● ● ● ● ● ● ● ● ● ● ● Scenario planning ● ● ● ● ● ● ● ● ● ● ● ● Deliberative valuation ● ● ● ● ● ● ● ● ● ● ● ● Degree of values captured by all methods ● adaptation to specific worldviews and decision contexts. Key aspects of this variability include (Figure 2): 1. The type of values elicited: methods fo- cusing on the value to individuals versus methods focusing on the value to society. Values can be considered at the level of the individual (what is considered useful, im- portant, good or morally acceptable by a person) and at higher levels of societal or- ganisation, including a group, a commu- nity or the society as a whole (Figure 2). The latter type includes social and cultural values and refers to the fact that societies hold shared principles and virtues, as well as a shared sense of what is worthwhile and meaningful. Shared social values influence individual values because all of us are part of and have been socialised within, a specif- ic community and social context. Valuation methods differ in terms of focusing on per- sonal (individual) understandings of value, or eliciting those value dimensions that are shared by a group of people and culturally embedded within a society. 2. The type of rationality attributed to partic- ipants (value providers): self-oriented versus others-oriented methodological approaches. We can distinguish between individual (I) and collective (We) rationality as the two main rules of thumb behind reasonable ac- tions (Figure 2). When following “I” ratio- nality, we consider individual benefits and costs of personal actions and choose the most beneficial option for ourselves. On the other hand, following “We” rationality means that before acting, we consider what is good and bad for our community/society and how our actions can impact others. Therefore, “I” ra- tionality refers to self-oriented actions and choices, while “We” rationality refers to oth- er-regarding actions and choices. 3. The process of including participants (value providers) in valuation: observation, consultation or engagement methods. There are three options to gain knowledge on pref- erences, depending on whether preferences (values) are considered as pre-existing or in the process of formation. Preferences can be observed and reported when participants have a direct relation with the subject of valuation (e.g. they frequently use or enjoy some ES). However, not having a direct re- lation with the subject of valuation does not necessarily mean that participants do not attribute value to it. To explore these social preferences, participants can be consulted or asked via questionnaires or interviews about their perceptions of ES. If preferences are not expected to exist a priori, or are in the process of formation (i.e. participants do not have a priori knowledge about, or have not faced others’ perceptions of certain ES), we can also engage participants in a joint preference formation process through de- liberative valuation, participatory scenario planning or PGIS. 4. The dominant approaches to handling data: predominantly quantitative, predom- inantly qualitative and mixed methodolog- ical approaches. All three types of methods can be used to collect quantitative, as well as qualitative data. Quantitative data can be collected in numerical form from large pop- ulations and, if representative, can provide results that are applied, in a general sense, from local to regional or even broader spa- tial scales. Quantitative data can be collected both at individual and group level and then aggregated to generalise the results from the sample to larger populations. Qualitative data allow an in-depth understanding of values and underlying motivations, but usu- ally for a much smaller (and often non-rep- resentative) sample. Qualitative data can be collected at the individual and group level in the form of narrative arguments (main- ly words, but also pictures, drawings, etc.). Due to the heterogeneity of types of data, aggregation is often impossible and other Mapping Ecosystem Services108 means of synthesis have to be used (e.g. nar- rative methods or deliberation). In practice, quantitative and qualitative approaches can be placed along a continuum (Figure 2) and, in many cases, they are used in a mixed and complementary approach. Figure 2. Variability among socio-cultural valuation methods in relation to three axes: type of values elicited, type of rationality attributed to value providers, and the dominant approach of handling data. SELF- ORIENTED INDIVIDUAL CONTACTING APPROACH CONSULTATION ENGAGEMENT OBSERVATION PREFERENCES SOCIAL OTHERS- ORIENTED Preference assessment Photo-elicitation Participatory-based GIS Narrative approaches Preference assessment Photo-elicitation Participatory-based GIS Time use Narrative approaches Preference assessment Participatory scenario planning Participatory-based GIS Narrative approaches Deliberative approaches Implementation of socio- cultural valuation methods in the decision-support Ecosystem service assessments have increas- ingly been called to support environmental planning, mainly based on biophysical and economic indicators. However, the expecta- tions of decision-makers in relation to how these assessments can support decision-mak- ing are not always fulfilled. Moreover, few studies have included the socio-cultural dimension of ES, despite its being consid- ered a research priority. Overlooking the socio-cultural dimension might obscure human-nature relationships and hinder the mainstreaming of ES across societal sectors and in decision-support. Integrated valuation aims to clarify the in- terdependencies between the multiple val- ues associated with different ES (see also Box 2 for an example). The biophysical dimension, i.e. an ecosystem’s capacity to supply services, determines the range of po- tential uses by society which also influenc- es its socio-cultural and monetary values. Socio-cultural values might also have an influence on monetary values because indi- Chapter 4 109 vidual and social motivations determine the ‘utility’ a person obtains from a particular service. Conversely, monetary values have social interpretations and the process of monetary valuation is value-articulating in itself. These interdependencies between val- ue dimensions and the different information provided by them, justify combining the different value domains to properly inform environmental decision-making processes. In this section, we formulate several propo- sitions regarding how socio-cultural valua- tion methods can provide support in deci- sion-making: 1. Socio-cultural approaches help broaden the valuation scope and capture multiple values that complement other valuation methods. Socio-cultural valuation methods can be used to identify how values and per- ceptions toward ES differ among stakehold- ers and offer insights into the motivations for conserving nature and the symbolic, cul- tural and spiritual values that are frequent- ly invisible in other valuation approaches. Further, socio-cultural valuation methods can address relational values that are prefer- ences, principles and virtues associated with nature-human relationships. For example, deliberative methods allow the consider- ation of ethical beliefs, moral commitments and social norms. 2. Socio-cultural valuation methods can cover different spatial scales. Values de- rived from large representative samples of a population can be transferred to oth- er locations when the social, cultural and ecological conditions are similar and ag- gregated to larger scales than the original study. Given the emphasis of socio-cultur- al valuation methods on social formation and context-dependency of values, some approaches such as value transfer, aggrega- tion and scaling are less common than in economic valuation where assumptions of pre-existing individual preferences encour- age comparisons across contexts. In addi- tion, a number of socio-cultural valuation methods are applied at local scales to assess certain values in depth. 3. Socio-cultural valuation methods are a useful tool to identify how plural values are interlinked. These help identify plural and heterogeneous values that are relevant for different people (e.g. different socio-demo- graphic profiles, different cultures or cos- mologies), at different temporal scales (e.g. seasons of the year) and different choice sit- uations (individual versus group). Socio-cul- tural valuation methods can reveal how plu- ral and heterogeneous values are interlinked and contribute to human wellbeing. 4. Socio-cultural methods are more appro- priate in situations of social conflict than other valuation methods. Aiming for an in-depth understanding of human-nature relationships, some socio-cultural methods integrate different forms of knowledge (e.g. expert or technical knowledge and experi- ential and local knowledge) held by differ- ent social actors. Sometimes, the interests of one stakeholder group might be in con- flict with the interests of other stakeholders and power relations might operate between them. In that case, socio-cultural valuation can support the identification of conflicts arising from different perceptions, needs and uses of ES, as well as power inequities in the access to ES. 5. Socio-cultural preferences can serve as in- dicators of the impact of different manage- ment options on the ecosystems’ capacity to deliver services. Socio-cultural preferences are often associated with ecosystem service bundles. They are helpful in identifying ecosystem service synergies and trade-offs resulting from stakeholders’ diverging inter- ests and knowledge. Social preferences for ES can be used as indicators of present and future pressures on landscapes and land-use Mapping Ecosystem Services110 change. For example, multi-criteria analysis can combine a biophysical ecosystem ser- vice assessment with people’s willingness to trade off one ecosystem service for another, establishing a ranking order of landscape management alternatives that can be used in priority-setting. In summary, socio-cultural valuation meth- ods can provide decision-support in the form of awareness-raising, value and knowl- edge recognition, value conflict identifi- cation and priority-setting. They also help bring different voices and stakeholders into the decision-making process. Box 1 . Participatory mapping of ecosystem service flows in a National Park (Sierra Nevada, Spain) Participatory GIS seeks to produce ecosystem service maps in regions of data scarcity while engaging stakeholders through the mapping process. These two aims were pursued in the process developed in Sierra Nevada to map ecosystem service flows. In a two day workshop, 20 participants mapped the supply and demand (i.e. Service Provision Hotspots and Service Benefiting Areas) of 11 ES. Results showed the importance of protected areas to deliver ES and allowed the elaboration of concrete policy proposals for the protected area and its surrounding landscape. Regarding ecosystem service supply, potential restoration areas and areas that require a value enhancement strategy were identified. Eco- system service demand maps showed the need of a multi-scale strategy for protected area management beyond protected area boundaries to be able to manage the demand that affects the ecosystem within the protected area. Participatory mapping of ES developed by experts (i.e. managers and scientists) in Sierra Nevada Protected Area. Chapter 4 111 Box 2 . Socio-cultural valuation of ES in Hungary (Homokhátság) The major aim of this ES study was to help local stakeholders and decision-makers move towards a more sustainable landscape management system. To this end, in-depth and semi-structured interviews and focus groups were applied. We carried out narrative methods to understand the institutionalised mechanisms affecting farmers’ choices that are often in conflict with nature conservation. Moreover, we carried out deliberative valuation to understand how farmers relate to biodiversity and whether it has different meanings and values to different groups of farmers. A preference assessment survey was carried out to mobilise community members and collect information on their knowledge, opinion and feelings related to ES. This was then channelled into a participatory scenario planning pro- cess, combined with modelling, to enable stakeholders and experts to explore alternative future options and choose the most desirable one(s) together. This long lasting research process was able to highlight multiple dependencies between local inhabitants and their surrounding environment. We could identi- fy plural and heterogeneous values and their possible changes across time and space. Focus group with local stakeholders using visual stimuli to elicit socio-cultural values of ES and their spatial distribution in the landscape. Mapping Ecosystem Services112 Further reading Chan K, Balvanera P, Benessaiah K, Chapman M, Díaz S, Gómez-Baggethun E, Gould R, Hannahs N, Jax K, Klain S, Luck G, Martín-López B, Muraca B, Norton B, Ott K, Pascual U, Satterfield T, Tadaki M, Taggart J, Turner NJ (2016) Why Protect Nature? Rethinking Values and the Envi- ronment. PNAS 113:1462-1465. García-Llorente M, Martín-López B, Inies- ta-Arandia I, López-Santiago CA, Aguilera PA, Montes C (2012) The role of multi- functionality in social preferences toward semi-arid rural landscapes: an ecosystem service approach. Environmental Science and Policy 19-20: 136-146. Gómez-Baggethun E, Martín-López B (2015) Ecological Economics perspective in eco- system services valuation. In: Martínez- Alier J, Muradian R (Eds) Handbook of Ecological Economics, 260-282. Edward Elgar, London. Iniesta-Arandia I, García-Llorente M, Aguil- era P, Montes C, Martín-López B (2014) Socio-cultural valuation of ecosystem ser- vices: uncovering the links between values, drivers of change and human well-being. Ecological Economics 108: 36-48. IPBES (2015) Preliminary guide regarding diverse conceptualisation of multiple val- ues of nature and its benefits, including biodiversity and ecosystem functions and services, IPBES 15 Deliverable 3 (d). Kelemen E, García-Llorente, M, Pataki G, Martín-Lopez B, Gómez-Baggethun E (2014) Non-monetary techniques for the valuation of ecosystem services. Open- NESS Syntehsis Papers No. 6. Kenter J, O’Brien L, Hockley N, Ravenscroft N, Fazey I, Irvine K, Reed M, Christie M, Brady E, Bryce R and Church A (2015) What are shared and social values of ecosys- tems? Ecological Economics, 111: 86-99. Kovács E, Kelemen E, Kalóczkai Á, Margóczi K, Pataki G, Gébert J, Málovics G, Balázs B, Roboz Á, Kovács E, Mihók B (2015) Understanding the links between eco- system service trade-offs and conflicts in protected areas. Ecosystem Services 12: 117-127. Martín-López B, Iniesta-Arandia I, García- Llorente M, Palomo I, Casado-Arzuaga I, Del Amo DDG, Gómez-Baggethun E, Oteros-Rozas E, Palacios-Agundez I, Wil- laarts B, González JA, Santos-Martín F, Onaindia M, López-Santiago C, Montes C (2012) Uncovering ecosystem service bundles through social preferences. PLoS ONE 7(6): e38970. Oteros-Rozas E, Martín-López B, González JA, Plieninger T, López CA, Montes C. (2014) Socio-cultural valuation of ecosys- tem services in a transhumance social-eco- logical network. Regional Environmental Change 14: 1269-1289. Plieninger T, Bieling C, Ohnesorge B, Schaich H, Schleyer C, Wolff F (2013) Exploring futures of ecosystem services in cultural landscapes through participatory scenario development in the Swabian Alb, Germa- ny. Ecological and Society 18(3): 39. Santos-Martín F, Martín-López B, García- Llorente M, Aguado M, Benayas J, Mon- tes C (2013) Unravelling the Relationships between Ecosystems and Human Wellbe- ing in Spain. PLoS One 8: e73249. Teddlie C, Tashakkori A (Eds) (2009)Founda- tions of mixed methods research: Integrat- ing quantitative and qualitative approach- es in the social and behavioural sciences. Sage Publications Inc. Chapter 4 113 4.3. Economic quantification Luke M. Brander & Neville D. Crossman Introduction Economic quantification of ES attempts to measure the human welfare derived from the use or consumption of ES. Economic quantification or valuation is one way to assess and communicate the importance of ES to decision-makers and can be used in combination with other forms of informa- tion (e.g. bio-physical or social quantifica- tion - see Chapters 4.1 and 4.2). The com- parative advantage of economic valuation is that it conveys the importance of ES directly in terms of human welfare and uses a com- mon unit of account (i.e. money) so that values can be directly compared across ES and across other goods and services that are consumed by society. The aim of this chapter is to introduce the key concepts underlying economic quantifi- cation of ES and to provide an explanation for the various economic methods that can be applied. The economic value of ES In this section, we provide definitions of the various concepts of economic value that may be encountered when quantifying ES. In neo-classical welfare economics, the eco- nomic value of goods or services is the well- being derived from its production and con- sumption, usually measured in monetary units. In a perfectly functioning market, the economic value of goods or services is de- termined by the demand for and supply of those goods or services. Demand for goods or services is driven by the benefit, utility or welfare that consumers derive from it. Sup- ply of goods or services is determined by the cost to producers of producing it. The left- hand panel in Figure 1 provides a simplified representation of demand (marginal benefit) and supply (marginal cost) for goods traded in a market at quantity ‘Q’ and price ‘P’. Area ‘A’ represents the consumer surplus which is the gain obtained by consumers because they are able to purchase a product at a market price that is less than the highest price they would be willing to pay (which is related to their benefit from consumption and represented by the demand curve). The producer surplus, depicted by ‘B’, is the amount that producers benefit by selling at a market price that is higher than the lowest price that they would be willing to sell for (which is related to their production costs and represented by the supply curve). The area ‘C’ represents production costs which differ among producers and/or over the scale of production. The sum of areas A and B is labelled the ‘surplus’ and is interpreted as the net economic gain or welfare resulting from production and consumption with a quantity of Q at price P. It is important to recognise that, when we make a decision to allocate resources to pro- duce particular goods or services, we are also deciding not to allocate those resources to produce alternative goods or services. The goods or services that we give up is called the “opportunity cost” of our decision. Oppor- tunity cost can be defined as the value of the Mapping Ecosystem Services114 foregone next best use of resources. This is an important concept in the context of ES since it is often the value of the alternative use of resources (e.g. agriculture, timber extraction, aquaculture) that drives ecosystem loss. In the case that ES are not traded in a mar- ket, the interpretation of the welfare derived from their provision can also be represented in terms of surplus. The right-hand panel of Figure 1 represents the supply and demand of a non-marketed ecosystem service. In this case, the ecosystem service does not have a supply curve in the conventional sense that it represents the quantity of the service that producers are willing to supply at each price. The quantity of the ecosystem service that is ‘supplied’ is not determined through a mar- ket at all but by other decisions regarding ecosystem protection, land use, manage- ment, access etc. The quantity of the eco- system service supplied is therefore inde- pendent of its value. This is represented as a vertical line. For the most part, bio-phys- ical indicators of ES measure the quantity supplied but not the welfare obtained. The demand curve for non-marketed ES is still represented as a downward sloping line since marginal benefits are expected to de- cline with quantity (the more that we have of a service, the lower the additional welfare of consuming more). In this case, consum- ers do not pay a price for the quantity (Q) that is available to them and the entire area under the demand curve (D+E) represents their consumer surplus. It is useful to keep this picture in mind when considering the economic quantification of ES. Total Economic Value The concept of Total Economic Value (TEV) of an ecosystem is used to describe the comprehensive set of utilitarian values derived from it. This concept is useful for identifying the different types of value that an ecosystem provides. TEV comprises of use values and non-use values. Use values are the benefits that are derived from some physical use of the resource. Direct use val- ues may derive from on-site extraction of resources (e.g. fuel wood) or non-consump- tive activities (e.g. recreation). Indirect use values are derived from off-site services that are related to the resource (e.g. downstream flood control, climate regulation). The op- tion value is the value that people place on maintaining the option to use an ecosystem resource in the future given the uncertain- ty that they would actually use it. Non-use values are derived from the knowledge that an ecosystem is maintained without re- Demand Supply A B C P Q Quantity Price Demand Ecosystem service provision D E V Q Quantity Value Figure 1. Demand and supply for ES. Chapter 4 115 gard to any current or future personal use. Non-use values may be related to altruism (maintaining an ecosystem for others), be- quest (for future generations) and existence (preservation unrelated to any human use) motivations. The constituent components of TEV are represented in Figure 2. It is im- portant to understand that the “total” in To- tal Economic Value refers to the aggregation of different sources of value rather than the sum of all values derived from a resource. Accordingly, many estimates of TEV are for marginal changes in the provision of ES but “total” in the sense that they take a compre- hensive view of sources of value. Exchange value The concept of welfare value is used in most economic assessments of ES but it is not used in the system of national accounts (SNA) that is used to calculate gross do- mestic product (GDP) and other economic statistics. The SNA uses the concept of ex- change value which is a measure of producer surplus plus the costs of production. In the left-hand panel of Figure 1, this is represent- ed by areas B and C or equivalently P times Q. Under the concept of exchange value, the total outlays by consumers and the to- tal revenue of the producers are equal. For national accounting purposes, this approach to valuation enables a consistent and conve- nient recording of transactions between eco- nomic units. In the context of comparing the values of ES with values in the system of national accounts, it is therefore necessary to value the total quantity of ES at the market prices that would have occurred if the ser- vices had been freely traded and exchanged. In other words, it is necessary to measure exchange value and not welfare value. The differences between the concepts of wel- fare value and exchange value are the inclu- sion of consumer surplus (A) in the former and the inclusion of production costs in the latter (C). The concept of welfare value cor- responds to a theoretically valid measure of welfare in the sense that a change in value represents a change in welfare for the pro- ducers and/or consumers of the goods and services under consideration. The concept of exchange value does not correspond to a theoretically valid measure of welfare and a change in exchange value does not necessar- ily represent a change in welfare for either producers or consumers. Quantifying economic values A variety of methods have been developed for quantifying the economic value of ES. These valuation methods are designed to span the range of valuation challenges raised by the application of economic analyses to Total Economic Value Direct Use Indirect Use Option Altruism Bequest Existence Non-Use ValuesUse Values Figure 2. The components of Total Economic Value. Mapping Ecosystem Services116 the complexity of the natural environment. An important distinction exists between methods that produce new or original in- formation generally using primary data (pri- mary valuation methods) and those that use existing information in new policy contexts (value transfer methods). Primary valuation methods Table 1 provides an overview of primary valuation methods, typical applications and limitations and indicates which primary val- uation methods can be used to value which ecosystem service. The reader should be aware of the important distinction between primary valuation methods, i.e. the differ- ence between revealed preference methods (those that observe actual behaviour of the use of ES to elicit values) and stated prefer- ence methods (those that use public surveys to ask beneficiaries to state their preferences for, generally, hypothetical changes in the provision of ES). Revealed preference meth- ods may be favoured since they reflect actual behaviour but are limited in their applica- bility to some ES. Stated preference meth- ods, on the other hand, rely on responses recorded in surveys or experiments but are more flexible in their application. Value transfer methods Decision-making often requires informa- tion quickly and at low cost. New ‘prima- ry’ valuation research, however, is general- ly time consuming and expensive. For this reason, there is interest in using information from existing primary valuation studies to make inform decisions regarding impacts on ecosystems that are of current interest. This transfer of value information from one context to another is called value transfer. Value transfer is the use of research results from existing primary studies at one or more sites or policy contexts (“study sites”) to pre- dict welfare estimates or related information for other sites or policy contexts (“policy sites”).1 In addition to the need for expeditious and inexpensive information, there is often a need for information on the value of ES at a different geographic scale from that for which primary valuation studies have been conducted. So, even in cases where some pri- mary valuation research is available for the ecosystem of interest, it is often necessary to extrapolate or scale-up this information to a larger area or to multiple ecosystems in the region or country. Primary valuation studies tend to be conducted for specific ecosystems at a local scale whereas the information re- quired for decision-making is often needed at a regional or national scale. Value transfer therefore provides the means to obtain in- formation for the scale that is required. The number of primary studies on the val- ue of ES is substantial and growing rapid- ly. This means that there is a growing body of evidence to draw on for the purposes of transferring values for informed deci- sion-making. With an expanding informa- tion base, the potential for using value trans- fer is improved. Value transfer can potentially be used to estimate values for any ecosystem service, provided that there are primary valuations of that ecosystem service from which to transfer values. The use of value transfer is widespread but requires careful application. The alternative methods of conducting val- ue transfer are described here. 1 Value transfer is also known as benefit transfer but since the values that are transferred may be costs as well as benefits, the term value transfer is more generally applicable. Chapter 4 117 Valuation method Approach Application to ES Example ecosystem service Market prices Prices for ES that are direct- ly observed in markets. ES that are traded directly in markets. Timber and fuel wood from forests; Recreation at na- tional parks that charge an entrance fee. Public pric- ing Public expenditure or mon- etary incentives (taxes/subsi- dies) for ES as an indicator of value. ES for which there are public expenditures. Watershed protection to provide drinking water; Purchase of land for protect- ed areas. Defensive expenditure Expenditure on protection of ecosystems. ES from protected ecosystems. Nutrient filtration by protected wetlands Replace- ment cost Estimate the cost of replac- ing an ES with a man-made service. ES that have a man-made equivalent. Coastal protection by dunes; water storage and filtration by wetlands. Restoration cost Estimate cost of restoring degraded ecosystems to ensure provision of ES. Any ES that can be provid- ed by restored ecosystems. Coastal protection by dunes; water storage and filtration by wetlands. Damage cost avoided Estimate damage avoided due to ecosystem service. Ecosystems that provide storm or flood protection to houses or other assets. Coastal protection by dunes; river flow control by wetlands. Net factor income Revenue from sales of envi- ronment-related good mi- nus cost of other inputs. Ecosystems that provide an input in the production of marketed goods. Filtration of water by wet- lands; commercial fisheries supported by coastal wet- lands. Production function Statistical estimation of production function for marketed goods including an ES input. Ecosystems that provide an input in the production of marketed goods. Soil quality or water quality as an input to agricultural production. Hedonic pricing Estimate influence of envi- ronmental characteristics on price of marketed goods. Environmental characteris- tics that vary across goods (usually houses). Urban open space; air quality. Travel cost Use data on travel costs and visit rates to estimate de- mand for recreation sites. Recreation sites Outdoor open access rec- reation. Contingent valuation Ask people to state their willingness to pay for an ES through surveys. All ES Species loss; natural areas; air quality; water quality landscape aesthetics. Choice modelling Ask people to make trade- offs between ES and other goods to elicit willingness to pay. All ES Species loss; natural areas; air quality; water quality; landscape aesthetics. Group / participatory valuation Ask groups of stakeholders to state their willingness to pay for an ES through group discussion. All ES Species loss; natural areas; air quality; water quality; landscape aesthetics. Table 1. Primary economic valuation methods. Mapping Ecosystem Services118 Unit value transfer uses values for ES at a study site, expressed as a value per unit (usually per unit of area or per beneficiary), combined with information on the quantity of units at the policy site to estimate pol- icy site values. Unit values from the study site are multiplied by the number of units at the policy site. Unit values can be adjusted to reflect differences between the study and policy sites (e.g. income and price levels). Value function transfer uses a value function estimated for an individual study site in con- junction with information on parameter val- ues for the policy site to calculate the value of an ecosystem service at the policy site. A value function is an equation that relates the value of an ecosystem service to the charac- teristics of the ecosystem and the beneficiaries of the ecosystem service. Value functions can be estimated from a number of primary val- uation methods including hedonic pricing, travel cost, production function, contingent valuation and choice experiments. Meta-analytic function transfer uses a value function estimated from the results of mul- tiple primary studies representing multiple study sites in conjunction with information on parameter values for the policy site to calculate the value of an ecosystem service at the policy site. A value function is an equa- tion that relates the value of an ecosystem service to the characteristics of the ecosys- tem and the beneficiaries of the ecosystem service. Since the value function is estimated from the results of multiple studies, it is able to represent and control for greater variation in the characteristics of ecosystems, benefi- ciaries and other contextual characteristics. This feature of meta-analytic function trans- fer provides a means to account for simul- taneous changes in the stock of ecosystems when estimating economic values for ES (i.e. the “scaling up problem”). By includ- ing an explanatory variable in the data de- scribing each “study site” that measures the scarcity of other ecosystems in the vicinity of the “study site”, it is possible to estimate a quantified relationship between scarcity and ecosystem service value. This parameter can then be used to account for changes in ecosystem scarcity when conducting value transfers at large geographic scales. These three principal methods for transfer- ring ecosystem service values are summarised in Table 2. The choice of which value trans- fer method to use to provide information for a specific policy context is largely de- pendent on the availability of primary valu- ation estimates and the degree of similarity between the study and policy sites (in terms of biophysical and socio-economic charac- teristics and context). In cases where value information is available for a highly similar study site, unit value transfer may provide the most straightforward and reliable means of conducting value transfer. On the other hand, when study sites and policy sites are different, value function or meta-analytic function transfer offers a means to systemati- cally adjust transferred values to reflect those differences. Similarly, in the case where value information is required for multiple different policy sites, value function or meta-analytic function transfer may be a more accurate and practical means for transferring values. Representing economic values on maps The representation of economic values on maps involves estimating variable combina- tions of supply and demand across spatial units and plotting the resulting values. Spa- tial units in a value map can include land parcels (e.g. polygons representing own- ership), ecosystem patches (e.g. polygons representing distinct ecosystems of different type), ecosystem units (e.g. raster grids of ecosystem type), grid cells (e.g. raster grids Chapter 4 119 with land use/land cover), or beneficiaries (e.g. people plotted using residential or ac- tivity location). In most cases, spatial units are used to represent the ecosystem that supplies the ecosystem service, but mapping values by the location of beneficiaries can be useful in some decision-making contexts (e.g. for representing the distributional con- sequences of changes in ecosystem service provision across communities; or for design- ing payment mechanisms for ES). In mapping ecosystem service values, each spatial unit is treated as a separate sub-mar- ket for the ecosystem service. Methods for mapping ecosystem service values can ad- dress spatial variations in supply, demand, or a combination of both determinants. In general terms, bio-physical methods (see Chapter 4.1) are used to estimate the spa- tially variable quantities of ES supplied (e.g. probability of flood damage, quantity of clean water, area of recreational space, tonnes of carbon stored) and economic methods are used to estimate spatially variable marginal values per unit of ecosystem service provided and consumed. Mapping economic values therefore necessarily involves linking maps of biophysical ecosystem supply with economic valuation methods. Production/consumption statistics In addition to the quantification of human welfare derived from ES in monetary units, economic quantification of ES encompasses the recording or estimation of production and consumption statistics in physical units. Indeed, the measurement of physical units of production and consumption of ES is a necessary a step in the process of quantify- ing economic value. Economic quantifica- Approach Strengths Weaknesses Unit value transfer Select appropriate values from existing primary valuation studies for similar ecosystems and socio-economic contexts. Adjust unit values to reflect differences between study and policy sites (usually for income and price levels). Simple Unlikely to be able to account for all factors that determine differences in values between study and policy sites. Value information for highly similar sites is rarely available. Value function transfer Use a value function derived from a primary valuation study to estimate ES values at policy site(s). Allows differences between study and policy sites to be controlled for (e.g. differences in population characteristics). Requires detailed information on the characteristics of policy site(s). Meta- analytic function transfer Use a value function estimated from the results of multiple primary studies to estimate ES values at policy site(s). Allows differences between study and policy sites to be controlled for (e.g. differences in population characteristics, area of ecosystem, abundance of substitutes etc.). Practical for consistently valuing large numbers of policy sites. Requires detailed information on the characteristics of policy site(s). Analytically complex. Table 2. Value transfer methods. Mapping Ecosystem Services120 tion may, however, stop short of estimating values and directly report production and consumption statistics as useful information to support decision-making. To a great extent, the production of ES is quantified using bio-physical methods (see Chapter 4.1). For most ES, however, it is also necessary to use insights and methods from economics to measure the quantities that are actually used (i.e. to quantify uti- lised services as opposed to potential ser- vices). This generally involves measuring the extent of demand for ES in terms of the population and preferences of beneficiaries. The physical units in which production and consumption statistics are reported are spe- cific to each ecosystem service. For example, non-timber forest products (NTFPs), such as wild honey, may be measured in kilo- grams; water extracted for consumption is measured in kilolitres or megalitres, carbon sequestration is conventionally measured in tonnes of carbon or CO2; and recreational use of natural open space may be measured in numbers of visits – all usually expressed per unit of time over which the flow of ser- vice is recorded (e.g. per year). In very few cases will the quantity of an ecosystem ser- vice be explicitly observed and recorded in a systematic and accessible way (i.e. there is generally not an equivalent of the business activity surveys conducted for the SNA). In most cases, it is necessary to estimate the level of production and consumption using some form of bio-economic modelling. Box 1 . Example valuation and mapping of freshwater ES Freshwater ecosystems provide a variety of ES that can be affected by changes in water quality. In this case study, projected future changes in water quality for the period 2000-2050 are quantified using the IMAGE-GLOBIO model. This information is combined with a meta-analytic value function to estimate the economic value of changes in water quality. The analysis is performed at the resolution of 50 km grid cells. The supply of ES from water bodies (rivers and lakes) is implicitly modelled within the meta-analytic value function. The results of this value transfer application are mapped in order to communicate the spatial distribution of benefits (losses) derived from improvements (declines) in wa- ter quality (see Figure 3). In this application, the spatial units used to map changes in value are benefi- ciaries (households aggregated within 50 km grid cells) rather than the rivers or lakes providing the ES. Figure 3. Value map for changes in water quality 2000-2050 (Annual willingness to pay; Million USD; 2007 price levels). > -435 -435 – -23 -23 – -8 -8 – -5 -5 – -2 -2 – 0 0 – 3 3 – 8 8 – 19 19 – 234 Chapter 4 121 Box 2 . Example valuation and mapping of carbon sequestration The regulating service of carbon sequestration by ecosystems represents a special case in which supply is spatially variable (dependent on vegetation type, soil characteristics etc.) but demand is entirely spatially disconnected (since CO2 is a uniformly mixing stock pollutant, the marginal benefit of se- questration is not related to where the sequestration takes place). Figure 4 represents an estimate of economic returns from planting trees to sequester carbon under different carbon price scenarios in the period 2010-2050 in South Australia. Annual rates of carbon sequestration were modelled based on climate, soil and land management actions and then an economic model was used to estimate the net present value of converting from existing agriculture (crops and livestock) to trees for carbon. Figure 4. Net present value of economic returns from carbon sequestration by carbon monoculture species under five different carbon price scenarios ($/tCO2-e). Source: Bryan and Crossman (2013). $10/tCO2-e $30/tCO2-e $50/tCO2-e $20/tCO2-e Economic returns ($NPV/ha) Less than 0 0 - 1,000 1,000 - 2,500 2,500 - 5,000 5,000 - 7,500 7,500 - 10,000 10,000 - 12,500 12,500 - 15,093 $40/tCO2-e Mapping Ecosystem Services122 Box 3 . Example estimation of production statistics for non- timber forest products (NTFPs) This case study provides an example of how bio-physical and economic modelling can be combined to quantify production statistics for a provisioning ecosystem service. The production of non-timber forest products (NTFPs) in Mondulkiri province, Cambodia, is quantified by combining a validated biophysical model (InVEST) of NTFP availability with an economic model of household decisions regarding the number of harvesting trips to be undertaken and a separate economic model of harvest yield per trip. The bio-physical model quantifies the availability of six NTFPs given spatial variation in forest cover and species diversity. The harvest-trip function, estimated using data from a household survey, quantifies how many trips each household makes given their income, household size and the availability of NTFPs within harvesting distance of their village. The harvest yield function, also esti- mated using data from a household survey, quantifies how much of each NTFP is harvested per trip given household characteristics, number of trips made and NTFP availability. Figure 5 represents the methodological framework for this economic quantification of NTFPs. Figure 5. Combination of bio-physical and economic models to estimate spatially variable production of NTFPs in Mondulkiri province, Cambodia. hhs: house hould survey. Source: Brander (2015). Chapter 4 123 Further reading Bagstad KJ, Johnson GW, Voigt B, Villa F (2013) Spatial dynamics of ecosystem ser- vice flows: a comprehensive approach to quantifying actual services. Ecosystem Ser- vices 4: 117-125. Bouma JA, van Beukering PJH (Eds) (2015) Ecosystem Services: From Concept to Practice. Cambridge University Press. Brander LM, Brauer I, Gerdes H, Gherman- di A, Kuik O, Markandya A, Navrud S, Nunes PALD, Schaafsma M, Vos H, Wag- tendonk A (2012) Using meta-analysis and GIS for value transfer and scaling up: Valuing climate change induced losses of European wetlands. Environmental and Resource Economics 52: 395-413. Brander LM (2013) Guidance manual on value transfer methods for Ecosystem Ser- vices. United Nations Environment Pro- gramme. ISBN 978-92-807-3362-4. Brander LM (2015). Economic valuation of Ecosystem Services in the Eastern Plains Landscape, Modulkiri, Cambodia. Report for WWF Cambodia. Brouwer R et al. (2009) Economic Valuation of Environmental and Resource Costs and Benefits in the Water Framework Di- rective: Technical Guidelines for Practi- tioners. AquaMoney. Bryan BA, Crossman ND (2013) Impact of multiple interacting financial incentives on land use change and the supply of Ecosys- tem Services. Ecosystem Services 4: 60-72. DEFRA (2013) Guidance for policy and de- cision makers on using an ecosystems ap- proach and valuing Ecosystem Services. Department for Environment, Food & Rural Affairs https://www.gov.uk/ecosys- tems-services. DEFRA (2007) An introductory guide to valuing Ecosystem Services (2007) De- partment for Environment, Food and Ru- ral Affairs. Freeman AMI (2003) The Measurement of Environmental and Resource Values. Re- sources for the Future, Washington D.C. Johnston RJ, Rolfe J, Rosenberger RS, Brou- wer R (Eds) (2015) Benefit transfer of en- vironmental and resource values: A hand- book for researchers and practitioners. Springer. ISBN 978-94-017-9929-4. Pascual U, Muradian R, Brander L, Gómez-Bag- gethun E, Martín-López B, Verma M (2010) The economics of valuing Ecosystem Services and biodiversity. In Kumar (Ed.) The Economics of Ecosystems and Biodiver- sity: Ecological and Economic Foundations. Earthscan, London and Washington. Pearce D et al. (2001) Economic Valuation with Stated Preference Techniques Sum- mary Guide. Department of Transport, Local Government and Regions, London. Schägner JP, Brander LM, Maes J, Hartje V (2013) Mapping ecosystem service values: Current practice and future prospects. Ecosystem Services 4: 33-46. Mapping Ecosystem Services124 4.4. Computer modelling for ecosystem service assessment Robert W. Dunford, Paula A. Harrison & Kenneth J. Bagstad Introduction Computer models are simplified represen- tations of the environment that allow bio- physical, ecological, and/or socio-economic characteristics to be quantified and explored. Modelling approaches differ from mapping approaches (Chapter 5) as (i) they are not forcibly spatial (although many models do produce spatial outputs); (ii) they focus on understanding and quantifying the interac- tions between different components of so- cial and/or environmental systems and (iii) by changing parameters within models, they are capable of exploring both alternative sce- narios and internal model dynamics. When applied to the assessment of ecosys- tem services (ES), models are important tools which can quantify the relationships that underpin ES supply, demand and flows and, in some cases, produce maps represent- ing these factors. Furthermore, as models can explore scenarios, trade-offs that result from different scenarios can be assessed. This chapter provides a broad overview of different types of models that have been ap- plied to ES assessments and discusses, with examples, the ways that these models have the potential to be used in practice. In the context of ES, there are a number of ways of distinguishing between different types of models. Here, we distinguish be- tween individual models focussing on single ES and modelling frameworks that can as- sess multiple ES within the framework of a single modelling tool. What is a model? A model is a simplification of reality that represents the relationship between two or more sets of factors and uses one set of factors to predict values of the other: y = x2 is a simple model where the variable y can be predicted from variable x by performing the square function on x. However, there are many different types of models, most of which are considerably more complex. When used in ES research, models are gen- erally used to predict either ES themselves, or underlying environmental aspects from which ES are derived. ES models use di- verse types of input variables, but common- ly include measurements of environmental parameters (e.g., tree heights, river flows, species counts), survey responses or scores given by scientists or stakeholders (e.g., from questionnaire responses or interviews) or the outputs from another model (e.g., outputs from a climate model may provide precipitation inputs to a water flow model). What sorts of models are useful for ES assessment? Computer modelling predates the popu- larisation of the ES concept and models have a decades-long history of use within the environmental sciences. As such, there are a large number of models that can be used to assist in ES assessment, though older Chapter 4 125 biophysical models may be less deliberately beneficiary-orientated than more modern ES models. It is beyond the scope of this chapter to provide a full overview of modelling per se. In the following text, we describe five gen- eral types of models that are used for ES assessment. Conceptual models Conceptual models, although rarely com- puterised, are the first stage of any com- puter modelling process. They are used to gain an understanding of linkages between different components of the system being studied. The carbon and water cycles, for example, provide the underlying concep- tual models for a number of more detailed computer models used to predict ES, such as carbon sequestration by vegetation or the role of vegetation in mediating floods. The first step of any new computerised model- ling process is to draw a conceptual diagram that illustrates how model components in- terlink, then to determine how to quantify those linkages. Statistical models Where there is no known quantified rela- tionship between components of the con- ceptual model, statistical models can help to establish relationships by drawing on collected data. For example, if a conceptual model suggests that freshwater provisioning is driven by rainfall and forest cover in a catchment area, corresponding data can be collected and regression-based approaches can be used to explore the strength of these relationships. Sufficiently strong relation- ships that are identified can then be used in a deterministic model to predict expected freshwater provisioning in areas for which data are not present. Deterministic models A deterministic model assumes links be- tween cause and effect. The y = x2 example above is a very simple deterministic model: for every example of x the value of y will be x2. Deterministic models are usually based on fundamental physical laws derived from a process-based understanding of the sci- ence (e.g., the physical sciences or, as above, a statistically derived relationship). An implicit implication of deterministic models is that there is only one possible out- put for a given set of inputs (x2 will always enumerate to x2). This can lead to a false impression of accuracy when modelling complex systems where uncertainty is com- mon. Probabilistic models have emerged to address these issues. Deterministic models underpin many com- mon ES assessment approaches. The Revised Universal Soil Loss Equation (RUSLE), for example, is a commonly used deterministic model developed to better understand the factors driving soil loss from agricultural land. Expressed as: A = R x K x LS x C x P, it relates soil loss (A) to rainfall erosivity (R), soil erodibility (K), slope steepness and length (LS), management practice (C) and conservation practice (P). Probabilistic models Probabilistic models recognise that random behaviour is often part of a system; they ex- press likelihoods of events occurring (e.g., the return period of a flood of a given magni- tude). Rather than using single values as in- puts, probabilistic approaches use probabil- ity distributions functions (PDFs) as input parameters. Instead of using mean rainfall, a probabilistic approach might use a range of inputs sampled from a normal distribu- tion around the mean up to the maximum and minimum recorded observations. These Mapping Ecosystem Services126 sample values can be selected systematically or randomly (the “Monte Carlo” approach) and then run through a deterministic model to explore the range of outputs that result. This allows the probability of a given output to be assessed, rather than implying that, in a complex system, a single output value can be expected for a given combination of inputs. Rule-based modelling Rule-based models can be applied, using Boolean (yes/no) decisions and if-then statements to construct a path from input to output. They are often represented as nested decision trees. These are common in remote sensing and biological classification keys and use if-then options to decide be- tween possible output classes (e.g., if vari- able x, representing tree cover, is above a given threshold then class y = forest, other- wise class y = grassland). Rule-based models can also be incorporated into context-aware artificial intelligence (AI)-supported model selection platforms that account for context by selecting from a library of possible data and models. Agent-based models (ABMs) are a special type of rule-based model that set (usual- ly simple) rules for individual “agents.” By allowing the individual agents to interact in a model, collective behaviour emerges. The “agents” within an ABM can represent anything from individual species or deci- sion-makers to institutions or countries at international levels. Another important as- pect of ABMs is that agents’ learning from experience can be simulated within the model as it runs, allowing ABMs to model aspects such as the transfer of ideas between individuals and other organic processes. The ABM approach is very different from deterministic approaches, where a clear path can be traced between model inputs and outputs. Within an ABM, results depend on the timing, identity and consequences of agents’ interactions. If there is an element of randomness in the guiding rules (i.e., 50% of the time an agent makes choice X and 50% choice Y), then the same outcome will not be replicated on repeated model runs although, over a large number of runs, com- mon emergent behaviour may be apparent. In ES assessment, ABMs offer significant opportunities to understand how interac- tions between individual actors may influ- ence ES. This could involve human actors (e.g., farmers) interacting with policies and institutional structures to determine the best crop types for their farm and the associated impact of this change for ES provision. Or, it could assess the effects of predator-prey species interactions on the ES provided by these species. Integrated modelling systems Models tend to be developed for specific pur- poses, to address a particular problem raised within a given sector. However, within a single-sectoral model, any number of indi- vidual models may be used to represent dif- ferent linkages within the conceptual model. Furthermore, the outputs of one model may be used to provide inputs to another model creating modelling chains. This can allow in- tegrated modelling systems to be set up that take into consideration cross-sectoral interac- tions, synergies and trade-offs including, for example, the implications for one ES (e.g., drinking water provision) as a result of chang- es in another (e.g., soil erosion regulation). How can models help us better understand ES? The previous section provided a brief over- view of the types of models that exist to provide information to help with ES assess- Chapter 4 127 ment. We next focus on how these models can be used, i.e., on which aspects of the en- vironment do models provide information and how does this help us better assess ES? In the following sections, we consider four main applications of models. First, models of the natural environment that do not pro- vide direct information about ES, but can assess underlying ecosystem structures and functions from which ES can be under- stood. Second, models that are focussed on ES, both on individual services and those intended to enable assessment of a suite of different ES. Third, modelling systems that take an integrated approach to ES, which allows for an assessment of trade-offs and synergies between groups of services. Final- ly, models designed to explore ES with deci- sion-makers or stakeholders. Models of the natural environment Many more models address the natural en- vironment than address the more recent field of ES. Such models may need to be extended to evaluate ES. However, as some have been used for many decades, they may be well known, understood and trusted by stakeholders. This may make them useful entry points to introduce the ES concept, or may introduce a barrier of inertia (“we have a tool that works, so why change it?”). Species distribution models, land use/land cover (LULC) models and general biophys- ical models are common natural systems models that can provide ES assessment. Species distribution modelling Species distribution modelling (SDM) is of- ten used to identify how plant and animal species respond to changing environmen- tal parameters such as atmospheric CO2, climate, or habitat availability. There is a wide range of approaches to SDM, such as simple, statistical “profile methods,” regres- sion-based techniques and approaches that use machine learning. Advanced SDM ap- proaches combine these maps with land use modelling (see below) to determine where habitats are available and, using dispersal and connectivity models, can project the abilities of species to colonise new habitats. The outputs of species distribution models are maps of species distributions for a giv- en scenario. These can be used to assess ES provision related to these species. For exam- ple, maps of charismatic or endemic species distributions can help assess whether partic- ular areas may maintain or lose species with particular religious, social or cultural value. Land use/land cover modelling Land use/land cover data are key inputs to many ES mapping approaches (see Chapter 5) and there are various ways to link LULC with additional datasets to map ES (e.g., the “matrix approach” see Chapter 5.6.4). Ini- tial land cover data are often derived from remote sensing or habitat mapping and land use can be modelled from this baseline in a wide range of ways. Given the impacts of LULC change on ES (i.e., through urban- isation, agricultural intensification, or eco- logical restoration), LULC data have obvi- ous value in understanding how ecosystem service flows are changing over time. Three common approaches are detailed below. First, “Lowry-type” models can quantify where the location of an attribute of interest (e.g., demographic data or recreational op- portunities) is a function of the attraction and travel costs associated with different locations (using, for example, the Rural Ur- ban Growth model (RUG) or the Ecosys- tem Service Mapping Tool (ESTIMAP)). Second, by assigning probabilities to transi- tions between land use types (for instance, Mapping Ecosystem Services128 50% of grassland will turn to forest), tran- sition probability approaches can project land use change into the future. These prob- abilities can themselves be driven by other spatial and/or scenario variables to produce more complex patterns of change. Third, state-and-transition models (STMs) are conceptual models that use simple, dia- grammatic approaches to address non-linear shifts in ecosystems in response to external environmental or anthropogenic disruption. State-and-transition models are typically created through a consultation process with experts and their diagrammatic approach makes them well suited to participatory work with stakeholders. A STM consists of a recognised number of possible states of an ecosystem and the factors driving tran- sitions between these states. Some, but not all, STMs are spatially explicit. Finally, agent-based modelling can also be used to understand how interactions be- tween groups of actors and their environ- ment (e.g., individual farmers or policy makers under changing environmental or socio-economic factors) lead to different LULC patterns. Such approaches allow LULC to evolve in response to the agents’ changing understanding of the environment as they adapt and learn. ABMs thus provide a powerful tool for ex- ploring emergent properties in LULC change. Biophysical models A large number of biophysical models ad- dress major environmental systems, includ- ing climatic, ecological, hydrological and geochemical models of key earth systems such as air, soil and water. Well-known examples include the Soil and Water Assessment Tool (SWAT), which can be used to assess water-related ES and the above-mentioned Revised Universal Soil Loss Equation (RUSLE). Such models tend to focus on a single aspect of the environment (such as the hydrologi- cal, soil, or biological subsystem) and may not be directly appropriate for assessing ES in their strictest sense. Often an addition- al modelling component will be needed to convert from a biophysical parameter (such as annual soil loss) to its ES (e.g., impacts of soil loss on drinking water quality), par- ticularly to connect these processes to their human beneficiaries. However, due to their long history many of these models are often trusted by environmental decision makers, sometimes making them more preferable than some more recent ES-specific tools. Modelling systems that explicitly focus on ES As interest in ES has grown, tools have been developed with an explicit focus on individu- al ES or suites of services. Some of these tools have been developed to be transferable across contexts whilst others are hard-wired into their local context. In the following sections, we discuss a number of these tools to illus- trate broad categories of available approaches and how they are used for ES assessment. Matrix-based approaches Matrix-based approaches sit on the border between ES mapping and modelling. They combine GIS (geographical information system) and spreadsheets analysis of LULC input data to produce maps of ES supply and/or demand. At their simplest, these are just mapping techniques (see Chapter 5.6.4 for a more complete description): they combine GIS LULC layers and scored val- ues for the provision of ES to provide ES provision maps across a study area. By us- ing standardised values, ES provision may be compared between regions or, by using Chapter 4 129 locally targeted ES values, more locally ap- propriate values can be generated. The pro- cess can be undertaken with stakeholders to allow maps of both ES supply and demand to be mapped. Additional GIS datasets can be included to improve the process—a pro- cess known as a multi-attribute lookup ta- ble—and these can be modified to reflect management scenarios. Matrix-based approaches can be applied with very limited technical expertise. How- ever, the more the matrix values rely on ex- pert knowledge rather than quantification with primary data, the more they are open to the critique of over-simplification and subjectivity, particularly when compared to primary data or more detailed modelling approaches. We include this technique here to stress that ES computational modelling need not always be complex. Transferable ES modelling frameworks A number of frameworks have been devel- oped with standardised methods designed to be transferred between contexts. Three of the most commonly applied modelling frameworks, InVEST, ESTIMAP and AR- IES are described below but there are nu- merous others (e.g., Co$ting nature, LUCI, MIMES; see Chapter 3.4) and still others which are under development. InVEST InVEST1 is a suite of modelling tools that provides a standard approach for applica- tion to varied contexts. InVEST includes 18 tools for assessing marine, coastal, terrestrial and freshwater ES. Each output is spatial- ly explicit and driven by user-input spatial datasets. Most InVEST models account for both ES supply and demand, in terms of the locations of people who would benefit from 1 http://www.naturalcapitalproject.org/invest/ these services; this allows supply-demand mismatches to be assessed. InVEST models are freely available and open-source. InVEST requires quantitative skills to be run; although experience with GIS is required to map outputs, coding is not required and the models can be run independently of specialised software (us- ing industry standard or open-source GIS platforms to visualise and prepare input and output data). Collecting and processing the datasets required can take time and effort. Each of InVEST’s tools is a separate model and can be run independently, depending only on the user’s needs. The outputs can be produced in biophysical units (i.e., tonne C km-2) or monetised, however the interactions between ES are not specifically modelled. InVEST has been used in a wide range of contexts including using ES metrics to as- sess sustainable coastal management in Be- lize, supporting decision-making over clean water supply in Latin America and national ecosystem planning in China. InVEST is a good example of a suite of tools that has gained ground through its use in multiple contexts. However, it does not yet assess the full range of ES and, like many biophysical models, is weaker on harder-to-assess cul- tural ES (see Chapter 6.6). ESTIMAP ESTIMAP is a collection of spatially explicit modelling approaches that assess the supply, demand and flow of ES. It is implemented within a GIS and is designed to be a stan- dardised, replicable system developed for use in the European Union (EU). It uses differ- ent methodologies for some ES and covers different ES than InVEST, focussing mainly on regulating ES (air quality regulation, pro- tection from soil erosion, water retention, pollination, habitat for birds, and recreation). Mapping Ecosystem Services130 Although the ESTIMAP approach was de- veloped to be applied at the EU scale for policy support, it is quite flexible and can be customised for application to local case stud- ies or specific problems in a way that is more difficult with InVEST’s pre-made models. ARIES ARtificial Intelligence for ES (ARIES) is a flexible modelling framework that uses AI to select the most appropriate modelling components (deterministic, probabilistic, or ABM) to map ES at context-appropriate scales. This approach moves away from the idea that one model should fit all circum- stances. The ARIES framework attempts to recognise the dynamism and complexity of environmental systems and balances this with the need for models that are simple enough to remain usable at a range of spatial scales and in a variety of contexts. ARIES has a strong focus on both the identification of beneficiaries and not oversimplifying ES to static values but instead focusing on dy- namic benefits that change with both space and time. It is cloud-based and semantic which allows diverse users to contribute data and models to a growing library that the AI system can select from, increasing its power and flexibility. In other ways, it shares many common attributes with InVEST (i.e., it is spatially explicit, open-source and produc- tion function-based). Integrated assessment models To combat the fact that many models are focussed on individual ES and may ignore or oversimplify key interactions, integrated assessment models have been developed that link sectoral models in a way that the out- puts of one are used as the inputs of anoth- er. This approach, though often technically challenging and time consuming to imple- ment, ensures that outputs have taken oth- er sectors into consideration in a way that comparing individual sectoral or ES models for the same scenario cannot. For example, an agricultural model may calculate water availability for irrigation based on rainfall, but without integrating a water allocation model that splits water availability between different sectors (e.g., irrigation, domes- tic supply, industry or power), it would be impossible to know whether that irrigation water was actually available for use. There are two main classes of integrated assessment models differentiated predomi- nantly by their application at global or re- gional scales. An example of each is illustrat- ed below. GLOBIO-ES GLOBIO-ES is an example of a dynamic global system model. It is a tool to assess past, present and future impacts of human activities on biodiversity and ES. Impacts on biodiversity are captured in terms of the bio- diversity indicator Mean Species Abundance (MSA) and ecosystem extent. Impacts on ES are included for 10 services. The model has been applied at both the national and global scales (see Chapter 5.7.3). GLOBIO-ES uses cause-effect relationships between environmental variables and ES identified by a literature review. It simulates future changes in ecosystem functions and services on a global scale. The methodolo- gy uses spatially explicit inputs on environ- mental drivers from the global climate and agriculture model IMAGE and the global land use model GLOBIO. The close link to the IMAGE-GLOBIO framework enables the assessment of inter- actions between human development (e.g., consumption patterns) and the natural envi- ronment (e.g., climate) based on key drivers like population growth, economic develop- ment, policy and governance, technology, Chapter 4 131 lifestyle and natural resource availability. The future directions of these drivers are quantified from different scenarios of future socio-economic developments. CLIMSAVE Integrated Assessment Platform (IAP) The CLIMSAVE IAP is an example of a re- gional integrated assessment model. It is a freely accessible web-based model that pro- vides options for ES assessment at a Europe- an scale. It is based on an integrated system of models for a number of different sectors including urban growth, freshwater, coast- al/fluvial flooding, biodiversity, agriculture and forestry. The model provides a number of output variables from the integrated models includ- ing indicators related to land use and a va- riety of ES. A wide selection of climate scenarios is in- cluded within the system as well as four stakeholder-defined socio-economic scenar- ios. The socio-economic input settings are able to be fully customised beyond the pre- set scenarios for a number of socio-economic drivers and adaptation options. This allows the IAP the ability to explore a very broad range of combined socio-economic and cli- mate scenarios to analyse their impacts on ES and allows adaptation options to be ex- plored. This enables ES synergies and trade- offs to be investigated at a European scale. MIMES (Multiscale Integrated Model of Ecosystem Services) MIMES is an integrated assessment system that models five distinct ‘spheres’: the litho- sphere, the hydrosphere, the atmosphere, the biosphere and the anthroposphere. Interac- tions between spheres are controlled using a matrix and ES are modelled by applying pro- duction functions that link ES to the system elements necessary to produce those services. Demand profiles, created for different societal groups are used to determine how environ- mental processes lead to production and use of ES. MIMES is also designed to assist in learning about system processes and the broad range of possible futures rather than provid- ing definitive maps of expected futures. How- ever, whereas CLIMSAVE uses interlinked process-based models, MIMES takes an ap- proach using production functions linked to an economic input-output model. Models to help with decision-making The ES concept provides decision-makers with a different way of looking at environ- mental management problems. A forest is no longer just a timber stock, but also a provider of climate regulation, habitat pro- vision, scenic beauty and recreation. Whilst this brings a broader lens to the value of eco- systems, it also brings new challenges: how do we decide which ES are more important? What are the implications if we choose to harvest the forest as timber? Modelling can help provide quantitative answers to many of these questions. In the following sections, we provide examples of how modelling can help decision-makers explore the implications of management al- ternatives. The line between previously men- tioned modelling tools and the decision sup- port elements, discussed below, is somewhat fuzzy and we recognise that previously men- tioned models and modelling tools can be integrated with the approaches that follow. Bayesian Belief Networks (BBNs) A Bayesian Belief Network is a type of mod- el that uses conditional probability to assign likelihoods to a suite of potential outputs given a known state of some or all of the in- puts (see Chapter 4.5 for more information about Bayesian Belief Networks). Mapping Ecosystem Services132 When applied to ES, BBN inputs are likely to be factors determining ES supply (such as land cover, soil types and other environmen- tal parameters) whilst the outputs will be ES supply, demand costs or benefits. BBNs have a number of advantages. First, they are very flexible in terms of the data that they can integrate. Both qualitative and quantitative values can be used, allow- ing them to be populated from field data, outputs from other models and expert opinion. They are also capable of integrat- ing more complex models within them. Second, if the conditional probabilities are not known, they can be inferred from exist- ing data using automated machine learning or a statistical approach. Third, their con- ditional probabilistic approach explicitly takes uncertainty into consideration so that neither inputs nor outputs are forci- bly treated as a deterministic single value. Fourth, BBNs can be embedded in a GIS or web-based platform to provide outputs that can be demonstrated spatially. Finally, they are well suited for exploring scenarios interactively with stakeholders as the mod- ification of inputs allows for a quick iden- tification of changing probabilities of the outcomes which can be performed directly with stakeholders. Multi-Criteria Decision Analysis (MCDA) Multi-criteria decision analysis is an umbrel- la term for a suite of flexible modelling ap- proaches designed to highlight the optimal choice in a situation with many decision alternatives. It breaks problems down into smaller components and analyses and values these relative to one another in terms of a number of consequences (e.g. costs, ecolog- ical and social impacts). When applied to ES assessment, MCDA can be used to evaluate trade-offs between multiple ES in a variety of different scenar- ios. MCDA is explicitly designed as a deci- sion support tool and has been used with both individual decision makers and groups of stakeholders to analyse preferences for different decision outcomes. Participatory modelling with stakeholders Modelling has traditionally been performed by experts in isolation from decision-makers and stakeholders. This has led to criticisms of elitism and has been shown to reduce stakeholder interest, understanding and trust in the modelling. Including stakehold- ers in the modelling process has, however, been demonstrated to increase the legit- imacy of the modelling in the eyes of the stakeholders. Furthermore, taking the stake- holder’s local knowledge into consideration often improves the quality of the modelling itself (see Chapter 4.6). In an ES context, the importance of partic- ular ES to local people can be paramount to their overall value. Participatory model- ling ensures that the modelling performed highlights ES that are of most importance to the local context rather than addressing a standard suite of service outputs that miss locally important ES. A “knowledge co-production” approach can be taken with any modelling approach which places interactions between the mod- eller and stakeholder on an even ground. Due to their iterative nature, such ap- proaches are often considerably more time- consuming. In fact, it may require modellers to develop entirely new models to address questions posed by stakeholders rather than the questions they pose themselves. This may mean that approaches which modellers would have planned to follow (e.g. expand- ing existing models) may not be appropriate for addressing stakeholder needs. Chapter 4 133 Considerations with modelling ES We conclude by discussing five general issues that should be considered when modelling ES. Which ES? Not all ES are as easy to model as others. In general, provisioning and regulating services have a longer history of being modelled than cultural ES. In fact, modelling cultural ES tends to be limited to analyses of services with relatively tangible physical aspects to their provision, such as recreation, tourism and, to some extent, aesthetic beauty. This is because factors with greater social or cultur- al meaning are considerably harder to tie to environmental parameters. It is in situations such as these that participatory approaches come to the fore (see Chapter 5.6.2 for a discussion on the use of participatory ap- proaches for mapping cultural ES). Care should be taken when interpreting model outputs as ES, as these outputs often represent proxies rather than the actual ES of interest. A clear example is carbon seques- tration which is often used as a proxy for the ES of climate regulation, but there are many others (e.g. the distance to locally accessible green space as a proxy for recreation pro- vision). It is very important to understand exactly what the output represents: is it evaluating the underlying ecosystem struc- ture and function only, or does it provide a direct benefit with concrete beneficiaries? Furthermore, does it quantify actual service provision (as directly used by beneficiaries) or potential ES provision (that could be tak- en up by the beneficiaries, if they had de- mand for and accessibility to the ecosystems supplying the service)? Whether ES supply or demand is modelled is another consideration. For some ES, both can be modelled and overlaid to identify mismatches between the two (e.g., air pollu- tion filtration by trees can be modelled using forest data and compared with a map of hu- man exposure to pollutant levels). However, it is often far better to apply another model that accounts for ES flows via service-spe- cific flow mechanisms, rather than to just identify in situ supply-demand mismatches. Though less commonly mentioned, it is of course possible with the same caveats to model ecosystem disservices or their con- verse – the natural benefits that control dis- services. Values How much is an ecosystem service worth? This is a key question in studies of ES – and can be a very loaded question. Model- ling studies are often capable of producing quantified outputs of ES (or their proxies) in biophysical (e.g., forest stock as tonne C/ ha) and monetary units (e.g., sale price of timber in £/$/€). However, value is a much more elusive concept particularly when weighing disparate services against one an- other. Questions such as “value for whom?” and “value as of when?” are key questions that also need to be considered by both modellers and those who use the outputs of models. This is because values are plural; they are not static and they vary depend- ing on which groups place a value on ES. However, models, particularly deterministic ones producing single outputs, do not usu- ally reflect these issues. This is particularly problematic for cultural services which are very socially determined, but even provi- sioning and regulating services will have different values in different social contexts in response to changing environmental, so- cio-economic or political factors such as a changing climate, political tensions, trade bans or new supply opportunities. Mapping Ecosystem Services134 Validation Validation is a key best practice in modelling; it is good modelling practice to test model validity against known data. In a statistical model, a measure of goodness of fit such as an R2 value in a regression or a kappa value for LULC classification can be used. However, to validate a model, it is necessary to know what the true values should have been. This is difficult for some ES, especially ones based on expert opinion and cultural services against which there are no objective values to test. This leaves such models more open to critique of their scientific robustness. Interpreting model results When dealing with models, it is important to remember that they (i) are man-made constructions, (ii) are just one way of access- ing information on the environment and (iii) need to be considered in context. It is easy to envision situations where decision- makers are led to the wrong conclusions if model outputs are taken as indisputable proof without understanding how well mod- el outputs represent the environmental issue in question, or because a modeller has ap- plied a pre-existing model to a new situation without adapting it to meet local conditions. The ES concept is designed to raise deci- sion-maker awareness of the benefits of- fered by nature. This decision-maker focus means that ES model developers need to be keenly aware of the implications of how their models are used. Uncertainty Uncertainty is a key aspect of model inter- pretation: how sure are we that the model output represents the real world phenome- non it seeks to quantify? There are multi- ple elements of uncertainty (see Chapter 6), for example: (i) to what extent do the in- put datasets used to train the model reflect the conditions for which they are intended (data uncertainty) (ii) to what extent does the model represent the processes that hap- pen in reality (model uncertainty) and (iii) for models forecasting the future, to what extent is that future likely to occur (scenario uncertainty)? Model validation is often used to address model uncertainty. Inter-model compari- son studies also reveal differences in outputs due to different model types. Probabilistic approaches and sensitivity analysis can also be used to address scenario and input data uncertainty by exploring the influence of input parameter changes on model outputs by performing multiple runs and identify- ing overall patterns. It is, however, rare that the full holistic uncertainty (that addresses all these factors) is addressed. A validation statistic may be produced that says, for example, “this model explains 80% of the variation in the dataset we tested it against,” but this provides no information about the confidence in this dataset (was it randomly sampled, or taken from locations easily ac- cessible by monitoring teams?); the factors within the model that provide the modeller with confidence in the approach taken (e.g., are there any subjectively selected adjust- ment factors?); or, the pragmatic factors such as time, expertise and funding that shaped the model development. We stress this because it is critically im- portant that the context of the modelling is considered when interpreting its outputs for decision making. This is not to say that models are any more inherently flawed than any other way of understanding the envi- ronment; there will be some models, partic- ularly those driven strongly by physical laws that can reliably and repeatedly reproduce real-world outcomes. We simply stress that models are simplifications of reality and should be interpreted with care. Whenever Chapter 4 135 possible, model interpretation should take place with the assistance of the modeller (or someone who understands the model) and local stakeholders who understand the con- text of its application. Conclusions Modelling is being widely applied in the field of ES. There are a large number of mod- elling approaches and a wide range of exist- ing models that can be used for ES assess- ment. Modelling has considerable potential to evaluate both the ecosystem structure and function underlying ES and the supply and demand for ES themselves. Furthermore, modelling provides the potential to explore the impacts of environmental change and management on the future provision of ES through scenarios, making them vital tools for ES decision support. Disclaimer Any use of trade, product, or firm names is for descriptive purposes only and does not imply endorsement by the U.S. or any other Gov- ernment or by the authors of this article. Further reading Bagstad KJ, Semmens DJ, Waage S, Winthrop R (2013) A comparative assessment of de- cision-support tools for ecosystem services quantification and valuation. Ecosystem Services 5: 27-39. Christin ZL, Bagstad KJ, Verdone MA (2016) A decision framework for identifying mod- els to estimate forest ecosystem service gains from restoration. Forest Ecosystems 3: 3. Dunford RW, Smith A, Harrison PA, Hanganu D (2015) Ecosystem Services in a changing Europe: adapting to the impacts of com- bined climate and socio-economic change. Landscape Ecology 30 (3): 443-461. IPBES (2016) Summary for policymakers of the methodological assessment of scenar- ios and models of biodiversity and eco- system services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. Ferrier S, Ninan KN, Leadley P, Alkemade R, Acosta LA, Akçakaya HR, Brotons L, Cheung W, Christensen V, Harhash KA, Kabu- bo-Mariara J, Lundquist C, Obersteiner M, Pereira H, Peterson G, Pichs-Madruga R, Ravindranath NH, Rondinini C, Win- tle B (Eds.) Secretariat of the Intergovern- mental Science-Policy Platform on Bio- diversity and Ecosystem Services, Bonn, Germany, 32 pp. Schröter M, Remme RP, Sumarga E, Barton DN, Hein L (2015) Lessons learned for spatial modelling of ecosystem services in support of ecosystem accounting. Ecosys- tem Services 13: 64-69. Wainright J, Mulligan M (Eds) Environmen- tal Modelling: Finding Simplicity in Com- plexity, 2nd Edition, Wiley, 494 pp. Mapping Ecosystem Services136 4.5. Bayesian belief networks Dries Landuyt, Adrienne Grêt-Regamey & Roy Haines-Young Introduction The complexity of natural systems and the interactions between nature and society im- pedes the use of state-of-the-art, data-driv- en, process-based techniques for ecosystem service (ES) modelling. Instead, simplified, pragmatic approaches can be used to pro- vide initial estimates of ecosystem service delivery. Although simplification leads to an increase in model output uncertainty, many modelling approaches, however com- plex, often do not take uncertainties into account. Despite their apparent simplicity, Bayesian Belief Networks (BBN) do take uncertainty into account and, as a result, are worthy of attention. Bayesian Belief Network models are graph- ical probabilistic models that conceptualise the system being represented as a chain of causal relations, visualised as a Directed Acyclic Graph (DAG). Such a graph con- sists of nodes that represent the system’s variables and arrows that represent causal relations amongst them. Variables are typ- ically discrete and relations amongst them are quantified through probabilistic rules, captured as conditional probability distri- butions. These distributions can be derived from data, from expert knowledge or a com- bination of both. An example BBN that enables an analysis of how our estimate of wood production would change, given information about land use and soil type, is provided in Figure 1. The first step of the model development process consists of selecting suitable variables and defining putative causal relations. By as- suming that land use and soil type are the most important drivers that determine the production of wood, the model’s variables are restricted to ‘soil suitability’, ’land use’ and ‘wood production’. By assuming that soil type and land use both influence wood production and that both variables are in- dependent, the structure of the graph is de- fined (Figure 1). To implement the model, probability distributions need to be defined: unconditional ones for the input nodes, conditional ones for the others. By combin- ing the information captured in the model’s conditional probability tables (CPTs) with the initial probability distributions of the network’s input nodes, probability distribu- tions for other nodes can be calculated based on Bayes’ theorem, which describes the con- ditional probability of an event. There are a number of software tools available that enable users to make these calculations au- tomatically. The calculated probability dis- tributions are represented as so-called belief bars in the model (Figure 1). The application of BBNs generally consists of inserting new information or evidence in one or more nodes of the model and, subse- quently, analysing the resulting belief chang- es. This new information can be determin- istic or probabilistic depending on whether the information implies that a state is exactly known or not. Figure 2 provides two exam- ples of inserting deterministic evidence in Chapter 4 137 the model that was introduced in Figure 1. When evidence is inserted in one of the in- put variables of the model, the model will be run in predictive mode and will predict ef- fects of input changes (Figure 2a). Knowing that the soil suitability is high, our belief in high wood production will increase substan- tially. Our belief in the ‘zero’ state, however, will not change, as we still know that 10% (this information has not changed) of the forests are reserves and, thus, do not produce wood. When evidence is inserted in the out- put node, models are run in diagnostic mode and will predict causes instead of effects (Fig- ure 2b). If we know that wood production is high, we can infer that soil suitability will be high with a high probability. Moreover, based on the inserted information, we can infer that the forest stand being considered is definitely not a forest reserve. Strengths and weaknesses Although BBN models have been used since the 1980s, applications were restrict- ed to medical diagnosis, where BBNs were used to combine probabilistic information on disease occurrence with probabilistic Figure 1. A model example illustrating the structural components of a Bayesian Belief Network: (a) the directed acyclic graph (DAG) and (b) the conditional probability tables (CPT). Figure 2. Predictive and diagnostic belief updating in a Bayesian Belief Network model. Mapping Ecosystem Services138 information on symptom development to support the process of reaching a diagnosis. Late in the 1990s, BBNs were introduced in the environmental modelling domain, predominantly because of their ability to ex- plicitly account for uncertainty, an import- ant aspect when natural processes are being modelled. A second important reason for their adoption was the potential of the tech- nique to integrate expert knowledge in the modelling process. Expert knowledge can be used to develop the network structure or to populate the model’s CPTs with sub- jective probabilities, which are also referred to as beliefs. This functionality is especially useful in case variables need to be included for which no supporting data are available, a frequently occurring problem in ecosystem service modelling. A final strength of the modelling technique is its graphical nature. Due to this feature, BBNs are transparent models that are relatively easy to grasp. This means that non-expert stakeholders can be involved in model development. Another important advantage in the context of ecosystem service modelling is that BBNs fit extremely well in the ‘ecosystem services cascade’ which has been used as a basis for many ecosystem service studies (see Chapter 2.3) (Figure 3). The idea that ecosystem ben- efits are generated through services, services through functions and functions through the biophysical structure of the environment, closely resembles a chain of causal relations which can be easily modelled in a BBN. Although a linear representation of the eco- system service production process might facilitate system understanding, in reality most ecosystem service delivery processes are non-linear and involve a range of feed- backs which BBNs cannot easily take into account. Developing several models for successive time steps of a system and chain- ing them afterwards is a ‘workaround’ that mimics feedbacks with BBNs. Such time- sliced models, however, often become very complex and lack transparency. Another im- Figure 3. General Bayesian Belief Network structure for ecosystem service modelling. Chapter 4 139 portant drawback of BBN is the obligation to discretise continuous variables so that they can be represented as having states in a node. To minimise information loss discreti- sation methods need to be chosen carefully. Nevertheless, for some applications, nodes with discrete states are more easily under- stood than continuous variables. If discreti- sation thresholds are chosen in accordance with the aim of the model, discretisation does not necessarily lead to information loss. Where information loss is not accept- able, more complex software packages are available that enable the use of continuous variables in BBNs. The drawbacks discussed above all suggest that BBNs are less suitable for modelling compli- cated processes than some other approaches. Thus, for well-studied services, the added val- ue of using BBNs instead of process-based, validated models is low. The real strength of BBNs lies, however, within the integration of well- and poorly-studied services in one inte- grated model. Such integrated models might deliver additional insights into trade-offs and synergies among services. In addition, their graphical nature can facilitate stakeholder in- volvement and social learning, two objectives that are difficult to attain using conventional modelling approaches. Model development guidelines To determine which variables need to be in- cluded in the model, a variety of knowledge sources can be used. Domain experts can be consulted to select variables that are import- ant for biophysical modelling of service pro- vision, while stakeholders can be consulted to include social interests, i.e. the ecosystem services that are considered important in the study area, or the values associated with them. The type of endpoint being modelled is also an important aspect to consider when including ES in the model. These endpoints can be biophysical quantities, scores that represent social values or monetary values of generated benefits. Modelling the full ES cas- cade, up to the final benefits, can be attained by integrating studies from different research fields such as economics and sociology. To select input variables, two things need to be considered, namely the management options that need to be evaluated by the model and whether spatial application of the model is desired. For spatial modelling applications, spatial data on the model’s input nodes or on their proxies need to be available. To make management evaluation possible, all variables that are influenced through management and that impact eco- system service delivery need to be included. Discretisation of continuous variables is the next step in the model development process. In general, the number of states needs to be kept as low as possible. As the number of states directly affects the com- plexity of the model’s CPTs, less effort needs to be invested in CPT quantification in case the number of states is kept low. Thus, a balance needs to be found between reducing model complexity and minimis- ing information loss. Additionally, in case CPTs are learned from data, the number of states needs to be restricted to ensure that sufficient information is available for all state combinations. To develop the structure of the model, ex- perts are often consulted. To integrate local knowledge in the model structure, stakehold- ers can also be consulted. As stakeholders and experts are generally not aware of a model’s technical restrictions (e.g. the fact that feed- back loops cannot be included), modellers need to guide the model development pro- cess and, if necessary, adjust the structure af- terwards. Although data can be used to create model structures and estimate probabilities Mapping Ecosystem Services140 using learning algorithms, relations that are defined through this process are not necessar- ily a result of causality and, therefore, they are sometimes hard to interpret. Quantification of CPTs is the final step to- wards implementing the model. A broad range of knowledge sources can be consult- ed for this, including expert and stakeholder knowledge, empirical equations, simulations with existing models, literature data and field data. Although data might seem the most objective way to quantify a CPT, data- sets are often not sufficient to fully quantify a model’s CPTs. In these situations, experts can be consulted to define prior CPTs and data can be used to update CPTs; this is a typical Bayesian workflow. Aside from CPT quantification, the probability tables of the input nodes also need to be quantified. If input nodes represent spatial variables, his- tograms of spatial datasets can be used to populate these probability tables. To increase the credibility of a Bayesian Be- lief Network, model validation needs to be performed. To evaluate a model’s predictive performance, a broad range of validation metrics are available, similar to those ex- tensively used in other modelling domains. The predictive performance of BBNs, how- ever, is generally low compared to other techniques. While most models only fo- cus on performing one specific task opti- mally, BBNs try to approximate the joint probability distribution over all variables, mostly at the expense of their predictive performance. Predictive performance is, therefore, generally not the most import- ant aim of a BBN model, especially in the field of ecosystem service modelling. Other evaluation criteria include the ability of the model to describe a system, to enhance so- cial learning and to facilitate decision-mak- ing. To evaluate those aspects, evaluation through experts and stakeholders might be more appropriate. The consulted experts and stakeholders for model evaluation are preferably not those consulted during model development. To perform the above tasks, a range of soft- ware packages are available. Frequently used software packages in the ecosystem services modelling domain are ‘Netica’ and ‘Hugin’. They both provide a user-friendly graphical user interface for model development that can potentially be used with stakeholders. Most packages also include algorithms that can be used to train and validate models us- ing existing datasets. Furthermore, through application programming interfaces (API), software packages can be extended with all kinds of tools. Following this approach, BBNs can, for example, be coupled to geo- graphical information systems (GIS). This is an important functionality when BBNs are used for ecosystem service mapping. Conclusions As illustrated in this chapter, BBNs have much potential for modelling and mapping ES. They operate at an intermediate level of complexity which makes them especially useful where the volumes of available data and knowledge are not sufficient for empir- ical or process-based modelling. Additional- ly, BBNs are useful tools to help structure the available knowledge into comprehen- sible ways that can support social learning and stakeholder participation in ecosystem service modelling and management studies. Further reading Cain J (2001) Planning improvements in nat- ural resource management. Guidelines for using Bayesian networks to support the planning and management of develop- Chapter 4 141 ment programmes in the water sector and beyond. Wallingford: Centre for Ecology and Hydrology. Haines-Young R (2011) Exploring ecosystem service issues across diverse knowledge domains using Bayesian Belief Networks. Progress in Physical Geography 35(5): 681-699. Landuyt D, Van der Biest K, Broekx S, Staes J, Meire P, Goethals PLM (2015) A GIS plug-in for Bayesian belief networks: To- wards a transparent software framework to assess and visualise uncertainties in eco- system service mapping. Environmental Modelling & Software 71: 30-38. Landuyt D, Broekx S, D’hondt R, Engelen G, Aertsens J, Goethals PLM (2013) A review of Bayesian belief networks in ecosystem service modelling. Environmental Model- ling & Software 46(0): 1-11. Potschin M, Haines-Young R (2016) Defin- ing and measuring ecosystem services. In: Potschin M, Haines-Young R, Fish R, Turner RK (Eds) Routledge Handbook of Ecosystem Services. Routledge, London and New York, 25-44. Glossary BBN Node Graphical representation of the system variables in a Bayesian Belief Network model. State A value, discrete class or qualitative level to which a variable can be assigned. Each variable in a Bayesian Belief Network model has a set of states it can manifest. Probability distribution The set of probabilities assigned to the states of a variable that express the prob- ability that the variable equals one of its states. This set of probabilities always sums up to 1. Arrow Graphical representation of the causal re- lations amongst the system variables in a Bayesian Belief Network. Each arrow flows from a parent node to a child node. Conditional probability The probabilities that quantify the model’s causal relations and express the probability distribution of a child node given the sta- tus of its parent nodes. Conditional probability table or CPT A table that contains probability distribu- tions over a node’s states conditional on all possible combinations of its incoming nodes’ states. Directed acyclic graph or DAG Graphical representation of the system being modelled by means of nodes that represent system variables and arrows that represent causal relations among the sys- tem variables. Mapping Ecosystem Services142 4.6. Applying expert knowledge for ecosystem services- quantification Sander Jacobs & Benjamin Burkhard Ecosystem services (ES) are a complex field of study. The application in practice poses several additional challenges. Although ES quantifications can built on existing experi- ence, methods and data (see Chapters 4.1- 4.5), specific human-environmental system settings, policy frameworks and characteristic ES need to be considered thoroughly. Expert involvement can provide information in cas- es where other sources are lacking, efficient- ly generating results and validating maps. Moreover, structural expert involvement in trans-disciplinary projects can improve effec- tiveness of projects which are geared at real world impact. This chapter provides basic considerations on expert involvement and puts forward some guidelines to tackle chal- lenges related to trans-disciplinary mapping. Why experts? Expectations towards ES science and appli- cation are very high. The global socio-eco- logical challenges which researchers are aim- ing to tackle are both urgent and important. Still, the amount of trust and public re- sources going to ES studies and mapping is relatively high compared to their current impact on solving real world problems. Applied ecology and sociology are con- sidered complex fields, combining several disciplinary frameworks, ways of thinking and related methods. ES, at the crossroads of applied ecology, economy, sustainability science and social sciences, can be defined as “super-complex”. Super-complex or so- called wicked problems require engagement of several theoretical disciplines and practi- tioners in actual implementation from the very onset of the problem-solving process. What looks like just a simple ES map is often a complex combination of selected quantitative data, proxies and expert esti- mates, qualitative judgements, theoretical assumptions, technical choices and commu- nicative visual goals (see Chapters 3.3 and 6.4). The quality of the actual mapping pro- cess directly determines the qualities of the map in all its aspects (credibility, relevance, clarity, usefulness; see Chapter 5.4). Creat- ing a map which lives up even to minimal real world application ambitions obliges the involvement of ‘experts’ to legitimise, clari- fy, improve and validate maps to be relevant for any specific application context. What makes an expert? Delineating who is an expert and who is not is not straightforward. From the above, it is clear that, when solving real world prob- lems, merits of diploma and discipline are not enough by far. A bright GIS technician is certainly a required expert, but without complementary input from the ecology ex- pert, the modeller, the economist and the social scientist, there is actually nothing to map or to interpret. Also, without the local Chapter 4 143 or topical experts to put the socio-economic and natural science theories into a specif- ic context, maps will be hard to validate. Moreover, without the expert who connects specific policy demands, cultures and know- how, implementation of the maps into actu- al solution strategies will rarely happen. And finally, deciding on societal importance of issues or values of specific ES to decide upon trade-offs requires input from policy makers and/or the direct end-users of these services. All these types of knowledge are indispens- able for the mapping process, and not nec- essarily related to education level or strictly technical skills. The central idea is that all experts - or knowledge-holders - need to be thoughtfully engaged. Selective expert engagement From the point of view of a technical map- ping project, involving experts is often re- garded as costly, tedious and complicated. We will show that structural expert involve- ment will add value to the whole process of map creation and effective problem-solving. Three examples of selective engagement are discussed here. The section, following this discussion, returns to address the more profound expert engagement of trans-disci- plinary research. 1 . Experts plaster the holes in your data The most commonly heard argument for en- gaging experts is to provide ‘educated guess- es’ and estimates of ES supply, locations or contexts where a given dataset or model is not providing quantitative information. In- deed, this is a highly effective way of filling in missing data to obtain a dataset which al- lows the creation of a map. The explicit as- sumption is that these estimates are ‘second choice’ and ‘less reliable’, and best replaced by model outputs as soon as these become available. Note that this technical argument disregards the fact that quantitative models (see Chapter 4.4) have originally been com- piled and designed by experts. Often they are applied/extrapolated to another context by implementing expert-based modifications and assumptions. In addition, many aspects of ES mapping are simply not quantitative in the natural science sense: economic data, val- uations, ecological quality estimates - they are all based to a large extent on qualitative expert estimates. Collaboration among diverse and multiple experts from the onset could help to avoid the disciplinary bias of the experts that happen to steer the mapping process. 2 . Experts generate quick results A second pragmatic reason to involve experts is that they provide quick access to a broad range of knowledge and comparable ES maps can be obtained in a relatively cost-efficient way. Indeed, with a minimum of resources, maps can be obtained, with known reliabil- ity and high credibility (provided that some basic rules are followed concerning which experts to select, the representativeness of this selection and how to evaluate expertise levels). A process model-based quantification (tier 3; see Chapter 5.6.1) does not necessar- ily deliver more useful or ‘true’ results than a tier 1 (expert-based relative scoring) or a tier 2 quantification. In an optimum case, sever- al approaches (tiers) can be applied for the same ES in one region and the results can be triangulated in order to cross-validate and in- crease reliability. There is a risk that an overly pragmatic approach ignores existing data and models already available. In addition, in- volved experts are frequently frustrated when the highly detailed and complex knowl- edge they hold is reduced, for example, to a comparable scoring format for predefined indicators. Much more potential lies in the combination and comparison of diverse ap- proaches from different mapping tiers (see Chapter 5.6.1) and quantification methods (see Chapters 4.1-4.4), from the start. Mapping Ecosystem Services144 3 . Experts fix your credibility A third common application of expert en- gagement is ensuring the local or topical validity of the maps created. This concerns local ecological knowledge or elicitation of societal values, but it can also entail spatial validation and adaptation of resulting maps. Although the type of validation can vary, this step is essential for any map which is meant to provide reliable and credible input to decision-making. The difficulty with such methods and relat- ed results is that these often do not come in before the end of a study. Experts are con- fronted with an end-product which is not always part of a clear process or linked to a recognisable problem. Maps represent high- ly complex and variable data types, combi- nations and technical choices in a single, static 2D representation (see Chapter 3.2). Apart from assessing the overall plausibility of the result and ‘recolouring’ local correc- tions, information to (re)calibrate models or assess credibility of assumptions made is very hard to obtain. Moreover, if a map turns out not to be useful at all, it is often far too late to change course. A stakeholder analysis, a knowledge-needs inventory and an engagement strategy at the start of an ES mapping project allows the involvement of key experts (including local/ topical experts) and guarantee validation and credibility in order to develop an effec- tive map product. All three selection-perspectives are prag- matic and instrumental to improve quali- ty, efficiency and effectiveness of mapping projects. Still, these perspectives regard the mappers as project owners, mandated to se- lect ‘other experts’ for a certain purpose and within a restricted window of engagement. In the next section, we show that a trans-dis- ciplinary approach not only combines the advantages mentioned above, but provides additional benefits for the effectiveness of a mapping project. Structural engagement of experts Mapping ES in the context of real world problem-solving needs to go further. Struc- tural engagement of experts departs from a different paradigm. The underlying prin- ciple is that there is no de facto distinction between experts and laymen, or between stakeholders and researchers. All people in- volved in, or potentially affected by, the ES mapping project are stakeholders as well as experts in a certain aspect. Such a trans-disciplinary viewpoint has two immediate consequences: first, the researchers mandated to perform the mapping project depart from a humble attitude (see Chapter 5.4). Second, experts/stakeholders outside of the actual project team are ‘promoted’ to the level of potentially indispensable knowl- edge-holders and project-owners. These in- clude people commissioning the project, top- ical experts on certain ES, technical experts on different methods, experts on local or themat- ic context into which the mapping project is framed and people actually depending on ES. The above does not mean, of course, that every mapping project should involve large numbers of experts throughout the project in order to be effective. The actual number of experts is not the issue here, but it is their competence, diversity, qualification and role they have in the project. In the following sec- tion, a theoretical illustration of a mapping project’s cycle is presented. This example imagines an ideal project without issues of policy restrictions or budgetary constraints. 1 . Scoping This first phase sets out clear project goals, adding requirements and conditions for Chapter 4 145 well-defined final map products as well as concerning inclusion of various viewpoints in the process. A broad and realistic selec- tion of experts is made to join the project team and co-design, conduct, steer and eval- uate the mapping project. Questions to answer: Why is the project needed? Which problem needs to be solved? Who are the end-users of the maps? What are the maps going to be used for exactly? Who will be affected by the envisioned solution? How dependent are different people/groups on the human-envi- ronmental system, how large is the potential impact on their well-being? What power or representation do they have, to what extent can they govern their own environment? Expertise needed to answer these questions: – Experts from policy and administration commissioning the project; – Experts from the end-user side con- cerning format and requirements of the map (see Chapter 5.4); – Technical expertise on policy and defin- ing client demand for product develop- ment; – Experts on various stakeholder points of view, directly or by representatives (e.g. NGOs); – Technical expertise on stakeholder anal- ysis and participation of special groups. 2 . Method selection and project design This phase develops an agreed-upon work plan, project governance structure and workload distribution. Questions to answer: What methods and data do we need to create the product? What methods and know-how do we need to set up the process accordingly? Expertise needed to answer these questions: – Experts from different disciplinary fields; – Technical mapping experts; – Specialist experts on detailed sub-topics (e.g. certain ES, habitats, land use prac- tices, stakeholder groups); – End-user experts to follow up on map usability; – Policy experts to follow up on relevance; – Stakeholder representation to follow up on different goals and conditions; – Technical expertise to design and facili- tate participation and feedback process between product developers, end-users, commissioning bodies and stakeholders. 3 . Creating reliable maps This phase produces maps with transparent reliability, conscious decisions affecting in- terpretation and best available knowledge, while safeguarding purpose, usability and local/thematic specificities. Questions to answer: How can we include and combine various data types? How can we determine reliabili- ty of different types of data and knowledge? How can we select data and communicate reliability? How do we make technical choices which impact the outcome (e.g. in- terpretation of maps)? Expertise needed to answer these questions: – All experts and stakeholders need to reach agreement on choices concerning reliability within the particular project; – Different experts on similar topics need to triangulate and cross-validate meth- ods and results; Technical experts need to design and facili- tate efficient decision processes and commu- nicate decisions. 4 . Implementation of the maps This phase ensures effective implementa- tion of the products as well as adherence to the agreed goals. Ideally, this phase runs throughout the project, in order to test early versions of the maps and adapt methods (or goals) based on these tests. Mapping Ecosystem Services146 Questions to answer: How can we ensure effective application of the maps in the envisaged solution/instrument? How can we evaluate distance to target? Expertise needed to answer these questions: – All experts and stakeholders need to agree on engagement in implementa- tion and criteria for evaluation; – End-user experts need to test applica- tion and provide feedback. Solutions and recommendations • Clear goals. Being effective requires the right product, produced in the right way. Clearly formulated goals are essential. • Diversity. The best people should be identified with the diverse skills and knowledge types needed. Consider them equal regardless of their diplomas and promote this attitude. • Facilitation. Do not think that a trans-disciplinary process will run itself. Project facilitation is a skill, and skilled people will be needed to keep the pro- cess running smoothly. • Parsimony. Do not overdo it. Weigh costs and efforts against stakes. Be pragmatic when needed, but without forsaking the project goals. Adapt unre- alistic goals to more realistic objectives. • Testing and evaluation. • Do not expect that your team will pro- duce a perfect product at the end of the project. Look for the weaknesses in the project and address them. Test maps as soon as possible and avoid the trap of self-evaluation. The sooner a weakness or failure is identified, the greater chance there will be of finalising your project with a high level of success and impact. Further reading Bradshaw GA, Borchers JG (2000) Uncer- tainty as information: narrowing the sci- ence-policy gap. Conservation Ecology 4(1): 7. Cornell S, Berkhout F, Tuinstra W, Tàbarae JD, Jäger J, Chabay I, de Wit B, Langlais R, Mills D, Moll P, Otto IM, Petersen A, Pohl C, van Kerkhoff L (2013) Opening up knowledge systems for better responses to global environmental change. Environ- mental Science & Policy 28: 60-70. Drescher M, Perera AH, Johnson CJ, Buse LJ, Drew CA, Burgman MA (2013) Toward rigorous use of expert knowledge in eco- logical research. Ecosphere 4: 1-26. Gunderson LH, Holling CS (Eds.) (2002) Panarchy. Island Press, Washington Covelo London. Hay I (2010) Qualitative Research Methods in Human Geography. 3rd Edition. Ox- ford University Press. Jacobs S, Burkhard B, Van Daele T, Staes J, Schneiders A (2015) The Matrix Reload- ed: A review of expert knowledge use for mapping ecosystem services. Ecological Modelling 295: 21-30. Seidl R et al. (2013) Science with society in the anthropocene. Ambio 42(1): 5-12. Voinov A, Seppelt R, Reis S, Nabel JEMS, Shokravi S (2014) Values in socio-environ- mental modelling: persuasion for action or excuse for inaction. Environmental Mod- elling and Software 53: 207-212. Chapter 5 147 CHAPTER 5 Ecosystem services mapping Mapping Ecosystem Services148 Chapter 5 149 5.1. What to map? Ralf-Uwe Syrbe, Matthias Schröter, Karsten Grunewald, Ulrich Walz & Benjamin Burkhard Introduction Ecosystem services (ES) originated as a con- cept that reflects the value of nature for hu- mans and provides additional reasons for protection and sustainable management of ecosystems (see Chapter 2.3). Many ES face spatially explicit pressures or rely on anthro- pogenic contributions such as technology, energy or knowledge. ES maps can help to uncover risks for ecosystem health, unsus- tainable use of potentials to provide a service, harmful impacts on a landscape, impaired spatial flows of ES as well as mismatches be- tween ES supply and demand (see Chapter 5.2). Such information can indicate where to improve ES provision and where to prioritise nature and biodiversity conservation. Multiple components play a role in ES pro- vision and use which can be mapped, as- sessed and monitored. ES can be mapped and assessed using quantitative indicators or qualitative estimations. ES mapping and as- sessment include ecosystem properties and conditions, ES potential, ES supply, ES flow and ES demand which we generically define in the next sections. ES mapping terms and their relationships The framework presented here aims to depict different aspects of ES important for map- ping. Our framework bridges variously in- terconnected ecosystems and socio-economic systems, including the interactions between their components. Figure 1 highlights aspects of ES which can be considered relevant for mapping. ES are generated in the context of different aspects or components, which are interrelated, but can be mapped separately. Figure 1. Mapping aspects of ES (own illustration, adapted version of the the ES cascade by Haines- Young & Potschin (see Chapter 2.3), Wolff et al. 2015, Bastian et al. 2013). Bold grey: subjects relevant for mapping; dashed: may be mapped; thin: additional aspects for which mapping could be developed. Ecosystem Ecosystem properties and conditions ES supply ES potential Socio-economic system ES demand Benets Human inputs Flow Mapping Ecosystem Services150 Ecosystem properties and conditions provide the ecological basis for ES potentials which, together with human inputs, form a capac- ity of a social-ecological system to provide ES (ES supply). ES flows (i.e. the actual use of ES) can be a fraction of this supply, or be higher in case stocks are depleted or ecosys- tems are unsustainably used. Demand for ES steers ES flows, i.e. without a demand for a service, there is no actual use. This demand can, however, be higher than actual flow, for example, in cases where societal preferences for specific services remain unsatisfied. With- in the socio-economic system, benefits arise from several kinds of ES use depending on the demands of concerned people. Feedbacks from the socio-economic system such as land use change, landscape maintenance or envi- ronmental pressures, affect the ecosystem and thereby the ES supply. The following sections explain these terms in detail. Ecosystem properties and conditions Definition: Properties describe the charac- ter, structure and processes of an ecosystem. Conditions refer to the integrity and health status of an ecosystem which determine its ability to generate ES (see Chapter 3.5). Land use or land cover provide the basis of many ES maps. Beyond that, ecosystem properties such as soil type, slope gradient and inclination, climate conditions and the position in relation to a shoreline or within a watershed are properties that essentially control the supply of many ES. Features of landscape structure like density of certain objects, edge conditions, connection and shape of areas can also be very important. Ecosystem conditions, however, comprise much more: for instance, the load of pollut- ants, species composition and health may be crucial preconditions for ES. Delimitation: Properties and condition re- flect both the natural ecosystem state and the type of ecosystem as result of a specific land use. Since the condition for ES supply differs between specific ES, the scope of related assessments has to be defined very carefully per ES. Necessity and applicability: Indicators for ecosystem properties and conditions should be applied to different protection goods or land use classes. They are relevant because they provide the spatial and physical precon- ditions for ES (see Chapter 2.2). ES poten- tials can, for example, give a reference point for planning and scenarios (see Chapter 7.2). Both the individual patches’ land use and land cover and the configuration and ar- rangement of such patches, are important for ES supply. Therefore, the landscape structure with its mosaic of patches should be consid- ered (see Chapter 5.2). Possible indicators: Land cover can provide an essential database for ES mapping. The CORINE land cover dataset is often used in European studies (see Chapter 3.5). At na- tional level, land use data from land survey or habitat mapping often are available. Ad- ditional data need to be integrated in more detailed evaluations (see Example 1). Ecosystem properties and conditions are di- rectly linked to the state of biodiversity. A high level of biodiversity – in most cases – underpins the supply of multiple ES (see Chapter 2.2). ES potential Definition: ES potential describes the nat- ural contributions to ES generation. ES ca- pacity is often used synonymously. ES po- tential measures the amount of ES that can be provided or used in a sustainable way in Chapter 5 151 Example 1 . Wood-dominated ecotones and non-fragmented forests Large contiguous areas of woodland are vital for nature protection by offering habitats for animals and plants and provide people with areas for relaxation. The size of uninterrupted woodland, not dissected by roads and railways, is an important criterion for ecosystem conditions. Ecotones are transitional areas between habitats. As such, they are home to a particularly rich variety of species, not only those of the adjacent communities but also species that have become specialised to the ecotone itself. In open landscapes, such elements are important as habitat for pollinating insects and for other beneficial organisms. At the same time, a landscape with high proportions of such elements is very attractive for human recreation. In this context, landscape configuration with ecotones is an indicator for ecosystem condition. The calculation of the perimeter of forest-dominated ecotones takes account of all hedges, tree rows and the margins of small copses as well as all forest margins (see Walz 2015). Mapping Ecosystem Services152 a certain region given current land use and ecosystem properties and conditions. It is recommended to regard this potential for a sufficiently long time period. Delimitation: The (natural) ES potential is often supplemented by human system in- puts to generate ES supply (see Section Hu- man inputs). The actual provision (co-pro- duction) of ES (flow) sometimes includes large human efforts, is strongly dependent on technological refinement and can be very difficult to determine. Necessity and applicability: In terms of ES potential, the ecological carrying capacity and resilience need to be considered. ES potential allows the distinction between a realised ES and the opportunities and limits of use which is often meaningful for planning purposes, scenarios and management issues. Some- times, an indicator for ES potential can help to better understand and calculate physical indicators for regulating ES supply. Possible indicators are, for example, the Muencheberg Soil Quality Rating (SQR), metrics for relief diversity and the share of water bodies as part of landscape aesthetics as well as proxies for processes such as ground- water recharge rates. ES potential is particularly applicable for planning, management and predictive re- search purposes. Since it is conceptualised hypothetically and for the long term, ES potential should not be assessed for short time periods (such as for only one season). Preferably, ES potential should be orien- tated on natural regeneration rates. Direct human interventions such as fertilisation, technical energy inputs or breeding and genetic engineering should not be consid- ered as contribution to ES potentials. In contrast, land use type (grassland, field, forest, settlement) and the consequences of long-lasting or very strong impacts such as mining have to be considered naturally. A distinction of a real ‘natural’ state that con- tributes to ES is not straightforward. ES supply Definition: Supply is the provision of a ser- vice by a particular ecosystem, irrespective of its actual use. It can be determined for a specified period of time (such as a year) in the present, past, or future. Delimitation: The amount of ES supply depends on natural conditions and often on human inputs (see below), such as land management contributions, knowledge and technology. Though there are some ES with- out human co-production, they may never- theless depend on ecosystem preservation. ES supply also includes stocks of natural assets as starting points of the flows of ma- terial, energy, information and organisms as results of both ecosystem potential and hu- man co-production. Necessity and applicability: ES supply is a central subject to be mapped and can be Example 2 . Crop potential To indicate the gross potential of crop produc- tion, the Natural Yield Potential from the Soil Atlas of Saxony, Germany was used. Compa- rable maps are available for most countries of the world. In a two-stage procedure, first the soil fertility was assessed using field capacity, capillary moisture, cation-exchange capacity and base saturation. Second, the ratio of ac- tual vs. potential evaporation, the length of the vegetation period and slope gradient were taken into account, resulting in five degrees in total. Technical measures such as fertilisation, liming, plant protection and irrigation were excluded here (see Bastian et al. 2013). Chapter 5 153 considered a complement to ES demand (see below). Possible indicators are average yields of crops, wood regrowth in forests, flood retention in catchments or floodplains, amount of carbon stored in soil and vegetation, relative reduction of noise or pollutants, aesthetics of scenery. ES flow Definition: Flow is a measure for the amount of ES that are actually mobilised in a specific area and time. Driven by a de- mand for a service, ES supply is turned into ES flows (Figure 1). In case both ES supply and demand are quantified using the same dimension and unit, a quantitative com- parison is possible (supply-demand budget calculation). Flow can, in a more tangible meaning, also involve a movement of ma- terial, energy or information across space. In case supply and demand are not spatial- ly congruent, flow maps can show spatial connections between Service Providing and Service Benefiting Areas (SPA – SBA; see Chapter 5.2). Delimitation: Service flow can be con- strained by an inadequate ES supply which would lead to exceedance of the ES potential. This again may lead to an over- use of given ES potentials, degradation of natural capital or to unmet ES demand. Necessity and applicability: ES flow maps can unfold spatial mismatches between ES providers and beneficiaries. If there are es- sential natural processes supporting these interactions between providers and benefi- ciaries, ES flow mapping gives insights to Service Connecting Areas (SCA; see Chap- ter 5.2). Their conditions such as possible barriers or other features shaping the flow are items that can be mapped meaningfully. Possible indicators are fish catch, timber logging, bioenergy gain, groundwater ex- traction (by wells), flood peak reduction, visitor numbers. Example 4 . Flood regulation Flood regulating ES provide excellent exam- ples for linkages of SPAs and SBAs via SCAs. Unlike many provisioning ES, flood regu- lating ES cannot be supplied and imported from remote areas. SPAs and SBAs need to be physically connected (e.g. by a water body or stream) or located in the same process unit (e.g. a watershed). The “flow” of flood regulat- ing ES takes place by spatial units that are able to capture excess water (e.g. from torrential rain) and to regulate the surface water run- off contributing to floods. Humans and their properties benefit from this regulating ES flow by lower amounts of floodwater reaching the SBA. The ES demand exceeds the supply in case of flood hazards. Land use change (e.g. afforestation) in the SPAs can help to increase flood regulating ES flows (see Nedkov and Burkhard 2012). Example 3 . Wood growth in Germany Forest stocks and wood growth are recorded by a forest inventory every 10 years in Ger- many. Wood regrowth as supply indicator results in 122 million m³ per year (compara- ble to a logging of 84 million m³ in 2013). It describes only the status quo; another wood re-growth could be realised at different stock levels, for example, by changing the tree spe- cies and age structures = “managed potential”. The wood stock in German forests, which may also be regarded as supply, is 3.7 billion m³, or 336 m³ ha-1. But since nobody could use them all, this number gives no meaningful indica- tion (Grunewald et al. 2016). Mapping Ecosystem Services154 ES flow should particularly be included in integrative supply-demand assessments. There is a broad range of process models (see Chapter 4.4), expert knowledge (see Chap- ter 4.6) or monetary valuation methods (see Chapter 4.3) which can be applied here. ES demand Definition: Demand is the need for specific ES by society, particular stakeholder groups or individuals. It depends on several factors such as culturally-dependent desires and needs, availability of alternatives, or means to fulfil these needs. It also covers prefer- ences for specific attributes of a service and relates to risk awareness. Demand links ES to particular beneficiaries. This means that without a demand for a service, there is no flow. Beneficiaries express demand and can have the power to translate this demand into an actual ES use. Demands for some ES (such as several regulating ES) might be uncovered, or certain groups of society might be unaware that they actually benefit from an ES. Delimitation: Demand can be different from flow which measures the actual ex- traction of a service within a region. De- mand can, for example, be higher than flow within that particular region. This means, when demand is realised, it could be fulfilled through services that come from another re- gion. For instance, many provisioning ES (e.g. food, timber, energy) can be import- ed. The demand for carbon sequestration (ES climate regulation) can be fulfilled by a region with a high potential to sequester carbon or cultural ES such as recreation can be actively used in another region through travel (see Chapter 6.2). The phenomenon of regionally-unmet demand is common to many ES and so far we have only started to understand the long-distance effects be- tween different regions caused by inter-re- gional ES use. (Regional) demand could also be lower than flow, in case ES are ex- ported. Demand is then expressed by other social-ecological systems while ES flow takes place in the region of interest. Necessity and applicability: Demand can change over time and can show an uneven pattern across space. As a result, it makes sense to map demand independently from potential, supply and flow. Regional demand can exceed the (regional) supply considerably and, through an increased flow, this could result in unsustainable regional levels of ex- traction or use of a service so that flow could exceed ES potential. As a consequence, local ecosystems are at risk of overuse or ecosys- tems in other parts of the world are degraded by land use change (ES footprint). Possible indicators are vulnerability of people or value of endangered assets for flood risk, desirable attributes for recreation, accessibili- ty and travel costs of visitors, socio-economic valuation and stakeholder perceptions. Demand involves human preferences which can be determined through questionnaires, but also involves basic needs (e.g. unpolluted air) and actually used ES (e.g. flood protec- tion at a riverside) even when people are not Example 5 . Demand for recreational use in Danish forest sites Using amongst others, travel costs, presence of viewpoints, distance to forest and coast, pop- ulation and income statistics, Termansen et al. (2013) mapped demand for recreation for Danish forest sites. They find spatial hetero- geneity in demand for recreation, with higher values in forests close to agglomerations such as Copenhagen and higher values for broad- leaved than for coniferous forests. Chapter 5 155 aware of them. Aspects of risk aversion can be based on assumptions, be modelled or by enquiries (stated preferences). In the case of provisioning ES, the beneficiary could be a farmer who benefits from an intact agricul- tural ecosystem. It could, however, also be the regional population that formulates the demand for locally-produced food. Human inputs Definition: Human inputs encompass all an- thropogenic contributions to ES generation such as land use and management (including system inputs such as energy, water, fertiliser, pesticides, labour, technology, knowledge), human pressures on the system (e.g. eutrophi- cation, biodiversity loss) and protection mea- sures that modify ecosystems and ES supply. Delimitation: Human inputs often emerge as harmful impacts to ecosystems caused by monocultural land use, land use change or intensification. Today, most ecosystems and the services they provide are used and influenced by humans. Necessity and applicability: Humans per- form multiple roles in ecosystems acting as managers, but also as co-producers, distribu- tors or beneficiaries of ES. Possible indicators are land use type and in- tensity, load of pollutants, material or energy input (such as nitrogen), effort of landscape maintenance, further contributions to ES. Human impacts are accompanied in many cases by substantial losses of biodiversity. Particular attention should be paid to hu- man inputs since they may alter ES supply considerably and this impact differs spatial- ly. Not only targeted land use activities in- fluence the integrity of ecosystems, but also the utilisation and improvement of ES can impact other services as well. Resulting ES trade-offs (see Chapter 5.7) are important to review, but are often hard to map. Conclusions and recommendations Depending on the scope of application, ES maps can show different contextual aspects of ES which are spatially heterogeneous in a different way and therefore relevant for ES mapping. Depending on data availability and the policy question or information needs at hand, mapping of one or two of these aspects might be sufficient. It is recommended to map only such aspects that can be derived from reliable data. When monitoring or sys- tematic balance over time is requested, data and indicators have to be double-checked for comparability which can also depend on methods or technology of data collection and on appropriate indicator selection. Example 6 . Nitrogen input in Europe The indicator Gross Nitrogen Surplus (GNS) indicates the potential surplus of nitrogen (N) on agricultural land. For EU-27, it re- mained relatively stable between 2005 and 2008 with about 51 kg N/ha/year. The GNS for the EU-15 reduced between 2001 and 2008 from 66 to 58 kg N/ha/year. The GNS was highest between 2005 and 2008 in coun- tries in the North-West of Europe (Belgium, the Netherlands, Norway, UK, Germany, Denmark) and the Mediterranean islands Malta and Cyprus, while many of the Med- iterranean (Portugal, Italy, Spain, Greece), Central and East European countries show the lowest N surpluses (Eurostat). Mapping Ecosystem Services156 Further reading Ala-Hulkko T, Kotavaara O, Alahuhta J, Helle P, Hjort J (2016) Introducing accessibility analysis in mapping cultural ecosystem ser- vices. Ecological Indicators 66: 416-427. Albert C, Galler C. Hermes J, Neuendorf F, von Haaren C, Lovett, A (2015) Applying ecosystem services indicators in landscape planning and management: The ES-in- Planning framework. Ecological Indica- tors 61, Part 1: 100-113. Bastian O, Syrbe R-U, Rosenberg M, Rahe D, Grunewald K (2013) The five pillar EPPS framework for quantifying, mapping and managing ecosystem services. Ecosystem Services 4: 15-24. Burkhard B, Kandziora M, Hou Y, Müller F (2014) Ecosystem service potentials, flows and demand – concepts for spatial localisa- tion, indication and quantification. Land- scape Online 34: 1-32. Fischer A, Eastwood A (2016) Coproduction of ecosystem services as human–nature In- teractions - An analytical framework. Land Use Policy 52: 41-50. Grunewald K, Herold H, Marzelli S, Meinel G, Syrbe R-U, Walz U (2016) Assessment of ecosystem services at the national level in Germany – illustration of the concept and the development of indicators by way of the example wood provision. Ecological Indicators 70: 181-195. Jones L, Norton Z, Austin AL, Browne D, Donovan BA, Emmett ZJ, Grabowski DC, Howard JPG, Jones JO, Kenter W, Manley C, Morris DA, Robinson C, Short GM, Siri- wardena CJ, Stevens J, Storkey RD, Waters G, Willis F (2016) Stocks and flows of natu- ral and human-derived capital in ecosystem services. Land Use Policy 52: 151-162. Liu J, Yang W, Li S (2016) Framing ecosystem services in the telecoupled Anthropocene. Frontiers in Ecology and the Environment 14: 27-36. Nedkov S, Burkhard B (2012) Flood regulat- ing ecosystem services - Mapping supply and demand, in the Etropole municipality, Bulgaria. Ecological Indicators 21: 67-79. Remme RP, Edens B, Schröter M, Hein L (2015) Monetary accounting of ecosystem services: A test case for Limburg province, the Netherlands. Ecological Economics 112: 116-128. Termansen M, McClean CJ, Jensen FS (2013) Modelling and mapping spatial heteroge- neity in forest recreation services. Ecologi- cal Economics 92: 48-57. Villamagna AM, Angermeier PL, Bennet EM (2013) Capacity, pressure, demand, and flow: A conceptual framework for analyz- ing ecosystem service provision and deliv- ery. Ecological Complexity 15: 114-121. Walz U (2015) Indicators to monitor the structural diversity of landscapes. Ecolog- ical Modelling 295 (1): 88-106. Wolff S, Schulp CJE, Verburg PH (2015) Map- ping ecosystem services demand: a review of current research and future perspectives. Ecological Indicators 55: 159-171. Chapter 5 157 5.2. Where to map? Ulrich Walz, Ralf-Uwe Syrbe & Karsten Grunewald The spatial-structural approach It is an important feature of natural and cul- tivated ecosystems that they are not evenly distributed across landscapes, coastal or ma- rine areas and they also vary over time. Eco- system services (ES) are usually generated by ecological processes within their area of influence such as catchments, habitats, nat- ural regions or land use units. This suggests the need for site-specific assessments. There- fore, each ecosystem service should not only be assessed through considering underlying ecosystem types but also with respect to: • underlying natural regional conditions (geology, landform configuration, soil, climate, etc.), • its positional relations to main types of landscape (urban, agrarian, near-na- ture), • the configuration (landscape structure) of the corresponding units (catchments, natural regions, etc.) with natural re- sources or land uses, • the relations between the ecosystem providers of a service and groups of people who make use of it (i.e. benefi- ciaries) and • the use, management and maintenance of the respective ecosystem. Spatial relationships, area types The holistic approach presented here pre- sumes that complex ecological systems un- derlie the production of most ES, which can be envisioned as SPUs, or Service Providing Units. For mapping purposes, such an SPU should be regarded as a spatial unit. This opens the way for applying landscape-scale geographic assessment methods based on landscape units corresponding to the area of influence (see above). In order to avoid terminological confusion with different SPU variants, we term the spatially defined complexes as Service Pro- viding Areas (SPA) (see Text Box 1). SPAs are a promising basis for an inclusive ap- proach of ES at the landscape scale. As the service providing areas defined above include entire ecosystems, their constituent populations and underlying biophysical characteristics, the best way of capturing them spatially is as ecological spatial units (e.g. living spaces, water bodies or soil areas) or as area of influence of the respective pro- cesses (e.g. catchment areas, flood plains). From this point of view, such biophysical- ly-delineated areas are more suitable for analysis than administrative units. In the spatial analysis framework, however, not only the SPAs are of interest, but also the regions to which their benefits accrue. For ex- ample, one might ask: where is the benefit of a given ecosystem service needed? In addition to the service providing areas (SPAs), Service Benefiting Areas (SBAs) should thus be de- fined in which beneficiaries receive the ser- vice (see Text Box 1). In a spatial framework, urban areas, rural settlement areas and espe- cially administrative units could be consid- ered as SBAs. Factors such as population den- Mapping Ecosystem Services158 sity, social facilities (e.g. schools, hospitals but also parks for recreation) and built structures (residential, commerce or industry buildings) or the number and size of the households, are important as indicators (e.g. per household measures of demand for specific ES). The service providing and service benefiting areas may overlap, but significant spatial differences are also possible (see example in Text Box 2). If the service providing and service benefiting areas are not adjacent, the properties of the connecting space can have an influence on the provision of the ser- vice (see Text Box 2). We include such an interstitial space between service providing and service benefiting areas in our consid- erations under the term Service Connecting Area (SCA) (cf. Fig. 1). The following fundamental types of relations between the service providing and the service benefiting areas can be distinguished (Fig. 1): a. ’in situ’: the two area types are identical, i.e. the ES are supplied and in demand in the same area (e.g. the population uses the groundwater of its settlement area), b. ‘central demand’: the surrounding area provides for / impacts on a central demand area (e.g. a settlement bene- fits from supply of fresh and cold air which is generated by open spaces in the surrounding), c. ‘omni-directional’: the service benefit- ing area surrounds a service providing area independent of direction (e.g., farmland benefits from hedges as a liv- ing space for beneficial insects), d. ‘directional without dependency on a slope’: the service benefiting area is sit- uated “behind” the service providing area, protected as it were with respect to the predominant impact direction (e.g. a residential area protected against traf- fic noise by a forest), Text Box 2 . Example The floodwater regulation service mainly depends on the character of the watershed that is upstream of beneficiaries, whereas the benefit from the reduced flood risk in the populous cities along the flood plains is presumably highest in the more built-up lower reaches. This raises the question of whether the residents at the upper reaches should unilaterally forego development options in favour of the downstream riparian beneficiaries and, if so, how much compensation should they be entitled to? Should the most vulnerable houses in a downstream settlement be resettled out from the flood plains or protected better? The service connecting area also plays an important role, since, for example, the channel geometry, tributary streams, natural floodplains and wetlands and reservoirs or other grey infrastructure can strongly modify the severity of a potential flood. Text Box 1 . Definitions Service Providing Area (SPA): spatial unit within which an ecosystem service is provided. This area can include animal and plant populations, abiotic components as well as human actors. Service Benefiting Area (SBA): spatial unit to which an ecosystem service flow is delivered to beneficiaries. SBAs spatially delineate groups of people who knowingly or unknowingly benefit from the ecosystem service of interest. Service Connecting Area (SCA): connecting space between non-adjacent ecosystem service-providing and service-benefiting areas. The properties of the connecting space influence the transfer of the benefit (also refer to Text Box 2). Chapter 5 159 e. ‘directional downslope’: the service ben- efiting area is situated downhill (down- river) from the service providing area, i.e. the service is dependent on gravi- tational processes (e.g. cold air, water, avalanches) and f. ‘spatially separated’: e.g. drinking wa- ter, food production, recreational areas. There can be different connective meth- ods, e.g. natural hydrologic flow within watersheds, infrastructure (pipes/aque- ducts) or road/trail networks. The relation types d, e and f can especially exhibit considerable service connecting areas. Analysing the spatial structures Once SPAs, SBAs and SCAs are defined for each ecosystem service (see Table 1 for examples), they can be described in greater detail according to their properties such as structure, type and characteristic of the spatial situation. The comprehensive characterisation of a service-providing area should contain at least the following information: 1. a site characterisation and classification of the potential for providing the ser- vice with respect to the required natural processes and their dynamics, Figure 1. Types of spatial relations of Service Providing Areas (SPA), Service Benefiting Areas (SBA) and Service Connecting Areas (SCA) (adapted and extended from Fisher et al. 2009; Syrbe & Walz 2012). SPA = SBA S P A SBA SBA SPA SBA SPA SPA SBA a) ‘in situ’: SPA and SBA are identical b) ‘central’: surrounding area supplies / acts on the central benefiting area c) ‘omnidirectional’: directed on all sides - to larger surrounding area d) ‘directional’ – spatially separated from each other: SBA lies ‘behind’ the SPA e) ‘directional’ – spatially separated from each other (e.g., slope dependent) f) “non-directional‘ – spatially separated from each other SCA SCA SPA SBA Mapping Ecosystem Services160 Ecosystem Services * Service providing area (SPA) Service connecting area (SCA) Service benefiting area (SBA) P groundwater recharge Arable land, wood, grassland, wetlands and other open land in a groundwater basin Groundwater flow paths (with possible contam- inated sites and risk areas for the protection of groundwater) Settlement areas, irrigat- ed areas P drinking water Headwaters and catchment areas Bodies of groundwater, streams, rivers, (pipe- lines); (see Chapter 6.2) Settlement areas, indus- try (for production, less for cooling) P fodder for grazing animals Grassland and forage crops Pastoral paths Farms R protection against snowdrift, storm Forest, road trees, shrubs, hedges Embankments at roads and railway lines E.g. roads, railway lines and runways R erosion prevention - by wind - by water Forest, hedges, bushes, trees and shrubs (grassland, permanent crops) Field edges, gullies Areas under cultivation, water reservoirs R flood prevention Forest, ponds, wetlands, etc. in flood generation areas Floodplains above benefiting areas Built-up area in the floodplain R local climate regu- lation (cold/fresh air) Open land, parks above cities Slopes (with or without obstacles) around a city City in valley R noise reduction Roadside greenery, wood, ramparts Areas (if appropriate buildings) around the source of noise Residential and recreational area R avalanche and landslide prevention Forest above residential or recreational areas Slope area Residential or recreational areas below steep slopes R pollination Nesting habitats of insects Radius of flight and foraging habitat Farms with crops requiring pollination R pest control Nesting habitats of predators Foraging habitat Crop land R stream water purification Surface water bodies, wetlands Water catchments Residential or recreational areas C appreciated scenery Viewsheds (areas which can be seen from a particular site) Line of sight, open country Settlements and touristic infrastructure C recreation activities Surface water bodies, mountains, wood Road and path network between SPA and SBA Touristic lodging units Table 1. Examples for Ecosystem Services, which depend on lateral or vertical landscape processes, with associated Service Providing Areas (SPA), Service Benefiting Areas (SBA), Service Connecting Areas (SCA) (adapted from Syrbe & Walz 2012). * Type of service: P – provisioning services, R – regulating services, H – habitat services, C – cultural services Chapter 5 161 2. an analysis of the human usage patterns also regarding their internal structure, for example, through landscape metrics and 3. the consideration of the conditions re- garding location and neighbourhood according to the respective processes. The comparison and the positional relations of the service-providing areas to the associat- ed service-benefiting and service-connecting areas form another focal point of the spatial investigation. The characteristics of the ser- vice-providing areas are primarily founded on the natural sciences, since they relate to the beneficial natural resources and - if ap- plicable - to those processes which ensure their regeneration. Furthermore, the anal- ysis needs to examine whether investments (protection or management measures) are necessary in order to preserve the service supply capacity. If so, the kind and frequen- cy of the maintenance measures should be determined and the necessary cultivation rules should be known. If the natural capital is reducible (by consumption), the natural capacity to regenerate has to be determined in order to adapt the consumption to the regeneration rate if a sustainable resource management is to be achieved. The characterisation of the service-bene- fiting areas also includes further analysis from the social sciences. Especially, users’ demands have also to be incorporated into analysis of SBAs. Depending on the area of investigation, the demands, preferenc- es and values of the benefiting population groups represent indicators for the demand for the ES. The size of a population group is an important basis for determining and assessing the service. However, whether threshold values should or must be defined is also crucial for the assessment. This can be the case, for example, with respect to people endangered by natural disasters (e.g. restriction of construction areas because of flood hazards) or to unsustainable resource harvest rates (see above). Moreover, limit- ed or widely demanded resources require clear rules for accessing them in order to avoid “free-rider effects” (benefiting from a service without contributing to it) and misinvestments. The type of access (pri- vate, common or public) to a resource and the possibility of excluding people from such access determine the marketability or non-marketability of the ES. Even if a service connecting area does not exist separately because service-provid- ing and service-benefiting areas overlap, an analysis of the connection properties is useful, since horizontal transfer processes are influenced by landscape characteristics. If there is an interstitial space between the service-providing and the service-benefiting area, this connecting space first needs to be determined more closely, which can at times be difficult. This can be modelled for exam- ple using the transport and transformation paths of substances, energy, biota and possi- bly also information. Spatial units as the basis of ecosystem services assessments Depending on the type of the ecosystem service that is being assessed, very different spatial units can be considered for service providing, benefiting and connecting areas (Table 2). Moreover, certain actors whose actions significantly contribute to the ben- efit may participate in the service provision or the transfer of benefits. Stimulating their economic interest (remunerating instead of disadvantaging them) is an essential goal of the ES approach. Examples of different types of spatial units for capturing and assessing individual eco- system services are: Mapping Ecosystem Services162 Conclusions A decided advantage of a spatial-structur- al approach is that it makes it possible to understand ES beneficiaries and flows. As soon as it is possible to determine the ben- eficiary of a service, the benefit of such a service can also be identified. This especial- ly applies when the provision and the use of such services do not spatially overlap. Only this knowledge makes it possible to design incentive systems and fair payment for the providers when they deliver this ser- vice (see Chapters 7.2 and 7.3). This is also the prerequisite for ES’ availability in the long term. Further Reading Bagstad KJ, Villa F, Batker D, Harrison-Cox J, Voigt B, Johnson GW (2014) From theoretical to actual ecosystem services: Mapping beneficiaries and spatial flows in ecosystem service assessments. Ecology and Society 19(2): 64. a. Single surfaces (patches), landscape el- ements such as arable field sections or forests provide the spatial basis of refer- ence most frequently used for the over- all assessment. b. Administrative units are useful if data are analysed which have such units as a reference basis, such as socio-economic data or legal framework conditions. c. Water catchment areas are typical mor- phological units which can be delimited well using GIS on the basis of a digi- tal elevation model (see Chapter 3.4). They therefore represent the common reference units for hydrologic services. d. Natural units should be drawn upon if natural properties such as soil condi- tions, surface forms, climate, geology or vegetation mainly determine a service. e. Landscape units, delimited not only ac- cording to natural conditions but also according to land use, are usable for most services, especially for site-scale assessment of ES. These units serve as a reference basis for assessing different aspects of diversity, for determining the spatial heterogeneity and as a spatial framework for practical management. Nature and origin of the service Spatial unit Generated by specific species Suitable habitats Based on biophysical resources Natural regions Depending on specific landscape mosaic Spatial unit with comparable spatial features Generated by an abiotic process Area of influence of this process Dependent on specific land management practices Management units Rooted in history and culture Units of the historic cultural landscape Hydrologic services Water catchment areas Demand for ecosystem services by people Administrative units Table 2. Example for the suitability of landscape units for designation of provision, benefiting and con- necting areas. Chapter 5 163 Bastian O, Grunewald K, Syrbe RU (2012) Space and time aspects of ecosystem ser- vices, using the example of the EU Water Framework Directive. International Jour- nal of Biodiversity Science, Ecosystem Ser- vices & Management: 1-12. Burkhard B, Kandziora M, Hou Y, Müller F (2014) Ecosystem Service Potentials, Flows and Demands - Concepts for Spa- tial Localisation, Indication and Quanti- fication. Landscape online 34: 1-32. Fisher B, Turner RK, Morling P (2009) Defin- ing and classifying ecosystem services for decision making. Ecological Economics 68 (3): 643-653. Grunewald K, Bastian O (Eds) (2015) Ecosys- tem Services. Concept, Methods and Case Studies. Berlin, 312 pp. Syrbe RU, Walz U (2012) Spatial indicators for the assessment of ecosystem services: Providing, benefiting and connecting areas and landscape metrics. Ecological Indica- tors 21: 80-88. Mapping Ecosystem Services164 5.3. When to map? Carlos Guerra, Rob Alkemade & Joachim Maes Setting the scene Mapping ecosystem services (ES) is often seen as a static three-dimensional problem where space (x, y) and the value of a given ecosystem service (z) are referred as the main factors of analysis. A wide group of examples that fol- low this approach populate current scientific papers, books and technical reports. The is- sue with these assessments is that they often consider that the value of a given ecosystem service in a particular place is (a) stable in time or (b) it already encapsulates the effects of the underlying ecological processes/cycles. Under a spatial notation (x, y, z), ecosystem service supply is represented by a magnitude, a spatial distribution or configuration and an extent. Although perceptive, this approach does not consider that specific ES are often supplied in different moments in time (e.g. pollination, food production and flood reg- ulation) and generate benefits that can be equally temporally displaced (e.g. in flood regulation there is a lag of time between the accumulated decrease of runoff [superficial water flow] by percolation and the actual reduction of the downstream flood plain). This results from the fact that ecological pro- cesses/cycles vary through time and, because most ES (namely, production and regulating services) depend on specific ecological pro- cesses/cycles, ecosystem service supply is also dynamic. These dynamics can be illustrated by focussing on a specific ecosystem service provider, e.g. a deciduous tree (Figure 1). Figure 1. Example of a within-year ecosystem service supply cycle considering a deciduous tree as the focus of ecosystem service supply. Chapter 5 165 From January to December, the life cycle of a deciduous tree allows for the supply of a relatively large number of ES. Start- ing in spring, this single tree represents an important support for bird nesting contrib- uting to habitat quality while, at the same time through photosynthetic processes, it captures carbon and other atmospheric pol- lutants, thus improving local air quality. As time passes, it gives shade for picnics in the summer but it also promotes heat absorp- tion and reduces the albedo effect which helps to reduce heat waves in cities or the probability of fires in forests. At the end of the summer, rain season starts and the same tree contributes to control soil erosion by re- ducing the erosivity power of precipitation. In the autumn, the leaves fall contributing to local soil fertility. The landscape changes to autumn colours which inspire poets, paint- ers and mountaineers. When winter arrives, the same tree that in summer absorbed heat, now lets the solar radiation pass and thus improves local heat regulation. When isolat- ed, this deciduous tree has a rather narrow potential to supply all of these ES but, when part of a community (e.g. integrated in a de- ciduous forest), this potential is multiplied and new ES can emerge. This example serves to show the dynamics and complementarity of ecosystem service supply through time. It also highlights the need to include temporal variations in the assessment of ES, as the likelihood of mis- representing ecosystem service supply in static assessments is considerable. In fact, time dependency of ES correspond to a very broad and complex issue that includes various time scales, ranging from very short timescales (within a day or a year) to sev- eral years, decades or centuries depending on the ecosystem service under assessment. Properly selecting the scale of assessment is fundamental and it mainly depends on the objectives of the assessment and the ecological cycle/process under study. For example, in flood protection, it is possi- ble to focus on hourly variations (if a peak flood is considered and the capacity of veg- etation to reduce runoff velocity is stud- ied), monthly variations (if the purpose is to identify hotspots of ecosystem service supply), or yearly variations which can be projected through centuries (if the purpose is to study probability of flood events or projection of trends). Another dimension of complexity is also the significant mismatch between the potential for ecosystem service supply and the actual ecosystem service supply. This mismatch is also linked to temporal issues. Consider soil protection as a regulating eco- system service. In this context, vegetation cover protects soils from being eroded. If vegetation is removed, for instance by har- vesting crops, there is an enhanced erosion risk. If we evaluate the temporal dynamics of vegetation cover (here representing the potential to supply soil erosion prevention) and the actual ecosystem service supply (avoided erosion), these variables have two very different temporal distributions result- ing in a supply and demand mismatch. As for the ecosystem service supply, the demand for ES is also dynamic. It usually correlates with the cycles of environmental impact (in the case of regulating services), production cycles (e.g. the requirement for pollination services according to crop cycles), specific consumer demands (e.g. the increase in codfish or turkey demand during the Christmas period), recreation cycles (e.g. the increase in the demand for hiking areas during the summer time), amongst others. The potential differenc- es between the dynamics of demand and supply of ES are among the drivers for over-exploitation of ecosystems making the evaluation of temporal dynamics even more significant. Mapping Ecosystem Services166 Ecosystem service dynamics Assessing ecosystem service potential, sup- ply and demand (see Chapter 5.1) requires a thorough understanding of ecological cycles and ecosystem service mechanisms. Both of these are dynamic and entail the recognition that an ecosystem service is dependent on multiple simultaneously occurring processes with different (often competing) objectives and that ecosystem service supply is secured by different ecosystem service providers with their own specific ecological cycles, targets and trends. This recognition is critical when assessing ecosystem service supply but it also depends on the objectives of the assessment and on the research question that is being addressed. Within a static approach, the indicators of ecosystem service supply portray a snapshot (an image of a single moment in time). These indicators often neglect the existence of eco- logical or environmental cycles and dynamics or assume that these are already encompassed within the results obtained. Although these indicators can eventually be used as state or impact indicators they often lack the ability to produce a good representation of ecosys- tem service supply that is suitable for policy support, land management assessments or other forms of decision-making. One of the reasons for this, is the inability of static indicators to capture the influence of particular management practices on the overall ecosystem service supply. Or at least this is often only true when using long cy- cles and when a direct relation between eco- system service supply and the accumulated effects of specific impacts (e.g. the effect of intensive ploughing on soil erosion) is effec- tively established. In this example, a static impact prevention indicator can be used to illustrate the spatial distribution of ecosys- tem service supply but it gives little informa- tion regarding the underlying process. To effectively assess ecosystem service supply, it is essential to implement methodological approaches that consider indicators that vary over time and space. Many examples of these approaches can be found in literature (see the “Further reading” section in this chapter) and more recently StDMs (stochastic dynamic methodologies) are being used to highlight the influence of specific land management strategies on the ecosystem condition and the related ecosystem service supply. Independently of the chosen method, there are three major dimensions to be considered when implementing a dynamic assessment of ecosystem service supply: i) the signifi- cant temporal amplitude of the underlying ecological cycles; ii) co-dependency process- es and their impact on the provision of mul- tiple ES; and iii) seasonality. Ecological processes develop within a wide range of temporal cycles from short- to long-term. Therefore, correctly assessing ES strongly depends on identifying the relevant temporal amplitude that allows the capture of the full extent of ecosystem service supply. Another aspect for consideration is the de- termination of the relevant temporal ampli- tude to identify the effect of specific drivers on ecosystem service supply. In some cases, within the same ecological process, one has to look at both the short- and long-term cy- cles in order to understand the contribution of ecosystem service supply to society and the influence of different drivers. Good examples come from assessing the contribution of ES to mitigate a particular flood event versus de- termining the mitigation effect in the case of extreme, long-term, events (e.g. a 0.01 prob- ability event such as a “100-year flood”). At the same time, many ecological process- es have “multiple” co-dependency relation- ships between themselves. This dependency is often determined by the cycle of one or more ecosystem components and it is also Chapter 5 167 reflected in ecosystem service supply. For example, both flood regulation and soil ero- sion prevention depend on the processes by which water percolates into the soil and is retained by vegetation. Although using dif- ferent processes and interactions, vegetation plays a significant role in the supply of these two different ecosystem services. As in this example, these co-dependence effects often do not necessarily happen at the same time and are therefore often overlooked by eco- system service supply analysis. As an example, crop yield depends strong- ly on water (from infiltration) and nutrient availability (e.g. from nitrification), but the service supply from these three services oc- curs and has to be quantified at different moments in time. Related to this is the seasonality of ecosys- tem service supply and its related benefits. Previously illustrated in Figure 1, the in- tensity and frequency of ecosystem service supply depends strongly on the seasonali- ty of the ecological processes underlying a given ecosystem. All of these different aspects contribute to undermine ecosystem service supply quan- tification and its analysis. When assessing disturbance or recovery dynamics, an as- sessment of ecosystem service supply should consider at least one or more of these differ- ent aspects in order to produce consistent results and to enable the illustration of spe- cific dynamics of change. Trend analysis Ecosystems evolve over time as they are af- fected by and react to different human and environmental drivers of change. This evo- lution can result in cumulative effects for the ecosystem (e.g. the cumulative effect of soil sedimentation in wetlands) but can also allow determination of the influence of specific drivers in relation to specific ecological functions. Here lies the value of trend analysis, the contribution for under- standing the past and current development pathways in order to create knowledge about the future of ecosystems. Current assessments of ES do not always fa- vour the use of time series. This often comes from data limitations regarding the use and availability of contemporary datasets for all system components but also and more im- portantly, the availability of temporal datasets with an amplitude and a frequency that is rele- vant for the processes under study. A common limitation is related to the availability of com- parable time series of soil datasets or the exis- tence and availability of biodiversity data with relevant thematic, temporal and spatial extent. Nonetheless, the use of trend analysis corre- sponds to one of the most valuable tools to identify the determinants of change. Exam- ples of this can be seen through literature (see Further reading for references) using long time series to illustrate the effects of policies, land management, forest fires, amongst oth- ers. Figure 2 presents an illustration of a time series of land cover and land use change for a montado landscape in the South of Portu- gal from which it is possible to calculate long term trends. Such data is of critical impor- tance for understanding changes in ES over time as a result of changes in management and policy implementation. Scenario analysis At the same time, trend analysis also presents a valuable opportunity to better design and describe future scenarios of ecosystem devel- opment. These scenarios are plausible repre- sentations of possible future states for one or Mapping Ecosystem Services168 more components of a system, or as alterna- tive policy or management options intended to alter the future state of these components. Scenario analysis in ecosystem assessments, policy support and decision-making aims at visualising future impacts on biodiversity and ES of global, regional or local changes such as land use change, invasive alien species, over-exploitation, climate change and pollu- tion. Scenario analysis also provides decision support for developing adaptive management strategies and exploring the implications of al- ternative social-ecological development path- ways and policy options. At the same time, scenario analysis and scenario planning have been successfully applied in many local stud- ies, in national assessment and for regional and global assessments (Chapter 5.7.3). Generally, scenario analysis includes three major phases. The initial step is to define the major tendencies for a specific region or for a specific subject and to analyse the drivers of change that are likely to be involved in the foreseen tendencies. This phase results in a few plausible scenarios. A second phase is to translate these scenarios quantitatively or qualitatively into variables that describe the major drivers of change, such as economic development or demography. These driv- ers of change are then the input for models that relate these changes to environmental change, such as land use change or climate change, and on biodiversity and ES. A third phase starts with analysing the outcomes of these models and formulate policy options to avoid undesired developments in key variables of biodiversity and ES. Models used in scenario analysis are typi- cally able to describe dynamic relationships amongst drivers, biodiversity and ES. Often a wide range of models is needed to perform an adequate scenario analysis. Not only Figure 2. Example of land use and land cover change over a period of 61 years in a montado area in the South of Portugal. Land use change Land cover change time Artificial surfaces Artificial surfaces Agro-forestry areas with 30- 50% of tree cover and >50% of shrubs Permanent pastures Mixed forest Production forest Shrubs and/or herbaceous vegetation associations Water bodies 1951 1969 1986 1995 2004 2012 Agro-forestry areas with 30- 50% of tree cover and <50% of shrubs Agro-forestry areas with <30% of tree cover and >50% of shrubs Agro-forestry areas with <30% of tree cover and <50% of shrubs Arable land and permanent crops Irrigated arable land Non-irrigated arable land Olive groves Orchards Agro-forestry areas with >50% of tree cover Agro-forestry areas Water surfacesForest areasPermanent pastures Chapter 5 169 models that quantify changes of ES based on changes of land use are needed, but also models that drive these land use changes, such as economic and demographic models. In addition, hydrological and other biophys- ical models in combination with biodiversi- ty interactions are required if more complex issues are under consideration. New approaches for scenario analysis are pro- posed and applied, where stakeholders and local knowledge holders are increasingly in- volved. Another recent development in mod- elling for scenario analysis is to understand the feed-back loops from changing ES provi- sion to a change in economic development. Issues with data quality for dynamic assessments Ecological modelling and particularly pro- cess-based ecological modelling, depend on a vast array of ecological, biophysical and anthropogenic datasets to generate relevant results. Although in recent years, earth ob- servation systems have evolved to the point of delivering continuous (temporally and spatially) data for particular ecosystem com- ponents (e.g. forest change and extent, tree density, elevation, human density, economic characteristics, precipitation, etc.), many of these lack the ability to be compared or used in a modelling environment due to different resolutions and/or methods/sensors. Additionally, there is a clear mismatch between the publication date of the vari- ables to be used in a given assessment (e.g. LUCAS soil data from 2009) and the ref- erence date for the assessment itself (for example using vegetation data from 2016 to assess the effect of soil erosion preven- tion without considering the 7 years’ dif- ference between these datasets). In several cases, if any modelling approach is to be implemented, these mismatches cannot be simply overcome and often error propaga- tion assessments should be implemented to minimise unwanted effects. Independently of the problems or potential caveats related to particular datasets, the tem- poral resolution (i.e. the amplitude and fre- quency of data collection) of a given dataset is an important determining factor for dataset selection in trend analysis. Therefore, future ecosystem service supply studies should in- clude the effects of data quality on their re- sults as it can produce important biases in the overall interpretation and decision-making support. Further reading Bateman IJ, Harwood AR, Mace GM, et al. (2013) Bringing ecosystem services into economic decision-making: land use in the United Kingdom. Science (80) 341: 45-50. Guerra C, Metzger MJ, Maes J, Pinto-Cor- reia T (2016) Policy impacts on regulating ecosystem services: looking at the implica- tions of 60 years of landscape change on soil erosion prevention in a Mediterranean silvo-pastoral system. Landscape Ecology. doi: 10.1007/s10980-015-0241-1. Kandziora M, Burkhard B, Müller F (2013) Mapping provisioning ecosystem services at the local scale using data of varying spatial and temporal resolution. Ecosystem Services 4: 47-59. doi: 10.1016/j.ecoser.2013.04.001. Koch EW, Barbier EB, Silliman BR et al. (2009) Non-linearity in ecosystem services: temporal and spatial variability in coastal protection. Frontiers in Ecology and Envi- ronment 7: 29-37. doi: 10.1890/080126. Mapping Ecosystem Services170 Nelson E, Mendoza G, Regetz J et al. (2009) Modelling multiple ecosystem services, biodiversity conservation, commodity pro- duction and tradeoffs at landscape scales. Frontiers in Ecology and Environment 7: 4-11. doi: 10.1890/080023. IPBES (2016) Methodological assessment of scenarios and models of biodiversity and ecosystem services, Ferrier S, Ninan KN, Leadley P, Alkemade R, Acosta-Michlik LA, Akçakaya HR, Brotons L, Cheung WWL, Christensen V, Allam Harhash K, Kabubo-Mariara J, Lundquist C, Obersteiner M, Pereira HM, Peterson G, Pichs-Madruga R, Ravindranath N, Ron- dinini C, Wintle BA (Eds.) Secretariat of the Intergovernmental Platform for Bio- diversity and Ecosystem Services, Bonn, Germany. Chapter 5 171 5.4. Why to map? Sander Jacobs, Wim Verheyden & Nicolas Dendoncker Meaningful mapping Maps for ecosystem services (ES) are made for a broad set of purposes. These include advo- cacy (awareness raising, justification, decision support), ecosystem assessment, priority set- ting, instrument design, ecosystem account- ing, economic liability and scientific spatial analysis. Figure 1 illustrates the theoretical relationship between mapping purposes and quality requirements. Requirements concern notably spatial and temporal resolution, sci- entific accuracy and reliability and ease of understanding. Additional methodological requirements not represented in Figure 1 are the extent of the mapping exercise, the repeat- ability, the theme of the mapping (e.g. supply, demand, conflict maps etc.) and basics of car- tography and mapping semantics (see chap- ters 3.1 and 3.3). These vary depending on the specific context of the mapping exercise (e.g. community development versus nation- al assessment, see Figure 1). Figure 1 can be interpreted across purposes for one specific requirement or across re- quirements for one specific purpose. For ex- ample, the expected clarity of a map meant for research use is lower than that aimed at policy advocacy. On the other hand, maps used by research should be highly reliable while those used for awareness raising (ad- vocacy) do not require such high reliability. Many current mapping applications focus on quantitative valuation and accounting. Typically, these maps are neither meant to be understood by a broad range of stakeholders nor do they necessarily require a high spa- tial resolution, but they should be highly accurate and reliable. This chapter illustrates this for two specific examples concerning re- gional assessment and priority setting. Good enough is just perfect Mapping quality requirements are bound by resource availability and by the risk of decisions based on them. The upper bound- ary of requirements is set by the principle of parsimony, stating that “among two good solutions, the simplest is always best”. This highlights the need for using the least re- sources or assumptions necessary to solve a problem. In other words, one should not spend excessive (project) time and/or (pub- Figure 1. Ecosystem services mapping requirements according to purpose. Parsimony Risk RELIABILITY MAP REQUIREMENTS ACCURACY RESOLUTION CLARITY Mapping Ecosystem Services172 lic) money to map at a greater level of detail than necessary. For example, land use based maps (see Chapter 5.6), that can be pro- duced repeatedly at relatively low costs (in terms of time and money) are sufficiently adequate for most purposes, while more reli- able data can sometimes only be obtained at excessively high cost, or involving complex assumptions. Moreover, the time spent on a specific map should be traded off against the urgency of the purpose. The lower limit of map quality requirements is determined by the societal impact of the decisions based on the mapping. Uncertain- ty (or absence of information on uncertain- ty) translates in a societal risk for adverse outcomes if decisions are based on wrong data. Public or policy advocacy for the im- portance of ES does not require highly ac- curate or detailed maps. However, commu- nication maps cannot be used for purposes which have more stringent requirements, such as ecosystem accounting or economic liability: the risk for unfair or undesired out- comes is too high or unknown. This brings us to the issue of the safe oper- ating space for each type of map. Maps with lower requirements cannot be used for pur- poses which have higher requirements. On Figure 1, this goes both ways: for instance, maps made for scientific purposes need sim- plification to be clear enough for priority setting, assessment or advocacy, while as- sessment maps have to be detailed further to obtain the accuracy and reliability required for some scientific purposes. Maps are means, not ends Maps are instrumental tools that are com- bined with other types of data and contex- tual information in order to achieve a certain purpose (see Figure 1). This information can be quantitative and qualitative and is rarely spatially explicit. Knowing how maps will be combined with these non-spatial data and used in a specific context is essential for the mapping process. We illustrate this below by showing how maps are used as part of the diverse information for two common eco- system service questions: a land use priority setting in a local context and a regional eco- system assessment. The modest mapper In this final section, we provide guidelines for critical map-makers to engage in effec- tive ES mapping. While most of these will seem evident, they are rarely applied in prac- tice. Following these guidelines will improve effectiveness of ecosystem service maps to impact actual decision-making and contrib- ute to scientific advance. • Clearly define the purpose for which mapping is needed. Plenty of maps are created without clear purpose and later applied for the wrong purpose. • Determine the minimum reliability, ac- curacy, resolution and clarity required. The risk for undesired outcomes grows if maps are used for higher impact de- cisions. • Assess the resources (time and money) needed to meet these requirements. Highly expensive, detailed or complex maps are not necessarily more effective. • Delineate the safe operating space of your maps. The map-maker, being aware of the power and limitations of maps, bears responsibility to caution against wrong or risky application (see Chapter 6.4). • Target the form and communication of maps fitted to the process they are used in. Maps are essential for many processes, but project purposes are never just maps. Chapter 5 173 Box 1 . Local example priority setting for land consolidation to optimise ES provision ES mapping at the local scale is often used to set priorities and guide decision-making to optimise ES provision. This example describes how ecosystem service maps were combined with biophysical models and valuation data to serve a participatory land-consolidation plan for three municipalities in Wallonia, Belgium. It is co-constructed by the administrations, scientists and local stakeholders. The project’s ob- jective is to design a replicable methodology, based on hands-on experience in a first case study. Figure 2 describes the methodological framework further. After selecting a list of locally relevant ES and, based on a typology of ecosystems, biophysical assess- ment and social valuation are carried out. The biophysical assessment includes mapping and quantifica- tion of selected ES based on indicators obtained from a hydrological model and scenario development of potential ecosystem service supply. Social analysis comprises stakeholder analysis, societal valuation according to these stakeholders, participatory validation of the biophysically mapped ES and partici- patory mapping of ecosystem service demand. These supply and demand maps are then used to guide participatory comparison of land-consolidation actions. For instance, maps of biophysical indicators were compared with demand maps to highlight locations for which there is potential improvement of supply. Technical experts of land consolidation then suggest potential measures (e.g. installation of new hedgerows, creation of new water retention basins, new flower strips along a walkway etc.) to be implemented in the final land consolidation plan. This example clearly demonstrates that maps are used as a central means in combination with various other data, methods and actions, to achieve a broader objective shared by various stakeholders and lead to improved decision-making. Figure 2. Methodological framework for integrated valuation of ES to set priorities for land consolidation in Wallonia: Maps are central, necessary parts of a yet broader process (from Baptist et al. 2016). Mapping Ecosystem Services174 Box 2 . Regional example - regional ecosystem assessment National and regional ecosystem service assessments seek to assess the state and trends of ES in their re- gion, with the purpose of monitoring their evolution and informing policies. The state of ES comprises information on the demand, the supply, the balance between demand and supply, the use of ES, eco- system functions underpinning them, drivers of change, impacts on human well-being and governance. Spatial data - also in regions with high data-density - are not available for all aspects of all services and for some aspects the spatial dimension is even irrelevant. The Flanders regional assessment has assessed demand, supply, balance between these two and interac- tions between use of services. These statements were based on a detailed review of all data and informa- tion in 16 ecosystem service chapters to obtain one single concise table on the state of ES with known reliability. Despite the focus of the chapters on maps, the data underpinning this assessment are only partly spatially explicit and range over different data types which are synthesised in key findings (Figure 3). Although the separate maps can be used to answer specific questions, the context of a regional as- sessment requires synthesising maps into short conclusive statements or non-spatially explicit indicators for policy communication. Therefore, the statements derived from the 78 maps to inform the regional state assessment were verified and reviewed by all the involved map-makers. In conclusion, maps which are integrated in communication, decisions or even research will be reduced to quantitative or qualitative findings and combined with other data and information to obtain final outcomes. Mapping will be more effective when engaging in the specific context, by targeting and communicating the maps to the specific purpose and by tuning maps to the diverse information they are combined with. Figure 3. Proportion of spatially explicit (distribution available on Flanders scale) data throughout the ecosystem service chapters (left panel) and per data type (right panel). In many cases, maps are a starting point for an open discussion about what the maps need to indicate and about the as- sumptions made in the underlying mod- els. Using maps top-down as ‘objective data’ often discards nuanced reality of a local context and is counterproductive in most real-life decision processes. To ef- fectively apply maps, the ES map-maker needs to involve: • Interdisciplinary engagement: learn from existing practices and cooperate with other research fields, such as envi- ronmental decision support, communi- cation science, participatory processes, etc. to avoid classic pitfalls. Chapter 5 175 • Trans-disciplinary engagement: consid- er the use of co-design approaches from the very start. Nowadays, stakeholder involvement is an essential indicator for end-user satisfaction and final uptake of the developed maps and the only reality check the ES-map-maker has. Ecosystem service mapping can be highly rewarding in terms of impact on real-world decision-making. This requires leaving the comfort zone of single disciplines and clear data layers and finding the right balance between scientific demands, user demands, functionality and available resources. for ev- ery mapping project again. Further reading Baptist F, Degré A, Grizard S, Maebe L, Pi- part N, Renglet J, Sohier C, Dufrêne M, Dendoncker N (2016) Elaboration d’une méthodologie d’évaluation des incidences sur l’environnement de l’aménagement foncier s’appuyant sur la notion des ser- vices écosystémiques. Rapport général. Di- rection Générale Opérationnelle de l’Ag- riculture, des Ressources Naturelles et de l’Environnement, 187 pp. Gómez-Baggethun E, Barton DN (2013) Classifying and valuing ecosystem services for urban planning. Ecological Economics 86: 235–245. Guerry AD, Polasky S, Lubchenco J, Chap- lin-Kramer R, Daily GC, Griffin R, Ruck- elshaus M, Bateman IJ, Duraiappah A, Elmqvist T, Feldman MW, Folke C, Hoek- stra J, Kareiva PM, Keeler BL, Li S, McK- enzie E, Ouyang Z, Reyers B, Ricketts TH, Rockström J, Tallis H, Vira B (2015) Natural capital and ecosystem services in- forming decisions: From promise to prac- tice. Proceedings of the National Academy of Sciences 112(24): 7348-7355. Hauck J, Görg C, Varjopuro R, Ratamäki O, Maes J, Wittmer H, Jax K (2013) Maps have an air of authority: Potential benefits and challenges of ecosystem service maps at different levels of decision making. Eco- system Services 4, 25-32. doi:10.1016/j. ecoser.2012.11.003. Jacobs S, Dendoncker N, Keune H (Eds.) (2013) Ecosystem Services: Global Issues, Local Practices, Elsevier, New York. Jacobs S, Spanhove T, De Smet L, Van Daele, T, Van Reeth W, Van Gossum P, Stevens M, Schneiders A, Panis J, Demolder H, Mi- chels H, Thoonen M, Simoens I, Peymen J (2016) The ecosystem service assessment challenge: Reflections from Flanders-REA. Ecological Indicators 61: 715-727. doi:10.1016/j.ecolind.2015.10.023. McIntosh BS, Ascough JC, Twery M, Chewe J, Elmahdi A, Haase D, Harou JJ, Hepting D, Cuddy S, Jakeman AJ, Chen S, Kassa- hun A, Lautenbach S, Matthews K, Mer- ritt W, Quinnm NWT, Rodriguez-Roda I, Sieber S, Stavenga M, Sulis A, Ticehurst J, Volk M, Wrobel M, van Delden H, El-Sawah S, Rizzoli A, Voinov A (2011) Environmental decision support systems (EDSS) development - Challenges and best practices. Environmental Modelling & Software 26: 1389-1402. Pavlovskaya M (2006) Theorising with GIS: a tool for critical geographies? Envi- ronmental Planning A 38: 2003-2020. doi:10.1068/a37326. Mapping Ecosystem Services176 5.5. Mapping specific ecosystem services Joachim Maes This chapter is one of the core chapters of this book. It contains guidance and exam- ples of how to map provisioning, regulating and cultural ecosystem services (ES). These three categories constitute a commonly used classification for ES (see Chapter 2.4) and thus for ES mapping. Different methods and models are used to map specific ES as indicators, used to quantify these three categories of ES, dif- fer remarkably. Provisioning ES are often quantified based on indicators for their ac- tual use/ES flow or demand (see Chapter 5.1) or their value. In contrast, assessment of regulating ES is usually based on supply indicators, such as the different ecological processes which are the basis of ecosystem regulation or avoided events (e.g. erosion or floods) and related hazards. Indicators for cultural ES have been mostly limited to recreation and (eco-)tourism for which both supply (popular ecosystems to visit) and de- mand (visitor numbers) are quantified. The use of provisioning ES involves the ex- traction of a product from the ecosystem (e.g. harvested biomass in tonne per ha per year; see Chapter 5.5.2). Mapping provi- sioning ES therefore relies often on data from statistical offices which collect statis- tics of water consumption, crop and timber harvests, fishery yields and livestock data. Sometimes these data are geo-referenced and are thus available as geospatial data lay- ers. If not available, statistical data can be spatially allocated over different ecosystem types, land use/land cover types or other spatial units such as watersheds or cadastral data to obtain mapped values. Regulating ES (see Chapter 5.5.1) are of- ten mapped by using biophysical models (e.g., ecosystem models, species distribu- tion models, water and air quality models; see Chapter 4.4). These models simulate the fate and transport of, for example, carbon, nitrogen, water or pollutants through the ecosystems and the environment. The eco- logical processes which are modelled can be used to infer values for regulating and main- tenance ES. Researchers mostly map poten- tial or flow of regulating ES (see Chapter 5.1). Demand for regulating ES is usually not mapped since it is conceptually less un- derstood (see Chapter 6.2). As already indicated, assessments of cultural ES (see Chapter 5.3.3), to date, are mostly limited to recreation and tourism. Actual use/ES flow needs to be mapped based on surveys, national accounts and data collec- tion (e.g. national park visitor statistics or entrance fees). These data can be combined with spatial data in order to map and assess the service and to provide detailed informa- tion on how ecosystems contribute to recre- ation and tourism. The remainder of this chapter goes into more detail for each of these. Each ES categories section contains a representative selection of ES for which mapping techniques and meth- ods are illustrated at various spatial scales. Chapter 5 177 5.5.1. Mapping regulating ecosystem services Joachim Maes, Chiara Polce, Grazia Zulian, Ine Vandecasteele, Carolina Perpiña, Inés Marí Rivero, Carlos Guerra, Sara Vallecillo, Pilar Vizcaino & Roland Hiederer Introduction Ecosystems regulate our environment by controlling or modifying the stocks and flows of material and energy that make up our ambient environment. Ecosystems help provide clean air and water by removing pollutants. They regulate the global and lo- cal climate through evapo-transpiration or simply by providing shade. They maintain habitats for insects and birds which support the production of crops or which suppress pests and diseases. They store carbon, buf- fer flows of water or maintain the fertility of soils. All these services are not directly con- sumed as goods by people but regulating ES provide many direct benefits by keeping a safe and habitable environment, supporting food production systems or processing and removing waste and pollution. Before mapping, it is important to under- stand first which ecosystem processes are at the basis of regulating ES and what the spatial characteristics are (scale and direction of dif- ferent flows of material and energy). Further- more, it is crucial to consider the difference between mapping capacity and mapping flow or use (Chapter 5.1). Actual use of a regu- lating service happens when there is a de- mand for it. Consider the protection of soils from erosion. Soil erosion in cropland occurs when wind or water remove fertile soils (top- soil). Vegetation, in particular grasslands and patches of forest, keep the soils fixed and thus avoid erosion. To provide the service, two conditions need to be met. First, there needs to be a demand for soil protection. Typically bare croplands on slopes are prone to erosion so farmers would benefit from enhanced ca- pacity of the ecosystem to protect soils. Sec- ond, the right ecosystems need to be present to provide the service wherever and whenever the service is needed. Understanding the different functions that underpin the delivery of regulating ES is thus the first step in a mapping process. In broad terms, ecosystems deliver regulating services by storing, capturing, absorbing or immobilising material such as carbon, wa- ter or pollutants, by maintaining or creating suitable conditions for species that provide regulating services (e.g. pollination, pest control, or soil quality regulation), or by buffering or mediating material and energy stocks and flows (regulation of waste and toxics, regulation of the atmosphere, water or soil erosion). The remainder of this chapter presents de- tailed examples of how different regulating and maintenance ES can be mapped. We frame ES mapping using the ES cascade model (see Chapter 2.3) and classify maps depending on whether they represent eco- Mapping Ecosystem Services178 system processes, functions (potential sup- ply), use or demand. The focus is on the biophysical mapping, not on mapping eco- nomic values. For every ecosystem service, we identify each time which underpinning functions can be mapped but we also de- scribe how to map actual use and demand. Although this chapter does not present all methods available for mapping, it gives the reader a flavour on how to map certain reg- ulating ES. Several other chapters provide other useful ways to map ES, for example, based on Bayesian statistics (Chapter 4.5) or matrix models (Chapter 5.6.4). The work presented here falls largely under the catego- ry of tier 3 maps (see Chapter 5.6.1). Such ecosystem service maps are based on models which are spatially resolved. Crop pollination Different ecosystems, particularly forest edges, flower rich grasslands or riparian ar- eas, offer suitable habitats for wild pollinator insects such as solitary or honey bees, bum- blebees or butterflies. As soon as these in- sects start foraging, the ecosystems that host these insect populations have the potential to increase the yield of adjacent crops which are dependent on insect-mediated pollina- tion. Fruit, vegetables, nuts, spices and oil crops profit from pollination. Mapping supply and demand of pollination services therefore involves mapping the suitability of ecosystems or habitats for pollinator insects, mapping flight distances between the nest and the crops that need pollination (which range from a few metres to a few kilome- tres) and mapping the occurrence of crops in need of pollination. Habitat suitability maps are usually based on a number of environmental layers or- ganised within a Geographic Information System (GIS). Examples of spatial layers rel- evant to pollination maps are land use/land cover, topography, distance from roads, or semi-natural vegetation. The choice of layers largely depends on which data are available and on knowledge about the ecological traits of the pollinator species. Habitat suitability maps, based on literature reviews and expert opinions, involve assigning a weight to each factor and then a suitability score to each class within a factor. Suitability scores, com- bined with an estimated foraging distance, are then combined to form a single (habitat) suitability map. Habitat suitability maps de- rived from empirical or statistical techniques require species occurrence data which can be either presence/absence or presence-on- ly records. The suitability is then derived by relating species occurrences to habitat factors by means of the chosen technique. Examples are regression methods, machine learning techniques and Bayesian statistics. Different packages and stand-alone software exist to implement these techniques; exam- ples include packages available within the software R, or stand-alone modelling tools such as Maxent or DIVA-GIS. The results of these models are then imported to GIS software to display maps of probability of species occurrence across the landscape of interest. Suitability maps for insect pollina- tors, regardless of the approach adopted to obtain them, can be interpreted as supply (potential services). Mapping the demand requires information on where crops that need pollination are grown in combination with information on crop dependency on insect pollination. Information on the pollinator-dependency can be obtained through literature and ex- pert knowledge. Crop location, on the other hand, can be obtained through a variety of resources. Examples of these resources are regional statistics on agricultural land and production, online databases, field samples and models (for instance when looking at fu- Chapter 5 179 ture potential crop distribution). The choice often depends on the extent of the area (e.g. regional vs. national vs. global data), on the crop type (e.g. perennial vs. annual) and on the agricultural practice (e.g. rotational agri- culture). Mapping the use can be based on overlay of supply (i.e. the habitat suitability) and demand (i.e. the crop distribution) or based on modelling the impact on yield in the absence of pollination. Soil protection The root network of grass, herbs, shrubs and trees physically keeps soil together; thus, it avoids soil from being eroded by the natural physical forces water or wind and flushed downstream to cause problems such as loss of fertile soil or siltation of watercourses. The demand for soil erosion control services is usually associated with farmland dedicat- ed to crop production on slopes. Rainfall on bare soils, for instance after harvesting, en- hances erosion. Mapping soil protection is largely based on mapping soil erosion. Five main factors contribute to soil erosion: rainfall, erod- ibility or soil type, absence of vegetation, slope and land management. These are usually modelled using the Revised Uni- versal Soil Loss Equation (RUSLE) equa- tion. By turning on or off the impact of vegetation or conservation practices, the contribution of ecosystems can be estimat- ed to avoid soil erosion which is then tak- en as an indicator for soil protection or soil retention. This is quantified by means of two indicators: the capacity of ecosystems to avoid soil erosion and soil retention (ac- tual ecosystem provision). The capacity or potential of a given land cover type to pro- vide soil protection can be mapped with a dimensionless indicator taking values be- tween 0 and 1. Capacity is assumed to be correlated with the amount of vegetation which, in turn, can be derived from re- mote sensing data such as the Normalised Difference Vegetation Index (NDVI). Soil retention can be calculated as the differ- ence between a model which calculates soil loss without vegetation cover and a model including the current land use cover pat- tern. A case study on mapping soil protec- tion is illustrated in Box 1. Climate regulation Ecosystems regulate our climate at various levels. In and around cities, urban forests provide shade during hot summer days and by evaporating water through their leaves, they cool down cities, thus delivering bene- fits in terms of saved energy costs or lowered ozone production and concentration. On larger spatial scales, forests, wetlands, coast- al systems and other ecosystems maintain comfortable atmospheric conditions and regulate climate. Yet, mapping ES which contribute to the regulation of climate, is often narrowed down to mapping carbon storage and carbon sequestration. Climate change science and policy is evidently the reason for this focus. Net primary produc- tion is at the basis of this and many other ES and therefore often mapped. Much useful information to map primary production is available through remote sensing, field ob- servations and modelling. Given the increase in atmospheric carbon and the consequences for climate, terrestri- al carbon pools are an important factor in the carbon balance. The terrestrial organic carbon pool (soil and vegetation) is estimat- ed to be 3500 Pg C, most of which (75%) is stored in soil. This is almost fivefold the amount of carbon in the atmosphere. The carbon stored in the soil mainly originates from dead organic material. The main gov- Mapping Ecosystem Services180 Box 1 . Mapping soil protection in Europe An assessment of soil protection in Europe in 2010: (a) Soil retention at European scale, b) Soil reten- tion in Central Portugal, c) Capacity to avoid soil erosion in Central Portugal and d) Structural impact in Central Portugal. Soil retention (Es) was calculated as soil loss without vegetation cover (structural impact, Y) minus soil loss including the current land use/cover pattern (the mitigated impact), mea- sured in tonns ha-1 year-1. The structural impact is the total soil erosion impact when no ecosystem service is provided. The capac- ity of a given land cover type to provide soil protection (e) is expressed using values ranging from 0 to 1 for every mapped grid cell. To estimate the capacity, the vegetation per land cover type was computed using the Normalised Difference Vegetation Index (NDVI), the environmental zones and the snow cover. The highest soil retention values corresponded to areas covered by forest, transitional woodland and shrubs (semi-natural vegetation areas) and pastures. Soil retention is also a function of structural impact (high potential erosion). Expressed differently, soil retention only occurs where soils run the risk of being eroded. In these places, vegetation cover protects the soil against water flows (surface runoff), reduces the structural impact and, therefore, effectively delivers a service. A close up is presented for the central part of Portugal (Alentejo and Centro Regions). In the Tagus river valley, soil retention is low (light orange areas) due to a low structural impact and the dominant land use type, mainly agriculture. High soil retention (high provision of the service, in dark blue) results from the combination of high structural impact and high capacity to avoid soil erosion, for instance in forested areas. In contrast, if the inherent structural impact is low, the provision of the service (soil protection) is low as well, thus lowering the role of vegetation in soil protection. a c b d Chapter 5 181 erning factors for the status of the soil or- ganic carbon (SOC) pool are land use/land cover and local climatic conditions. Chang- es in land use and management practices can lead to imbalances in the flux between carbon pools. Depending on environmental conditions, the SOC pool can act as either a source of atmospheric carbon or a sink, i.e. removing carbon from the atmosphere. Mapping changes to the SOC pool can be based on the methodology of the Interna- tional Panel on Climate Change (IPCC). The method uses type of climate, soil, cat- egory of land use, management and input practices as factors influencing SOC stocks. For each factor, the relative effect of changes to the SOC pool is provided for different cli- mate/soil regions. When all factors remain unchanged, an equilibrium in the SOC pool is assumed to be reached after 20 years. Given the factors influencing SOC content, the spatial distribution of SOC stocks is very variable (Figure 2). Most of the global SOC is stored in the northern hemisphere where cool and moist conditions favour plant decom- position into soil organic matter. However, under wet conditions and high productivity of vegetation, organic material may also ac- cumulate in tropical regions, such as in peat lands of south-east Asia. In tropical forests, the amount of carbon stored in the above- ground vegetation exceeds the carbon stored in the soil with the exception of peat lands. Water regulation Forests, grasslands and wetlands are ecosys- tems with a high capacity for regulating the flow of water. This is particularly import- ant for ensuring the supply of a sufficient quantity of water to support the immediate environment whilst avoiding extreme fluc- tuations in water flows. Where water is not properly regulated by the ecosystem (e.g., in cities, where the natural water cycle is often interrupted by impermeable surfaces), there is a much higher risk of such fluctuations, potentially leading to flooding or water Figure 2. Spatial distribution of global soil organic carbon density (t C ha-1). Source: FAO and ITPS. Mapping Ecosystem Services182 shortages. The provision of water regulation can be mapped by breaking down the pro- cess into its various components. Ideally, the landscape should naturally retain and store an adequate amount of water for its needs, whilst limiting the amount of surface run- off - an excess of which may cause flooding further downstream. Water flow through a landscape may be influenced by the follow- ing natural processes, all of which contrib- ute to the storage of water and therefore the reduction of surface runoff: interception by vegetation, storage in surface water bodies, infiltration and retention in soil and perco- lation to groundwater stores. In addition to these processes, the amount of water which can be retained will also be affected by the slope of the landscape and by the degree of permeability of the soil. Steeper slopes will promote faster surface runoff, whilst flatter areas allow greater time for infiltration of water. Impermeable sur- faces (e.g. artificial infrastructure such as roads and buildings) represent a barrier to the infiltration and retention of water, thus promoting surface runoff. Figure 3 gives an overview of the parameters taken into account to map the water reten- tion as a proxy for the water regulation ca- pacity of the ecosystem. The retention of water in vegetation, surface water bodies, soil and bedrock (groundwa- ter stores) are considered landscape storage factors. Additionally, the influence of slope and surface imperviousness are considered as physical factors altering the actual water re- tention capacity of the landscape. The contri- bution of each process to the final indicator is approximated using one or more parameters or characteristics of the landscape. The pa- rameters shaded in grey are those which are changeable over time. The various factors are combined to give the final composite indica- tor representing relative landscape water re- tention or, rather, the capacity of the ecosys- tem to provide water regulation as a service. Pest control Agricultural ecosystems are often harmed by pests such as insects (i.e. caterpillars) and small mammals (i.e. moles), significantly re- ducing the harvested share of crop produc- tion. However, nature offers natural fight- ers against these pests, thus saving farmers billions of dollars annually by protecting crops and reducing the need for chemical control. There are different groups of natu- ral enemies known to play a key role in pest Water Retention Index Landscape Storage Factors Vegetation Rv Water Bodies Rwb Soil Rs Water Holding Capacity Organic carbon content Relative Bedrock PermeabilityBedrock Rgw Slope Surface imperviousness Leaf Area Index Share of surface water bodies Physical Factors Figure 3. Schematic overview of the structure of the indicator for mapping water retention. Parameters in grey are dynamic and thus change over time. Chapter 5 183 control, such as birds, mammals, spiders, lady bugs and other types of organisms. So, mapping pest control clearly relies on spatial information on the distribution of predator species (species distribution models, see also section on pollination). We show below an example of mapping potential pest control by birds in agricul- tural systems (Figure 4). The example is based on species distribution models of 49 bird species, recognised as pest-control pro- viders. Modelled species include the Little Owl (Athene noctua), a known hunter of mice, voles, shrews, moles and rabbits and the Hoopoe (Upupa epops) which has an in- sect-rich diet. Species distribution models map the prob- ability of species occurrence based on field observations. A probability threshold can be defined for instance at 50% to assume the presence of a certain species. By over- laying all the species occurrence maps of the 49 modelled species, a map of potential pest control by predatory birds is obtained. Figure 4. Spatial distribution of predatory bird species richness in the European Union. The close-up around Paris shows that species richness is lower near urban areas (mapped in red). Mapping Ecosystem Services184 Higher species richness corresponds to a more diverse community of natural preda- tors and is assumed to exert a greater control on pest populations. Figure 4 shows poten- tial pest control by bird species across Eu- rope. The inset is a close-up around Paris, showing spatial differences in bird species richness, with low values (areas in yellow) in and around the urban areas (in red). Air quality regulation Air pollution is one of the main environ- mental risks for human health and is the main cause of premature deaths. In this con- text, abatement of pollution has become of major concern especially in areas with high pollutant concentrations, typically urban areas. Maintaining and developing green ur- ban areas can be part of an integrative strat- egy to help increase air quality in European cities. Trees reduce temperatures in cities by evaporating water and they remove air pol- lutants and particulate matter via their leaves through dry deposition. Urban trees, green areas and forests surrounding cities have the capacity to remove significant amounts of pollutants thereby increasing environmental quality and human health. Mapping air quality regulation is based on three types of information: the dry deposition velocity (supply), the removal of air pollut- ants (flow) and human exposure (demand). The pollutant dry deposition velocity by vegetation is considered often as a proxy to assess the ecosystems capacity to remove pollutants from the atmosphere. This quan- tity measures the rate at which pollutants are collected from the atmosphere by tree leaves. The contribution of vegetation is often mapped and modelled using spatial- ly explicit data of the leaf area index (LAI). The LAI is defined as the one-sided green leaf area per unit ground surface area. The larger this area, the more pollutants are cap- tured by trees. Furthermore, the pollutant removal flux by vegetation, which is estimated as the prod- uct of pollutant dry deposition velocity by vegetation and pollutant concentration, is usually considered as a good measure for the ecosystem service flow. Finally, demand for the service can be mapped using population exposure to pollutant con- centrations beyond the limit established within the legislation currently in force. Maps of the atmospheric concentration of pollutants are essential inputs to map air quality regulation as an ecosystem service. Mostly, they rely on a network of monitor- ing stations where different pollutants are measured. The measurements collected by different monitoring stations can then be interpolated to obtain maps of concentra- tions. Several GIS techniques exist to per- form interpolation by, for example, kriging and spatial regressions. Figure 5 presents an example for the Barce- lona metropolitan region. In this case, con- centrations of NO2 were estimated using Land Use Regression (LUR) models. The LUR model was built using NO2 concentra- tion measurements for the year 2013 from the operational monitoring stations as de- pendent variables and a set of spatial predic- tor parameters (independent variables) that were considered to be the most relevant for distribution of NO2 concentrations, related to land cover type, geomorphology, climate and population. The map of unsatisfied de- mand for air quality regulation was generat- ed from the population living in areas where annual mean concentrations exceed the EU limit value (40 µg/m3 for NO2). Chapter 5 185 Conclusions Much progress has been achieved in map- ping regulating ES. Data, maps and models are often available from other scientific dis- ciplines such as research on air quality, hy- drology, climate change or biodiversity and they need relatively minor adaptations for applications in an ES context. Mapping regulating ES is often based on mapping the capacity of ecosystems to pro- vide these services rather than mapping the actual use of the service. One possible rea- son is that it is not always clear what the use is of a regulating service in comparison with provisioning services. Mapping the capacity of ecosystems to pro- vide regulating services can be based on the combination of different data layers to arrive at a composite index between 0 and 1 where 0 stands for no capacity to deliver a service and 1 stands for maximum capacity. Where species provide a regulating service, capacity is often approximated based on species oc- currence which can be mapped. In the case of air quality regulation, mapping capacity is based on mapping the dominant physical process (deposition). The case of soil protection demonstrates how actual flow or use of regulating services can be done. Avoided erosion is modelled as the difference between erosion in the ab- sence of vegetation and erosion with protec- Figure 5. Indicators for the assessment of air quality regulation of NO2 in the Barcelona Metropolitan Re- gion: concentration, deposition velocity, removal capacity and population affected. Mapping Ecosystem Services186 tive cover. This technique can be applied to other regulating services such as pollution and excess nutrient control. Demand for regulating services can be mapped if spatial data are available which identify use, users or beneficiaries. Examples are crops which need pollination, farmland exposed to erosion or people exposed to low levels of air quality. Further reading and resources Civantos E, Thuiller W, Maiorano L, Guisan A, Arajo MB (2012) Potential impacts of climate change on ecosystem services in Europe: The case of pest control by verte- brates. BioScience 62: 658-666. FAO and ITPS (2015) Status of the World’s Soil Resources (SWSR) – Main Report. Food and Agriculture Organisation of the United Nations and Intergovernmental Technical Panel on Soils, Rome, Italy. 650pp. Guerra CA, Maes J, Geijzendorffer I, Metzger MJ (2016) An assessment of soil erosion prevention by vegetation in Mediterranean Europe: Current trends of ecosystem ser- vice provision. Ecological Indicators 60: 213-222. Nowak DJ, Crane DE, Stevens JC (2006) Air pollution removal by urban trees and shrubs in the United States. Urban Forest- ry and Urban Greening 4: 115-123. Polce C, Termansen M, Aguirre-Gutiérrez J, Boatman ND, Budge GE, Crowe A, Gar- ratt MP, Pietravalle S, Potts SG, Ramirez JA, Somerwill KE, Biesmeijer JC (2013) Species Distribution Models for Crop Pol- lination: A Modelling Framework Applied to Great Britain. PLoS ONE 8: e76308. Scharlemann JPW, Tanner EVJ, Hiederer R, Kapos V (2014) Global soil carbon: Un- derstanding and managing the largest ter- restrial carbon pool. Carbon Management 5(1): 81-91. Zulian G, Maes J, Paracchini M (2013) Link- ing Land Cover Data and Crop Yields for Mapping and Assessment of Pollination Services in Europe. Land 2: 472. Chapter 5 187 5.5.2. Mapping provisioning services Marion Kruse & Katalin Petz Introduction Material and energy outputs from ecosys- tems are usually classified as provisioning services. These are tangible goods or services that are directly used, traded or exchanged by all human beings. They can be grouped into nutrition (e.g. cultivated crops, seafood from aquaculture, wild food), materials (e.g. fibres and genetic materials) and energy ser- vices (e.g. fuel wood). Some of them, such as cultivated crops and animal outputs, are amongst the most mapped ecosystem ser- vices (ES), whereas others, such as genetic materials and energy provided by animals, have been studied or mapped less frequently to date (Figure 1). Provisioning services are often produced and consumed or used in different places. They are generally transported from the place of pro- duction (i.e. supply) to the place of consump- tion (i.e. demand). It is more common and eas- ier to map the supply, as it is spatially explicit and directly depends on the ecosystem’s structure and functioning, whereas the demand is a function of socio-economic drivers. The economically im- portant crop and animal products, as well as tim- ber and fish products, can be closely associated with agriculture, forestry and fishery/aquaculture and consequently, their related land cover/use types. As they represent traditional economic activities and research focuses that have existed for a very long time, a large body of subject-spe- cific knowledge and data sets are available to quantify ES supply based on these economic sectors. This fact also gives the opportunity to analyse changes and trends of these ES in many regions. These production systems are usually monocultures and require a large amount of human input (Chapter 5.1). On the contrary, wild plants, water and genetic resources are less clearly associated with one specific land cover/use class and are generated in more diverse and semi-natural or natural ecosystems and landscapes. The production of ES is not only loca- tion-specific, but it is also dynamic over time Figure 1. Schematic overview of least and most mapped provisioning services. Mapping Ecosystem Services188 (Chapter 5.3). Examples include crops main- ly being grown and harvested in spring-sum- mer period in the temperate zones. In the tropical zone, the growing season lasts all year around. Another example is the dynamic supply of drinking water from mountain re- gions, which follows the seasonal changes in hydrological and climatological conditions. Mapping methods for provisioning services Based on the CICES classification (Chapter 2.4), provisioning services can be grouped into the classes reported below. Differ- ent methods and data sets are available for mapping these classes. A short overview of selected mapping methods and data sets is provided in the following sections. Cultivated crops & reared animals and their outputs (e .g . cereals, vegetables, meat, milk) These provisioning services are mainly com- mercially valued and traded as the direct output of agriculture from arable land and pastures. They are amongst the best mon- itored ES and their level of production is documented in agricultural statistics or ac- counting in many areas. Therefore, they can be easily mapped using land cover/land use maps in combination with indicators of crop or animal production (e.g. t/ha/year crop yield, number of animals/ha, l/ha/year milk production) from national or other sta- tistics. This corresponds to a tier 2 mapping approach (see Chapter 5.6.1). This method has minimal data requirements and is there- fore easy and quick when the corresponding data are at hand. With such data sets, maps for these provisioning services can be generat- ed for local up to global scales. Use of a single indicator, however, neglects the effects of the management regime and the environmental characteristics of the agricultural ecosystem (e.g. soil texture, climatic and hydrological conditions), all of which influence the level and quality of ES generated. It is also possible to include anthropogenic system inputs and environmental effects as indicators instead of only crop yield or animal numbers. Due to the commercial character, there is a large addi- tional input (e.g. fertilisers, pesticides) in most agro-ecosystems. Furthermore, it is important to notice that in the case of reared animals, land cover/land use does not automatically correspond to the area of supply. Livestock is often kept in buildings, resulting in point ac- cumulation within the respective map. There are also crop or animal production models (e.g. Common Agricultural Policy Regionalised Impact (CAPRI) model or Agri- cultural Production Planning and Allocation (APPA) model) accounting for the ecosys- tem’s capacity, environmental effects and hu- man inputs to obtain more accurate results. Nevertheless, these models are time and data intensive. They are suitable if the aim is to bet- ter understand a certain production system or create a crop and animal production map for a certain location under a specific socio-eco- nomic scenario or environmental constraint. For general purposes and the mapping of multiple provisioning services, look-up ta- bles are in common use (Chapter 5.6.4). For some outputs from reared animals (meat, milk), only aggregated or average data exist (slaughtering for a defined reference date). On the local scale (e.g. farm), detailed anal- yses can be included in maps, such as varia- tions over a season. Considering the growing season of cultivated crops, the supply does not always match the continuous demand. Over the entire growing season, up until the moment of harvest, crops can be considered as only potential provisioning services. The real use (flow) is connected to harvest, pro- cessing and consumption. Chapter 5 189 Wild plants and wild animals and their outputs (e .g . wild berries, mushrooms, wa- ter cress, game, fish, honey from wild bees) This class includes both commercial and subsistence berry and mushroom collec- tion, fishing and hunting for food. These less marketed ES can provide subsistence especially in less developed countries, while they are considered by some stakeholders and researchers as hobby or recreational ac- tivities in other regions. Only few examples exist for the mapping of these provisioning services because of their individual charac- ter. Statistical data are available for hunting and recreational fishing in some regions or countries. Data and information are often available for commercial fishing as regu- lations (e.g. exclusive economic zone and catch quota) have to be respected in many regions. However, mapping the exact area of where fish and seafood are extracted in- cludes uncertainties due to the mobility of most species. Mapping fishing grounds re- quires GPS data or interview data. Usual- ly, statistical data are grouped into specific areas (e.g. Baltic Sea or North Sea) or on an administrational level (e.g. states and countries) making the mapping less spatially explicit. For some individual or subsistence activities, such as berry/mushroom picking, recreational fishing or hunting, licences are needed. This can be used as a proxy to quan- tify the amount of potential users. Most im- portant though, is the exact area of supply and the respective amount of provisioning service. This requires laborious and possibly expensive, field studies and interviews. Data from random sampling is often used for ex- trapolation. Methodological studies need to reveal which natural and socio-economic settings, extrapolation or value-transfer are accurate enough for reliable results. Mapping of berry and mushroom collection and game hunting is possible by combining land cover/land use and species habitat maps with other biophysical layers (e.g. manage- ment intensity, climatic factors) and with accessibility or travel time from settlement areas. Another approach is participatory mapping (Chapter 5.6.2) of indicators such as kg/ha berries collected or the location of honey collection, relying on the knowledge and wild food collection habits of local in- habitants and stakeholders. Wild food collec- tion is also closely related to cultural services (such as cultural diversity and traditions) and is affected by the human-environmental re- lationship and societal conditions (e.g. laws, regulations, property rights). This informa- tion can be included in the mapping process- es through, for example, overlaying protected areas or area with restricted access. Figure 2 presents an example for medicinal plants. Few studies are available on regional or na- tional scales. Mapping of these services (ex- cept commercial fishing) is most suitable for the local or regional scale, making it possible to include high resolution data and informa- tion needed for sustainable management. Fibres and other materials from plants and animals (e .g . timber, cotton, grass as fodder) This class includes both consumptive and ornamental uses and both commercial (e.g. industrial timber production) and subsis- tence (e.g. local wood collection) uses. Tim- ber, grass and fodder production have been widely mapped, whereas other materials, such as cotton and silk are rarely mapped. Mapping methods range from the use of a single indicator (e.g. m3/ha wood harvested, kg/ha grass collected) to complicated forest or vegetation production models e.g., the European Forest Information SCENario Model EFISCEN (see Figure 3 for an ap- plication), the Global Forest Model (G4M). Subsistence use can be mapped using partic- ipatory techniques as well, especially at local scales. An application example for mapping Mapping Ecosystem Services190 Figure 3. Mapping timber harvest (m³/ha/yr) in 2050 under the VOLANTE A2 business-as-usual scenario modelled with the EFISCEN (European Forest Information SCENario) model on the European scale. Yellow indicates no harvest and grey indicates non-forest areas on the map (Source: Schelhaas & Hengeveld pers. comm). Figure 2. Participatory mapping of medicinal plants in the Bereg region, Hungary: Local stakeholders were questioned if, where and which medicinal plants grow and if they are collected. The growth and col- lection of medicinal plants were related to different land cover types. Although, the study’s objective was an assessment of ES, the results could be translated into a map showing the location of this provisioning service in a further step (Source: Petz et al. 2012). Chapter 5 191 local fuel wood supply is shown in South Af- rica in Figure 4. For timber, statistical data are available. However, separation from fuel wood is difficult as sometimes several prod- ucts are manufactured from the same source (e.g. timber, woodchips). In contrast to an- nual or seasonal supply from some fibres and fodder (e.g. cotton, hay), wood products are usually only harvested over longer time pe- riods (> 40 years) due to growing phases in the temperate zones. Fast-growing species are harvested in the (sub-)tropical regions at shorter intervals. Similarly with cultivated crops and reared animals, there is a mismatch between the supply and demand for (some of ) these services, as timber and wood products are marketed across the globe. Some fibres (e.g. cotton) belong to the most important provi- sioning services and are heavily traded glob- ally. Very few studies in the context of the ecosystem service concept exist, although some (global) statistics on production are available. Recycling and multiple uses or purposes of materials result in possible un- certainty of these assessments. The provisioning services of cultivated crops, reared animals and their outputs and materials from plants and animals are often produced on the same farm. In this case, double-counting is possible. Nowadays, fodder is however imported to many inten- sive farming regions, making the assessment of provisioning services more difficult or uncertain. Local mismatches of supply and demand result from this, which make the mapping of these services incomplete. Figure 4. Mapping fuel wood supply in the Baviaanskloof Catchment in South Africa using local data consultations. Combining multiple indicators (left) results in a fuel wood yield map (right). Several data sets were combined to show the spatially diverse supply of fuel wood (including topography, accessibility and also conservation areas), (Source: Petz et al. 2014). Mapping Ecosystem Services192 Genetic materials from wild plants and animals (e .g . medicines and wild species used in breeding programmes) Genetic material from wild plants can be used for biochemical industrial and pharma- ceutical processes (e.g. medicines, fermenta- tion), as well as bio-prospecting activities (e.g. wild species used in breeding programmes). Genetic materials have been mapped infre- quently. On a similar basis, wild food and medicinal plants have also a close link with cultural traditions and societal conditions. The occurrence or supply of medicinal plants could be mapped similarly to wild food by combining species richness and land cover/ land use data or applying participatory map- ping studies. Biodiversity models could pro- vide useful information about the occurrence of different wild species. Suitable habitats for and spatial dynamics of mobile species, such as insects or mammals, can be explored with agent-based models (Chapter 4.4). These provisioning services have been rarely mapped, although there is ongoing research (considering different species and ecosys- tems ranging from the tropical rainforest to marine environments). Usually, this covers only limited areas. Animals and plants from aquaculture Mapping provisioning services from aquat- ic ecosystems is usually more difficult. In- formation on water bodies is often not as detailed from land cover/land use maps as for terrestrial ecosystems (see Chapter 7.4). More detailed information about protected areas, different habitats, or spawning areas is needed to map animals and plants from aquatic ecosystems. An application example for mapping fishery areas in the Baltic Sea is shown in Figure 5. Wild caught fish in marine and freshwater ecosystems is an important food resource globally. Due to declining stocks and regu- lations (e.g. EU fisheries regulations) aqua- culture is employed more and more to meet the demand for seafood and algae. These data are available on different spatial scales and, in most cases, over a very long time period as they are important for the economy. Here also single indicators (e.g. fishing statistics in t/year) are available. In- frastructure from aquaculture, such as cages, basins, ropes, is visible in the field and can be used to identify the extent of the provi- sioning area. Water for drinking and non-drinking purposes Water extraction is usually undertaken in single spots where the conditions are suitable (i.e. infrastructure and water quantity, qual- ity and intensity). Groundwater is recharged over a larger area and depends on ecosystem conditions, such as substrate and vegetation cover. Surface water is used in many regions where ground water extraction conditions are not suitable. Maps can show groundwater yield and the amount of water (m3) that can be extracted without declining the yield. Hydrological models can be applied to simulate the effects of changes in consumption and hydrologi- cal and climatic conditions (Chapter 4.4). Some statistical data are available for aver- age water consumption (drinking water, non-drinking water) and the water incor- porated or locked in products (e.g. food, clothes). Large regional mismatches occur on the global scale due to trading of prod- ucts. The price for water could be used as an indicator for mapping as well. Temporal changes in water demand is an important aspect in management, especially in areas with high usage (e.g. from tourism). Chapter 5 193 Figure 5. Mapping fishery areas in coastal ecosystems of the German Baltic Sea. The example shows that a land cover based approach results in an over-estimation of supply area. Here additional informa- tion is included to depict spatially explicit areas of restriction due to protection within a National Park. Boat traffic is partially prohibited. Temporal fishing restrictions exist in the spawning areas. Recreational fishing (at land) is only allowed in the designated spots and areas where boat traffic is allowed. continous urban fabric pasture, meadows and other permanentgrasslands under agricultural use discontinous urban fabric airports construction sites green urban areas sport and leisure facilities non-irrigated arable land industry or commercial units broad leaved forest mixed forest natural grasslands moors and heathlands transitional woodland shrub beaches, dunes, and sand plains coniferous forest sparsley vegetated areas inland marshes salt marshes water courses water bodies coastal lagoons sea and ocean peat bogs CORINE LAND COVER 2012 Darss Zingst Bodden Chain, Northern Germany Mapping Ecosystem Services194 In some cases, water needs to be pumped from other areas to fulfil the demand. These regional and temporal mismatches make mapping of water as an ES challenging. Plant and animal-based energy resources (e .g . fuel wood, crops and labour provid- ed by animals) It is more common and easier to map plant- based energy resources than animal-based ones. The mapping of energy crops, such as oilseed rape, is similar to food crops. Sometimes they are also competing in cultivated areas. The supply of fuel wood can be mapped with a single indicator (e.g. forest areas, oc- currence of certain tree species), forest pro- duction models (e.g. EFISCEN, G4M) for commercial use and participatory approach- es for subsistence use. Large regional differ- ences exist. In some regions, fuel wood is the only energy source for cooking and heating whereas, in other regions, it is only a supple- mentary source or even unnecessary (e.g. ur- ban areas with good energy infrastructure). Labour provided by animals as an ES has not been mapped yet. It could be mapped, for example, using statistics involving the quantity of animals. In some areas, la- bour provided by animals is important in agriculture, but also for transportation. However, due to mechanisation in many sectors, this provisioning service is of less and less importance. Challenges and solutions for mapping provisioning services There are several maps and data sets avail- able that facilitate the mapping of provi- sioning services. As usual, all methods have advantages and disadvantages regarding un- certainty and the objective/purpose of the maps (see Chapter 6). When using statistical data, the maps are not always spatially explicit. These data sets are often generated at administrative levels (e.g. municipality, regions), which do not neces- sarily match the case study area. Although farm land might be stretched over sever- al administrative units, the respective data (e.g. number of animals, yields) are only assigned to the location of the farm. Addi- tionally, wild animals which hunt and fish are mobile and forage in areas which do not always match with reporting units. Furthermore, when using statistical data, it is often not possible to distinguish between the different uses of the product (e.g. rapeseed for human nutrition, biodiesel or fodder). Many provisioning services are supplied in larger areas that can be represented by poly- gons. However, there are sometimes import- ant (hot) spots which affect only parts of an area. Besides static services providers (e.g. forests), there are some mobile ones, such as fish and seafood (see Chapter 5.2). Though the ecosystem might be restricted to aquatic ones, there are several factors that might de- termine the size of catches (e.g. in recreation- al fishing) or the exact location of the actual service (e.g. exact location of caught fish). Temporal aspects are generally difficult to integrate on a map but several maps can be used to show the change of the supply or de- mand of the provisioning service over time (i.e. seasonal maps; see Chapter 5.3)). Maps on wild food (mushrooms and berries) can change significantly between years due to cli- matic variations or silvicultural management. The quality of the modelling results depends on the input data and the research questions. All participatory mapping approaches are Chapter 5 195 impacted by the number of stakeholders and their background and how they are instructed. Following the grouping and classification of provisioning services from CICES, it be- comes clear that the main challenge is the large and detailed amount of possible pro- visioning services supplied in an area or de- manded by consumers. Globalisation makes it more challenging to track down detailed information on a spatially explicit scale. For many of the examples of provisioning ser- vices at the CICES class level, no data or in- formation exist or the ES are part of a larger supply chain. Therefore, the map-maker must carefully decide to which detail pro- visioning services should be analysed or if a distinction into broader classes is necessary (e.g. wooden biomass from forests instead of subdividing into type and use, such as cellu- lose, timber, fuel food). The question of the purpose of the map and the necessary details are also relevant. For a coarse (first) mapping of provisioning ser- vices, data sets and methods are available from local-global scale, especially when land cover/land use and statistical data can be used. The specific (policy) question guides the work and detail needed to create a prop- er map of provisioning services (Chapter 5.4). For the least mapped services, direct mapping based on sampling can overcome the lack of suitable data. Another challenge is that many provisioning services outside of markets, which are mainly for private use/subsistence, have no detailed or comprehensive data sets for proper map- ping. Many people are not aware of these “benefits”, such as ornamental use, and do not keep detailed records or data sets. This also applies for mushroom or berry picking, or recreational fishing for personal con- sumption. What is most important in this circumstance, is raising awareness to show the interlinkages of ES and the need for near-nature ecosystems for the supply. The purpose and importance of these provision- ing services need to be taken into account to decide whether or not a provisioning service should be classified as cultural ES or not. The main challenge of incorporating the temporal dynamics of provisioning services in maps remains. Many studies are limit- ed to a conceptual description of mapping provisioning services. A larger body of ap- plications for all provisioning services would result in progress in closing the knowledge gaps, which lead to incomplete assessments of ES. A final question remains: should we map the area and spatial extent of provision- ing services, which is comparatively easy with land use/land cover maps, or should we also include information on the amount, quality and benefits? Conclusion Maps of provisioning services are essential, as provisioning services play a key role in economic activities from local to global scale and from the past to the future. Information on the distribution and intensity of provi- sioning services supply and demand is need- ed for sustainable land use management and policy-making. The more important an eco- system service is (e.g. food), the more data or information are available. As provisioning services are diverse and are delivered by different ecosystems, several methods are needed in the assessment and mapping process, ranging from simple indi- cators or land cover/land use data to mod- elling and participatory approaches. Many details should or can be integrated in pro- visioning services maps, but the purpose of the map guides the information content. Mapping Ecosystem Services196 As a close link between provisioning services, regulating services and cultural services ex- ists, it is therefore advisable to cross-refer- ence the respective maps. Further reading Brown G, Fagerholm N (2015) Empirical PPGIS/PGIS mapping of ecosystem ser- vices: A review and evaluation. Ecosystem Services 13: 119-133. García-Nieto AP, García-Llorente M, Ini- esta-Arandia I, Martín-López B (2013) Mapping forest ecosystem services: From providing units to beneficiaries. Ecosystem Services 4: 126-138. Kandziora M, Burkhard B, Müller F (2013) Mapping provisioning ecosystem services at the local scale using data of varying spa- tial and temporal resolution. Ecosystem Services 4: 47-59. Karabulut A, Egoh BN, Lanzanova D, Griz- zetti B, Bidoglio G, Pagliero L, Bouraoui F, Aloe A, Reynaud A, Maes J, Vandecasteele I, Mubareka S (2016) Mapping water provisioning services to support the eco- system–water–food–energy nexus in the Danube river basin. Ecosystem Services Ecosystem Services 17: 278-292. Petz K, Minca EL, Werners SE, Leemans R (2012) Managing the current and future supply of ecosystem services in the Hun- garian and Romanian Tisza River Basin. Regional Environmental Change 12: 689-700. Petz K, Glenday J, Alkemade R (2014) Land management implications for ecosystem ser- vices provision in a South African rangeland. Ecological Indicators 45: 692-703. Rasmussen L V, Mertz O, Christensen, AE, Danielsen F, Dawson N. Xaydongvanh P (2016) A combination of methods need- ed to assess the actual use of provisioning ecosystem services. Ecosystem Services 17: 75-86. Schulp CJE, Thuiller W, Verburg PH (2014) Wild food in Europe: A synthesis of knowl- edge and data of terrestrial wild food as an ecosystem service. Ecological Economics 105: 292-305. Verkerk PJ, Levers C, Kuemmerle T, Lindner M, Valbuena R, Verburg PH, Zudin, S (2015) Mapping wood production in Eu- ropean forests. Forest Ecology and Man- agement 357: 228-238. Chapter 5 197 5.5.3. Mapping cultural ecosystem services Leena Kopperoinen, Sandra Luque, Patrizia Tenerelli, Grazia Zulian & Arto Viinikka Introduction Cultural ecosystem services (CES) bind ele- ments between social and ecological concepts. They are seen as nature’s intangible benefits related to human perceptions, attitudes and beliefs. People obtain spiritual enrichment, cognitive development, reflection, recreation and aesthetic experiences from ecosystems (Table 1). People’s perceptions can differ sig- nificantly, not only person by person, but also from one area and culture to another. There- fore, CES are not readily transferrable from one place to other environments. CES have both use and non-use values in- cluding existence, bequest, option and in- trinsic values. Relational values referring to cultural identity and well-being derived from people’s relationships with both other people and nature and mediated by particular places are also typical of CES. The focus of CES can be on individual needs and values or those fulfilled and possessed at a collective level. At both levels, CES concretely contribute to human well-being, public health and psy- chological experiences. As a result, CES are greatly appreciated by people and, in many instances, they are even better acknowledged than other ES. In more traditional commu- nities, CES are often essential for cultural identity, livelihoods and even survival. The problem is, however, that many CES are difficult to quantify or their value too com- plex to assess and map. That has led to an over-emphasis on recreation and ecotourism which are empirically and conceptually easi- er to identify and measure while, at the same time, neglecting other important CES that matter to people but which are not as easy to measure (e.g. spiritual services). This chapter aims to present how to sur- vive the challenge of mapping non-materi- al services, what examples of methods exist to map the potential provision of CES at different spatial levels, how to involve stake- holders in the mapping activity and what are the options that social media provide for CES mapping. Many methods useful for mapping CES are also presented elsewhere in this book. What is specific about mapping cultural ES? As CES are considered non-material bene- fits, their quantification can be rather chal- lenging: how to get hold of values linked to human perceptions compared to, for exam- ple, provisioning services where the actual stock of material can be quantified using different units of measure? Rapid quanti- tative mapping might not be easy for com- plex CES but it is possible to map them by combining knowledge and (also qualitative) methods from different disciplines, includ- ing not only natural and environmental sci- ences but also psychology, anthropology and other social sciences. Mapping Ecosystem Services198 In order to map CES, methods to capture cultural norms and to express plurality of values in a spatially-explicit way are needed. Some researchers consider CES and their value measurable since they are expressed in human actions. Values ascribed to CES can be identified, for example, using the presence of certain products of an area, vis- ible manifestations of CES in the physical landscape, or the number of studies, artis- tic representations etc. about an ecosystem as proxies. Spatial datasets giving location to certain socially or culturally normative values of the environment (e.g. inventoried cultural heritage or valuable landscapes; green areas of sufficient size and location) can also be used as indicators of areas pro- viding CES. However, if a more detailed and precise pic- ture of CES is to be gained in a specific area, local people must be involved in mapping. Thus, mapping CES is inherently participa- tory if it is to be done properly. Division Group Class Examples Physical and intellectual interactions with biota, ecosystems, and land-/ seascapes [en- vironmental settings]    Physical and experiential interactions   Experiential use of plants and animals . Experiential and physical use of land- / seascapes in different environmental settings In-situ whale and bird watching, snorkelling, diving and other experiential enjoyment of nature. Walking, hiking, climbing, boating, leisure fishing (angling), leisure hunting and other physical activities in nature. Intellectual and repre- sentative interactions    Scientific Subject matter for research both on location and via other media. Educational Subject matter of education both on location and via other media. Nature as a location for education. Cultural heritage in connection to nature Historic records, cultural heritage e.g. pre- served in water bodies and soils, interplay of nature and culture, traditional uses of nature, cultural identity. Entertainment Ex-situ viewing / experience of natural world through different media, such as photographs, films, literature. Aesthetics Beauty of nature and land- / seascapes, artistic representations of nature. Spiritual, symbolic and other interactions with biota, ecosystems, and land-/ seascapes [en- vironmental settings]    Spiritual and/ or emblem- atic   Symbolic Emblematic plants and animals e.g. national symbols such as American eagle or Welsh daffodil, sense of place, place identity. Sacred and / or reli- gious Spiritual values, ritual identity e.g. ‘dream paths’ of native Australians, holy places, sacred plants and animals and their parts. Other cultur- al outputs   Existence Enjoyment provided by the pure existence of wild species, wilderness, ecosystems and land- / seascapes. Bequest Willingness to preserve plants, animals, eco- systems, land- / seascapes for the experience and use of future generations; moral / ethical perspective or belief. Table 1. CES according to Common International Classification of Ecosystem Services v. 4.3. Chapter 5 199 As a result, mapping CES capacity and de- mand are interwoven. What is considered as potential capacity of an area is depen- dent on what brings people well-being and what people perceive they need and value in terms of CES. This needs to be understood first. When there is knowledge of this, dif- ferent datasets can be used to identify where valued environments, features, species, or opportunities for specific experiences having the capacity to provide CES are located (Ta- ble 2). Mapping actual demand for CES is frequently done by using participatory map- ping methods or indirectly utilising con- tents of social media. Participatory mapping means involving stakeholders, locals, etc. to identify, assess or otherwise value and point out on a map, areas or spots where they en- joy or feel CES (see more about participato- ry mapping in Chapter 5.6.2). Social media based methods include, for example, asking people to take photos of perceived CES in an area, involving people in scoring photos of different landscape types or, for example, analysing geo-tagged photos uploaded on social media. The latter is an indirect meth- od to reveal people’s preferences and locate their activities. Other geo-referenced con- tents of social media can also be analysed and used for the same purpose. CES ca- pacity and demand, as well as the flow, can also be mapped using deliberative mapping methods where a group of people discuss, compile knowledge and finally build a con- sensus on these in a certain area on a map. In a tiered approach for mapping, Tier 1 does not easily fit for the spatial representation of CES (see Chapter 5.6.1 for explanation of dif- ferent tiers). For some CES, for example ‘eco- systems as sites for activities’, land cover can be used as a proxy for the landscape’s suitabil- ity for different use types from the perspective of potential capacity. The demand side is eas- ier to map in a Tier 1 approach as people can be asked to score or value different land cov- er types with regard to their appreciation in terms of CES (for example, ‘forests bring me feelings of sanctity’ or ‘meadows are aesthet- ically important for me’). Still, the outcome remains quite vague as it is usually not purely the land cover that adds the cultural meaning but a combination of different attributes. Tier 2 suits better for CES mapping as more detailed and specific data can be used to give variation to the characteristics of an area. Types of data beneficial for mapping poten- tial CES capacity include data on cultural heritage sites, sites with events combining culture and nature, spiritual or religious sites, habitats of symbolic species (caution in visualising sites of threatened species on a map must be applied), or recreational fa- cilities, such as trails or campsites. The selec- tion of data depends naturally on the CES in question. In the demand side, statistics of recreational visitors or the number of fishing licences in an area, for example, can be used for recreational ES demand, number of vis- itors of a religious event linked to a specific natural site for spiritual ES, or number of photos taken of beautiful scenery as a mea- sure of aesthetic appreciation of a place. Tier 3 mapping, based on process-based models, could be understood as modelling the availability of, the accessibility to and the demand for CES that are needed in a certain place. The exceptions are intrinsic and be- quest values of nature as CES. They can be understood as services which people need not necessarily be able to experience or to see by themselves but which they want to preserve because of their value for current and future generations and for which they feel joy and thus receive a non-material benefit. People can identify these kinds of places on the map or they can name specific species or habitats that they value after which they can be placed on a map based on other data. In the following, we give examples of some available CES mapping methods that repre- Mapping Ecosystem Services200 sent Tier 1 (mapping CES demand using a matrix; Chapter 5.6.4), Tier 2 (photo series analysis and ESTIMAP-recreation model) and Tier 3 (viewshed analysis). Table 2. A non-exhaustive list of methods suitable for mapping CES. Level of needed expertise refers to the degree of needed skills in GIS and / or statistics. Method type Method name For mapping: Capacity = C, Flow = F, De- mand = D Which CES can be mapped with the method Level of needed expertise: Low = L, Medium = M, High = H Characteristics of the method in regard to CES mapping M od el s, m ap pi ng m et ho ds ESTIMAP C, D Potentially all CES M The GIS processes are relatively easy to implement, re- quiring only a medium level of GIS expertise. The model allows simulation of different scenarios and evaluation of different policy options; it is flexible and can be downscaled and modified in order to fit local needs and conditions. Expert opinion is needed for inputs variables selections and scoring. Scientific evidence for the used thresholds is scarce and they thus mainly rely on expert opinion, too. InVEST - recreation module C Recreation, nature tourism H Predicts the spread of person-days of recreation and tourism, based on the locations of natural habitats, accessibility and built features that factor into people’s decisions about areas for recreation. Regression mapping that uses photos as a dependent variable. http://data.nat- uralcapitalproject.org/nightly-build/invest-users-guide/ html/recreation.html MIMES   Recreation, nature tourism H Suits ideally the examination of trade-offs under various economic, policy and climate scenarios in space and over time. Allows for testing management scenarios that would be socially unacceptable. ARIES / k.LAB C, D Aesthetics, poten- tially all CES H Is based on probabilistic modelling using Bayesian frame- work. Requires expert-level modelling skills. http://www. integratedmodelling.org/ GreenFrame C, D All CES M The GIS processes are relatively easy to implement, requiring only a medium level of GIS expertise. Makes use of a multitude of GIS datasets combined with both scientific and local expert scorings. Data on harmful phe- nomena that diminish the CES potential can be included in the analysis. Both quantitative and qualitative data can be used. Gives an overall picture of the relative spatial variation of CES provision potential. Land cover / land use based map- ping C, D Recreation, aesthetics, edu- cation L Gives only a very rough proxy with high uncertainty level. Suits best for quick mapping of specific recreational or experiential activities. So ci al m ed ia b as ed m ap pi ng Photo-series analysis C, (F), D Physical and intellectual interactions with biota, ecosystems, and land- / sea- scapes (including recreation, nature tourism, land- scape aesthetics, cultural heritage and education) M-H It represents a cost-effective way of gathering space-and time-referenced data on observed people’s preferences. It does not directly allow for obtaining information related to the user characteristics (socio- and psycho-cultural). Inherent bias is related with the interpretation of pic- tures. The photo-sharing community may not be repre- sentative of all social groups (the represented population will be dependent on the level of access to information technology, education and age and the user’s ability / willingness to correctly geo-tag the photos). Other analyses of social media content C, (F), D Recreation, nature tourism, cultural heritage, potentially all CES M-H Specialised social media communities can produce data on, for example, sites suitable for specific activities but which are not commonly known and do not exist in databases. Communities may not be representative of all social groups (the represented population will be depen- dent on the level of access to information technology, education and age, and the user’s ability / willingness to correctly geo-tag locations. Chapter 5 201 Method type Method name For mapping: Capacity = C, Flow = F, De- mand = D Which CES can be mapped with the method Level of needed expertise: Low = L, Medium = M, High = H Characteristics of the method in regard to CES mapping Pa rt ic ip at or y m ap pi ng (O n- sit e an d off -s ite m ap pi ng ) On-site map surveys using paper maps F, D All CES L Easy to implement anywhere and anytime. Time-con- suming and laborious. Only restricted amount of infor- mation can be collected unless plenty of workforce is available. Collected information may be better in quality as any problems in mapping can be solved immediately. Good social skills are needed. On-site map surveys using electronic device F, D All CES L Easy to implement anywhere and anytime. Time-con- suming and laborious. Only restricted amount of infor- mation can be collected unless plenty of workforce is available. Collected information may be better in quality as any problems in mapping can be solved immediately. Good social skills are needed. Malfunction of electronic device can happen any time. Interviews for the elicitation of values F, D All CES L Laborious and time-consuming and thus a limited number of people can be reached. Gained knowledge is much more detailed and much deeper understanding of the local CES can be derived in addition to maps. On-line map surveys F, D All CES M Several companies providing opportunity to implement on-line map surveys exist (paid service). Service includes usually basic reporting tools. Planning a workable sur- vey can be demanding. All population groups can be difficult to achieve (access to computer, skills of using it), digitising is not always easy for laypeople for several reasons (ability to locate places on maps, etc.). Usually low response rate. With simple point mapping, lots of data can be derived. Background information of the respondent and additional information can be collected together with the map markings. Used for spatial plan- ning purposes to gather knowledge and feedback. Deliberative mapping in a group on paper maps or using device, e.g. computer, visual table or landscape theatre C, F, D All CES L-M Demands good facilitation skills as the data on CES is mapped in a face-to-face setting and can involve partic- ipants of varying map reading skills and with opposing views. Sensitive to malfunction of electronic device if those are used. Mobile phone appli- cations F, D All CES L Suitable for mapping CES related activities, values and perceptions of a target group at local scale. Works also for environmental awareness-raising simultaneously with mapping. La nd sc ap e an al ys is Viewshed analysis C. F, D Landscape aesthetics H Combines social media and physical landscape analysis. It represents a cost-effective way of gathering space-and time-referenced data on observed people’s preferences. The viewshed is an approximation of the real visible surface. Quality of assessments depends on the resolution of the digital elevation model. Analysis includes compu- tational complexity. GIScame C Landscape aesthetics M Landscape aesthetics; aesthetical aspects can be character- ised by analysing landscape structure or the distribution of land use types with the help of landscape metrics. http://www.giscame.com/giscame/english_home.html Mapping Ecosystem Services202 Mapping CES using a matrix- based approach A matrix-based approach can be used as a quick and relatively easy way to map supply of or demand for CES. In its most simple form, only land cover data or similar one dataset is sufficient for this purpose. If sup- ply is mapped, experts can be asked to score each land cover type based on its capacity to provide different CES. On the other hand, residents of the study area can be asked to do the same based on how important they personally perceive the different land cover types in terms of CES, i.e. for which land cover types they have demand. As a result, a number of scored matrices are produced in both cases. After some basic statistical operations, such as calculating variance and medians of the given scores, a result matrix is produced. This can be easily transferred to a GIS and combined with the land cover data to produce a map (see Chapter 5.6.4). An example of a result matrix and a map produced from it is presented in Table 3 and Figure 1. The example stems from a re- al-life planning process in the city of Järven- pää, Finland, where an open participatory workshop was arranged for the residents to map the demand for CES. Participants of the workshop were given clear instructions for the matrix scoring task both orally and on paper. In addition, written explanations of different CES classes were also given as guidance. The CICES classification was used as a basis but the CES classes were simplified and broken down to sub-classes in a way that was easily understandable for laymen. The previously created green in- frastructure (GI) typology for the city was used as land cover data (see the GI typology map in Chapter 7.3.1). Participants scored individually each GI type (= environment type) based on how important it was to them personally in terms of different CES. In the Järvenpää case, the matrix task was followed by a spatially-explicit map exercise in another room. This allowed for both a general overview of the demand for differ- ent CES in different environments, as well as spatially-explicit knowledge of locations that people value. When simple matrix-based maps are used, the restrictions of the method must be kept in mind. The demand map, such as in the given example, reflects the perceptions of people in a given location and they are seldom transfer- rable to other locations. They are also coarse generalisations and, in reality, there can be several factors that either improve or dimin- ish the demand for certain locations even if the type of environment is important in general. For example, a forest may be located next to an industry with problematic emis- sions or the quality of water in a certain lake is poor and even aesthetically unpleasing. Figure 1. Mapped demand for CES based on scored matrices by individual residents in the City of Järvenpää, Finland. Artistic inspiration from nature according to the GI typology Not important 0 2 Quite mportant Verymportant Chapter 5 203 Photo series analysis There is limited access to spatially explicit data in relation to cultural activities. Yet, there is a growing need for territorial planning to incorporate the perception of numbers of visitors who could be attracted by landscape aesthetic or cultural heritage amongst other key cultural values. As representative field data are expensive and time consuming to gather, gaining understanding on how CES can be spatially defined and visualised is still challenging. A novel way to overcome this is to use crowdsourcing information. Until recently, user generated contents are providing volunteered geographic informa- tion in different place-based applications. The very fast rate of image uploading on popular social media platforms offers poten- tial for a new mapping paradigm based on a crowd of observers. Recent studies have used geo-located photographs retrieved from on- line platforms to explore place perception. Public image storage analysis has already been applied in studies for assessing CES. These studies suggest an empirical approach based on the location of visitors, assuming that vis- itors are attracted by the location where they take photographs. This approach opens up opportunities to directly analyse the presence of beneficiaries on the provision site which provides a proxy for actual service provision. The main limitation of using public image storage analysis to retrieve geo-located infor- mation is given by the representativeness of the social media platforms or in relation to specific groups. However, the taken photographs can be considered as observed people’s preferences which are less vague than declared preferences. The spatial distribution of visitors’ preferences provides an indicator of CES, allowing a local analysis of service providing areas and address- ing the lack of quantitative indicators of CES. Application Programming Interfaces (APIs) are publicly available for Web 2.0 applica- tions, such as Flickr and Panoramio, allowing accessing the data, including the photograph’s metadata, tags and geographic position. Different spatial analysis methods can be applied to analyse the specific patterns and CES sub-class / GI type Fo re sts C ro pl an ds G ra ss la nd s Al lo tm en ts Al lo tm en ts w ith h ut s U rb an p ar ks H ou se g re en G re en b uff er zo ne s M ire s a nd w et la nd s La ke s R iv er s C re ek s Recreation in nature 2.0 0.8 1.4 0.8 0.6 1.9 1.9 1.3 1.3 2.0 1.8 1.1 Nature as a subject mat- ter and site for education 1.9 1.3 1.9 1.0 1.0 1.5 1.5 1.4 1.6 1.8 1.8 1.5 Natural aesthetics 1.9 1.3 1.9 1.1 1.3 1.8 1.8 1.5 1.6 2.0 1.9 1.8 Artistic inspiration from nature 1.9 1.4 1.6 0.6 0.7 1.7 1.4 1.1 1.4 1.7 1.7 1.6 Identity value of nature 1.7 1.0 1.4 0.7 0.4 1.7 1.6 0.7 0.8 1.7 1.4 1.3 Place for obtaining em- powerment from nature 2.0 0.9 1.3 0.7 0.4 1.6 1.6 0.7 1.0 1.9 1.6 1.3 Feeling of holiness in nature 1.7 0.7 1.1 0.1 0.3 1.0 1.0 0.6 0.9 1.3 1.3 0.9 Intrinsic and bequest values of nature 1.7 1.0 1.1 0.8 0.7 1.6 1.3 0.7 1.7 2.0 2.0 1.6 All CES together 1.8 1.0 1.4 0.8 0.7 1.5 1.6 1.0 1.3 1.7 1.6 1.3 Table 3. Demand for different types of green and blue environments as a source of CES based on scored matrices by individual residents in the City of Järvenpää, Finland. Mapping Ecosystem Services204 identify the landscape settings which shape the actual service provision. A systematic vi- sual analytic process, based on expert knowl- edge, also allows the identification of different CES categories and their relative importance (Figure 2). Photographs of animal and plant species can, for instance, be classified as “ex- periential use and enjoyment of wildlife”, while photographs of sport and recreational activities, such as skiing, climbing, hiking and camping, represent “physical use of land- scape”. Other categories such as “landscape aesthetics” and “cultural heritage” can also be identified by photo-content analysis. Moreover, the temporal attributes of the photo-series (date), available on most public photo-archives, can be used to analyse the seasonality of CES (Figure 3). Specific time and location may show over supply, there- fore conflict and trade-offs between differ- ent ES can also be mapped. The photo-series analysis can be applied at different spatial scales, ranging from municipality to national, according to the context, photo-density and positional ac- curacy of the photographs. The final ser- vice provision map can therefore inform stakeholders and policy makers at different institutional levels on priority areas (Figure 4). Finally, the analysis of community-con- tributed photographs can be used to design location-based interviews, questionnaires or focus groups in order to take into ac- count socio- and psycho-cultural aspects which are related to CES values. Figure 2. Photo location and count by CES cate- gory in Quatre Montagnes Region as case study demonstration (French Alpine Mountain Range). Figure 3. Example: Seasonality of CES catego- ries in the study site of Quatre Montagnes. Legend Enjoyment of wildlife Recreation: winter sport Recreation: summer sport Other recreation activites Landscape aesthetic Au tu mn Winter Sum mer Sp rin g Wildlife Natural landscape Anthropic landscape Other recreation Sport-summer Sport-winter Rcreation Aesthetic Chapter 5 205 Mapping landscape aesthetic service through viewshed analysis The aesthetic value of landscapes, such as scenic beauty, represents a specific category of CES which has received growing attention in the socio-ecological research. Although the visual aesthetic quality of landscapes has been researched for centuries, standardised and quantitative assessment approaches are so far scarce. Geo-tagged photographs uploaded on on- line photo-sharing platforms can be used to locate aesthetically attractive areas and derive the frequentation rate. Together with the bio- physical and built-up characteristics of the landscape, the photo-series allows the anal- ysis of complex visual landscapes which are associated with scenic beauty as it is perceived by beneficiaries at specific locations. In open areas, as scenic beauty is especially related to panoramic view, photographs capturing pan- oramas can therefore be used as spatially ex- plicit data of actual service provision. These data can be related with biophysical factors of the landscapes seen from the respective view- points. The visible area and the respective visual indicators can be calculated for each theoretic viewshed, derived from a Digital Elevation Model, corresponding to the pho- tograph location. The viewshed is thus con- sidered as a Service Providing Area from the perspective of the beneficiary (Chapter 5.2). The landscape aesthetic theory allows the linking of landscape visual indicators to the landscape’s visual characteristics. Those in- dicators represent the landscape structures related to the information functions of the landscape which contribute to enjoyment of scenery as a final service. A quantitative framework can thus be applied to identify the landscape variables contributing to the visual landscape attraction. The procedure for capturing and mapping the visual character of the landscape has been applied in the same study region as the pho- to series analysis (Figure 5). The analysis al- lowed the evaluation of the visual landscape Figure 4. Methodological framework for exploratory analysis of CES through Geographically Weighted Regression (GWR). Given the high diversity of habitats and ecosystems in the study site of Quatre Montagnes, we assume that CES delivery is context-specific and we expect a significant geographical variation in the relations between the photo-count and explanatory variables. The variables correspond to the physical (environmental settings) and infrastructure (opportunity settings) characteristics of the landscape whose spatial variation may affect the CES provision. The spatially weighted regression showed that specific variables correspond to prominent drivers of CES at the local scale. Dominant habitat, accessibility, diversity of habitat and proximity to view points were identified as the variables having a major impact on CES. Database preparation Geographically Weighted Poisson Regression Flickr API Explanatory variables Model calibration Results diagnostic Mapping Ecosystem Services206 preferences by considering the information from the users’ source and assuming the re- lationship between the mental landscape perceptions and the visual scale. Different visual indicators were considered which refer to six different components of the landscape: depth, relief, land cover, landform, geolo- gy and habitat. Each indicator was linked to nine visual concepts, describing different landscape characters and landscape aesthetic theories. The visual indicators were finally used to run a cluster analysis in order to iden- tify spatial patterns and geographical regions (Figure 6). This approach provides a framework for performing a systematic analysis of scenic beauty aspects and facilitates interpretation of the landscape information function. By expressing the actual service provision in a spatially explicit way, we can learn about the beneficiaries’ perception and the landscape’s visual character providing integrated infor- mation which can support landscape moni- toring and regional planning. Modelling CES supply using the opportunity spectrum approach: ESTIMAP recreation Public, nature-based, outdoor recreational ac- tivities include a wide variety of practices rang- ing from walking, jogging or running in the closest green urban area or at the river/lake/sea Depth Relief Landcover Landform Geology Habitat Distance zones Viewshed calculation Flickr photo classication 150 m zone 6 km zone 30 km zone 150 km zone Variable reduction Cluster analysis Figure 5. Methodological framework for visual landscape assessment. The visible area (viewshed) was calculated for each location of photo representing panoramic views. The four distance zones were set to respect the degrading visual properties with increasing distance from the viewpoint. Figure 6. Landscape clusters based on the visual indicators. Four different typologies of landscapes – corresponding to groups of viewsheds – emerged as distinct clusters. The spatial distribution of the landscape groups showed a clustered pattern, allowing a regionalisation of the landscape characters. Chapter 5 207 shore, bike riding in nature after work, pic- nicking, observing flora and fauna, organising daily trips to enjoy the surrounding beauty of the landscape, amongst a myriad of other possibilities. These activities have important roles in human well-being and health. While tourism is an occasional activity, local outdoor recreation affects the daily life of people. The ESTIMAP recreation model (see also chap- ter 4.4) assesses the capacity of ecosystems to provide nature-based outdoor recreational opportunities which can be enjoyed on a daily basis. The model consists of three parts: (1) Recreation Potential (RP); (2) Recreation Op- portunity Spectrum (ROS); (3) The number of potential trips. The Recreation Potential (RP) (Figure 7) rep- resents a composite dimensionless indicator that estimates the potential capacity of a group of identified landscapes and features to provide opportunities for local outdoor recreation. The provision varies according to four main components: (1) the suitabil- ity of land to support recreational activities; (2) the blue-green infrastructure in urban areas; (3) the presence and typology of nat- ural protected areas and natural features; (4) the presence and quality of water bodies and coastal areas (inland and sea). The Recreation Opportunity Spectrum (ROS). People can benefit from the opportunities provided by nature for recreational activi- ties if they are able to reach them. The ROS was chosen as a method to map different degrees of service available according to their proximity to the people. First, a prox- imity map is computed by combining Eu- clidean distance from urban and Euclidean distance from roads. A final map of recre- ation opportunities is then computed by a cross tabulation between the RP and the Proximity using a second set of parameters with thresholds for the degree of recreation opportunities provided by nature and the degree of proximity and remoteness. Pa- rameters can be based on national standards or law (normative) or on observed data. Number of potential trips The potential flow of the service to visitors can be estimated by computing the share of potential trips that can theoretically be undertaken in order to reach the different ROS zones. As mentioned above, the pres- ent model addresses daily recreation there- fore, according to literature, two reference distances were identified for close-to-home and daily maximum travelled distance: 8 and 80 km. A moving window with a kernel file is ap- plied to a raster grid of population density to compute an estimate of potential trips per each pixel in the grid per day. Figure 8 shows a map of potential close-to-home trips. The percentage of potential trips per ROS zone can be calculated by dividing the sum of potential trips per ROS zone by the total of all possible trips, see graphs in Figure 9. Figure 7. Potential nature-based opportunities for recreation in Europe. very low Nature Based Recreation Opportunities 2010 low medium high very high Non-EU Member States Mapping Ecosystem Services208 Figure 8. Potential close-to-home trips in Europe. The graph represents the shape of a distance decay function which can be used to model the close–to-home trips. Y axis represents the decay function, X axis the distance. Figure 9. The Recreation Opportunity Spectrum in Europe. More details are provid- ed for two cities (Naples and Helsinki); for both cities, an estimate of close-to-home potential trips was computed. Pie charts represent the percentage of potential trips to all ROS categories. Potential local trips for ROS categories (%) Chapter 5 209 Conclusions The intangible CES are extremely import- ant for people’s well-being in many ways. Mapping them might seem difficult but it is worth the effort. Knowledge of CES ca- pacity, demand and the flow from service providing areas to beneficiaries is crucial in spatial planning, nature tourism develop- ment and sustaining and enhancing, for ex- ample, people’s physical, mental and social health. CES can frequently be overlooked if they are not analysed and visualised in a spatially-explicit way. Mapping provides a means to bring them into discussion along with more easily understood provisioning and regulating services. Further reading Casalegno S, Inger R, DeSilvey C, Gaston KJ (2013) Spatial co-variance between aes- thetic value & other ecosystem services. PLOS ONE 8 (6): e68437. Di Minin E, Tenkanen H, Toivonen T (2015) Prospects and challenges for social media data in conservation science. Frontiers in Environmental Science 3: 63. doi: 10.3389/fenvs.2015.00063. Kopperoinen L, Viinikka A, Zulian G, Yli-Pel- konen V, Niemelä J (2016) Developing cultural ecosystem service mapping for spatial planning purposes – Sibbesborg, Finland, as a case study. Ecosystem Ser- vices (in review). Martínez Pastur G, Peri PL, Lencinas MV, García-Llorente M, Martín-López B, (2015) Spatial patterns of cultural ecosys- tem services provision in Southern Patago- nia. Landscape Ecology 31: 383-399. Ode Å, Tveit MS, Fry G (2008) Capturing Landscape Visual Character Using In- dicators: Touching Base with Landscape Aesthetic Theory. Landscape Research 33: 89-117. Paracchini ML, Zulian G, Kopperoinen L, Maes J, Schägner JP, Termansen M, Bido- glio G (2014) Mapping cultural ecosystem services: A framework to assess the poten- tial for outdoor recreation across the EU. Ecological Indicators 45: 371-385. http:// doi.org/10.1016/j.ecolind.2014.04.018. Richards DR, Friess DA (2015) A rapid in- dicator of cultural ecosystem service usage at a fine spatial scale: content analysis of social media photographs. Ecological Indi- cators 53: 187-195. Schirpke U, Tasser E, Tappeiner U (2013) Pre- dicting scenic beauty of mountain regions. Landsc Urban Plan 111: 1-12. Tenerelli P, Demšar U, Luque S (2016) Crowdsourcing indicators for cultural eco- system services: A geographically weighted approach for mountain landscapes. Eco- logical Indicators 64: 237-248. Tenerelli P, Püffel C, Luque S (2016) Spatial assessment of aesthetic services in Alpine region: combining visual landscape with Volunteered Geographic Information. Landscape Ecology (in review). Zulian G, Paracchini ML, Maes J, Liquete Garcia MDC (2013) ESTIMAP: Ecosys- tem services mapping at European scale. (E. U. R.-S. and T. R. Reports, Ed.). Euro- pean Commission. Retrieved from http:// publications.jrc.ec.europa.eu/repository/ bitstream/111111111/30410/1/lb-na- 26474-en-n.pdf. Mapping Ecosystem Services210 5.6. Integrative approaches Benjamin Burkhard Ecosystem services (ES) are an integrative multi-, inter- and trans-disciplinary field of study per se (see Chapter 2.1). Therefore it is necessary to integrate multiple approach- es, methods and data of varying quality and quantity (see Chapters 4 and 5) as well as experts from multiple backgrounds (see Chapter 4.6) in ES mapping and assessment projects. Depending on the purpose of the map product, the most suitable methods and available data need to be chosen and in- tegrated accordingly (see Chapter 5.4). Integration takes place on different spheres such as different ES (regulating, provision- ing, cultural) spatial and temporal scales, domains, (biophysical, social, economic), methods and data (e.g. direct measure- ments, modelling, interviews) and levels of application (i.e. global, national, regional or local decision- making). The enormous complexity of ES maps and the processes of producing them require a broad range of approaches - from rather simple to complex - that can be integrated in order to harness the advantages of each and to deliver the most applicable and reliable results. How- ever, a more complex approach does not always deliver more robust or more appli- cable outcomes. For some applications, less can actually be more (or at least sufficient) as was previously stated in the 14th centu- ry: “It is futile to do with more things than which can be done with fewer” (cf. Occam’s Razor and Chapter 5.4). The following four sub-chapters introduce different integrative ES mapping and assess- ment approaches. All approaches can be ap- plied in combination with the concepts and methods described in the preceding chapters. The tiered ES mapping approach (see Chap- ter 5.6.1) provides a suitable conceptual framework to combine different levels of complexity from tier 1 to tier 3. Participa- tory GIS (PGIS; Chapter 5.6.2) is another highly integrative approach combining var- ious kinds of knowledge perspectives with spatial information in a straightforward manner. Harnessing citizens’ knowledge and willingness to voluntarily contribute to data gathering is the idea of citizen science as described in Chapter 5.6.3. The ES ‘ma- trix’ (see Chapter 5.6.4) is based on spread- sheets that link geospatial units to ES supply or demand providing relatively quick out- puts in a spatially explicit manner. Further reading Burkhard B, Kroll F, Nedkov S, Müller F (2012) Mapping supply, demand and budgets of ecosystem services. Ecological Indicators 21: 17-29. Maes J, Crossman ND, Burkhard B (2016) Mapping ecosystem services. In: Potschin M, Haines-Young R, Fish R, Turner RK (Eds.) Routledge Handbook of Ecosystem Services. Routledge, London, 188-204. Chapter 5 211 5.6.1. A tiered approach for ecosystem services mapping Adrienne Grêt-Regamey, Bettina Weibel, Sven-Erik Rabe & Benjamin Burkhard Introduction: The need for a tiered approach in ES mapping Understanding strengths and weaknesses of the different ecosystem services (ES) map- ping methods is crucial for understanding what information can be derived from a map and how applicable it eventually will be. Particularly, information about reliabil- ity, accuracy and precision of ES maps is important for users to determine their suit- ability in a specific context (see Chapters 3.7 and 6.3). ES mapping approaches can broadly be classified into five categories: 1. A simple and widely used approach directly links ES to geographic infor- mation, mostly land cover data and is often referred to as the “lookup table” approach. The land cover data are used as proxies for the supply of (or demand for) different ES. The ES in the lookup table can be derived from statistics such as crop yield for agricultural production. 2. Approaches, mainly relying on expert knowledge (see Chapter 4.6), include expert estimates of ES values in lookup tables but also other methods such as Delphi surveys. 3. The “causal relationship” approach es- timates ES based on well-known rela- tionships between ES and spatial in- formation retrieved from literature or statistics. For example, timber produc- tion can be estimated using harvesting statistics for different areas, elevations and forest types provided in a national forest inventory. 4. Approaches that estimate ES extrapo- lated from primary data such as field surveys linked to spatial information. 5. Quantitative regression and socio-eco- logical system models that combine field data of ES as well as information from literature linked to spatial data. To provide guidance in the choice of the appropriate ES mapping method and to enhance comparability between different ES assessments, tiered approaches can be used. The methods can be categorised into tiers with increasing complexity between the different levels such as, for example, in the TEEB1 tiered approach. This idea has also been implemented in the InVEST model (see Chapter 4.4) where a simple (tier 1) and more complex (tier 2) approach is suggested. Usually the tier 1 approach relies on widely available data and the tier 2 approach includes more specific information for the case study area. Another well-established example is the IPCC tiered approach which structures and facilitates the reporting on climate change at 1 TEEB stands for The Economics of Ecosystems and Biodiversity; http://www.teebweb.org/ Mapping Ecosystem Services212 global and national scales. Inventory reports on national greenhouse gas refer to different tiers when describing the methods used and changes in methods from one report to an- other are related to the tiers defined. A tiered approach for ecosystem services mapping Similar to the approaches mentioned above, a tiered approach for ES mapping is proposed in this chapter: it is most use- ful to define the tiers according to the goal of the mapping exercise (see Chapter 5.4) to make sure the information relevant for the related decision-making process is pro- vided. This supports the efficiency of the mapping process avoiding far too complex approaches where rough estimates would be sufficient. In a first step, the different components of the analysed human-environment system should be described which include the eco- systems and ES as well as the beneficiaries and institutions involved and their inter- actions. For example, for microclimate regulation in urban areas, the considered ecosystems are usually green urban areas, the service they provide is microclimate reg- ulation, beneficiaries are residents and in- stitutions are city planning agencies. These system components can be described at different levels of detail, for example, the ecosystem can be described in terms of its condition and structure (see Chapter 3.5), the service provided can be quantified in different units (see Chapter 2.4), the ES demand can be structured according to different beneficiary groups (Chapter 5.1) and different instruments of institutions including NGOs or businesses (see Chap- ter 7), for example, can be identified. This description of components should make the boundary of the considered system and the spatial and temporal scale explicit. ES ben- eficiaries and institutions represent relevant stakeholders who could be considered in the decision-making process. Once these components have been de- scribed, the appropriate tier and associated ES mapping method can be selected. To guide this selection, we present a decision tree in Figure 1. The first question addresses the process-understanding of the human-en- vironment system. If interactions between the system components are relevant and a deeper understanding of processes is need- ed (e.g. to understand how management of ecosystem components can influence the provision of ES), a tier 3 approach would be required. Otherwise, if the purpose of the map is mainly to provide a rough overview of ES values in a certain area, their abun- dance, presence and absence, a tier 1 ap- proach can be selected. If information about different ES is required at a certain level of detail but not linked to an explicit manage- ment question tackling the human-environ- ment system components and processes, a tier 2 approach may be suitable. However, if the ES map is to be used to explicitly eval- uate management measures, again a tier 3 approach should be considered. After the most suitable tier has been identified, the availability of resources for the ES map- ping should be evaluated. In case resources are severely limited, a method involving a lower tier can be applied. Yet, efforts should be made to identify the most suitable tier to provide information that is useful for deci- sion-makers. We associated the five different categories of ES mapping methods (see above) with the different tier levels: while most methods are applicable at all tier levels, they usually have a focus at a certain level as indicated in Figure 1 with the shading. ES quantifica- tion and mapping methods are described in more detail in Chapters 4 and 5. Chapter 5 213 How to choose the appropriate tier Are a deeper understanding and analysis of underlying socio-economic and/or geo-bio-physical processes needed? Is the mapping purpose exclusively a rough overview of ES in space? Do the planned actions require information on the system behaviour? Process-understanding necessary? Explicit measures needed? Rough overview? Are data in sufficient quality, quantity, scale and resolution available to conduct an ES assessment in this tier? Are there enough technical, human and financial resources available? Are data and resources available? Tier I Tier II Tier III Look-up tables (e.g. linking ES values to land-cover classes) Expert knowledge (e.g. Delphi survey: experts rank land-cover types) Causal relationship (e.g. BBN: incorporate combined knowledge about ES) Extrapolation of primary data (e.g. field survey data linked to spatial information) Regression and socio-ecological system models (e.g. field and statistical information linked to spatial data) Figure 1. Decision tree guiding the selection of tiers for ES mapping. Mapping Ecosystem Services214 Box 1 . Illustrating the tiered approach: Microclimate regulation In this example, we illustrate the tiered approach for mapping microclimate regulation within urban areas with ES mainly provided by green space and important in the context of heat island effects. The components of the human-environment system include green urban spaces as ecosystems, microclimate regulation as provided ES, residents as the main user group and city planning agencies as main institu- tions. If the purpose is to provide a rough overview, i.e. to compare cities or city districts, no detailed process-understanding is required and a tier 1 approach would be most suitable. Using a lookup table approach, the microclimate regulation can be estimated based on the amount of green space as illustrat- ed in Figure 2. Alternatively, experts could also rank the different land use/land cover (LU/LC) classes according to their suitability for providing microclimate regulation. If the map is to be used to analyse microclimate regulation in more detail without providing informa- tion for an explicit management measure targeting system components or processes, a tier 2 approach can be applied. Here, we present a causal relationship approach, where the green volume is estimated by combining high resolution remote sensing data with LU/LC information: Green areas are estimated from the remote sensing information based on the normalised-difference-vegetation-index (NDVI), which allows, for example, identifying single trees. Additionally, the remote sensing data provides infor- mation about the height of these identified green areas to estimate the volume. As reducing the urban heat islands by increasing microclimate regulation requires an understanding of how certain measures such as changes in the amount and/or structure of green area quantitatively affect the cooling potential, a process-understanding is needed guiding us to a tier 3 approach. Figure 2. Illustrating the tiered approach for microclimate regulation. Chapter 5 215 Further Reading Grêt-Regamey A, Weibel B, Kienast F, Rabe S-E, Zulian G (2015) A tiered approach for ecosystem services mapping. Ecosys- tem Services 13: 16-27. Martinez-Harms MJ, Balvanera P (2012) Methods for mapping ecosystem service supply: a review. International Journal of Biodiversity Science, Ecosystem Services & Management 8: 17-25. Conclusions The suggested concept and decision tree provide guidance in the selection of the ap- propriate tier and associated methods for mapping ES. The presented tiered approach distinguishes the different tiers according to their purpose i.e. the intended use of the ES map. Thus it ensures that ES maps provide information useful to decision-makers in the specific context avoiding either the applica- tion of over-complex and resource intensive methods resulting in high costs at a level of complexity of methods which might not be required or over-simplified assessments which could mislead decision-makers. If we want the concept of ES to be used by decision-makers in the next decades, ES map- ping needs to be of high quality and provide precise and reliable information. To provide a solid ground for decision-making, the se- lection of ES maps should not only be based on methods and data available, but also on the ES that are assessed, because the lack of consideration of relevant ES can significantly change ES trade-off assessments and the se- lection of alternative policy options. In a tier 3 approach, the cooling effect is estimated based on a model combining ecological information, i.e. the cooling potential of various vegetation types with the given green infrastructure and their green volumes: the volume of green infrastructure can be derived from a detailed land use typology at the cadastral level based on field surveys with classes such as private yards, sport facilities and infrastruc- tural green. Each class of the typology is related to the amount of trees, grasses, shrubs and settlement or infrastructure present. For the categories tree, grass and shrubs, the volume is estimated based on well-known geometric relations and combined with remote sensing information. The potential cooling effect for high, middle and low green infrastructure can then be modelled considering climate infor- mation such as precipitation, temperature and solar radiation. Finally, the effect of infrastructure such as roads or buildings on the cooling potential is considered for estimating the resulting cooling effect. Mapping Ecosystem Services216 5.6.2. Participatory GIS approaches for mapping ecosystem services Nora Fagerholm & Ignacio Palomo Introduction Participatory mapping is the process where individuals contribute to the creation of a map. It can be applied to ecosystem ser- vices (ES) assessment by engaging various stakeholders to identify and map a range of ES that originate from location-based knowledge. These approaches are com- monly known as Public Participation GIS (PPGIS) or Participatory GIS (PGIS) (in this chapter, the acronym PGIS is used) and refer to the use of spatially explicit methods and technologies for capturing perceptions, knowledge and values of individuals or groups via surveys and/or workshops, with the aim of using this spatial information in land use planning and management process- es. PGIS approaches represent a spatially ex- plicit socio-cultural assessment of ES. The location-specific mapping communicates the assigned environmental values, i.e. the judgement regarding the worth of objects such as places, ecosystems and species. Since the early 2000s, when PGIS ap- proaches addressing community values for ES appeared, this field has increased expo- nentially for pragmatic and practical reasons such as: the idea of crowd wisdom to create knowledge from the masses, the lack of spa- tial data in specific contexts or for certain services, the need to include socio-cultural perceptions for ES assessment, technolog- ical development allowing sophisticated mapping solutions (e.g. web-based partici- patory mapping) and the democratic aim of bringing stakeholders to participate in the assessment of the value of nature’s services and related decision-making. The increased use of PGIS has resulted in its application in multiple contexts and with different aims including informing land use planning, rural landscape planning, protected area management, conservation planning, ur- ban planning and coastal zone management amongst others. Collecting data through participatory mapping approaches Data collection with PGIS approaches rep- resents pluralism. Common data collection methods include self-administered surveys, either web- or paper-based, face-to face surveys and workshops (see Table 1 for a comparison of these methods). Mapping ac- tivity typically engages the lay-public such as residents or visitors to an area but also various stakeholders including land hold- ers, environmental professionals, planning practitioners and other experts. Random and meaningful sampling, on site recruit- ment and volunteered open participation or different methods for stakeholder prioritisa- tion can be applied to select participants for the mapping process. Chapter 5 217 PGIS mapping of ES involves either dig- ital mapping interfaces, often web-based, with zoomable background maps or print- ed map layouts commonly presenting one given scale. Information, given on the background maps typically includes aerial/ satellite image overlaid with basic map el- ements, or topographic maps showing, for example, relief and basic natural and man- made features. The most often applied method for marking has been point place- ment (e.g. movable plastic discs or stickers) followed by drawing polygons presenting areas or using predefined land units as a basis for assigning values. Most commonly applied typologies for map- ping include ES classifications (MA, TEEB, CICES, see Chapter 2.4), adaptations of these to specific contexts or landscape values and landscape services typologies developed in case studies and based on research on so- cial values. Direct ES identification and val- uation through an inductive approach, not deriving from a given typology, have been rarely applied. Characteristics Web-based surveys Face-to-face surveys Workshops Participants type Often lay public Lay public and experts Often experts (e.g. local inhabitants with ecological knowledge, environmental specialists, planning practitioners) Time, cost and facilitation requirements Time efficient but resources needed for inviting participants, no facilitation needed and participation not restricted by specific time and place Time consuming as each person needs to be met individually, resources also needed for inviting participants and training interviewers Time efficient as it allows all data gathering during the workshop, but demanding for preparation and training facilitators Sampling method Random sampling, volunteered open participation Random sampling, on site recruitment, meaningful sampling, stakeholder prioritisation Meaningful sampling, stakeholder prioritisation Sample size and representativeness Easier to reach a larger and more representative sample, although survey respondent rate often remains low (under 15%) Depends on available resources, possibility to control representativeness Remains often low, statistical representativeness not targeted Type of participation and its effect on data quality Instrumental and self- administered, difficult to analyse the level of understanding of the participant and data quality Self-administered but allows facilitation and a detailed exploration of the issue analysed, contributes positively to data quality Allows communication among participants and detailed exploration of the issue analysed (deliberative mapping), contributes positively to data quality Table 1. Three common data collection approaches in PGIS (web-based surveys, face-to-face surveys and workshops) for ES assessment and their characteristics. Mapping Ecosystem Services218 Analytical process A wide variety of analytical approaches can be applied to PGIS data on ES (Figure 1). Typically the analytical process starts by describing the characteristics of informants who participated in ES mapping. Spatial analysis often begins with description of the spatial patterns and characteristics of ES through testing the level of clustering or dispersion, intensity/density estimation, diversity of ES, identification of hotspots and by calculating distances, for example, to the respondent’s home. Between pairs of ES, the spatial overlap has been studied through correlation analysis to look at the co-existence of ES. In spatial analysis of mapped ES, areas have a predefined precise boundary but points are treated as repre- senting the centroid of the spatial occur- rence of a specific ES extending outwards to an unknown distance. Spatial concurrence is commonly studied through overlay analysis to explain the re- lationship to physical land features (such as land cover, land use, management units, land change) or ecological data. In addition, spa- tial indices such as landscape metrics derived indices are common to quantify the distribu- tion across different land use or management units within the study area. Clustering tech- niques have been found useful for exploring the potential relation between the mapped ES and, for example, land use and socio-de- mographic characteristics of informants. In- terest has also been paid to extrapolate and model the distribution of the participatory mapped ES to locations where data were not collected through value transfer methods. Analysis of ES bundles has not been frequent but is gaining more attention as is the analysis of ES flows and trade-offs. Opportunities and challenges for future research and practice Several case studies show that socio-cultural valuation of ES through PGIS has success- fully facilitated the identification of spatial 1. Description of informant characteristics Descriptive statistics 2. Description of ES spatial patterns Spatial arrangement (clustering/dispersion) Intensity/density estimation Identification of hot spots Distance analysis (e.g. from mapped locations to informant homes 3. Spatial overlap between mapped ES Correlation analysis 4. Relationship to land use and other physical or administrative land properties Overlay analysis 5. Quantification of ES distribution across land use or management units Spatial indices (e.g. social landscape metrics) 6. Relation between mapped ES, land use and socio-demographic characteristics Clustering techniques (e.g. redundancy analysis (RDA)) 7. Extrapolation and modelling Value transfer methods PGIS data PGIS data integrated with other spatial datasets Figure 1. Example of an analytical process for PGIS data from basic descriptive steps to more advanced spatial and statistical analysis where PGIS data is integrated with other spatial data sets. Chapter 5 219 areas of ES supply and demand and how these vary between stakeholder groups. PGIS data may also be integrated with spatial data on ES produced through oth- er methods, advancing trans-discipline and more comprehensive ES assessment. Some opportunities and challenges for PGIS ap- proaches may be identified and formulated around the following three points. Firstly, PGIS allows addressing certain as- pects of ES that cannot be evaluated without participation. PGIS approaches have poten- tial to enhance the appreciation of abstract, symbolic and intrinsic values that landscapes and ecosystems provide to humans. Insuffi- cient acknowledgement of these values has been addressed in literature as one of the re- curring critiques of the ES framework. Cer- tain ES categories such as cultural services (e.g. landscape aesthetics, cultural identity, place attachment, etc.), might naturally better fit in PGIS than non-PGIS methods, since PGIS can directly capture the percep- tions and values individuals have towards ES. Cultural services are also often inferred from proxy data underestimating the multi- ple socio-cultural benefits widely recognised as critical for human well-being. Secondly, new opportunities arise through information technologies (ITs). The mush- rooming new ITs and increasingly available open source spatial data sets can facilitate the application of PGIS methods through citizen science (e.g. via smart-phones and use of open source high resolution imagery or topographic maps) opening new possi- bilities for open public access of ES map- ping for decision-making (Chapter 5.6.3). In data-scarce regions, PGIS has also been proposed as an alternative to complex and expensive data-building processes to map ES. In this case, depending on the use of the data, it is important to evaluate the accura- cy of information outputs (e.g. to compare participatory mapped ES data with physical landscape features mapped by participants, or with modelling approaches). Thirdly, it should be emphasised that partic- ipation should be in the core of PGIS. Ca- pacity building and social learning should always be seen as important aims of partic- ipatory activities. Another important aspect of mapping ES through PGIS is to actually integrate this data into land use planning and decision-making regarding ecosystem protection, conservation and management and to communicate to the participants how it was applied. This would enhance public participation in practice and it would not only be seen as consultation which, unfortunately, is prevalent in the current PGIS practice. Hence, integration of the gathered spatially explicit knowledge into decision-making remains a challenge for the future and requires also significant commit- ment by researchers and facilitators as well as resources to appreciate PGIS as a process in ES assessment. Further reading Bryan BA, Raymond CM, Crossman ND, Macdonald DH (2010). Targeting the management of ecosystem services based on social values: Where, what and how? Landscape Urban Planning 97: 111-122. Brown G, Fagerholm N (2015) Empirical PPGIS/PGIS mapping of ecosystem ser- vices: A review and evaluation. Ecosystem Services 13: 119-133. Brown G, Pullar DV (2012) An evaluation of the use of points versus polygons in public participation geographic information sys- tems using quasi-experimental design and Monte Carlo simulation. International Journal of Geographical Information Sci- ences 26: 231-246. Mapping Ecosystem Services220 Burkhard B, Kroll F, Nedkov S, Müller F (2012) Mapping ecosystem service supply, demand and budgets. Ecological Indica- tors 21: 17-29. Fagerholm N, Käyhkö N, Ndumbaro F, Khamis M (2012) Community stakehold- ers’ knowledge in landscape assessments– Mapping indicators for landscape services. Ecological Indicators 18: 421-433. García-Nieto AP, Quintas-Soriano C, García- Llorente M, Palomo I, Montes C, Martín- López B (2014) Collaborative mapping of ecosystem services: The role of stakeholders’ profiles. Ecosystem Services 13: 141-152. Klain SC, Chan KM (2012) Navigating coast- al values: participatory mapping of ecosys- tem services for spatial planning. Ecologi- cal Economics 82: 104-113. Palomo I, Martín-López B, Potschin M, Haines-Young R, Montes C (2013) Na- tional Parks, buffer zones and surrounding lands: mapping ES flows. Ecosystem Ser- vices 4: 104-116. Raymond CM, Bryan BA, MacDonald DH, Cast A, Strathearn S, Grandgirard A, Kali- vas T (2009) Mapping community values for natural capital and ecosystem services. Ecological Economics 68(5): 1301-1315. Raymond CM, Kenter JO, Plieninger T, Turn- er NJ, Alexander KA (2014) Comparing instrumental and deliberative paradigms underpinning the assessment of social val- ues for cultural ESs. Ecological Economics 107: 145-156. Sherrouse BC, Semmens DJ, Clement JM (2014) An application of Social Values for ES (SolVES) to three national forests in Colorado and Wyoming. Ecological Indi- cators 36: 68-79. Chapter 5 223 5.6.3. Citizen science Joerg A. Priess & Leena Kopperoinen What is citizen science? The term “citizen science” (CS) already sug- gests that citizens somehow are involved in science. Synonyms occasionally used are crowd science or crowd wisdom (with col- lective intelligence considered superior for solving social or environmental problems). A citizen, in this context, refers to amateurs or non-scientists voluntarily contributing to or participating in data gathering (such as observations of natural phenomena or species) or in scientific projects. Scientific projects involving citizens are often called participatory research. In many instances, volunteers are collecting, for example, ad- ditional biological or astronomical data, with the most popular and well-known citizen science activity probably being bird watching. In other cases, citizens may be more deeply involved in defining research questions and in designing and running the projects in which professional scientists may or may not be involved at all. Such proj- ects may take place in a citizen association focusing on, for example, regional history, language, landscapes. Nowadays, research is dominated by pro- fessionals but only two centuries back, am- ateur researchers such as Benjamin Frank- lin or Charles Darwin were more the rule rather than the exception. During the last decades, CS and participatory research have increased tremendously in various fields of science such as astronomy, biology, environ- mental science, history or the observation of weather phenomena such as cyclones. In recent discussions about the quality of data generated in CS projects, expertise, motiva- tion and honesty of CS contributors have been questioned by scientists. While data quality criteria usually are avail- able, potential conflicts of interest may be harder to detect and address (see suggested reading at the end of this chapter). What has been shown is that CS may improve deci- sion-making, generate new knowledge and innovations, empower citizens and generate political discourse and concern. In the context of ecosystem services (ES) mapping, CS implies that public partic- ipation can well go beyond participatory monitoring of ES in a research project. For instance, groups of urban gardeners could map their ES use with the objective of iden- tifying their main interests or the diversity of their contributions to food production or their recreational activities. In the rest of this chapter, we focus on CS contributing to the mapping and assessment of ES and present some of the methods available for CS / participatory approaches. CS approaches in ES mapping and assessment Different types of ES can be distinguished (see Chapter 5.5) and different methods and approaches are available to map and assess them (see Chapter 3.2). Many cultural ES are especially difficult to address via scientific mapping and modelling tools not involving the broader public, as cultural ES often de- Mapping Ecosystem Services224 pend on the preferences of users which may vary considerably locally or regionally. Ex- cepting touristic activities, for which much visitor or overnight stay data are available, large information deficits still exist about other cultural ES such as gardening, outdoor activities, appreciation of cultural heritage or intellectual experiences which are much more difficult to assess without asking or involving citizens. Thus, in the context of cultural ES, CS projects have a huge potential to increase our knowledge base and contribute to im- proving decisions and management. Further- more, citizens increasingly contribute to pub- lic debates and decision-making, especially concerning the governance of regulating or provisioning services by, for instance, discuss- ing and defining environmental thresholds such as the use of water resources. Which methods and tools are available for participatory mapping? In this section, we briefly present four of the participatory mapping methods available for CS in the context of ES mapping and assess- ment, covering work with conventional paper maps or tables, or digital tools such as geo- graphic information systems (GIS) or smart- phone apps. All methods can be used with different levels of citizen and scientist involve- ment. The higher the level of involvement of citizens, the higher the level of knowledge needed for citizens, for instance, about differ- ent ES, to handle spatial data in a geographic information system, to prepare paper or digi- tal maps of the study area, or to evaluate infor- mation generated during the CS project. In every ES mapping approach, citizens and scientists need to define the mapping per- spective, i.e. whether they want to address the (potential) supply of, the demand for, or the actual use (= flow) of ES (see Chapter 5.1). Additionally, in most approaches, spa- tial and temporal units / coverage also need to be clarified (see Chapter 5.7.5). ES use has been mapped by citizens and scientists using the MapNat smartphone ap- plication. Colours of flags indicate different types of ES use. The selected ES use also in- dicates the frequency of use and the impor- tance, both reflecting the view of the person who mapped the ES use (Figure 1). Four CS-compatible mapping examples Mapping ES with paper maps: Identifying and locating ES on topographic or thematic Figure 1. ES use in Europe. Chapter 5 225 maps can be carried out as an indoor or an outdoor approach, the latter enabling par- ticipants (at least partly) to view and observe the area of interest. Mapping units may be predefined, for example, using units of land cover, or may be identified during the proj- ect as, for example, spatial units are assumed to differ between ES. For a quick qualitative ES mapping/assessment (tier 1; see Chap- ter 5.6.1), one workshop or one field visit may be sufficient, while the generation of more detailed information (tiers 2-3) can be expected to require more time and/or ad- ditional sources of information. ES identi- fication and mapping can be carried out in- dividually, either resulting in calculated ES means or ranges or both. Alternatively, ES can be mapped based on a group consensus (called deliberative mapping). The approach can be used to map (potential) supply of and demand for ES, trade-offs, mismatches etc. Data collected with paper maps can be digitised afterwards in GIS (Figures 2-3: Ex- ample from Sipoo, Finland). Mapping ES with GIS: This approach is comparable to the first, the difference be- ing the replacement of paper maps by PCs, laptops or tablet computers (see also PPGIS Chapter 5.6.2). The great advantage of dig- ital mapping is that different types of spa- tial information can be linked or combined, for example, to derive appropriate mapping units, as well as evaluating ES mapping re- sults. However, at least one citizen or scien- tist with GIS software experience is needed. The approach can be used to map (poten- tial) supply of and demand for ES. The technical threshold might, however, invoke a selection bias dependent on the knowledge of the involved participant. The Matrix method to map ES: This meth- od is presented in Chapter 5.6.4. In CS projects, both paper and digital versions may be used. As explained in the previous examples, the matrix method can also be applied individually or as a group exercise and the approach is also suitable for map- ping (potential) supply of and demand for ES and, additionally, actual ES use. Use of a smartphone app such as MapNat for ES mapping: Similar to mapping ES with GIS, MapNat is a participatory GIS Figure 2. Using paper maps in a local master plan exhibition to collect cultural ES related values, in Sipoo, Finland. Although a digital map of the planning area (see computer in front) was available for citizens, paper maps were preferred by them. Figure 3. A map presenting the opportunity spec- trum of the CES group ‘Aesthetics and cultural heritage’ in the background and residents’ point and polygon markings of the same CES group. Examples of open-ended explanations of the markings have been added on the map. The bor- der of the local master plan area is shown as well. CULTURAL ECOSYSTEM SERVICE SUPPLY AND ACCESSIBILITY CULTURAL ECOSYSTEM SERVICE SUPPLY AND ACCESSIBILITY IN SIPOO AREA Analysis includes 25 meter grids that presents potential aesthetic and cultural heritage areas. For every grid also their accessibility has been calculated. Digiroad © Finnish Transport Agency/Digiroad2013 Corine Land Cover 2006 © SYKE, EEA © EuroGeographics for the administrative boundaries Low provision - easily accessible High provision - easily accessible Medium provision - easily accessible Low provision - accessible High provision - accessible Medium provision - accessible Low provision - not easily accessible Highprovision - not easily accessible Eriknäs master plan area Medium provision - not easily accessible 0 0.25 0.5 1 Kilometers Mapping Ecosystem Services226 approach (see Chapter 5.6.2). Mapping ES uses the GPS (geo-positioning) unit of the smartphone or a tablet to locate ES at the current position of the user. Alternatively, app users can just use their fingers to map ES directly on the device’s screen. In this tool, the mapping perspective is focusing on recording the actual use of ES, either directly during ES use, or afterwards identifying the location in the app’s map view. This meth- od can also be applied individually or as a group exercise. Contrasting with the previ- ous examples, this tool also provides access to the ES records and valuations of all other app users worldwide (Figure 1), because all records are sent to and redistributed by an internet server. Further reading Bela et al. (2016) Learning and the transfor- mative potential of citizen science. Conser- vation Biology 00:0, 1-10. doi: 10.1111/ cobi.12762 Dickinson and Bonney (Eds) (2013) Citizen Science: Public Participation in Environ- mental Research. Cornell University Press, Ithaca & London, 304 pp. Editorial (2015) Rise of the citizen scientist. NATURE 524, 265. CS organisations Citizen Science Alliance: http://www.citizensciencealliance.org/ European Citizen Science Association: http://www.citizen-science.net/ CS definitions http://www.openscientist.org/2011/09/fi- nalizing-definition-of-citizen.html https://en.wikipedia.org/wiki/Citizen_sci- ence#Definition CS platforms ZOONIVERSE (https://www.zooniverse.org) Main German platform (in EN and DE): http://www.buergerschaffenwissen.de/en CS example Cyclones http://www.cyclonecenter.org/#/about Tools qGIS: http://www.qgis.org/de/site/index.html MapNat, ES mapping App: http://www.ufz.de/index.php?en=40618 Harava: https://www.eharava.fi/en/ Maptionnaire: https://maptionnaire.com/en/ Chapter 5 225 5.6.4. Ecosystem services matrix Benjamin Burkhard Introduction Ecosystem services (ES) are spatio-temporal explicit phenomena. Thus, ES supply, flow and demand (see Chapter 5.1) can be linked to units in space and time. One mapping method is the ES ‘matrix’ approach, which links ES to appropriate geo-biophysical spatial units. Thereafter, their supply, flow and/or demand are ranked using a relative scale ranging from 0 to 5 (not relevant to very high, see Figure 1). Based on this nor- malisation of ES rankings, various ES are made comparable and different points in time (including scenarios) can be assessed. Therefore, the approach has the potential to integrate all kinds of ES-related data based on diverse scientific disciplines or ES quan- tification methods (see Chapter 4) and of varying quality and quantity in illustrative matrix tables and maps. It can be applied in data-poor as well as in data-rich study areas, fulfilling mapping purposes from first ES screening studies and awareness-raising to very comprehensive integrated trans-disci- plinary ES assessments. In this Chapter, the ES matrix approach is described and related uncertainties are discussed. Approach The ES matrix provides a very flexible ES mapping methodology that can be applied on all spatial and temporal scales (see Chap- ter 5.7.5), for all ES (see Chapter 2.4), vari- ous multidisciplinary ES quantification ap- proaches (see Chapter 4) and for different mapping purposes (see Chapter 5.4). As shown in Figure 1, the basic steps of ap- plication include: 1. Selection of ES study area; 2. Selection of relevant geo-biophysical spatial units (forming the assessment matrix lines/y-axis); 3. Collection of suitable spatial data (e.g. land cover/land use (LULC) data, hab- itat map, soil map, hydrological map); 4. Selection of relevant ES (assessment matrix columns/x-axis); 5. Definition of suitable indicators for ES quantification; 6. Quantification of ES indicators (using various methods); 7. Normalisation of ES indicator values to the relative 0-5 scale; 8. Interlinking geospatial units and scaled ES values in the ES matrix; 9. Linkage of ES 0-5 rankings to geospa- tial units to create ES maps; and 10. Interpretation, communication and ap- plication of resulting ES maps. Figure 1 gives an overview of the key com- ponents that are typically involved in the process. Steps 1-6 are strongly related to the purpose of the ES mapping exercise (see Chapter 5.4) and available mapping capacities (data, methods, time and labour). Relevant stake- holders should be involved in the process as much as possible and when necessary. Steps 7-9 are specific for the ES matrix but also other ES mapping approaches and each step is related to characteristic uncertainties (see below). Step 10 refers to the map-maker to Mapping Ecosystem Services226 map-user communication (see Chapter 6.4) and applications of ES maps for different purposes (see Chapter 7). Data sources and quantification methods In its simplest form of application, the ES matrix assessment uses spatial LULC data as proxies for ES supply. The advantage of LULC data is, besides its availability in many regions of the world, that many pro- visioning ES (see Chapters 2.4 and 5.5.2) can be specifically and uniquely linked to single LULC types. Timber, for example, is harvested from forests, crops grow on ag- ricultural fields and fish and seafood occur only in water bodies, rivers and the ocean. Regulating ES (see Chapter 5.5.1) and cultural ES (see Chapter 5.5.3) are usually supplied in well-functioning and not too far degraded ecosystems which can be related to more natural LULC types. ES maps, based on LULC information, provide important spatial landscape information which already can be helpful for the identification and awareness-raising of ES and their supply and demand patterns. In case the ES mapping purpose goes be- yond providing a rough overview of ES sup- ply or demand in space, further data and ES quantification approaches can be integrat- ed. The tiered ES mapping approach (see Chapter 5.6.1) helps select the appropriate method based on the mapping purpose, the necessary process-understanding and need- ed explicit measures and, last but not least, the data and resources availability. ES data from all three mapping tiers can be integrated into the ES matrix. The use of expert knowledge for ES quantification and qualification has, for example, become very popular and increasingly accepted within the scientific community. More compre- hensive ES assessments would otherwise de- mand large resources in terms of time and personnel. Data from statistics, for instance about agricultural or forestry production or existing studies with relevant information - Map with geospatial units Not relevant Very low Low Medium High Very high Scale for ranking ES supply, flow or demand ES matrix linking geospatial units with ES rankings ES ranking based on different ES quantification methods ES1 ES1 ES2 ESN X0 0 1 2 3 4 5 0 2 1 2 1 2 1 1 1 5 5 X X X X X U1 U2 U3 U4 U5 U6 ES2 Figure 1. Overview of the ES matrix approach, based on geospatial map data, the actual matrix and resulting ES maps. Chapter 5 227 if available and appropriate - should further be integrated in the ES matrix assessment. Model outcomes provide further useful data applicable for ES mapping (see Chapter 4.4). In an optimum case, data from all tiers can be acquired for the same area, time and spatial scale and in comparable resolution. These data can then be triangulated in or- der to be cross-checked and to find the most suitable, reliable and useful (for the specific mapping purpose) ES quantification and mapping method (see also Chapter 4.6). Figure 2 shows how such triangulation could take place based on data from the three tiers, normalised to the relative 0-5 scale on which the ES matrix approach is based. Data normalisation As mentioned above, the ES matrix ap- proach is based on a normalisation of ES indicator values to a relative scale ranging from 0-5. “0” represents no relevant ES sup- ply or demand. The term “relevant” is men- tioned here because “0” does not necessarily mean absolute zero (0.000.....) for all types of ES. It is supposed to reflect the fact that, in natural systems, several ES are constantly supplied, but this supply is not relevant (or not yet perceived to be relevant) for human well-being. At the other end of the scale, “5” represents the maximum ES indicator val- ue. It is important to be clear about what the reference for this maximum value is. In most cases, it is not useful to use glob- al reference values and compare them with regional studies (e.g. using tropical forests’ primary productivity as reference for bore- al forests). One pragmatic solution is to use only ES indicator values that can be found in the study area, i.e. class 5 represents the maximum amount of ES supplied or de- manded in a region. The data normalisation to the six categories needs to be based on a sound data classi- fication method using appropriate class breaks. Usually the equal intervals (see Chapter 3.3) classification methods should be used to group the data into the 0-5 classes. Outlying values in the maximum value class can be included in the 5-class (i.e. integrating values that are larger than the last equal interval maximum value). In addition, for the lowest values (0-class), a data range smaller than the respective equal interval may be suitable. Classification methods, other than equal intervals such as natural breaks or quantiles, might affect results and are less suitable to make the dif- ferent classes and their values comparable with each other. Class “4”, for instance, in- tuitively indicates twice as much ES supply or demand than class “2”, which also needs to be shown by the data. The relative data normalisation approach is comparable to the commonly used Likert scale. This scale uses five categories of de- creasing (or increasing) values to indicate, for example, frequency (very frequently – frequently – occasionally – rarely – never), agreement, importance or likelihood. Figure 2. Data triangulation across ES mapping tiers illustrating the outcomes of different meth- ods on the relative 0-5 scale. The ES map-maker needs to decide which value is most reliable, real- istic and useful for the actual ES mapping purpose. Tier 3 Tier 2 Tier 1 5 4 3 2 1 0 Mapping Ecosystem Services228 Uncertainties of the ES matrix The most appealing aspect of the ES matrix approach is perhaps its simplicity of appli- cation. The matrix delivers tangible results of ES supply and demand patterns in look- up tables and resulting maps by integrating data from various sources. However, the approach and especially its integrative char- acter include several uncertainties (see also Chapter 6) which are presented in the fol- lowing, relating to the 10 steps of applica- tion shown above: 1 . Selection of ES study area The case study area needs to be representa- tive for the addressed question and region. It needs to reflect the specific local, natural and cultural settings, land management and changing socio-ecological system condi- tions. 2 . Selection of relevant geo-biophysical spatial units Generalisation (see Chapter 3.2) and cat- egorisation of complex landscapes into a limited number of classes (e.g. LULC types) include simplification and uncertainties. Spatial units are also dependent on spatial data resolution and study area size. 3 . Collection of suitable spatial data Information availability (e.g. appropriate biophysical data on soils, hydrology and vegetation) and data access often limit com- prehensive ES studies. In some regions, not all necessary data sets are available (e.g. habitat maps). Further uncertainties can be based on inaccuracies in spatial and themat- ic data and unsuitability of spatial and tem- poral scales. 4 . Selection of relevant ES Which ES are really relevant in the case study area and which user groups are ben- efitting? Are ES imported and exported to/ from the region? Especially for data-driven studies, many ES are neglected due to data availability. 5 . Definition of suitable indicators for ES quantification ES indicators need to be robust, scalable and sensitive to changes. Furthermore, appropri- ate indicator-indicandum (i.e. the subject to be indicated) relations need to be identified and defined. Various indicators are needed for ES trade-off and synergy assessments. 6 . Quantification of ES indicators Uncertainties can be due to the lack of ap- propriate data for ES quantifications and the use of surrogate indicators, model, mea- surement and statistical data uncertainties, mismatches between geo-biophysical data and statistical data spatial units or limited knowledge about complex ecosystem func- tions. 7 . Normalisation of ES indicator values Comparability of data from different sourc- es, varying quality and quantity and across various ES categories is not always given. Moreover, subjectivity in the scoring proce- dures and data classification include uncer- tainties. 8 . Interlinking geospatial units and ES in the ES matrix The averaging of ES data over space and time is difficult (a weighting system could help but would complicate communication of results). Usually, ES supply takes place spatially and heterogeneously and aggrega- tion of data, models and indicators without losing relevant information is not easy. 9 . Linkage of ES 0-5 rankings to geospa- tial units Mismatches of selected spatial units and ES (e.g. difficulties in allocating cultural ES to land cover data), including definition of ap- propriate service providing areas (see SPA; Chapter 5.2) and ES flows can lead to un- Chapter 5 229 certainties of ES maps. Limited knowledge about complex socio-ecological system linkages, data extrapolation to different or larger regions, the proper representation of multiple ES (2D maps usually only allow the presentation of one ES or ES averages/ sums) and GIS software/data issues also add further uncertainties. 10 . Interpretation, communication and application of resulting ES maps Badly designed maps and insufficient end-us- er interfaces might cause interpretation problems (see Chapter 6.4). Data and map misinterpretation can also be due to lacking knowledge of the study area or general lack of expert knowledge, for example, concerning interactions between landscape management and ES supply. ES information is often too complex and too aggregated for easy and fast understanding. Model and map validation (see Chapter 6.3) and respective uncertainty or reliability measures are, in most cases, not provided with the ES map. Conclusions The ES matrix approach has become a very popular methodology for ES mapping. Combined with the tiered approach, various data and ES quantification methods can be used and integrated. The key advantage is the high flexibility and applicability of the ES matrix at various levels of complexity. Less complex applications relatively quickly deliver results that, for example, are useful for awareness-raising or first ES screening studies. The ES matrix is currently applied in several case studies on different spatial scales all over the world and for different mapping purposes. The methodology is increasingly improved by this approach, until the full potential eventually can be harnessed. Further reading Burkhard B, Kroll F, Müller F, Windhorst W (2009) Landscapes’ capacities to pro- vide ecosystem services – a concept for land-cover based assessments. Landscape Online 15: 1-22. Burkhard B, Kroll F, Nedkov S, Müller F (2012) Mapping supply, demand and budgets of ecosystem services. Ecological Indicators 21: 17-29. Burkhard B, Kandziora M, Hou Y, Müller F (2014) Ecosystem Service Potentials, Flows and Demands - Concepts for Spa- tial Localisation, Indication and Quantifi- cation. Landscape Online 34: 1-32 European Commission Science for Envi- ronment Policy (2015) Ecosystem Ser- vices and the Environment. In-depth Report 11 produced for the European Commission, DG Environment by the Science Communication Unit, UWE, Bristol. Available at: http://ec.europa.eu/ science-environment-policy. Hou Y, Burkhard B, Müller F (2013) Uncer- tainties in landscape analysis and ecosys- tem service assessment. Journal of Envi- ronmental Management 127: 117-131. Jacobs S, Burkhard B, Van Daele T, Staes J, Schneiders A (2015) The Matrix Reload- ed: A review of expert knowledge use for mapping ecosystem services. Ecological Modelling 295: 21-30. Kandziora M, Burkhard B, Müller F (2013) Mapping provisioning ecosystem services at the local scale using data of varying spa- tial and temporal resolution. Ecosystem Services 4: 47-59. Mapping Ecosystem Services230 Sohel SI, Mukul SA, Burkhard B (2015) Landscape’s capacities to supply ecosystem services in Bangladesh: A mapping assess- ment for Lawachara National Park. Eco- system Services 12: 128-135. Stoll S, Frenzel M, Burkhard B, Adamescu M, Augustaitis A, Baeßler C, Bonet García FJ, Cazacu C, Cosor GL, Díaz-Delgado R, Carranza ML, Grandin U, Haase P, Hämäläinen H, Loke R, Müller J, Stanisci A, Staszewski T, Müller F (2015) Assess- ment of spatial ecosystem integrity and service gradients across Europe using the LTER Europe network. Ecological Model- ling 295: 75-87. Chapter 5 231 5.7. Mapping ecosystem services on different scales Susanne Frank & Benjamin Burkhard Mapping of ecosystem services (ES) is in- herently related to the topic of scales. Var- ious scale aspects need to be taken into account in order to consider key aspects which are driving decisions in the context of land use management. First of all, the spatial scale is of importance. It is crucial to identify the appropriate spatial scale which refers to the structures, process- es, functions and services which are pro- vided or demanded in a spatial unit. This unit might have a local, regional, national, continental, or global extent. Addition- ally, spatial scale is characterised by the “grain”, i.e. the spatial resolution of a map. The higher the spatial resolution (and the smaller the Minimum Mapping Unit, MMU), the more detailed statements may be derived from a map. Once the spatial scale is clear, the crucial information for successful application of the ES concept is provided by the map content. The thematic resolution of maps should reflect the subject of interest. For some basic statements, for example on soil sealing, the distinction of two thematic classes, sealed and un-sealed, can be suffi- cient. When it comes to the mapping of more complex processes, functions or ser- vices, higher thematic resolution would be required. This is especially true for high- ly specialised systems. To distinguish the erosion potential of specific crop rotation types, for example, numerous thematic classes are required which reflect the num- ber of crops, the length of a rotation peri- od, the soil management type etc. The spatial scale, however, should not only refer to grain and extent. The third dimen- sion should also be considered. Slope in- clination, relief intensity and the elevation above sea level significantly affect the quan- tity, quality and distribution of ES. Mapping of ES at a specific spatial scale re- veals insights of the current situation. For the application of the ES concept in policy making or spatial planning, the monitoring of changes, as well as visualisation and eval- uation of possible futures is of great impor- tance. Not only spatial “hot spots” and “cold spots” of ES supply and demand need to be considered (see Chapter 5.2), but also “hot moments” and “cold moments” (see Chap- ters 5.3 and 5.7.5). Temporal scales range from short-term, sea- sonal, annual, medium-term, to long-term considerations. Again, depending on the subject and the purpose of a study, the appro- priate scale needs to be identified. Cross-sec- toral, integrated spatial planning (see Chap- ter 7.2) at the regional scale, for example, typically refers to the medium term perspec- tive (10 to 20 years in the future). In con- trast, sectoral forestry planning (see Chapter 7.3.3) requires both operational short-term planning and long-term consideration of a couple of hundred years to reflect the forest development from planting to final harvest- ing (see Chapter 5.7.5). Looking backwards, maps of land use changes can reveal insights of past developments which are fundamen- tal for the estimation of future trends in a region’s development. Mapping Ecosystem Services232 These various dimensions of scale are inter- linked. Usually, global considerations take large-scale and long-term topics at lower thematic resolution into account (see Chap- ter 5.7.3). On the other hand, local or re- gional studies typically are characterised by deeper understanding of processes and functions and availability of high resolution data regarding spatial and temporal scale (see Chapter 5.7.1). One of the major chal- lenges is the up-scaling of local knowledge on higher scales (see Chapter 3.7). Without the understanding of local structures and processes, the regional, national and global mapping and assessment of ES would run the risk of neglecting essential information which determines the ES performance. Chapter 5.7 gives an overview of different scales of ES mapping, covering regional, national and global perspectives. Marine areas and the interactions of spatial, the- matic and temporal scales are specifically addressed. Further reading Reid WV, Berkes F, Wilbanks T, Capistrano D (Eds.) (2006) Bridging scales and knowl- edge systems: concepts and applications in ecosystem assessment / Millennium Eco- system Assessment. Island Press/World Re- sources Institute, Washington, DC. Chapter 5 233 5.7.1. Regional ecosystem service mapping approaches Marion Kruse Introduction Human activities and therefore ecosystem services (ES) act on different scales – not only temporally but also spatially. Consid- eration of these different spatial scales is especially important for a meaningful and precise mapping of ES (see Chapter 5.7). Results can be very different depending on the investigated ecosystem service(s) and the available data sets. The first barrier is the fact that no clear defi- nition exists about what regional or local mapping approaches mean or include. Most often the area that is mapped is considered as the spatial scale. However, the applied data sets also have different resolutions regarding the technical or thematic scale (ranging from very fine to very coarse) and have therefore the ability to identify important ES or not. The aim of this chapter is to give a short overview of some necessary requirements that need to be kept in mind when mapping at local or regional scale. Spatial scales Local scale The term local is mainly connected with a specific (geographic) position. Local scale can range from single farms to villages/ communities and to smaller administration units (e.g. municipalities). This depends, of course, on the administrative/political or historical conditions of the case study area. However, small protected areas or spe- cific ES that only act over a very narrowly defined extent (e.g. sacred/holy features of nature such as trees) can also be considered local in mapping approaches. In addition, some ecosystems cover only a limited area; for example, species-rich wetlands. In coast- al or marine ecosystems, harbours, steep coasts or reefs are, for example, in a specific location and provide many ES. Local case studies are particularly suitable for more labour-intensive data compila- tions and method testing. Many participa- tory mapping studies and direct stakehold- er-involved assessments (e.g. focus group surveys) have been undertaken at the local scale. Beyond that, there are several ES that can sometimes only occur at defined extents; scenic views, for example or wild food such as mushroom picking. Mapping these is best done on the matching scale for a precise re- sult of the human preferences and activities contributing to human well-being. Data sets must be of high resolution to ad- dress the peculiarities of local mapping stud- ies. Applying data that is too coarse (e.g. ag- gregated land cover or land use) will blur the findings. Data mismatches can have a strong misleading effect on land use management and decision-making. Local case studies are important for many participatory steps in- cluding communication and raising aware- Mapping Ecosystem Services234 ness for ES. Decisions on ecosystem man- agement need to meet local requirements and fine resolution data and information. Regional scale A region is an area of indefinite size that is different to the adjacent areas. It can range from a part of a country (e.g. Northern Ger- many) to a part of the globe (e.g. Scandi- navia). This means the term can act as an administrative unit or describe an area based on similar characteristics (e.g. subarctic re- gions with similar climatic conditions or the Amazon basin). Therefore, regions contain either similar natural or cultural/economic characteristics. Due to the similar features within a region, this spatial scale is a suitable mapping unit for many ES. In addition, there are specific connotations, such as a tourism region (e.g. the Alps, the Baltic Sea) or a region is important or known for its specific features (e.g. the breadbasket of a country like the Great Plains of the US). Based on these different criteria, regions can also overlap with each other or certain areas within a region could be excluded if they do not possess the functional or homogenous criteria. The term is also very specifically used in some languages, fostering further challenges in the assessment and mapping of ES by delineating case study areas. In many cases, several data sets are available in aggregated format ranging over a great ex- tent. Land cover or land use data sets can act as an appropriate (first) approach for map- ping regional ES (see Chapter 5.6.4). Mapping methods and data requirements Local mapping approaches can be quick- ly supported by direct (participatory) mapping and data acquisition/measuring or ground-truth checks when applying available data sets or other methods. Spe- cific data sets, such as detailed habitat or biotope maps, are available on local scales with high resolution. Supply and demand budgets can be accounted for and mapped more easily. Additionally, web or smart phone based data acquisition (e.g. citizen science; Chapter 5.6.3) are suitable for smaller case studies and stakeholder con- sultations (interviews, workshops). There are also models that work on the site scale or farm scale, especially for regulating and provisioning services. On the other hand, statistical data are often not available at high resolution information levels due to privacy protection or highly time-con- suming acquisition. Regional mapping approaches contain all available methods (see Chapters 4 and 5). Single indicators, statistical data and mod- elling can be applied together with stake- holder assessments (e.g. expert interviews). Spatial data resolutions are often > 100 m. Cultural ecosystem services Many cultural ES can best be mapped on lo- cal or regional scale, allowing the inclusion of specific aspects of preferences and activi- ties. Accessibility is an important point for recreation and tourism, as well as for land- scape aesthetics. Points of interest, hiking paths, roads, streams and other landscape features must be included for a comprehen- sive analysis. In regional maps, aggregated information (for instance, different beach types) is needed to give a more general over- view of cultural ES. Surveys in tourist locations are most often un- dertaken for a specific purpose to understand the motivation of tourists for visiting a certain place (e.g. beach vs. cultural attractions). Chapter 5 235 Regulating services Many of the underlying natural processes in- cluded in regulating services are not restricted to small areas and are complex. Besides pri- mary data collection which is usually difficult and resource-dependent, secondary data sets are often included for modelling regulating services (cf. InVEST, ARIES; Chapter 4.4). Water-related regulating services (e.g. nu- trient retention, erosion regulation, flood protection, water flow regulation) should be considered on the river basin/catch- ment scale. This is already implemented in river basin management (as required, for instance, under the EU Water Framework Directive; see Chapter 7.1). Similarly, landscape features (same soil, cli- mate and flora/fauna) or cultural landscapes (same land use in the past and today) can in- fluence the supply of ES and could be con- sidered as the mapping unit. Some models require different base layers of the natural conditions for the mapping of ES which are best quantified at the regional scale. Provisioning services Provisioning services are, in many ar- eas, well documented and monitored (see Chapter 5.5.2) and statistical data is often applied together with matching land use/ land cover data. For Europe, regional statistical data sets are available for many different subdivisions of countries (cf. NUTS and EUROSTAT). Furthermore, many countries have official land cover/land use data sets along with oth- er data sets (e.g. statistical data) which allow the comparison of ecosystem service supply and demand. However, as this is often ag- gregated and generalised, this approach is best applied in larger case study areas. Challenges and solutions Given the fact that ES act on different scales, it is not always possible to have all data sets available for all scales. For data-poor regions, value-transfer (Chap- ter 4.4) or look-up tables (Chapter 5.6.4) from similar biomes or ecosystems are often utilised. These should be carefully selected and checked. Many models incorporate a sensitivity or uncertainty analysis. Mapping on different scales can also support each other by testing methods and data sets for applicability and transferability. Analysing and mapping single ES that con- sider temporal aspects are time-consuming. However, knowledge gained on the local scale is important to further verify and improve conceptual and methodological issues. Re- gional scales are suitable for trade-off analysis between ES based on land use scenarios, as realistic supply-demand budgets can be calcu- lated and mapped. Furthermore, bundles of ES or synergies can be assessed and mapped. Conclusions Neither the term regional nor local can be broken down to a simple and clear defini- tion. What is clear is that diverse spatial scales are important for mapping ES be- cause ES (inter-) act over different scales. Considering the different spatial effects, it is necessary to carefully select the correspond- ing extent and data sets for mapping. Not all methods and data sets are easily trans- ferable between scales. A local scale is often appropriate for cultural services, whereas many regulating services are best modelled at the regional scale. Data available from statis- tics are, in most cases, a good source for map- ping provisioning services at regional level. Mapping Ecosystem Services236 Further reading Burkhard B, Crossman ND, Nedkov, S, Petz K Alkemade R (2013) Mapping and Mod- elling Ecosystem Services for Science, Pol- icy and Practice. SpecialIssue. Ecosystem Services 4: 1-146. Malinga R, Gordon LJ, Jewitt G, Lindborg R (2015) Mapping ecosystem services across scales and continents – A review. Ecosys- tem Services 13: 57-63. Pagella TF, Sinclair FL (2014) Development and use of a typology of mapping tools to assess their fitness for supporting man- agement of ecosystem service provision. Landscape Ecology 29: 383-399. Willemen L, Burkhard B, Crossman ND, Palomo I, Drakou E (Eds.) (2015) Best Practices for Mapping Ecosystem Services. Special Issue 13: 1-184. Chapter 5 237 5.7.2. National ecosystem service mapping approaches Sharolyn Anderson, Alberto Giordano, Robert Costanza, Ida Kubiszewski, Paul Sutton, Joachim Maes & Anne Neale Introduction The creation of any comprehensive mapping instrument at the national level requires the careful consideration of a set of issues, with components that range from the scientific to the technical and from the economic to the organisational. Wealthier countries, such as the United States and many European countries, have a long tradition of national level cartography, analogue and then digital, dating back centuries - with the first com- prehensive and ‘modern’ example being the Cassini Maps of 18th century France. In the United States, the ‘National Map1’ is the dig- ital version and the continuation of efforts to map the country at a variety of scales and for multiple purposes was started in the late 1800s by the United States Geological Sur- vey. One of many efforts to provide nation- al maps for the US was the ‘National Map’ which includes data layers on elevation, hy- drography, geographic names, transporta- tion, structures, boundaries, ortho-imagery and land cover. Another example, the ‘Aus- tralian National Map’2, includes not only the same data layers as the U.S. national map but also layers on communication, environ- ment, framework, groundwater, habitation, infrastructure, utility and vegetation. For the world in general, the quality and quantity of information related to ecosys- 1 http://nationalmap.gov/ 2 https://nationalmap.gov.au/ tems and ecosystem services (ES) has been growing and it is expected that it will contin- ue to do so as a result of increasing awareness of our fundamental dependence on natural capital and the value of ES. In this context, national maps may function as providers of reference cartographic data (see Chapter 7.1). Action 5 of the EU Biodiversity Strate- gy to 2020 calls for European Union’s mem- ber states to map and assess the state of eco- systems and their services in their national territory. In the United States, a memoran- dum was issued in October 2015 directing Federal agencies to factor the value of ES into planning and decision-making activities at the federal level (see Chapter 7.1 for more details). The mapping of ecosystems is an es- sential first step in conducting an inventory of that portion of our common wealth that manifests as natural capital. In this chapter, we briefly touch - from the perspective of the mapmaker - on a small set of topics related to the national mapping of ecosystems and ES. This discussion is by no means exhaustive and additional topics may be worth reviewing. Our objective is to inform the reader and to pique his or her curiosity; for further information, vast liter- ature exists on all of these topics. Mapping Ecosystem Services238 Peculiarities of national mapping scale and projections The term “scale” is often used loosely and casually in lay conversation and may take different meanings depending on the tradi- tions and conventions of individual fields. For example, some ecologists use the ex- pression ‘large scale’ when referring to large areas. In cartography, scale is defined as the ratio between distances on the map and corresponding distances on the ground (see Chapter 3.1). Thus, a 1:1,000 map is at a larger scale than a map with a scale of 1:10,000, because the value of the ratio of the former (0.001) is larger than the value of the latter (0.0001). Thus, for a cartogra- pher, a map at large scale shows a smaller area than a map at a smaller scale. Large scale maps show detail, as a map of one’s backyard might be. Although guidelines for the classification of maps, according to their scale, have been developed and are in use, what constitutes a ‘large’ or ‘small’ scale map is a matter of convention. In classical hand- books of cartography, maps have been classi- fied as ‘large scale’ (1:50,000 and less; for ex- ample, 1:25,000) or ‘small scale’ (1:500,000 and more, for example, 1:1,000,000), with medium scale maps somewhere in between. Individual countries may impose their own guidelines based on local situations, conven- tions and needs. Although national maps are typically at a larger scale than maps showing continents or the entire world, it is the size of the coun- try mapped that puts limits on the scale of its national maps and therefore on the level of detail for the cartographic representation. For example, a national map of ecosystems and ES for South Africa would be very dif- ferent from a comparable map for Belgium, not only because ecosystems are more varied in the former than in the latter, but also be- cause the level of detail at which thematic layers (land use, vegetation, infrastructures, etc.) that can be shown in the map of Bel- gium are much higher than in the South Af- rican example. Concerning projections, the cartographic representation of real-world 3-D objects on a 2-D map necessarily introduces distor- tion (see Chapter 3.1). The larger the object mapped, the higher the amount of distortion. Regarding the national mapping of ecosys- tems and ES, we would argue that distortion in the size of the objects mapped and their relative distance are of special concern, as quantitative errors affect measurements, both linear and areal. Distortion in shape or direc- tion may affect the cartographic representa- tion and should be taken into consideration - the latter would be especially serious in case of nautical maps. The good news is that the way distortion varies across a map is predict- able and tools exist (e.g., the Tissot’s Indica- trix) to measure it accurately. Another good news is that all countries have established co- ordinate systems (which also describe projec- tions, datum, etc.) for mapping their territo- ries at various scales with the explicit purpose of minimising distortion. Resolution In the cartographic context, a concept relat- ed to ‘scale’ is that of ‘resolution.’ The two differ in that scale is measured linearly, while resolution is a measure of size. Thus, a re- mote sensing image at a resolution of 100 metres shows an area of 10 by 10 metres (assuming a square pixel). Such a resolution level would be coarser than an image at a resolution of 30 metres. This is relevant to the map-making process at any scale, in- cluding the national scale, in the sense that images at higher resolutions give the cartog- rapher the option of making maps at larger scales. To return to the example made earli- Chapter 5 239 er, creating a map of one’s backyard would be impossible using an image at a resolution of 100 metres, but feasible with an image at 1-metre resolution. Thus, the spatial res- olution of available primary sources is one of the principal factors affecting map scale. One complicating factor is that, as it per- tains to satellite imagery, the term ‘resolu- tion’ has dimensions that are not spatial, including radiometric (e.g. how many levels of brightness; 6 bit, 8 bit, 12 bit, etc.), tem- poral (e.g. data acquisition frequency) and spectral (e.g. number of bands, bandwidths, etc.) resolutions. Note that the higher the resolution - in all of the above senses - the more expensive the primary source tends to be per size of the area mapped. Generalisation Cartographic generalisation, defined as the reduction of spatial and thematic detail needed to map the real world, is related to scale and resolution. In general, the smaller the scale of the map, the higher the amount of reduction needed (see Chapter 3.2). Note, however, that different levels of generali- sation can be applied to the same primary source. Generalisation is a decision-making process measured along a continuum from low to high, with the high limited by the resolution of the image (recall the backyard example). This example also makes anoth- er important point: the cartographer works with the expert (in this case, an ecosystem expert) to determine the level of generalisa- tion needed to answer specific research and/ or policy-related questions (see Chapter 4.6). Accuracy and currency of data In cartography, ‘accuracy’ is defined as the closeness of a measurement to its true val- ue. This is different from the definition of precision which pertains to the instrument used to make this measurement. To under- stand this idea, consider reading the latitude and longitude of the point at which you are standing from a GPS receiver. The position is estimated with a certain distance accura- cy (for example, 2 metres); if the signal is scrambled- as might be undertaken in areas of conflict by the country that controls the GPS (US, Russia, China, etc.) - the unit will continue to indicate the same level of accu- racy, even though its precision has been de- graded. In addition to its spatial dimension, measured in quantitative terms, accuracy has another dimension which is particular- ly important in the context of the national mapping of ecosystems and ES. This is the- matic accuracy, which is usually measured in terms of categories and therefore quali- tatively - for example, consider a land cover layer in which a vegetated area is incorrectly classified as urban area. As it is for spatial accuracy, methods and tools exist for mea- suring thematic accuracy both at the level of feature and for the entire map. Equally important is the currency of the in- formation used. In addition to the obvious consideration that having up-to-date infor- mation is to be preferred to having outdated information, a crucial factor to consider is whether individual layers are current rel- ative to each other. For example, consider deforestation which has progressed in some countries very quickly over the last 20 or 30 years: a layer of forested areas in, for exam- ple, Guatemala ca. 2000 would look very different than a corresponding layer from 2016. According to an old adage in cartog- raphy, a map is only as current as the newest data source that was used to create it. Creat- ing a composite map from layers that show the situation on the ground at different dates would lead to erroneous conclusions. Note, though, that currency is of concern for certain types of information but not for Mapping Ecosystem Services240 others: for example, a geologic map does not need to be updated as frequently as a map of urban areas (see also Chapter 5.3). In practical terms, accuracy and currency are dealt with in relative rather than abso- lute terms. This is the idea of ‘fitness for purpose’: because maps, especially at the national scale, are expensive to produce, up- date, maintain, distribute and, in legally liti- gious countries, the responsible agency can be brought to court for inaccurate represen- tations, governmental cartographic agencies should and, usually do, use metadata to de- scribe how the maps should be used, their limitations, accuracy levels and currency (in other words, their ‘fitness for purpose’). Related to this discussion, in the last thirty years many countries and international or- ganisations such as the ISO, have developed standards for the accuracy of geographic information. Note that, in the cartographic field, standards have been in long use, for example, the US National Mapping Accu- racy Standard (NMAS) dates back to 1947. Data Sources There are myriad sources of data that can potentially inform and contribute to the production of maps for ecosystems and ES (see Section 4). A non-exhaustive list might include various types of satellite imagery, human population census data, agricultural productivity statistics, soil maps, vegetation maps, air quality measurements, biological census data, transportation and other infra- structure maps and climate station data and maps3. These data can be applied to the pro- duction of different kinds of ecosystems and ES mapping. A key question to answer is how to structure and organise the representation of ES? This 3 http://biodiversity.europa.eu/maes question applies to all cartographic represen- tations ranging from the local to the region- al, to the national and to the international. One approach is to create a separate layer for every ecosystem service (e.g. one layer for carbon sequestration, one for erosion control, one for spiritual values etc.). This approach is convenient from a taxonomic perspective but can be problematic, as varia- tions in most of these services are driven by land cover proxy measurements (e.g. boreal forests sequester X kg/ha/year whilst deserts sequester Y kg/ha/year), but, in others, they vary as a function of spatial interactions with other spatially variable information (e.g. spiritual value will likely vary as a func- tion of proximate population density, the income of that population and the spiritual values of the proximate population). Car- bon sequestration provides a salient example of the relevance of these issues. It is increas- ingly regarded as a policy-relevant ecosys- tem service as a result of climate change. At a national level, authoritative, verifiable and valid ground-based measures of carbon sequestration which include direct measure- ments of vegetation and soil would likely be needed to produce a comprehensive, coun- try-wide map of carbon sequestration. Scientific accuracy, transparent methods of measurements and reliable and independent interpretation and dissemination of results would be needed to ensure the legitimacy of the process, both internally at the country level and in the international arena. Here, again, we run into the problem of economic costs, in the sense that valid and authori- tative maps representing real and dynamic phenomena may be expensive to produce, maintain and update at the required levels of cartographic detail, accuracy and currency. For example, the 2010 United States Census of the Population cost approximately $13 billion to conduct, or over $40 per person counted and mapped. The degree to which large investments can be made by individual Chapter 5 241 Box 1 . Mapping ecosystem services at national scale in the European Union In the EU, countries have started initiatives to map their ecosystems and ecosystem services (ES) on their national territory. The principal objective is to create a national knowledge base on ecosystems which can be used for planning purposes such as the selection of areas for ecological restoration, the development of new infrastructure projects or land and water management. The European Commission is providing guidance to countries on how to map ecosystems and ES through the MAES initiative and collects information of countries on the biodiversity information system for Europe4. Two examples for Cyprus and The Netherlands illustrate nation-wide mapping of ES in the EU. Cyprus is an island in the Mediterranean Sea. The map illustrates the recreational potential of the traditional landscape and nature. The map was made in a training workshop where country officials from the min- istry worked together with scientists to map recreational services on the island. The Netherlands create maps of ES which are publicly available via their Atlas of Natural Capital5. 4 http://biodiversity.europa.eu/maes 5 http://www.atlasnatuurlijkkapitaal.nl/en/home| recreation potential traditional landscape and nature < 0,090 0,090196078 - 0,098 0,098 - 0,17 0,17 - 0,50 > 0,50 0 25 50 kmGeoarchaeologySites A map of recreation potential offered by the traditional cultur- al landscape and nature. This map is based on the recreation opportunity spectrum approach. The red dots are places of ar- chaeological interest. Map of the water storage capacity of soil (expressed in mm) in the Netherlands is derived from the Atlas of Natural capital which collects spatially explicit data of ES at national scale. Mapping Ecosystem Services242 Box 2 . Mapping ecosystem services at the national extent for the conterminous United States In the US, the Environment Protection Agency leads a multi-organisation effort to develop and host a suite of nationwide maps of ecosystem services (ES) indicators and indices in EnviroAtlas6. This open access tool allows users to view, analyse and download a wealth of geospatial data and other resources related to ecosystem goods and services. More than 160 national indicators of ecosystem service supply, demand and drivers of change provide a framework to form decisions and policies at multiple spatial scales, educate a range of audiences and supply data for research. A higher resolution component is also available, providing data for finer-scale analyses for selected communities across the US. The ecosystem goods and services data are organised into seven general ecosystem benefit categories: clean and plenti- ful water; natural hazard mitigation; food, fuel and materials; climate stabilisation; clean air; biodiversi- ty conservation; and recreation, culture and aesthetics. EnviroAtlas incorporates many data sources with multi-resolution (i.e., 1 m and 30 m) land cover data providing fundamental information. The data are updated at 5 year increments, subsequent to US National Land Cover Dataset updates. 6 https://epa.gov/enviroatlas This map shows the kind of data layers that are available in EnviroAtlas. For one of the indicators in the climate stabilisation category, this map shows the amount of carbon stored in the above- ground tree biomass. Like most of the national maps in EnviroAtlas, the data are summarised by medium sized watershed drainage basins known as 12-digit hydrological unit codes (HUCS). There are approximately 85,000 of these HUCS in the conterminous US, with each being approx- imately 104 km². Users of EnviroAtlas can also overlay demographic maps to gain the perspec- tive of proximity and population dynamics of beneficiaries. Chapter 5 243 countries in order to map ecosystems and ES remain to be seen. Perhaps the solution is partnerships between countries - exam- ples include the European Union’s Joint Research Centre (JRC) and the United Na- tions Environmental Programme (UNEP) - as well as efforts by individual countries to create, maintain and share primary envi- ronmental data, including initiatives by US government agencies (for example the Na- tional Aeronautic and Space Administration (NASA) and the National Oceanic and At- mospheric Organisation (NOAA)). Conclusions For the public, national maps can provide benefits that exceed their costs of produc- tion, assuming the maps are soundly exe- cuted, regularly updated and distributed to the public at a reasonable cost. When mapping ecosystems and ES at national levels, careful consideration should be giv- en in the very early planning stages to the scale, accuracy and level of generalisation needed for the explicit and specific purpose the map is intended to serve. This is cru- cial when one considers that the degree to which a country acquires up-to-date and reliable knowledge of its ecosystems and ES will determine its ability to manage them. Mapping should not only provide information on the quality and quantity of ES but also on their distribution among the population within a country which is key to issues of equality and social justice. Usually, the loss of ES has the greatest im- pact on the poorest communities which, as a group, are the first to feel the effects when those ES begin to disappear. In this sense, the mapping of ecosystems at the national scale is essential to understanding the mag- nitude and spatial distribution of such ser- vices and for the development of policies to protect and restore them. Finally, we stress that the most important investment a country can make when ad- dressing these issues is on its human capi- tal. The creation, maintenance, update and distribution of a national mapping initia- tive require trained, skilled, committed and motivated personnel, with technolog- ical considerations important but second- ary. The human capital should have the highest priority. Further reading Bailey RG (2009) Mapping Regional Eco- systems. Springer 2nd ED. DOI: 10.1007/978-0-387-89516-1. Burkhard B et al. (2009) Landscape’s capacity to provide ecosystem services – a concept for land cover based assessments. Landscape on-line 151-22 DOI: 10.3097/lo.200915. EU biodiversity strategy to 2020 Mapping and Assessment of Ecosystems and their Ser- vices http//biodiversity.europa.eu/maes. Robinson AH et al. (1995) Elements of Car- tography. New York: John Wiley and Sons, sixth edition. Schmidt S, Manceur A, Seppelt R (2016) Un- certainty of Monetary Valued Ecosystem Services – Value Transfer Functions for Global Mapping PLOS ONE March 3. Pickard BR, Daniel J, Mehaffey M, Jackson LA, Neale A (2015) EnviroAtlas: A new geospatial tool to foster ecosystem services science and resource management, Ecosys- tem Services 14: 45-55. Mapping Ecosystem Services244 5.7.3. Global ecosystem service mapping approaches Katalin Petz, Clara J. Veerkamp & Rob Alkemade Introduction The global mapping of ecosystem services (ES) helps to diagnose management and con- servation problems and develop solutions for them, as well as to analyse the impact of man- agement decisions on biodiversity and ES. It enables the identification of synergies, trade- offs, hotspots of ES delivery and spatial mis- matches between ES supply and demand or within world regions or sectors. Global initia- tives (e.g. Convention of Biological Diversi- ty1 and Millennium Ecosystem Assessment2) make use of global ES maps to investigate the state and trends of global biodiversity and ES in order to formulate international policies. There is, consequently, an increasing demand for accurate maps of ES supply, demand and values. ES mapping is applied both for bio- physical assessment of services and for valu- ation of these services. The history of glob- al mapping of ES and their values globally dates back to the 1990s, concentrated on the monetary value of ecosystems. In the new millennium, global ES mapping studies were expanded to more biophysical descriptions. Although the number of publications target- ing the mapping of ES has rapidly increased in the last years, global ES mapping remained limited to a few provisioning and regulating services (e.g. food provision, water availabil- ity and carbon sequestration). Obstacles for global ES mapping include the resolution of the available data, the uncertainty involved in upscaling local phenomena and the lack of knowledge of global ecological processes (see Chapter 6). 1 https://www.cbd.int/ 2 http://www.millenniumassessment.org/ Various mapping approaches A common approach for mapping ES is to quantify the relationships between eco- system conditions (see Chapter 3.5) and ecosystem functions (i.e. the ecosystem’s potential to provide a service, see Chapter 2.3) or services (i.e. the actual use of the function by humans; see Chapter 5.1). The mapping of ES often starts with maps of ecosystem types, land cover and land use. ES are then derived by applying models, quantifying each ES for each type of land use or land cover within each ecosystem. These models can either be simple correl- ative or expert-based models (see Chapter 4.6) or more complex process-based mod- els (see Chapters 4.4 and 5.6.1). Develop- ing these models is one of the main chal- lenges for mapping global ES. Global models are suitable tools for interna- tional science-policy assessments and deci- sion-making support by assessing the impact of socio-economic drivers on the environ- ment and ES. The Millennium Ecosystem Assessment used already-published individ- ual models to assess the global ES trends and patterns. Others link sector-based global models to simulate the interaction between environmental processes and certain ES. Ex- amples for global models are the Integrated Model to Assess the Global Environment (IMAGE3) developed by the PBL Nether- lands Environmental Assessment Agency 3 http://themasites.pbl.nl/models/image/index. php/Welcome_to_IMAGE_3.0_Documenta- tion Chapter 5 245 and the Global Unified Metamodel of the Biosphere (GUMBO4) by the University of Maryland. The International Institute for Applied Systems Analysis (IIASA) has also developed several global models used in pol- icy support, such as the Global Biosphere Management Model (GLOBIOM5). Other efforts being applied to making decisions about ES in various case studies across the globe are the Natural Capital Project’s In- VEST6, the ARtificial Intelligence for ES7 and The Earth Genome8 (see Chapter 4.4 for an overview of ES models). Another common application of ES map- ping is the creation of maps of monetary val- ues (Chapter 4.3). Such approach is supposed to draw attention to the relative importance and the potential economic benefit that can be gained from ES, for example, when mak- ing choices on land management. The Bene- fit Transfer method is the simplest approach for ES value mapping. It estimates economic values by transferring existing estimates from studies already completed for (another) loca- tion. Values of various ES are aggregated to a constant value applied for an ecosystem or land cover type. The TEEB Valuation Data- base9 provides a Total Economic Value (TEV) for ES per global ecosystems or land covers. Global ES modelled by IMAGE and GLOBIO-ES IMAGE is one of the few integrated global models describing the impacts of socio-eco- nomic drivers on the environment. IMAGE has been used in combination with the glob- 4 http://www.sciencedirect.com/science/article/ pii/S0921800902000988 5 http://www.globiom.org/ 6 http://www.naturalcapitalproject.org/invest/ 7 http://aries.integratedmodelling.org/ 8 http://www.earthgenome.org/ 9 http://www.fsd.nl/esp/80763/5/0/50 al biodiversity model GLOBIO10 to assess impacts of human activities on biodiversity captured by the Mean Species Abundance. Later, the model was extended with addi- tional ES modules into the GLOBIO-ES model. IMAGE provides information about environmental drivers (e.g. climatic fac- tors and land use allocation) that feed into GLOBIO and GLOBIO-ES. These models map biodiversity and ES at 0.5°x 0.5° spatial resolution and apply cause-effect relation- ships between the environmental variables, biodiversity and ES derived from literature. Currently, biodiversity and eleven ES can be assessed with the IMAGE-GLOBIO mod- eling framework. Although the models are strong in simulating the effects of changing socio-economic drivers and consequent bio- physical and climate pressures on biodiver- sity and ES; the modelling of interactions between biodiversity and ES and between the various ES as well as the policy response to states of ES are missing links. The models have been applied for assessing biodiversity and ES at regional and global scale11. Two concrete application examples of these mod- els are presented in Boxes 1 and 2. Challenges of global mapping Mapping ES becomes more challenging with increasing extension of the mapped area, since less quantitative data and poorer knowledge of ecological and other processes are available and higher level of aggregation and simplification is necessary compared to regional and local scales. Data availability and quality Global ES modelling relies highly on land cover and land use data. Only few standard datasets exist and information on landscape 10 http://www.globio.info/ 11 https://www.cbd.int/gbo4/ Mapping Ecosystem Services246 structure, land use intensity and land man- agement is poor or lacking. A widely used ecosystem or biome map is provided by the World Wildlife Fund12. A commonly used land cover and land use dataset is the Glob- al Land Cover (GLC) 2000 map13, which is also used in the Millennium Ecosystem Assessment and the IMAGE and GLO- BIO-ES models. The TEEB Valuation Database uses the GlobCover dataset14. This dataset provides a higher-resolution alternative to the Global Land Cover, but it also has a lower thematic accuracy. There are also other databases available targeting certain ecosystem or land covers, such as the Global Lakes and Wetlands Database15, the World Database of Protected Areas16, the livestock density database of the Food and Agriculture Organisation of the Unit- ed Nations (FAO)17 and forest cover data- sets18. Due to the limited data availability, the same datasets are often used for mul- tiple purposes, which can lead to autocor- relations. Global data include increased uncertainty as they are often estimated or modelled (e.g. FAO livestock data). Un- certainty can be addressed with sensitivity analyses (see Chapter 6.3), but is not of- ten done in practice. Last but not least, it remains difficult to validate global datasets due to differences in temporal and spatial consistencies and classification systems, amongst others. 12 http://www.worldwildlife.org/biomes 13 http://forobs.jrc.ec.europa.eu/products/glc2000/ glc2000.php 14 http://due.esrin.esa.int/page_globcover.php 15 http://www.worldwildlife.org/pages/glob al-lakes-and-wetlands-database 16 http://www.protectedplanet.net/ 17 http://www.fao.org/ag/AGAInfo/resources/en/ glw/GLW_dens.html 18 e.g. http://www.globalforestwatch.org Ecological processes and ES: knowledge and scale at which they operate The knowledge of ecological and other pro- cesses becomes more limited with increasing extension of the mapped area. ES that operate based on well-known global processes, such as the hydrological or carbon cycle, are easier to map globally. Furthermore, global maps are more easily generated if an ES can be ag- gregated across time or space. This is the case for several provisioning services, such as crop, timber or livestock production. For these ES, monetary value maps can also be prepared, as their products are traded on markets. ES that operate locally are, however, more difficult to map globally. ES such as pest control and air quality regulation are rare- ly considered globally because of the lack of generalised knowledge and the local scale at which they operate. Pollination and pest control are dependent on small-scale land- scape elements making it difficult to map them accurately globally. Furthermore, cul- tural services such as aesthetic value, rec- reation and tourism have a subjective and local character which makes them difficult to generalise. As these ES do not have a di- rect market value either, it is more difficult to prepare a monetary value map for them. Little generalised information is available about the degradation of ecosystem func- tions over time and the inter-linkages be- tween biodiversity and ES and between the various ES. Degradation is, therefore, not fully addressed and biodiversity and ES are mainly modelled and valued separate- ly at global scale. An approach to address these inter-linkages between ES is to create hotspot maps (i.e. highlighting areas where multiple services are provided). Chapter 5 247 Conclusions Various approaches exist for the global mapping of ES, the most common ones being the biophysical and monetary value maps. Despite the limited data and knowl- edge available at global scale, global ES maps remain an important input for inter- Box 1 . Example soil erosion prevention on global rangelands Figure 1 provides an example map for the current state of soil erosion prevention on global rangelands. A further developed version of the Universal Soil Loss Equation (USLE) applied in the IMAGE model was used for mapping this service. Erosion prevention was mapped with an index (0-100) based on soil erodibility, rainfall erosivity, both derived from IMAGE and a refined land use/cover index derived from the vegetation cover fractions of the Global Land Cover map. Low erosion prevention (i.e. high erosion risk) is the result of steep slopes, sensitive soil and scarce vegetation cover (e.g. in the Mediter- ranean, Central Australia and Chile). Figure 1. Soil erosion prevention ES on global rangelands (Petz et al. 2014). national science-policy assessments and for awareness-raising. Global models have the capacity to simulate ES trends across space and time and to identify ES synergies, trade-offs and values. This makes them essential tools for decision-making about resource management and nature conser- vation across the globe. Mapping Ecosystem Services248 Box 2 . Global crop production under two extreme scenarios With the help of scenarios, the trends of ES delivery can be projected over time. In this example, the global crop production is simulated with the IMAGE and GLOBIO-ES models for two future sce- narios. The production of cereals, rice, maize, pulses, root and tubers is taken as an indicator of crop production. The demand for crops is driven by changing lifestyle and population, whereas technology, environmental factors and management determine the production efficiency hence the crop yield. The two scenarios are adjusted SSP scenarios (i.e. new IPCC scenarios) used in the OpenNESS EU project. The ‘Wealth-Being’ (WB) scenario stands for economic growth, while the ‘Eco-Centre’ (EC) scenario promotes sustainable management around the globe. Figure 2 illustrates the potential change in crop yield in 2050 in comparison to the base year of 2010. Crop yield increases in developing countries (e.g. Africa, India) in the EC scenario, while the WB scenario projects lower crop yield in these countries, but higher yield increase in US and Brazil. Figure 2. Change of crop production under two extreme scenarios (PBL 2016). Chapter 5 249 Further reading Alkemade R, van Oorschot M, Miles L, Nellemann C, Bakkenes M, ten Brinket B (2009) GLOBIO3: A Framework to Investigate Options for Reducing Global Terrestrial Biodiversity Loss. Ecosystems 12(3): 374-390. De Groot R, Brander L, van der Ploeg S, Cos- tanza R, Bernard F, Braat L, Christie M, Crossman ND, Ghermandi A, Hein L, Hussain S, Kumar P, McVittie A, Portela R, Rodriguez LC, ten Brink P, van Beuker- ing P (2012) Global estimates of the value of ecosystems and their services in mon- etary units. Ecosystem Services 1: 50-61. Dickson B, Blaney R Miles L, Regan E, van Soesbergen A, Väänänen E, Blyth S, Har- foot M Martin CS, McOwen C, Newbold T, van Bochove J (2014) Towards a global map of natural capital: key ecosystem as- sets. UNEP, Nairobi, Kenya. Kok M, Alkemade R (Eds.) (2014) How Sec- tors can contribute to sustainable use and conservation of biodiversity, CBD Techni- cal Series. Naidoo R, Balmford A, Costanza R, Fisher B, Green RE, Lehner B, Ricketts TH (2008) Global mapping of ecosystem services and conservation priorities. Proceedings of the National Academy of Sciences 105(28): 9495-9500. Petz K, Alkemade R, Bakkenes M, Schulp CJ, van der Velde M, Leemans R (2014) Mapping and modelling trade-offs and synergies between grazing intensity and ecosystem services in rangelands using global-scale datasets and models. Global Environmental Change 29: 223-234. Turner WR, Brandon K, Brooks TM, Costan- za R, Da Fonseca GA, Portela R (2007) Global conservation of biodiversity and ecosystem services. BioScience 57(10): 868-873. Schägner JP, Brander L, Maes J, Hartje V (2013) Mapping ecosystem services’ val- ues: Current practice and future prospects. Ecosystem Services 4: 33-46. Schulp CJE, Alkemade R (2011) Consequenc- es of uncertainty in global-scale land cover maps for mapping ecosystem functions: an analysis of pollination efficiency. Remote Sensing 3: 2057-2075. Schulp CJ, Alkemade R, Klein Goldewi- jk K, Petz K (2012) Mapping ecosystem functions and services in Eastern Europe using global-scale data sets. International Journal of Biodiversity Science, Ecosystem Services & Management 8(1-2): 156-168. Stehfest E, van Vuuren D, Bouwman L, Kram T (2014) Integrated assessment of global environmental change with IMAGE 3.0: Model description and policy applications. Netherlands Environmental Assessment Agency (PBL). Verburg P, van Asselen S, van der Zanden E, Stehfest E (2012) The representation of landscapes in global scale assessments of environmental change. Landscape Ecology 28: 1067-1080. Verburg PH, Neumann K, Nol L (2011) Challenges in using land use and land cov- er data for global change studies. Global Change Biology 1: 974-989. Mapping Ecosystem Services250 5.7.4. Mapping marine and coastal ecosystem services Evangelia G Drakou, Camino Liquete, Nicola Beaumont, Arjen Boon, Markku Viitasalo & Vera Agostini Introduction The marine environment, from the coasts to the open ocean, is closely tied to human well-being; from small-scale artisanal fisher- ies providing local communities with food, to large-scale regulating benefits like pro- tecting coasts from erosion and regulating global climate. Intense human intervention in these areas, for example, through mari- time transport, fishing and aquaculture, oil extraction, tourism and coastal land use, alter these ecosystems, hence impact- ing human well-being. Several treaties and policy instruments have been enacted from the local to global level to regulate human influence on the marine realm and to sus- tain these ecosystems (for example, the UN Convention of the Law of the Sea, the UN High Seas Treaty). In addition, the EU Ma- rine Strategy Framework Directive and that on Maritime Spatial Planning require an ecosystem-based approach to the manage- ment of human activities. Mapping of ES can help decision-makers define critical areas for intervention and aids regulation of activities. Although mapping methodologies are rapidly advancing for the terrestrial and inland water ecosystems, ma- rine and coastal ecosystem service (MCES) mapping is still limited. This chapter gives an overview of MCES mapping principles. We present below the major ES provided by marine and coastal habitats, the particularities and differences of MCES mapping compared to the terres- trial realm and its major requirements and limitations. ES provided by marine and coastal habitat types Each marine or coastal habitat type can gen- erate different ecological functions which can then generate ES for the benefit of hu- man beings. In Table 1, we list the major marine and coastal habitats and the MCES they provide according to what has been documented in the literature. The missing links between habitats and ES highlight the areas with the largest knowledge gaps, but not the lack of a link. It is worth mention- ing here that very few of these ES have been actually mapped. Mapping marine and coastal ecosystem services To map ES provided by marine and coastal ecosystems similarly to the terrestrial eco- systems, one has to understand the pro- cess of ES provision, from the ecosystem components, functions and processes to the actual ES. For each component of the ES provision chain, data need to be ac- quired and quantification methods applied Chapter 5 251 throughout. This information can be used to spatially represent the ES distribution. In Figure 1 we illustrate the process of gen- erating a map of MCES with a hypotheti- cal example. In the oceans and coastal seas, many eco- system functions occur within the water column which adds a third spatial dimen- sion to the system. These functions change with depth, water temperature, solar irra- diance, salinity and other factors and are extremely variable in space and time. This makes it difficult to capture this informa- tion in two-dimensional maps. MCES maps are delivered by: Analysis of primary data, for example, high resolution remote sensing of the coastal and pelagic zone, field sampling and socio-eco- nomic surveys. It can be very accurate, but it is also time and resource consuming. Habitat maps can be used to translate sea- bed habitat maps into capacity to deliver ES based on scoring factors. This method can be feasible and quick if the seabed habitat maps of the study area are already available. However, the scoring system can be subjec- tive and the results reflect only the services provided by benthic habitats. Provisioning Regulating and maintenance Cultural Fo od p ro vi sio n W at er st or ag e / p ro vi sio n Bi ot ic m at er ia ls/ Bi of ue ls W at er p ur ifi ca tio n Ai r q ua lit y re gu la tio n* C oa sta l p ro te ct io n C lim at e re gu la tio n O ce an n ou ris hm en t Li fe c yc le m ai nt en an ce Bi ol og ic al re gu la tio n* Re cr ea tio n To ur ism Sy m bo lic /A es th et ic v al ue s C og ni tiv e eff ec ts Beach and dunes ü ü ü ü ü ü ü ü ü ? ü ü ü Coastal wetland ü ü ü ü ? ü ü ü ü ? ü ü ü Estuary ü ü ü ? ü ü ? ü ? ü ü Mangrove ü ? ü ü ? ü ü ü ü ? ü ü ü Coral Reef ü ? ü ü ? ü ? ü ü ? ü ü ü Maerl bed* ? ? ? ? ? ? ? ? ü ? ? ? ? Oyster reef ü ? ? ü ? ü ? ü ü ? ? ? ? Macroalgal bed ü ? ? ? ? ü ü ? ü ? ? ü ? Seagrass meadow ü ? ? ü ? ü ü ü ü ? ü ? ? Unconsolidated sediments ü ? ? ü ? ? ü ü ü ? ? ? ? Open ocean/ pelagic ü ü ü ü ü ? ü ü ü ? ü ü ü Table 1. Major marine and coastal habitat types and their links with ES as documented in the literature. The (ü) symbol represents the relationships between habitat types and ES that have been assessed and documented in the literature. The (?) is there to represent the lack of sufficient knowledge to assess and hence quantify and map this relationship. * These habitats and ES are still very poorly analysed. Mapping Ecosystem Services252 Modelling Models such as those below can be used: a. Ecosystem models optimally integrated with socio-economic data, or bio-eco- nomic models. They can be relatively accurate with quantifiable uncertainty and capture three-dimensional (3D) processes across spatial scales. Still they require a lot of data, time and expertise. Model outputs may not be usable as such; composites or proxies often need to be generated for MCES mapping. b. Already available MCES mapping tools (see the following section). Most MCES maps depict the ES capacity and very few address the actual flow of, or the de- mand (Chapter 5.1) for MCES. The analysis of all these ES aspects is essential, especially for MCES whose use is often distant from the source of ES provision (e.g. the nutritional value of globally consumed tuna or climate regulation by mangroves in South-East Asia). Required data for MCES mapping The possibility of creating MCES maps is often limited due to scarcity of spatial data. For proper ES mapping, data should ideally be available for: • Habitats’ spatial distribution (or their model-derived proxies); • ecological state of the habitats; • water quality affecting ES provision (e.g. eutrophication or amount of harmful substances); • species distribution of dominant, hab- itat forming and keystone species that either provide or support ES; • biomass of fish and other seafood; • human activities affecting the produc- tion of ES or those which could be used as indicators for ES use (e.g. fishing ac- tivity, tourism etc.). Collecting such data is laborious and expen- sive, mostly because of the methodological Figure 1. The figure depicts the way data and ecological models contribute to the different components of a basic ecosystem service generation framework (ES cascade at the bottom of the figure) in order to generate ES maps. In an example of whale watching tourism as an ES provided by whales, species and habitat distribution models are used to describe the basic ES components. Then models are used to describe the ecosystem functions. The outputs of all these models are then combined along with so- cio-economic parameters (in the example we refer to the number of whale watchers, but it could also be revenues from whale watching) in order to generate a final map of the benefit or value from whale-watch- ing tourism. The arrows show the flow of information within the elements of the ES cascade. Chapter 5 253 challenges. Some examples are given in the following text. Data on benthic habitats need to be collect- ed with echo-sounding methods and tedious geological analysis of the sonar data. Spe- cies data need to be collected with a suite of methods that vary in spatial coverage and taxonomic accuracy. Data on sea bottom substrate and larger species can be collected with underwater cameras, while information on smaller species can be derived with un- derwater surveys (e.g. through scuba diving) and benthic sampling. Species identification often requires microscopic analysis. Some proxies for ES can be created for more cost-effective methods. The new satellite in- struments provide high resolution data (e.g. WorldView3 images have a resolution of 30 cm) that can be used to create proxies for some ES, like habitats essential for fish pro- duction. Semi-automatic in situ mapping devices, such as robot gliders, have been developed for collecting sea bottom data instead of cruises on research vessels. Such methods can complement, but never entire- ly replace, the traditional methods. Spatial data on certain human activities can easily be derived from public databases, but in most cases data are scarce. Proxies need to be calculated although these create uncer- tainties in the mapping. MCES mapping tools Different online tools, models and method- ological frameworks allow practitioners to assess and map different components of the MCES generation chain (Figure 1). Amongst the most popular and well-established ones, are the models from the InVEST1 toolkit that use ecological production functions to assess the supply and demand of MCES. These can assess wave energy, coastal pro- 1 http://www.naturalcapitalproject.org/invest/ tection, marine fish aquaculture, marine aesthetic quality, fisheries and recreation and marine habitat provision. ARIES2 has also been applied for MCES assessment to generate maps mostly in coastal areas, using artificial intelligence networks and expert opinion. In most of these models, data avail- ability and quality are the major issues that make their application difficult. Several initiatives focus on publishing spa- tially explicit information regarding or po- tentially supporting MCES mapping. The SeaAroundUs3 project has released a map server showing time series of the spatial dis- tribution of fisheries around the globe. The EU has recently released a new tool for map- ping fishing activities (MFA)4 for the Euro- pean seas which is based on AIS (Automatic Identification System) data acquired by fish- ing vessels. AquaMaps5 also provide maps of marine species distribution globally. The Baltic Sea data and map service6, by the Hel- sinki Commission, provides spatial data on biodiversity and human activities on sea. The Ocean Health Index Project7 provides a glob- al map of ES provided by the sea and how sustainably the countries are using them. Challenges of MCES mapping There is a high level knowledge pool on the functioning of the marine ecosystems and high expertise on ES mapping methods. Yet these two only recently started converging in an interdisciplinary manner. Hence the number of MCES assessments that actually provide maps is still very limited. Challeng- es to MCES mapping include: 2 http://ariesonline.org/ 3 http://www.seaaroundus.org/data/#/spa tial-catch 4 https://bluehub.jrc.ec.europa.eu/mspPublic/ 5 http://www.aquamaps.org/search.php 6 http://maps.helcom.fi/website/mapservice/in dex.html 7 http://www.oceanhealthindex.org Mapping Ecosystem Services254 • The dynamic three-dimensional (3D) nature of the marine environment, es- pecially in the pelagic zone, makes it difficult to produce two-dimensional maps. Averaging over time and space is necessary and hence the level of spatial accuracy is low. • Information on the distribution of hab- itat is scarce or entirely lacking making it difficult to map MCES based on these habitats. • As the ecological functions and process- es behind many ES, such as biological regulation, are not known or not easily quantified, their mapping is difficult. • Cultural ES, such as recreation, aesthetic information or inspiration, are based on human experiences which may be very variable. Linkage of such experiences to a specific habitat is difficult. • Data on ES demand or use is sensitive thus hard to obtain for some ES with high commercial value (e.g. food provi- sion from fisheries). • Uncertainty in data and maps is too high to be useful in a policy context, therefore having often a negative feedback effect on momentum to create these maps. Future recommendations Given the limited number of MCES maps, there is a need to: • Adapt the current ES methodologies and frameworks that have been developed based on terrestrial ecosystems to the specificities of the marine environment. • Improve the quality and spatial resolu- tion of data and improve data availabil- ity; advance initiatives such as the Eu- ropean Marine Knowledge 2020; and feed data into harmonised databases like the EMODNET8 data portal. 8 http://www.emodnet-biology.eu/ • Adopt a holistic view of the ES provi- sion chain focusing on the intermediate steps (from the ES to the benefit). In particular, the valuation of regulating services and the ecological processes supporting provisioning and cultural services should be reinforced. • Communicate the uncertainties in MCES maps. Explain how much of the spatial detail shown on maps is reliable. Recommend for which purpose the maps can – and cannot – be used. Further reading Böhnke-Henrichs A, Baulcomb C, Koss R, Hussain SS, de Groot RS (2013) Typolo- gy and indicators of ecosystem services for marine spatial planning and management. Journal of Environmental Management 130: 135-145. Boonstra WJ, Ottosen KM, Ferreira ASA, Richter A, Rogers LA, Pedersen MW, Kok- kalis A, Bardarson H, Bonanomi S, Butler W, Diekert FK, Fouzai N, Holma M, Holt RE, Kvile KØ, Malanski E, Macdonald JI, Nieminen E, Romagnoni G, Snickars M, Weigel B, Woods P, Yletyinen J, Whit- tington JD (2015) What are the major global threats and impacts in marine envi- ronments? Investigating the contours of a shared perception among marine scientists from the bottom-up. Marine Policy 60: 197-201. Liquete C, Piroddi C, Drakou EG, Gurney L, Katsanevakis S, Charef A, Egoh B (2013a) Current Status and Future Prospects for the Assessment of Marine and Coastal Ecosystem Services: A Systematic Review. PLoS ONE 8: e67737. Chapter 5 255 Liquete C., Zulian G., Delgado I., Stips A., & Maes J. (2013) Assessment of coastal pro- tection as an ecosystem service in Europe. Ecological Indicators 30: 205-217. Townsend M, Thrush SF, Lohrer AM, Hewitt JE, Lundquist CJ, Carbines M, Felsing M (2014) Overcoming the challenges of data scarcity in mapping marine ecosystem ser- vice potential. Ecosystem Services 8: 44-55. Mapping Ecosystem Services256 5.7.5. Spatial, temporal and thematic interactions Susanne Frank & Christine Fürst The role of spatial and temporal scales in ES mapping; scale dependencies of different (groups of) ES Ecosystem services (ES) are scale-depen- dent. While a single tree can have a positive impact on the micro-climate (local scale), it does not necessarily impact the climate regulation at global scale. Interactions of spatial and temporal scales make the map- ping of ES more complex. Therefore, with- out taking into account the age of a tree and its relations with other trees or other land uses (landscape scale/regional scale), no precise statement on its contribution to climate regulation can be derived across scales. Additionally, scale dependence is re- lated to different perspectives, including the ES provider (supply) and the ES beneficiary (demand), as well as ES assessment (expert) and ES management (stakeholder). In this chapter, we assess and clarify the various as- pects of scale interactions and perspectives in the context of ES mapping. The difficulty of scale interactions lies in many aspects. The mapping exercise as such can be conducted in a straightforward man- ner at many scales; the greater challenge is the data availability. Regarding scales, the assessment of ES at the local scale is some- times easier, because systems are smaller and thus better investigated, understood and supported by data. At the regional scale, spa- tial interactions between various ecosystems make the ES investigation more difficult, as boundary phenomena between ecosystems or land use types have been less investigated. Knowledge and data about the influence of composition and configuration of land use types remains limited (see Chapter 3.6). For assessing ES at continental or global scales (see Chapter 5.7.3), ES data from local, re- gional and national assessments (see Chap- ters 5.7.1 and 5.7.2) need to be up-scaled. Therefore, mapping and assessment of ES at global scale might involve high uncertainty. Additionally, indicators which are used to assess ES, are in many cases scale-sensitive. Furthermore, the beneficiaries as well as the perception of ES benefits change across scales: supply and demand are largely scale-dependent. At local scale, individuals might be directly dependent from provi- sioning services such as food or water (“pri- vate ES”) so that their activities (land man- agement, purchase) are directed towards harmonising the spatio-temporal variation between supply and demand. Global ser- vices, such as global climate regulation, are relevant for humanity (“public services”). Their spatial and temporal dynamics, also called “spatio-temporal lag”, are huge (see Chapter 5.2). Consequently, their percep- tion and appreciation have the character of a shared value which complicates their as- sessment and the application of financial in- struments such as Payments for Ecosystem Services (PES) to boost them (Table 1). In land management and land use planning, the ES concept contributes to the assess- ment of sustainability from a highly inte- grative perspective that covers regulating, Chapter 5 257 provisioning and cultural ES and allows the assessment of the value of biodiversity as a supportive backbone to enable ES supply (see Chapter 2.2). A multitude of indicators (e.g. in the context of CICES 4.31, Chapter 2.4) has been introduced for the different service groups. However, many of them ad- dress a specific scale so that the subsequent assessments require intense data collection, analysis and aggregation. Taking regulating services as an example, “mediation of smell/ noise/visual impacts” relates to local or re- gional scale, while “dilution by atmosphere, freshwater and marine ecosystems” refers to regional, national or even global scale. In this chapter, we explore how to inte- grate data from different scales in a com- prehensive manner. Using the results from the project RegioPower2 as an example, in Boxes 1-3, we show how local data can be up-scaled for supporting decision-making at the regional level. Scale interactions To move from local data to regional decision support, various data need to be collected, harmonised and integrated. Data might en- compass measured data from field studies, empirical data from surveys, modelled data, 1 http://cices.eu/ 2 www.eli-web.com/RegioPower/ or expert judgements if quantitative data is not available. Hence, the first challenge is the identification of adequate indicators. Regarding spatial reference, a cross-scale approach might be necessary, for example, the collection of local data, in order to re- gionalise them for an ES assessment at the regional or national scale (Box 1). Once the status quo of ES is assessed and mapped, the next challenge is the consider- ation of the temporal scale (Box 2). Provision of and demand for ES change during time. If available, historic data should be used as a basis for the development of future land use and management alternatives which should support decision-makers in finding the most sustainable planning strategies. In addition to space and time, thematic in- teractions need to be taken into account to avoid unexpected trade-offs (Box 3). With the term thematic, we refer to thematically heterogeneous ES, for example, provision- ing, regulating and cultural services. Various ES, which are relevant for a specific study in terms of spatial scale and management chal- lenges, should be mapped and assessed. At least, some ES from each category (provision- ing, regulating and cultural services) need to be considered for a reliable analysis of ES synergies and trade-offs. Depending on the case study framework, ES that are relevant for decision-making, should especially be con- sidered. However, neglecting one thematic ES group might lead to unforeseen trade-offs. Table 1. Generalised scheme of the antagonism of ES awareness across spatial scales. Scale Measurement Perception of benefits Beneficiary Payment for ES Local Easy Good Land owner* Feasible Regional National Continental Global Difficult Poor Human kind Difficult *and further local actors and stakeholders Mapping Ecosystem Services258 Box1 . Bridging spatial scales In RegioPower, we focussed on exploring regional biomass provisioning capacities and focus here on the service “timber production”. Measured or modelled data, as well as stakeholder experience or expert opinion can serve as the basis for the assessment of this service. We made use of forest inventory data and regional statistics (harvesting, trade) and included empirical data when no specific information could be obtained. Through normalisation, this quantitative information basis can be adjusted for trade-off analyses with other services, such as “Aesthetics” or “Carbon sequestration”. Subsequently, with the help of the software GISCAME3, the effective capacity of providing services bundles and their balance can be assessed in a spatially explicit manner or as summary information at regional scale. This approach of local data collection and subsequent normalisation for up-scaling to larger scales (Figure 1) can be applied for many ES and for various spatial scales (regional, national or larger). 3 www.giscame.com Figure 1. Upscaling of local stand data (growing stock in m³ per ha, right map) to the regional scale (relative scale from 0-100) using a normalization apporach (left map). Chapter 5 259 Box 2 . Temporal variations in ES provision Temporal variations in ES provision can be included through studies of historical land use change, by monitoring data, or by using models (see Chapter 5.3). In the RegioPower project, changes in growth and yield of forest stands were taken either from yield tables or models. Yield tables are based on long- term field trials that describe different forest management models and their impacts on stand properties. Though mapped information on the current timber provisioning capacities is helpful, more in-depth analysis of stand dynamics starting from current properties (age, stand density, tree species composition) provides valuable information on future capacities or limitations in the availability of timber and thus helps to adjust regional development strategies or investments (e.g. in power plants or saw mills). The example in Figure 2 shows the development of the regionally harvestable volume over time as a response to current forest management models considering rotation periods, harvesting, recreation and tending (business as usual). It reveals that reducing the assessment on the currently available timber would underestimate the amount of harvestable timber in the near future, while it would neglect the risk of an undersupply in the longer term. Changes in forest management, such as forest conversion, but also external impacts, such as climate change, would alter the harvestable volume. Consequently, such long-term analyses of the variability in ES supply need to be interpreted cautiously as they include high uncertainties. Even or especially the communication of the degree of uncertainty is highly valuable information in the context of deci- sion-support for spatial planning. Furthermore, ecological, biophysical and social/legal parameters influence the regional availability of ES such as timber production. We included information on the type and status of ecosystems to calculate the natural capacities of each land use type to contribute to the supply of ES such as timber. Topographical data (slope) were considered as limiting factors in the accessibility of forest resources due to technical lim- itations in harvesting, so that areas with steep slopes were counted with a lower potential for timber sup- ply. Additionally, information on ownership types (state, communal, private forests) and their particular mobilisation rates were used to adjust the potentially harvestable volume. The mobilisation rate in private forests is, for instance, only 60 % of the harvestable volume. Finally, forests in national parks and nature protection areas were calculated with only 10 % of the potentially harvestable volume. Figure 2. Temporal variation of the provisioning service "timber provision" considering three timber assortments. !" #$%&'())* +$&,'-.*',-,%/0).12/$34).'())* 5%-/1 6 7789:8 :;9<=: 8989= > 7;=9>; :8?6>6 >:87; :6 78<<<6 :>7=9> >;<79 :> 7>9;>: :>=8=> >>8>= 76 7??>:> :?6>>8 >?898 7> 7=7989 :?:<=9 >?<8> >6 7=877> :8967; >:>:6 => 7;:6:: :88:87 >6<=7 :66 :?778> ::7<:7 86?=6 :7> :98>66 ::;7== 8696 78>=7= :;6?<8 8?97? :=> 7=669> :8<7<: >6?;< 766 7>> 15 % tree cover is considered a forest in GLC2000 while in CLC a 30 % threshold is used to distinguish forests among other land cover types. Some of the land cover maps include a few details on land use by, for example, distinguishing pastures from natural grasslands. Thus, the choice of a specific land cover map for map- ping an ecosystem service to a certain extent defines the output. Model definition A key parameter for pollination services is the distance between a pollinator habitat (nesting site) and the crop which needs pol- lination. For calculating the distance to pol- linator habitat using a land cover map, one should decide whether each land cover type provides habitat or not. This introduces new Category Definition Landscape-based indicators Capacity of the landscape to support pollinator communities Percentage area of potential pollinator habitat Distance to pollinator habitat Probability that a location is visited by pollinators Species-based indicators Abundance of pollinators Abundance of specific pollinator species Species richness of pollinators Crop-based indicators Yield quantity Financial benefits of yield of pollinator-dependent crops Percentage yield loss upon absence of pollinators Table 1. Overview of indicators for the ecosystem service “pollination”. Mapping Ecosystem Services284 uncertainties. While for many individual pollinator species, habitat requirements are known, these often do not match the level of detail displayed in land cover maps. One can only assume that the specific vegetation type and structure or host plant which a pol- linator community requires, is present in a land cover type that is only described with a level of detail like “natural grassland” or “deciduous forest”. Secondly, a single land cover type can include both suitable and unsuitable areas, dependent on the types of vegetation included in the land cover or de- pendent on the management. Based on the knowledge of pollinator behaviour and of the land cover map, the map maker has to distinguish between presence or absence of a habitat. Depending on background knowl- edge and the exact nature of the uncertain- ties described above, different map makers can decide differently in similar situations. This introduces a new uncertainty. More uncertainty arises when calculating the distance to habitat, depending on the resolution of the data. Each pixel gets as- signed one distance value which, especially when using coarse resolution data, can devi- ate. In addition, land cover data often fail to properly represent small landscape elements such as tree lines, hedgerows or individual trees. An alternative approach that can help overcome this is, for example, using a map of the density of such elements. But deriving distance to habitat from a density map also introduces a new uncertainty. Relations between distance to habitat and the effectiveness of animal pollination have been previously established. Close to natural habitat, bee abundance and species richness tend to be high while richness and abun- dance decrease upon increasing distance to habitat. While this general principle is well known, in different situations the decrease of pollinator abundance or theoretical visi- tation rates can be hugely different. Figure 1 demonstrates the overall relation between distance to habitat, the probability that a lo- cation is visited by pollinators in temperate regions and also shows the uncertainty in these estimates. The impacts of input data uncertainty on the output uncertainty are demonstrated in Figure 2. Figure 2 compares four dif- ferent maps of the distance to nature (left) and the percentage yield reduction based on distance to nature (right) against a base map. The base map is a high-resolution land cover map specifically for the Netherlands, while the other maps are maps covering a larger area and with a coarser resolution. The left map demonstrates that mapping of distance to nature strongly depends on the resolution of the input map: in most of the area of the Netherlands, none or one of the maps properly represents the distance to nature as defined by the base map. The right map demonstrates that the yield loss as a function of distance to nature shows more similarity with the base map. This is because, upon large distances to nature, the yield reduction levels off, making deviations between different maps less important. Figure 3 provides a comprehensive overview of the impact of indicator choice, input data and the model to quantify the ecosystem 1.00 0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.00 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 0.10 Figure 1. Relation between distance to pollinator habitat and probability that a location is visited by pollinators. The black line is an average value for temperate regions while the uncertainty range is indicated in grey. Based on (Ricketts et al. 2008). Chapter 6 285 service on the final ecosystem service map. The figure compares four different maps of the ecosystem service pollination, mapped at European scale using four different defi- nitions of the service, slightly different but mostly overlapping input data and different methods to quantify the indicator. Figure 3 presents evidence that the four different out- comes disagree on the relative provision of the service (purple areas) for about one third of the study area (EU). In ca. 30 % of the EU territory, there is some agreement on high values of pollination provision (green areas) while, in just over 20 % of the EU territo- ry, there is some agreement on low values of pollination provision (blue areas). A similar type of analysis for the ES climate regulation, flood regulation, recreation and erosion pre- vention revealed comparable patterns. The level of (dis)agreement between different Figure 2. Impact of input data on mapping ecosystem structure and function. For distance to nature (left) and percentage yield reduction as a function of distance to nature, maps based on four EU / global scale maps were compared with a detailed reference map (based on Schulp and Alkemade 2011). Figure 3. Agreement between different maps of the ecosystem service “pollination” (from Schulp et al. 2014). Number of maps that are similar to the base map 0 4 coldspots 1 hotspot 1 3 coldspots 2 hotspot 2 2 coldspots 3 hotspot 3 1 coldspots 4 hotspot 4 No extreme values Disagreement about hotspot / coldspot Mapping Ecosystem Services286 maps of the same service depends on the level of understanding of the particular service and on the range of input data used. Dealing with uncertainties As ecosystem service mapping will always involve uncertainties, it is important to deal with these uncertainties in the best possible way. Dealing with uncertainties in ecosystem service maps means (1) improve methods for ecosystem service maps so as to reduce un- certainties to the largest extent possible; (2) quantify and communicate uncertainties and (3) account for uncertainties when using eco- system service maps in policy and practice. Improving measurements Firstly, for several ES, there is a lack of clarity about how to define the service, a lack of pro- cess understanding and a limited measuring accuracy. In all of these three sources of un- certainty, there is scope for improvement. For each case of mapping, it should be carefully decided how a service can be best quantified. Furthermore, for several ecosystem service models, the underlying measurements can be expanded and better stratified. Process understanding for some services needs to be better underpinned by field studies. Quantifying uncertainties Regardless of the scope for improvement of ecosystem service models, it is important to realise that uncertainties in ecosystem service maps cannot be completely ruled out. Sen- sors will never be 100 % accurate and the provision of ES is a complex and multifac- eted process where multiple datasets have to be combined, always involving some kind of expert judgement. It is, therefore, important to be transparent on uncertainties in ecosys- tem service maps. If ES are mapped using complex process-based models (Tier 3 ap- proaches; see Chapter 5.6.1), an uncertainty analysis or a Monte Carlo approach can be used. In a Monte Carlo approach, the indi- cator is calculated several thousand times. Each time, actual input values for calculation are drawn from a probability distribution of each input, resulting in different but realistic representations of the indicator. From these different representations, an average value of ecosystem service provision can be calculated, as well as indicators that quantify the uncer- tainty, such as a probability range, a standard deviation, or a probability that a specific tar- get or threshold value is met or not. A simpler uncertainty analysis includes making an inventory of the range of each in- put. Next, for each input, one should identi- fy if it increases or decreases provision of the service. Finally, the ecosystem service map should be calculated with the combination of inputs that provides a minimum, a max- imum and an average indicator value. This provides the possible range of the indicator. For methods that completely rely on expert judgement (Tier 1 approaches; see Chapter 5.6.1), it is important not to rely on a single expert, but instead to take stock of a wider range of expert knowledge in the field. Rat- ings by different experts on the capacity of the landscape to supply ES can, for example, be translated into a measure for the “agree- ment” of different experts and, with that, provide an indicator for the uncertainty. Intermediate approaches that combine expert knowledge with additional data or simplify process-based models (Tier 2 approaches; see Chapter 5.6.1) can use an intermediate approach for uncertainty quantification as well. Bayesian Belief Networks, as discussed in chapter 4.5, are typical examples of models that can account for a broad range of uncer- tainty types and can assess the effects of these uncertainties on model outputs. Chapter 6 287 Dealing with uncertain maps At the same time, scientists and policy mak- ers use maps of ES on which to base analyses and decisions. Such decisions should be ro- bust, meaning that they should not work out differently from what they are supposed to, because of the uncertainties in the maps. To ensure that uncertainties in ecosystem service maps do not impede decision-making, they must be quantified and communicated by map makers. Here, an uncertainty analysis, as described above, is essential and clear re- porting of uncertainties is compulsory. On the other hand, policy makers and oth- er users of ecosystem service maps should account for the level of certainty in their decision-making. To do so, a dialogue be- tween policy makers and mappers is essential (Chapter 6.4) to ensure that the indicator mapped actually reflects the request by poli- cy makers, given that the indicator is a major source of uncertainty. Users of ecosystem ser- vice maps should also be careful when mak- ing planning decisions based on ecosystem service maps. Also, for map users, it might be important to take stock of the broad knowledge in the field rather than relying on a single map upon decision-making. Finally, policy makers should be cautious upon using ecosystem service maps for decision-making. Further reading Grêt-Regamey A, Brunner SH, Altwegg J, Bebi P (2013) Facing uncertainty in eco- system services-based resource manage- ment. Journal of Environmental Manage- ment 127: 145-154. Jacobs S, Burkhard B, Van Daele T, Staes J, Schneiders A (2015) The Matrix Reload- ed: A review of expert knowledge use for mapping ecosystem services. Ecological Modelling 295: 21-30. Schulp CJE, Alkemade R (2011) Conse- quences of Uncertainty in Global-Scale Land Cover Maps for Mapping Ecosystem Functions: An Analysis of Pollination Effi- ciency. Remote Sensing 3: 2057-2075. Schulp CJE, Burkhard B, Maes J, Van Vliet J, Verburg PH (2014) Uncertainties in Ecosystem Service Maps: A Comparison on the European Scale. PLoS ONE 9: e109643. Schulp CJE, Lautenbach S, Verburg PH (2014) Quantifying and mapping ecosys- tem services: Demand and supply of polli- nation in the European Union. Ecological Indicators 36: 131-141. Stoll S, Frenzel M, Burkhard B, Adamescu M, Augustaitis A, Baeßler C, Bonet FJ, Car- ranza ML, Cazacu C, Cosor GL, Díaz-Del- gado R, Grandin U, Haase P, Hämäläinen H, Loke R, Müller J, Stanisci A, Staszewski T, Müller F (2015) Assessment of ecosys- tem integrity and service gradients across Europe using the LTER Europe network. Ecological Modelling 295: 75-87. Ricketts TH, Regetz J, Steffan-Dewenter I, Cunningham SA, Kremen C, Bogdanski AK, Gemmill-Herren B, Greenleaf SS, Klein AM, Mayfield MM, Morandin LA, Ochieng A, Viana BF (2008) Landscape effects on crop pollination services: are there general patterns?, Ecology Letters 11(5): 499-515. Mapping Ecosystem Services288 6.4. Map interpretation/end-user issues Christian Albert, Claire Brown & Benjamin Burkhard Introduction Maps are very powerful tools to communi- cate complex geographic information from the map-maker (the cartographer) to the end-user (such as a decision-maker). As in all communications, there are information losses and/or modifications during the trans- mission from the sender to the receiver. Eco- system service maps are a specific case due to their high thematic complexity, adding fur- ther potential for (mis)interpretation of the intended messages. It is therefore essential that the end-users not only have access to the map, but are also aware of any interpretation issues such as the categorisation of ecosystem services (ES) used (Chapter 2.4), choice of spatial scales (Chapter 5.7) and uncertainties or scientific errors (Chapter 6). Map communication model Maps can usefully be understood as a form of visual communication to describe spatial phenomena and their relationships. In early understandings, building, for example, on the transmission model of communication put forward by Shannon and Weaver in 1949, maps were seen as media for spatial information sent from the mapper to the map-users (Fig. 1). Since then, it has be- come increasingly apparent that many issues complicate the function of maps as com- munication devices including, for example, the technical question of how accurately information is actually transmitted, the se- mantic problem of how well the meaning of the map is conveyed (Chapter 3.3), the in- terpretation problem of maps by map-users and problems of power relations. Specifics of ecosystem service maps Ecosystem service maps are complex as they reflect the level of difficulty we find ourselves in managing the environment and ensuring equitable benefits across society. The science underpinning the actual mapping of ES is still unresolved despite recent advances. The issue of mapping ES is further compounded by the need to bring together and display the supply of a service and its demand (in- cluding flow; see Chapter 5.1) using envi- ronmental, economic and societal factors, a difficult endeavour when working in con- ventional two-dimensional space (Chapter Figure 1. A simple map communication model (adapted from Dent 1985). Feedback Transmission I Transmission II Map Coding Language, Symbols, Legend Map Making, including information losses Map Use, including interpretation issues Map UserMap Author Chapter 6 289 3.7). With the concepts of ES and natural capital rising high up the political agenda through processes such as the Intergovern- mental Platform for Biodiversity and Eco- system Services (IPBES), CBD Aichi Target 14, the UN SDGs or the EU’s Biodiversity Strategy, there is a need to ensure ecosystem service mapping is scientifically robust and easily understood (Chapter 7.1). Particular challenges surrounding the com- munication of ES maps to end-users include: • The existence of diverging categorisa- tions (Chapter 2.4) and conceptuali- sations of ES (e.g. as potentials, flows or benefits of ecosystems; Chapter 5.1) requires clearly specifying the exact meaning of what is being illustrated on the maps. End-users may be aware of different categorisations and conceptu- alisations of ES but will not be aware of the associated interpretation issues. • The possibility of spatial misfit between the areas that supply ES and the areas in which the benefits are consumed (Chapter 5.2). Communicating the choice of spatial scale or the mismatch between supply and beneficiaries can conceptually be difficult to understand. • The complicated spatial overlap of the provisioning and/or benefitting areas concerning several ES at the same site. Communicating this overlap spatially on maps can build upon, for example, hotspot and cold-spot analyses (Chap- ter 5.7). • The difficulty to communicate the un- certainties inherent in the delineation, quantification and evaluation of ES provision, supply and benefits despite the connotation conveyed by maps as authoritative spatial information (Chapter 6.3). Ecosystem service map-makers In recent years, ecosystem service mapping has gained prominence as a scientific field and as an output from research. It has permitted different disciplines (such as ecology, econom- ics and social sciences) to analyse different types of information together, often resulting in highly complex and specialised maps. With the advent of desktop Geographic In- formation Systems (GIS) and a number of ecosystem service mapping tools (see Chap- ter 3.4), creating maps has become easier and seemingly without the requirement of having specific cartography training. How- ever, the ease at which ecosystem service maps nowadays can be created needs to be balanced with the danger of creating a badly designed map. Maps that are not well designed or lack cartographic logic (Chap- ter 3) increase the risk of misinterpretation by decision-makers or even the deliberate abuse of ecosystem service information in non-sustainable environmental resource management. Therefore, (at least basic) car- tography training and knowledge are neces- sary in order to avoid typical technical and thematic pitfalls of map-making. Ecosystem service map end- users The end-users of ecosystem service mapping products vary in nature and in their purpose for wanting a map. End-users could be: • Decision-makers working at different scales who wish to make a specific land- use decision such as approval for a dam, road or land use change (e.g. forest to agriculture; see Chapter 7). The types of questions which are asked are highlight- ed in Table 1; Mapping Ecosystem Services290 • A decision-maker or NGO wanting to engage a group of stakeholders or the public in a specific issue such as demonstrating their links and benefits obtained from a particular site; • A practitioner synthesising information to present to a decision maker around a specific issue; • Policy makers on different levels want- ing to illustrate progress towards certain policy goals; • The scientific community, students and teachers. Although all these user groups may find eco- system service maps useful, once again the risk of misinterpretation is high. Many of the individuals involved will often not have a specific cartography education needed to ‘read’ and understand a map. Moreover, the map may be only one piece of evidence that their decision-making is based on. Sources of uncertainty in map interpretation Even if the best data, best model or best avail- able methods have been used by a very skilled map-maker, the applicability of a map can be hampered by limited map reading/interpreta- tion skills of the actual map end-user. Maps are generalising models of reality (Chapter 3.2) with inherent uncertainties related to all steps of map production. Lack of expert knowledge concerning ecosystem service supply and demand schemes can cause map misinterpretations. Much of the information included in ES maps is very complex and the information is highly aggregated to be di- rectly used in practical applications. On the other hand, even a highly trained map-user with comprehensive expert knowledge can- not overcome weaknesses in data and map compilation. In particular, the challenges can be sum- marised as providing maps: Policy questions Policy & research actions What are the status and trends of the EU’s ecosystems and the services they provide to society? Biophysical mapping of ES using data and models.What are the drivers causing changes in the EU’s ecosystems and their services? How might ecosystems and their services change in the EU under plausible future scenarios? Mapping and valuation of ES as part of an integrated and stakeholder-based approach to sustainable land management and use of natural resources. How can we secure and improve the continued delivery of ES? Can we set priorities for ecosystem restoration within a strategic framework at sub-national, national and EU level? Can we define where to strategically deploy green infrastructure in the EU in urban and rural areas to improve ecosystem resilience and habitat connectivity and to enhance the delivery of ES at member state and sub-national level? How can we foster synergies between existing and planned initiatives at local, regional or national levels in member states, as well as how to promote further investments, thereby providing added value to member states action? Table 1. Example of policy questions from the EU that ecosystem service mapping might address (adapted from Maes et al. 2012). Chapter 6 291 Policy questions Policy & research actions What are the status and trends of the EU’s ecosystems and the services they provide to society? Biophysical mapping of ES using data and models.What are the drivers causing changes in the EU’s ecosystems and their services? How might ecosystems and their services change in the EU under plausible future scenarios? Mapping and valuation of ES as part of an integrated and stakeholder-based approach to sustainable land management and use of natural resources. How can we secure and improve the continued delivery of ES? Can we set priorities for ecosystem restoration within a strategic framework at sub-national, national and EU level? Can we define where to strategically deploy green infrastructure in the EU in urban and rural areas to improve ecosystem resilience and habitat connectivity and to enhance the delivery of ES at member state and sub-national level? How can we foster synergies between existing and planned initiatives at local, regional or national levels in member states, as well as how to promote further investments, thereby providing added value to member states action? • at the scale appropriate for planning and management, • at the right point in time to make in- formed decisions, • in an accessible manner and • in communication formats appropriate for diverse user groups. Solutions In recent years there has been a call from the end-user community for more scientif- ic outputs to be policy-relevant and for the co-development of outputs. However, poli- cy-relevance and scientific integrity need to be balanced accordingly. Therefore the ques- tion is: how do you create a fit-for-purpose map that is scientifically robust? The first solution is to improve the communication between the map-maker and the end-user. Secondly, there is a need to improve the transparency of how the map was created and the uncertainties embedded in the map. Lastly, the reproducibility and the compre- hension of the results need to be improved (e.g. better maps produced). Engaging the end-user before the map is developed will allow the map-maker to un- derstand how the map is going to be used, i.e. what question will the map be used to answer (Chapters 4.6 and 5.4)? The map- maker can then use this information to de- termine the degree of precision required, as it is not always necessary to use the high- est data resolution with the most complex methods (Chapter 5.6.1). Often, simpler easy-to-comprehend approaches (Chapters 4.6 and 5.6.4) may deliver results that are easier to communicate. The scientific community is also required to continuously improve the methods that are used to quantify, measure, monitor, model, map and value ES. These methods should not only be accepted by their peers but also be communicated clearly to those practitioners who are frequently generating maps or interpreting maps for different de- cision-making contexts. While the most desired outcome would be to have the user-community trained to un- derstand spatial information generally and, more specifically, the interpretation of eco- system service maps, this is not really feasible due to the high resources required. However, the user-community’s capacity can be contin- ually enhanced over the long-term through dialogues with scientists and practitioners. Further reading Dent Borden D (1985) Principles of Thematic Map Design. Addison-Wesley Publishing. Reading, Mass. Hou Y, Burkhard B, Müller F (2013) Uncer- tainties in landscape analysis and ecosys- tem service assessment. Journal of Envi- ronmental Management 127: 117-131. Maes J, Egoh B, Willemen L, Liquete C, Vi- hervaara P, Schägner JP, Grissetti B, Drak- ou EG, La Notte A, Zulian G, Bouraoui F, Paracchini ML, Braat L, Bidoglio G (2012) Mapping ecosystem services for policy support and decision-making in the European Union. Ecosystem Services 1: 31-39. Monmonier M (1996) How to lie with maps. 2nd ed. The University of Chicago Press. Muehrcke PC (2005) Map Use: Reading, Analysis, and Interpretation. 5th ed. J P Pubns. Wood D, Fels J, Krygier J (2010) Rethinking the Power of Maps. Guilford Pubn. Chapter 7 293 CHAPTER 7 Application of ecosystem services maps Mapping Ecosystem Services294 Environmental restoration planning is one practical application where ecosystem services map are needed (Photo: Benjamin Burkhard 2008). Chapter 7 295 7.1. Mapping ecosystem services in national and supra-national policy making Joachim Maes, Benis Egoh, Jianxiao Qiu, Anna-Stiina Heiskanen, Neville D. Crossman & Anne Neale Introduction Despite the global efforts taken to conserve biodiversity it was clear in 2010 that the global ‘‘2010 target’’ of preventing the loss of biodiversity had not been met. The Mil- lennium Ecosystem Assessment, the various subsequent sub-global assessments and The Economics of Ecosystems and Biodiversity study have increased awareness of the neg- ative impacts of biodiversity loss on human welfare by addressing the value of ecosystems and biodiversity for sustaining livelihoods, economies and human wellbeing. Failing to incorporate the values of ecosystem services (ES) and biodiversity into economic deci- sion-making has resulted in investments and activities that degrade natural capital. In 2010, the tenth meeting of the Confer- ence of Parties (COP 10) to the Conven- tion on Biological Diversity (CBD) led to the adoption of a global Strategic Plan for biodiversity for the period 2011–2020. The ‘‘2020 Aichi targets’’ complement the previ- ous conservation-based biodiversity targets with the addition of ES. Anticipating the COP10, the European Union (EU) adopted a communication on “Options for an EU vision and target for biodiversity beyond 2010”. For the first time, explicit reference was made to the practice of mapping ES in a high level policy document. Maps of ES were expected to help define the scope of the maintenance and restoration ef- forts needed to achieve the new biodiversity targets. Eventually, the mapping of ES was retained in the EU Biodiversity Strategy to 2020 as one of 20 actions to be implemented by the EU member states. Well before 2010, South Africa had already pioneered ES research including mapping to support policy on biodiversity, restoration and poverty reduction. Thus, this chapter will start with the achievements in that coun- try to illustrate how mapping can contribute to policy support or, vice versa, how mapping entered into various policies. In later sections, developments on mapping for policy in other parts of the world will be presented. Mainstreaming ecosystem ser- vices into policy: South Africa When the concept of ES came into the lime- light in the mid to late 1990s, South Africa was one of the first countries to embrace it. In 1995, South African scientists carried out a ground-breaking study showing that invasive alien plants had a negative impact on water supply. The results were communicated to the then Minister of Water Affairs (Mr Kader Asmal) who later established a very success- ful Working for Water (WfW) programme Mapping Ecosystem Services296 aimed at removing invasive alien plants to improve water quantity in rivers, conserve biodiversity and provide jobs for local peo- ple. The WfW programme was so popular, its budget grew from $5M to about $50M and created about 35,000 jobs in just 2 years. This success has inspired other programmes such as “working for wetlands”. This example shows that the concept of ES can be a very powerful tool in developing policies that pro- mote sustainable land use and improving the livelihoods for poor people. South African scientists have written many influential papers on the mainstreaming of ES into policy, most of them inspired through their experience in the implemen- tation of biodiversity plans in their country. These lessons were incorporated within a new grassland initiative1 led by the South African National Biodiversity Institute in Pretoria (SANBI). As an implementation strategy within the programme, stakeholders, such as mining companies and the agricultural sec- tor, were brought in as partners in order to help them understand the value of ES in their business, how they can practise sustainable land use and minimise cost. The grassland programme was a huge success as stakehold- ers were able to directly see the benefits of conservation through the lens of ES. Since the grassland programme, much prog- ress has been made in integrating ES into pol- icy and practice. In 2013, the Department of Environmental Affairs (DEA) in South Africa set up the GREEN FUND (GF) to support green economy initiatives. As examples, this GF has supported the service of climate regu- lation through low-carbon initiatives such as the planting of trees in Durban and a study of the importance of ecological infrastructure in delivering ES. Scientists in South Africa are investigating the use of ES as a key entry point into developing the Strategic Environ- mental Assessment (SEA) for Thekwini mu- 1 www.graslands.org nicipality. The SEA is a key policy instrument in guiding development plans for the city of Durban. ES have direct links to the well-be- ing of people living in the city and are attrac- tive to policy makers. These examples show that ES are being integrated into the national, regional and local policy and practice. Mapping and Assessment of Ecosystems and their Services in the European Union (MAES) - A dedicated action of the EU Biodiversity Strategy The mapping and assessment of ES is an es- sential part of the EU Biodiversity Strategy to 2020 and a necessary condition in mak- ing ES key parameters for informing about planning and development processes and de- cisions. In particular, Action 5 of the Strategy requires member states, with the assistance of the European Commission, to map and as- sess the state of ES in their national territory by 2014, assess the economic value of such services and promote the integration of these values into accounting and reporting systems at EU and national level by 2020. The European working group on Mapping and Assessment of Ecosystems and their Ser- vices (MAES), which includes experts of the European Commission, the member states and the research community, has been instru- mental in providing an analytical framework, a typology of ecosystems and ES and a first set of indicators for mapping and assessment. Im- portantly, the EU supports dedicated research under its framework programme for research (Horizon 2020) to support the member states of the EU with the implementation of this policy. The project ESMERALDA2, for ex- ample, provides detailed guidance to various stakeholders for mapping and assessing ES. 2 www.esmeralda-project.eu Chapter 7 297 The work being carried out on the mapping and assessment of ecosystems and ES is not only important for the advancement of bio- diversity objectives, including the develop- ment of Europe’s green infrastructure, but also to provide information for the develop- ment and implementation of related policies on water, climate, agriculture, forest and re- gional planning. Box 1 presents a special case on how the MAES initiative could profit from ongoing assessments in the frame of the EU’s ma- rine policy. BOX 1 . Mapping and Assessment of Ecosystems and their Services (MAES) in Europe’s seas and oceans Under the present EU regulatory frameworks and the pressure to foster sustainable Blue Growth (COM(2012)494) in the marine regions of the EU, it is necessary to undertake more accurate, policy-driv- en research able to map marine ES. Competing uses of marine resources need to be analysed from a holistic perspective to enable achievement of the environmental goals and socio-economic needs that are often competing. ES maps are needed to provide information about the supply and demand of essential services in different coastal and marine regions. These services can be used by different sectors (such as fisheries or tourism and recreation) and supplied in variable scales: commercial fish are catches derived from large marine areas, while recreation destinations such as scenic and pristine beaches can be spatially quite re- stricted. Therefore, maps showing the marine hotspots of ES can be very useful for the EU Marine Spatial Planning Directive (MSPD; Directive 2014/89/EU) and should be disseminated to decision-makers, wider key stakeholders and the general public for both use and validation. Mapping of marine ES is a prerequisite for assessing ES and hence, for preparing environmentally and societal-relevant plans for usage of marine resources, i.e. maritime spatial plans. In the same manner, ES valuation can be used for estimation of the benefits of the EU Marine Strategy Framework Directive (MSFD; Directive 2008/56/EC) programme of measures when the target “good environmental status” is reached. Therefore the economic assessment that is a part of the MSFD assessment (in article 8 of the Directive 2008/56/EC) can directly provide informa- tion for the EU Biodiversity Strategy Action 5, should the ES approach be used in the assessment. The MAES framework consists of 4 different process steps: 1) mapping the ecosystems, 2) assessment of the conditions of ecosystems, 3) assessing ES and 4) integrated assessment based on these three com- ponents (Figure 1). The MAES process can potentially use information from the assessment processes carried out as part of the implementation of the MSFD, the Water Framework Directive (WFD; Directive 2000/60/EC) the MSPD and the Habitats Directive (HD; Directive 92/43/EEC). In Figure 1, a general overview of the linkages between the MAES framework, the MSFD, the MSPD and also the WFD and HD processes is presented. There is a win-win situation for the EU member states, if the data is collected diligently and subsequently used in assessment and reporting for all these directives as well as the MAES process. Here the principle “measure only once and report for several purposes” could be a gold mine for simplifying the reporting procedures of member states. The current EU directives that govern the use and protection of marine environment, namely MSFD and WFD, together cover all marine waters (including transitional waters). MSFD, WFD and HD include assessment of ecological status and pressures and impacts that will provide information for the MAES process step 2 ‘assess the conditions of ecosystems’. MSFD and HD also provide data and information on the distribution of species and habitats for process step 1: mapping the ecosystems. MSPD can potentially provide data and information to assess the use of marine space and to derive indicators on demand of the ES for process step 3: assessing the ES. However, the data flow from the directives’ reporting might still not be sufficient and additional environmental and socio-economic data could be needed to assess the supply of ES and to provide information for the MAES process step 4 ‘integrated ecosystem assessment’. Mapping Ecosystem Services298 China: A unique opportunity to mainstream ES into policies of a rising country economy China, as the world’s most populous nation and amongst the largest in geographic ex- tent, is endowed with immense reserves of biological resources, natural capital and the supply of ES. However, escalating anthro- pogenic pressures, including growing pop- ulation, rapid economic development and ineffective governance, have led to substan- tial degradation and loss of a wide range of ES from local to national scales with mas- sive impacts on human welfare. Such ef- fects can sometimes even be ramified into natural disasters (e.g. devastating flooding, drought and sandstorms) and cascade with global implications through globalisation, international trade, pollution and resource exploration. The increasing public and government aware- ness of environmental problems has triggered a series of large-scale and pervasive nation- al policies to protect natural resources and safeguard the sustainability of ES. Amongst the most prominent policies are the Natural Forest Conservation Programme (NFCP), Grain-to-Green Programme (GTGP), Nat- ural Reserve System (NRS) and Forest and Grassland Eco-Compensation Programmes. Most of these policies, such as NFCP and Figure 1. Overview of the linkages between the Mapping and Assessment of Ecosystems and their Services (MAES) framework and the EU directives that govern environmental status and biodiver- sity in marine and coastal areas: MSFD, WFD, the HD and the MSPD (see text for the explanations for the abbreviations). The linkages illustrate how the information and data from the assessments’ components (from the implementation process of these directives) can feed into the MAES process and its modules 1-3. Optimally, such environmental and socio-economical flow of data could allow the use of the same information in multiple reporting purposes, if undertaken diligently. MAES assessment framework modules Module 4: Integrated ecosystem assessment MAES assessment framework modules Information and data needed for MAES in marine and coastal waters Assessment components included in the EU directives Module 1: Mapping of ecosystems Module 2: Assess ecosystem condition Module 3: Assess ecosystem services HD Distribution and conservation status of species and habitats WFD Coastal and transitional waters Ecological status Pressures and impacts Distribution and status of marine and coastal ecosystem (species, habitats) Human pressures on marine and coastal ecosystems Data and indicators for demand of ecosystem services Human activites and use of marine and coastal ecosystems MSFD (sall marine waters) State of marine ecosystems Socio-economic assessment of human activities and pressures MSPD Human activities and their locations in the sea Chapter 7 299 GTGP, are reliant on the scheme of pay- ment for ecosystem services (PES), in which subsidies and compensation are provided to participants as incentives to promote conser- vation and safeguard ES. Specifically, NFCP conserves and restores natural forests through logging bans and afforestation with incentives to forest enterprises, whereas GTGP converts erodible and steep croplands to forest and grasslands through offering grain and cash subsidies to farmers. These programmes are thus, by far, two of the largest programmes in both China and worldwide in terms of scale (i.e. altogether encompassing 97% of China’s counties), amounts of payment (i.e. investment exceeding 700 billion yuan at $1 = 6.6 yuan as of 2016) and duration of ef- fectiveness (i.e. ca. 20 years and continuing). NRS, on the other hand, is a series of action and policies, primarily regulatory, to restrict economic development and prohibit regular human activities (e.g. gathering, poaching) in designated reserve areas in order to protect all forms of biological diversity that underlies the provision of ES. This NRS effort has been in place for many decades and has resulted in the establishment of 319 nature reserves across China covering ca. 93 million ha. Research on mapping and assessing ES over the past several decades has played a criti- cal role in supporting these policy efforts in multi-faceted ways. First, it provides the scientific basis for valuation of ES and the foundation on which the PES-related pol- icies were implemented (e.g. calculation of subsidies or compensation). Secondly, monitoring and quantifying changes, in particular long-term changes in ES through mapping, can adequately assess effectiveness and support the continued implementation of these policy efforts. The provision of tan- gible effects of these policies on natural cap- ital and the provision of ES can help raise public, economic and institutional support for future policy implementation. Thirdly, most of current policy efforts are not holistic and tend to be piecemeal or system-specif- ic (e.g. forest- or grassland-centred). Map- ping and incorporation of multiple ES and their complex interrelationships call for the need to consider multiple services and also encourages future policies to broaden their scope through coordinated management which potentially could improve the effec- tiveness and efficiency. Last but not least, comprehensive monitoring and assessment of ES can help provide timely feedback for adjusting and refining these programmes, helping to identify current gaps and provide information about areas where future policy efforts and funding need to be prioritised. The United States: Growing evidence of a commitment to consideration of ecosystem services in decision-making In October 2015, the US White House Council on Environmental Quality issued a landmark Executive Office Memorandum to all US Federal government agencies calling on them to incorporate ES into federal plan- ning and decision-making. The memoran- dum “directs agencies to develop and institu- tionalize policies that promote consideration of ecosystem services, where appropriate and practicable, in planning, investment and reg- ulatory contexts”. It establishes a process for the Federal government to develop a more detailed guidance on integrating ES assess- ments into relevant programmes and aims to help maintain ecosystem and community resilience. It also required Federal agencies to develop work-plans describing how their current and future efforts will meet the re- quirements of this new policy. Leading up to the 2015 Executive Mem- orandum in July 2011, the US President’s Council of Advisors on Science and Tech- Mapping Ecosystem Services300 nology (PCAST) published a list of recom- mendations to President Obama in the Re- port on Sustaining Environmental Capital: Protecting Society and the Economy. This report was developed as a sequel to the 1998 PCAST report to President Clinton entitled “Teaming with Life: Investing in Science to Understand and Use America’s Living Capital”. PCAST is an advisory group of the nation’s leading scientists and engineers who directly advise the President and the Executive Office of the President. The 2011 report recommended a suite of ambitious solutions related to ES, two of which having particular and direct relevance to national mapping of ES. The PCAST recommend- ed that the US Government establishes an Eco-informatics-based Open Resources and Machine Accessibility (EcoINFORMA) initiative. This recommendation was aimed at improving existing data collection efforts related to biodiversity, ecosystems and ES and maximising their accessibility and in- ter-operability. Although the PCAST also recommended that the US conduct a qua- drennial ES trends’ assessment, this has not yet come to fruition. Even prior to the 2015 Executive Memo- randum, ES were already becoming evident in US national policies, regulation and de- cision-making (e.g. 2008 Farm Bill, 2008 update for compensatory mitigation under Section 404 of the Clean Water Act, 2012 Forest Planning Rule, ongoing Environ- mental Protection Agency efforts to incor- porate ES into secondary air quality stan- dards). These legislative actions have helped to open the door for markets and payments for ES schemes to emerge with the US De- partment of Agriculture and the US Envi- ronmental Protection Agency entering into a joint partnership to support water quality trading and other market-based approaches for ES consistency, where applicable, with the protection of water quality pursuant to the Clean Water Act (CWA). The Federal Resource Management and Ecosystem Services Guidebook, developed by The US National Ecosystem Partnership and led by the Duke University Nicholas School of the Environment serves as an on- line training resource for incorporating ES in decision-making and includes a number of case studies in which ES were incorporat- ed into Federal decision-making. All of the above culminate in a growing need for better data and tools to support an ES approach to decision-making. EcoINFOR- MA, recommended by the 2011 PCAST report, was launched in late 2014. At the time of writing, EcoINFORMA includes three major data resource hubs: 1) Biodi- versity Serving Our Nation (BISON) con- taining millions of records of species obser- vations, 2) EnviroAtlas, the ES hub and, 3) Multi-Resolution Land Cover Consortium, providing land cover data. Additional hubs will likely be forthcoming. The EnviroAtlas3 is a web application serv- ing hundreds of open access geo-spatial data layers to technical as well as non-technical audiences (see Chapter 5.7.2). This tool is built on an ES framework with every lay- er described in terms of its relevance to production, delivery, or driver of change of ecosystem goods and services. The data span the continental US with wall-to-wall coverage of many indicators as well as with a consistent suite of indicators for selected communities across the US. Australia Australia is the world’s driest continent, has many unique ecosystems and endemic flo- ra and fauna and, since European arrival in the late 1700s, has witnessed intensive and widespread modification of land and water 3 https://epa.gov/enviroatlas Chapter 7 301 resources. Australia is particularly vulner- able to further declines in its natural cap- ital which will be exacerbated by climate change. Since the 1980s, Australian govern- ments have invested many billions of dollars in restoring its natural capital through sig- nificant policies such as the National Land- care Programme, Natural Heritage Trust (1 and 2) and the national water reform pro- cess. The roles of ES assessment to provide information on investments under these programmes are varied. For example, under the Australian Govern- ment’s 2011 Water for the Future Plan, about AU$10 billion is being invested in water li- cence buy-backs and irrigation infrastructure improvements to reduce by about 3,200 gi- galitres the annual volume of water taken from river ecosystems for irrigation. ES assess- ments are an important part of the knowledge base for decisions about where to allocate this investment that will provide the greatest en- vironmental and socio-economic benefits. A study by CSIRO showed that the social and economic benefits from the return of this wa- ter to the environment, via enhanced flow of ES, could be worth an amount similar to the Australian Government’s investment. In the State of Victoria in south-eastern Aus- tralia, recent analysis by the State Government has estimated the value of the ES benefits pro- vided by the State’s protected areas4. They con- clude that nearly 4 million hectares of protect- ed areas provide, annually, up to AU$1 billion in recreational values, up to AU$200 million in avoided health costs, AU$134 million in water quality improvements, plus a number of other ES benefits. This information will be used to support protected area planning, in- vestment and management decisions as well as to provide information for policy decisions about maintaining the natural capital in Vic- toria’s protected areas. 4 http://parkweb.vic.gov.au/about-us/news/valu ing-victorias-parks Conclusions Current design and implementation of many national and regional policies require spatially explicit information on ES. This is particularly evident for supporting policies on restoration, agriculture, spatial and ur- ban planning or marine spatial planning. Many countries recognise this and have ini- tiated programmes to mainstream quantifi- cation and mapping of ES in policies. These commitments, once they are effectively implemented, will contribute significantly to the global and regional assessments which are part of IPBES, the International Platform on Biodiversity and Ecosystem Services. Further reading Cowling RM, Egoh BN et al. (2008) An op- erational model for mainstreaming ecosys- tem services for implementation. PNAS 105: 1983-9488. Crossman ND, Bark RH, Colloff MJ, Hatton MacDonald D, Pollino CA (2015) Using an ecosystem services-based approach to measure the benefits of reducing diver- sions of freshwater: a case study in the in the Murray-Darling Basin, Australia. In: J. Martin-Ortega, R. C. Ferrier, I. J. Gordon & S. Khan (Eds.). Water Ecosystem Ser- vices: A Global Perspective. Cambridge: Cambridge University Press. Hasler B, Ahtiainen H, Hasselström L, Heiskanen A-S, Soutukorva Å, Martin- sen L (2016) Marine ecosystem services in Nordic marine waters and the Baltic Sea – possibilities for valuation. TemaNord 2016:501. Nordic Council of Ministers. http://dx.doi.org/10.6027/TN2016‐501. Mapping Ecosystem Services302 Liu JG, Diamond J (2008) Science and gov- ernment - Revolutionizing China’s environ- mental protection. Science 319: 37-38. Liu JG, Li SX, Ouyang ZY, Tam C, Chen XD (2008) Ecological and socioeconomic ef- fects of China’s policies for ecosystem ser- vices. Proceedings of the National Acad- emy of Sciences of the United States of America 105: 9477-9482. Lu YH, Fu BJ, Feng XM, Zeng Y, Liu Y, Chang RY, Sun G, Wu BF (2012) A Policy-Driv- en Large Scale Ecological Restoration: Quantifying Ecosystem Services Changes in the Loess Plateau of China. Plos One 7. Sousa et al. (2015) Ecosystem services provid- ed by a complex coastal region: challenges of classification and mapping. http://www. nature.com/articles/srep22782. Maes J, Egoh B, Willemen L, Liquete C, Vi- hervaara P, Schägner JP, Grizzetti B, Drakou EG, Notte AL, Zulian G, Bouraoui F, Luisa Paracchini M, Braat L, Bidoglio G (2012) Mapping ecosystem services for policy sup- port and decision-making in the European Union. Ecosystem Services 1: 31-39. Memorandum for Executive Departments and Agencies on Incorporating Ecosystem Ser- vices into Federal Decision Making, https:// www.whitehouse.gov/sites/default/files/ omb/memoranda/2016/m-16-01.pdf. National Ecosystem Services Partnership (2016) Federal Resource Management and Ecosystem Services Guidebook. 2nd ed. Durham: National Ecosystem Services Partnership, Duke University, https://ne- spguidebook.com. Partnership Agreement between the United States Department Of Agriculture And The United States Environmental Pro- tection Agency Regarding Water-Quality Trading (2013) https://www.epa.gov/sites/ production/files/2016-05/documents/im- age2016-05-23-125618.pdf. Pickard BR, Daniel J, Mehaffey M, Jackson LE, Neale A (2015) EnviroAtlas: A new geospatial tool to foster ecosystem services science and resource management. Eco- system Services 14: 45-55 http://dx.doi. org/10.1016/j.ecoser.2015.04.005. President’s Committee of Advisors on Science and Technology (2011) Sustaining En- vironmental Capital: Protecting Society and the Economy (Executive Office of the President, Washington, DC). https:// www.whitehouse.gov/sites/default/files/ microsites/ostp/pcast_sustaining_environ- mental_capital_report.pdf. Schaefer M, Goldman E, Bartuska AM, Sut- ton-Grier A, Lubchenco J (2015) Nature as capital: advancing and incorporating ecosys- tem services in United States federal policies and programs. Proceedings of the National Academy of Sciences of the United States of America 112(24): 7383-7389 http://dx. doi.org/10.1073/pnas.1420500112. Van Wilgen BW, Le Maitre D, Cowling RM (1998) Ecosystem services, efficiency, sustainability and equity: South Africa’s Working for Water Programme. TREE 13(9): 378. Chapter 7 303 7.2. Application of ecosystem services in spatial planning Christian Albert, Davide Geneletti & Leena Kopperoinen Introduction Spatial planning and landscape planning are generally concerned with the spatial arrange- ment and management of land but differ in focus and disciplinary orientation. Spatial planning, according to the European Re- gional/Spatial Planning Charter, “gives geo- graphical expression to the economic, social, cultural and ecological policies of society”. It includes various instruments, such as com- prehensive planning, zoning and Strategic Environmental Assessments (SEA). Land- scape planning, in contrast, has been defined by the European Landscape Convention as “a strong forward looking action to enhance, restore or create landscapes”. In many EU member states, landscape planning is an in- tegral part of spatial planning. The aims of this chapter are to introduce the current spatial and landscape planning practice concerning the integration of envi- ronmental information, to present options for applying ES maps in planning and to discuss related opportunities and challenges. Current practices of integrating environmental information in planning Assessing and addressing environmental is- sues is not new to the fields of spatial and landscape planning. Depending upon the planning instrument under consideration, different types of environmental informa- tion and approaches for integration are already in use. SEA, particularly, aims to provide a high level of protection for the en- vironment by systematically integrating en- vironmental considerations during planning preparation and adoption. The environmen- tal issues explicitly mentioned by the Euro- pean SEA legislation include biodiversity, population, human health, fauna, flora, soil, water, air, climatic factors, material assets, cultural heritage (including architectural and archaeological heritage) and landscape. Landscape planning also illustrates various approaches for taking account of environ- mental information. The German ‘Land- schaftsplanung’, for example, analyses the current state of the landscape concerning a set of landscape functions, defined as “the capacity of a landscape […] to sustainably fulfil basic, lasting and socially legitimised material or immaterial human demands”. As such, it considers the capacities (or po- tentials) of ecosystems to deliver ecosystem services (ES) as demanded by society, re- gardless of their actual and current use. The measures, against which landscape plan- ning assesses and evaluates these landscape functions, are legally derived environmental development objectives and expert-based as- sessments of rarity and value. Importantly for useful application, mapping approaches need to be adapted to the specif- Mapping Ecosystem Services304 ic objectives and interests of decision-mak- ers, planners and stakeholders involved in the planning processes. Furthermore, the delineation of maps often relates to jurisdic- tional boundaries whereas ecosystems and ES provisioning and benefiting areas easily transcend them. To this end, a multi-level approach to mapping with eventually dif- ferent degrees of mapping detail (Chapter 5.6) are required to provide decision-mak- ers with information on how external effects influence their decision-making and how their decision-making in the respective ju- risdiction may influence ES provision and delivery in other jurisdictions. Options for applying ES maps in planning Various options exist for applying ES maps in support of spatial planning and deci- sion-making. The way in which the ES maps can be used depends upon the specif- ic planning instrument in use, the need to fulfil statutory requirements for the imple- mentation of the respective instrument, the needs and interests of instrument users and decision-makers, as well as the time and re- sources available for developing ES maps (in addition to what is already legally required). Consider the following examples. ES maps can be used as an information source for investigating impacts of proposed planning decisions and for comparing pos- sible alternatives. Recent publications have addressed the question of how ES maps can be used to support SEA of spatial planning (see Chapter 7.8). ES maps can help to identify where areas of particular environmental sensitivity or high potential for ES delivery or for demand for ES are located. Such information is useful for developing comprehensive and strate- gic development plans. For example, areas which have particular environmental sensi- tivity against impacts, provide particularly important ES, or provide opportunities for exploiting synergies by delivering several ES simultaneously, should be safeguarded, en- hanced or restored. Maps of green and blue infrastructure rep- resenting the spatial variation in ES supply potential, coupled with spatially explicit data on people’s values and actual use of ES, help spatial planners identify mismatches between supply and demand, as well as trade-offs or compensation actions to be undertaken in planning decisions. In addition, the flow of ES from supplying areas to the beneficiaries can be illustrated with ES maps, especially when using participatory mapping methods. ES maps can enhance stakeholders’ and de- cision-makers’ engagement by better com- municating the benefits and shortcomings associated with proposed planning options. ES maps visualise the trade-offs that can be caused by land-use changes and urban man- agement alternatives for ES provision. ES maps support valorisation, for ex- ample, by selling agrarian and touristic products with price premiums as a way to co-finance environmentally sensitive land use management. ES maps contribute to understanding the spa- tial relationships between the planning area (which typically corresponds to a jurisdiction, for example, at the regional or national level) and the areas where ES are supplied and used. A proper recognition of these relationships al- lows addressing situations where the benefits of planning decisions accrue at one scale, but costs are borne at another scale. By using open access data and methods for mapping, similar approaches can easily be Chapter 7 305 made available for scientific review, practical application, comparison between different regions and further development. Case example of applying ES maps in spatial planning, city of Järvenpää, Finland The small and relatively compact city of Jär- venpää, Finland, decided to take positive ac- tions for land improvement by placing infill development in city-owned land parcels that were mainly green areas of varying quality. To understand the values of the potential infill development sites, the green infrastructure of Järvenpää was mapped based on natural values, ecological connectivity and ES supply (Figure 1) and demand (Figure 2). The maps covering the whole city area were then used to assess the importance of each potential site. The values of the sites were described in detail and this information helped the spa- tial planners to make an informed decision about which areas could be used for con- struction while causing least harm to both nature and people. Requirements of ES maps to be usefully applied in planning In order to be useful in planning, ES maps need to fulfil a number of requirements: They need to be specifically attuned to the context and purpose of the planning study and the interests and concerns of the pop- ulation. To be actually useful, the mapping exercise needs to begin with a joint decision of map-makers, users and decision-makers concerning the spatial scale and resolution applied, the ES considered, the indicators used, the approaches used for assessing and valuing, as well as the format of the mapping outputs. As a consequence, the information needs and requirements of po- tential users and decision-makers need to be investigated and addressed in the design and implementation of the mapping exer- cise from the very outset. The ES classes selected and examined need to be specifically attuned to the issue at stake. Figure 2. Demand for ES, assessed by a map survey in a workshop organised for local residents in Järvenpää. The dots represent markers placed by residents and the different colours of dots represent different cultural ES-related values of the respondents. The potential development sites within the urban fabric are delineated with a red line. Black areas are buildings, white areas are impervious land. Other colours of the areas show to which class in the created green infrastructure typology the area belongs. Figure 1. The variation in the cultural ES supply potential in and around the potential infill develop- ment site of the eastern and western Aittokorven- puisto (delineated with a red line) in Järvenpää. The darker the green, the greater the supply potential. Black areas are buildings, white areas are impervious land. Mapping Ecosystem Services306 Mapping of ES supply is only a part of the planning process. It needs to be comple- mented with spatially explicit information on ES demands, stakeholder interests etc. Users and decision-makers need to be sys- tematically involved in the development of the ES maps. Feedback from local and re- gional experts is also essential in verifying the maps because no spatial data is perfect and without gaps. The timeliness and longer term appropri- ateness of the maps should be ensured. The maps need to be prepared in the timeline with the planning decision that is to be made. In addition, ES maps should be de- veloped and delivered in a way that allows them to be updated once changes have been made to land uses and management. Opportunities and challenges of applying ES maps in planning Several challenges exist concerning the ap- plication of ES maps in planning. ES maps, as with any kind of environmental information, are only one part of the vari- ous information and concerns that planning needs to take into consideration. They may illustrate and, thus, helpfully support efforts to integrate environmental considerations in decision-making, but the actual potential to influence decision-making is limited (es- pecially within statutory planning). Incorporating ES in decision-making can make the planning process more complex. This is a significant challenge that might be alleviated by developing assessment stan- dards, the provision of ES maps by national institutions, simple but robust methods and tools for the creation of maps. ES maps appear to represent true infor- mation, but they most often have inherent uncertainties attached to them (Chapter 6). Communicating this uncertainty to the au- dience and appropriately addressing the un- certainty by planning- and decision-makers is an enduring challenge. The opportunities provided by using ES re- late to the provision of essential and import- ant information for planning. The use of the ES concept, versus other con- cepts such as landscape functions, has the potential to relate well to diverse groups of users and stakeholders through the notion of ‘services’ provided by nature and land- scape to people. As such, they can facilitate cooperative landscape and spatial planning and implementation in practice. ES maps can complement existing envi- ronmental information and approaches by providing more differentiated information on the actual provision and use of ES (and not just ES potentials as hitherto the case), trade-offs and synergies of land use options concerning the delivery of various ES and the spatial allocation of the supply of and demand for ES. ES maps can provide a useful basis for quan- tification and economic valuation of ES which in turn may provide additional added value for planning and decision-making. Conclusions Maps of ES supply and demand are useful for planning- and decision-support in providing information concerning ES provisioning and benefiting areas as well as synergies and trade- offs between several ES. This information can relate to the status quo or in alternative land use options. Outcomes of ES maps can then Chapter 7 307 be used to identify areas that need to be safe- guarded, enhanced or developed. To harness these opportunities for applying ES maps, planning practitioners need to apply the mapping techniques and maps in ways carefully adapted to the specific user, governance and decision-making context. Further reading Albert C, Aronson J, Fürst C, Opdam P (2014) Integrating ecosystem services in landscape planning: requirements, ap- proaches and impacts. Landscape Ecology 29: 1277-1285. Albert C, Galler C, Hermes J, Neuendorf F, von Haaren C, Lovett A (2016) Applying ecosystem services indicators in landscape planning and management: The ES-in- Planning framework. Ecological Indica- tors 61, Part 1: 100-113. Albert C, Hauck J, Buhr N, von Haaren C (2014) What ecosystem services informa- tion do users want? Investigating interests and requirements among landscape and regional planners in Germany. Landscape Ecology 29: 1301-1313. Geneletti D (2011) Reasons and options for integrating ecosystem services in strategic environmental assessment of spatial plan- ning. International Journal of Biodiversity Science, Ecosystem Services & Manage- ment 7(3): 143-149. Hansen R, Pauleit S (2014) From Multifunc- tionality to Multiple Ecosystem Services? A Conceptual Framework for Multifunc- tionality in Green Infrastructure Planning for Urban Areas. AMBIO 43: 516-529. Hauck J, Görg C, Varjopuro R, Ratamäki O, Maes J, Wittmer H, Jax K (2013) “Maps have an air of authority”: Potential benefits and challenges of ecosystem service maps at different levels of decision-making. Eco- system Services 4: 25-32. Hauck J, Schweppe-Kraft B, Albert C, Görg C, Jax K, Jensen R, Fürst C, Maes J, Ring I, Hönigová I, Burkhard B, Mehring M, Tiefenbach M, Grunewald K, Schwarzer M, Meurer J, Sommerhäuser M, Priess JA, Schmidt J, Grêt-Regamey A (2013) The Promise of the Ecosystem Services Con- cept for Planning and Decision-Making. GAIA - Ecological Perspectives for Science and Society 22: 232-236. Kopperoinen, L., Itkonen, P., Niemelä, J. (2014) Using expert knowledge in com- bining green infrastructure and ecosystem services in land use planning – an insight into a new place-based methodology. Landscape Ecology 29: 1361-1375. DOI 10.1007/s10980-014-0014-2. von Haaren C, Albert C (2011) Integrating ecosystem services and environmental planning: limitations and synergies. Inter- national Journal of Biodiversity Science, Ecosystem Services & Management 7: 150-167. von Haaren C, Albert C, Barkmann J, de Groot R, Spangenberg J, Schröter-Schlaack C, Hansjürgens B (2014) From explanation to application: introducing a practice-oriented ecosystem services evaluation (PRESET) model adapted to the context of landscape planning and management. Landscape Ecology 29: 1335-1346. Mapping Ecosystem Services308 7.3. Land use sectors Benjamin Burkhard The human utilisation of a piece of land for a certain purpose is called land use. Land use is often closely related to land cover, but it is not the same. Land cover represents the features that cover the earth‘s surface as they would be viewed from above, for example, from an aeroplane or a remote sensing sat- ellite. Land use clearly refers to activities of people and how they are using the land. In today’s heavily cultivated and modified world, it is difficult to find wilderness areas without any human impact on land cover. Therefore both terms are often used in a combined way such as land use/land cover (LULC). All forms of land use are causing impacts on ecosystem functions by alter- ing ecosystem structures and processes and related ecosystem services (ES) supply (see Chapter 2.2). Land use intensification and increased technology use will enhance these impacts in future if no sustainable strategies can be found. Traditional and typical land use sectors are agriculture (see Chapter 7.3.2), forestry (see Chapter 7.3.3), tourism, mining, industry (Chapter 7.5), infrastructure, military areas or urbanisation (see Chapter 7.3.1). The most widespread form of land use today is agriculture, currently covering more than 37 % of the earth’s terrestrial areas. Graz- ing land accounts for about 26 % and crops grown for animal fodder account for about 33 % of all cultivated land. In addition, non-use forms such as nature protection areas (see Chapter 7.3.4) are claiming land and can be considered a land use sector, for instance, when it comes to landscape plan- ning. Each available (and reachable) piece of land can be utilised by human beings for a limited number of uses only. Some forms of land use are mutually exclusive, such as con- ventional agriculture and forestry or military areas and tourism. Other forms of land use can create synergies amongst each other, for example, agricultural tourism, agroforestry or urban gardening. Some forms of land use can be exclusive such as mining or military areas. Such ‘hard’ forms of human activi- ties (but also nature reserves) often cause conflicts due to their exclusivity or rivalry. Studies on land use conflicts and related ES gains and losses are highly relevant for envi- ronmental management and complex trade- off decisions between land use development and conservation. LULC changes can affect ES on various spatial and temporal scales. Therefore it is important to know about the effects that different land use sectors have on ES and to map them. Land use data can be used as basic geospatial map units to up- or down- scale aggregated models (see Chapter 4.4) or statistical data to quantify and map ES. Re- spective statistics such as agricultural yields, forestry harvests, fish catches or tourist numbers are available for most land use sec- tors. Land use data can provide spatial units to start the mapping until more suitable spa- tial data in finer scales (such as watersheds, field blocks) are available. In Europe, the land cover classes of the Eu- ropean CORINE1 project are applied fre- quently for ES mapping. Comparable ap- proaches exist in North America (NALC2) 1 http://www.eea.europa.eu/data-and-maps/figures/ corine-land-cover-types-2006 2 https://lta.cr.usgs.gov/pathfinder/nalc_project_ campaign Chapter 7 309 and on a global scale (GlobCover3). The data originate from remote sensing. Thus they provide a logical combination of land cover and land use as ‘seen’ from space and as it can be found in reality - a combination of natural conditions and human activities. Information from ES maps has a high appli- cation potential for land use planning and management. They can contribute to the development of site-specific, optimised and sustainable land use strategies. 3 http://due.esrin.esa.int/page_globcover.php Further reading Hassan R, Scholes R, Ash N (Eds.) (2005) Ecosystems and human well-being: Cur- rent state and trends: findings of the Con- dition and Trends Working Group. The Millennium Ecosystem Assessment Series, Vol. 1. Island Press. Maes, J, Crossman ND, Burkhard B (2016) Mapping ecosystem services. In: Potschin M, Haines-Young R, Fish R, Turner RK (Eds.) Routledge Handbook of Ecosystem Services. Routledge. London: 188-204. Mapping Ecosystem Services310 7.3.1. Mapping urban ecosystem services Grazia Zulian, Inge Liekens, Steven Broekx, Nadja Kabisch, Leena Kopperoinen & Davide Geneletti Introduction Globally, more people live in urban areas than in rural areas, with 54 % of the world’s population living in urban areas in 2014. As the world continues to urbanise, sustainable development challenges will be increasingly concentrated in cities. The UN Sustainable Development Goals well summarise this concept with goal 11: “Make cities inclu- sive, safe, resilient and sustainable”. Maintaining functioning, healthier and equally accessible urban ecosystems and services is thus an essential point for future urban policies and planning. Urban ecosystems can be defined as an integrated ensemble of connected built (sharing built or paved infrastructures) and green infrastructures (GI). The tangi- ble integration of GI in urban policies re- quires awareness-raising amongst planners, stakeholders and citizens as well as tools to monitor progress of policy objectives and to support local planning. Nevertheless urban environments are very peculiar and a general framework for the mapping of urban ecosystem services (ES) cannot be directly adopted. This chapter illustrates how urban ES can be mapped according to a tiered approach (see Chapter 5.6.1). This chapter introduces a selection of ES particularly relevant in cities. Next it provides concrete examples on map- ping urban GI and urban ES applying a tier 1 approach, based on Urban Atlas landcover data provided by the European Environment Agency and local data. The chapter presents two tier 3 models, for mapping regulating and cultural services. Finally a web-based tool for an analysis of urban ES is introduced. Ecosystem services relevant in cities Trees, parks, gardens and (peri-)urban for- ests help improve the quality of the air, re- duce noise and mitigate extreme summer temperatures or peak flood events. They also provide non-material benefits, such as recreation, education, cultural and aesthetic values and contribute to social interactions. Table 1 presents a list of key urban ES. Cit- ies also depend on ecosystems beyond city limits and, in this case, we refer to indicators described in other sections of this book. Mapping urban ecosystems and urban green space as the base layer for assessing urban ES A detailed map of urban GI can serve as the basis for mapping urban ES supply and Chapter 7 311 demand. This requires detailed spatial data for identifying the service providing units of GI. Depending on the context and purposes of the study, the analysis can cover a variety of spatial extents (from large metropolitan areas to small compact cities) and can be based on different data sources. In Järvenpää, Finland, GI was identified and a typology of GI was created based on fair- ly detailed spatial data (municipal biotope data) and areal units, including even the smallest green spaces. All permeable sur- faces were considered as areas potentially providing ES. Therefore, the land use and land cover data were masked by all sealed areas including mainly streets, railroad, oth- er traffic areas, landfills and buildings. This was undertaken by using several national and municipal spatial datasets. At the final stage, the most recent available aerial photo- graphs were used to check the validity of the digitised features. The final outcome of the spatial representa- tion of the GI typology in Järvenpää is pre- sented in Figure 1. GI was classified according to land cover and land use type. Public and private land were both considered as potential service providing units for urban ES provision. In fact, private yards and gardens can be very important for provision of regulating and cultural services (e.g. stormwater retention, pollination and adding to aesthetics of an area). Public green and blue areas, on the other hand, are very important from an environmental justice point of view. The benefits delivered by these areas should be available and accessible easily and evenly to different population groups to improve the well-being of residents. In Leipzig, Germany, the Urban Atlas land cover data set, provided by the European Environmental Agency, was used to show CICES Section CICESClass Provisioning Cultivated crops Surface water for drinking Groundwater for drinking Surface water for non-drinking purposes Groundwater for non-drinking purposes Regulation & Maintenance Filtration/sequestration/storage/accumulation by ecosystems Global climate regulation by reduction of greenhouse gas concentration Micro and regional climate regulation Mediation of smell/noise/visual impacts Hydrological cycle and water flow maintenance Flood control Pollination and seed dispersal Cultural Physical and intellectual use of land-/seascapes in different environmental settings Scientific/ Educational Heritage, cultural Aesthetic Table 1. Key urban ES organised according to the CICES classification. Mapping Ecosystem Services312 spatial patterns of urban ES indicators and their performance1 . 1 (http://www.eea.europa.eu/data-and-maps/data/ urban-atlas). Urban ES values for carbon storage and recre- ation services for the 20 different Urban Atlas land cover classes were derived from empiri- cal studies. For the assessment of recreation, the per capita green space in 63 districts of Leipzig was used as proxy. Population data reflect the district population in 2014. The results are urban ES performance maps based on the different land cover classes. Figure 2 shows the resulting map for carbon storage and per capita green space for the city of Leipzig. The use of secondary land cover and popula- tion data may limit the opportunities for sta- tistical analysis. Using land cover data always means generalisation but this provides an overview of city-wide urban ES performance. Figure 2. Carbon storage in Leipzig (left) and per capita green space in the districts (right). Carbon storage is highest in the riparian forest areas in Leipzig. The per capita green space is highest in districts near the floodplains and in the southern, north-western and north-eastern districts near the city border where the population number is comparatively lower than in the inner city districts. Figure 1. Map of green infrastructure in Järvenpää for the assessment of urban ES provision. Built-up areas are shown in white. Chapter 7 313 Mapping regulating and cultural services: A tier 3 approach Assessing the cooling capacity of urban GI Assessing the urban ES provided by GI is of- ten too data-demanding for being routinely conducted in urban planning. A method, based on literature data, has been developed to assess the cooling effect provided by GI. This method can be employed by planners to support the design and management of these infrastructures. First, the main func- tions involved in the cooling capacity of GI were identified: shading and evapo-transpi- ration. These functions were assessed in- dividually and then combined in order to estimate the overall cooling capacity of GI. The assessment of the shading function was based on an analysis of the tree canopy cov- erage which is one of the key elements in- fluencing the shading effect. The assessment of evapo-transpiration considered soil cover, tree canopy coverage and climatic area of the GI which are the three main compo- nents involved in providing this function. Each function was classified into categories and the categories were then combined into an overall cooling capacity value which also considered the size of GI. This capacity was then classified into six classes from “E” to “A+”, adopting the European Union Energy Label classification, where A+ represents the highest cooling performance. Finally, decay models were also applied to assess the effect beyond the boundaries of GI. The overall purpose was to provide planners with a rel- atively simple model to predict the cooling capacity of GI and to support their design and inclusion in urban plans. Figure 3 pro- vides an example of cooling capacity assess- ment in the city of Trento, Italy. Assessing the social value of public parks and playgrounds Public parks and playgrounds are key re- sources for urban citizens since they provide recreational, cultural and educational oppor- tunities. Nevertheless these opportunities are not only related to the amount of public green surface per capita but also to other as- pects, for example, type of facilities available or the presence of bicycle paths to reach the park. This problem was addressed by devel- oping a model to estimate the amount of Figure 3. Map of the cooling capacity of the urban GI in the city of Trento. Cooling capacity is expressed in classes from A+ (highest capacity) to E (lowest capacity). Mapping Ecosystem Services314 service provided by urban parks. The model consists of two parts: 1) it estimates the So- cial Value of Public GI (SVPGI); 2) it calcu- lates a potential accessibility measure which accounts for user’s characteristics (the age). Figure 4 presents the structure of the model; Figure 5 shows the amount of service poten- tially available in Padua (Italy) amongst the population younger than 11 years old. Planning for green infrastructure in cities: The “Nature Value Explorer for Cities” tool The online Nature Value Explorer tool2 aims to value the impact of nature develop- ment projects on ES. The tool is currently being extended with an urban version. The purpose of this version is to support cities, administrations and planners in providing an equal and adequate supply of urban GI, paying attention to the quality and the func- tions of the GI and the trade-offs between different urban ES. Users can estimate the effects of the existing and planned GI on reaching different sustainability goals. The urban context requires a specific typology of urban green and valuation methodologies specifically suited for urban environments. Urban ES which can be valued include ur- ban farming, air pollution and urban heat stress reduction, carbon sequestration, water retention, health and wellbeing. The maps below (Figure 6) are produced for the city of Antwerp (Belgium) and represent the actual demand, supply and potential for green vegetation to reduce urban heat impacts. Demand maps are based on population densi- ty. The urban heat map for Antwerp is a com- bination of UrbClim model simulations with in-situ validation and satellite images, whereas 2 www.natuurwaardeverkenner.be Figure 4. Overview of the structure of the model. The SVPGI (A) depends on the green area sur- face and the presence of playgrounds-sport-rec- reational facilities (A.2) and key contextual factors (proximity to bicycle paths, safety) (A.3). To calculate the social services map (B), the SVPGI is allocated amongst all citizens (or amongst defined user groups), giving each one an amount proportional to a distance decay function (B.1). The parameters of the function can be adjusted, according to the users’ age or other characteris- tics. The distance can be estimated through the local road network. Figure 5. The estimated social service per population younger than 11 years old. Chapter 7 315 the supply maps represent the cooling effect of the existing vegetation and water system. The potential for additional trees to reduce urban heat impacts depends on the mismatch of supply and demand, the impact of trees on urban climate and the spatial boundary con- ditions for additional trees (we assume trees cannot replace existing buildings). Figure 6. Urban ES maps for heat stress in Antwerp. Supply from existing vegetation and water is scored from zero (0) to maximum (5). Based on a heat map of the city and population densities, the demand is mapped leading to zones with varying degrees of impact vegetation. Taking into account the current supply and demand, the potential for green measures is calculated and scored from no potential (0) to maximum potential (20). Demand: avarage radiation temperature (oC) 2012 Supply: impact existing vegetation (0:none; 5: high) Potential impact tree row (0:none; 20: high) < 40 40 - 60 >60 0 1 2 3 4 5 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Further reading Broekx S, Liekens I, Peelaerts W, De Nocker L, Landuyt D, Staes J, Meire P, Schaafs- ma M, Van Reeth W, Van den Kerckhove O, Cerulus T (2013) A web application to support the quantification and valuation of ecosystem services. Environmental Im- pact Assessment Review 40: 65-74. Geneletti D, Zardo L, Cortinovis C (2016) Pro- moting nature-based solutions for climate adaptation in cities through impact assess- ment. In: Geneletti, D (Ed) Handbook on biodiversity and ecosystem services in im- pact assessment. Edward Elgar, 428-452. Kabisch N, Larondelle N, Artmann M (2014) Urban Ecosystem Services in Berlin, Ger- many and Salzburg, Austria: Climate Regu- lation and Recreation function. In Kabisch N, Larondelle N, Artmann M (Eds.) Hu- man-Environmental Interactions in Cities - Challenges and Opportunities of urban land use planning and green infrastructure. Cambridge Scholars Publishing, 66-80. Haase D, Kabisch N, Strohbach M, Eler K, Pintar M (2015) Urban GI compo- nents inventory. Milestone 23. GREEN SURGE project (2013-2017), EU FP7 (ENV.2013.6.2-5-603567) 16 pp. http:// greensurge.eu/working-packages/wp3/ files/MS23_update_19022015.pdf. Mapping Ecosystem Services316 Larondelle N, Haase D, Kabisch N (2014) Mapping the diversity of regulating eco- system services in European cities. Global Environmental Change 26: 119-129. Maes J, Zulian G, Thijssen M, Castell C, Baró F, Ferreira AM, Melo J, Garrett CP, Da- vid N, Alzetta C, Geneletti D, Cortinovis C, Zwierzchowska I, Louro Alves F, Souto Cruz C, Blasi C, Alós Ortí MM, Attorre F, Azzella MM, Capotorti G, Copiz R, Fusa- ro L, Manes F, Marando F, Marchetti M, Mollo B, Salvatori E, Zavattero L, Zingari PC, Giarratano MC, Bianchi E, Duprè E, Barton D, Stange E, Perez-Soba M, van Eupen M, Verweij P, de Vries A, Kruse H, Polce C, Cugny-Seguin M, Erhard M, Nicolau R, Fonseca A, Fritz M, Teller A (2016) Mapping and Assessment of Eco- systems and their Services. Urban Ecosys- tems. Publications Office of the European Union, Luxembourg. Secco G, Zulian G (2008) Modelling the so- cial benefits of parks for users. In Carreiro MM, Song Y-C, Wu J (Eds.) Ecology, Planning and Management of Urban For- ests: International Perspectives. New York, Springer, 312-335. Chapter 7 317 7.3.2. Ecosystem service maps in agriculture Louise Willemen, Sarah Jones, Natalia Estrada Carmona & Fabrice DeClerck Introduction Agricultural ecosystems are the largest eco- systems in the anthropocene. To produce food, fodder and fuels, these agricultural systems strongly depend on a reliable flow of ecosystem services; examples include water, pollination, pest control, soil fertility and the gene pool of wild crop relatives. At the same time, it is well known that many agricultur- al practices and the expansion of agricultural areas are a major threat to well-functioning healthy ecosystems. However, the inverse can arguably be just as true; agriculture, if well managed, can become an important means by which to secure and safeguard ecosystem services (ES). Agriculture has been the most direct way humans altered their natural sur- roundings and has brought major increas- es in well-being and income to humans. It is important to realise that most ES result in human benefits only after human input or activities, such as seeding and harvesting crops, travelling to attractive locations, or re- directing water (Chapter 5.1). Agricultural systems are intensely managed by humans and are more controlled and reg- ulated than most other ‘ecosystems’. Many governance systems are in place to manage and distribute excludable and rival goods (e.g. water board for irrigation water, fishing quota, timber extraction licences). This high level of human management and regulation creates opportunities for securing and safe- guarding ES for agriculture and non-agri- cultural production uses. ES in agricultural landscapes operate across different spatial and temporal levels: before an ES reaches the field, it may have moved over various distances from different land cover types in the surrounding areas. For ex- ample, soil conservation practices on slopes reduce the negative impact of sedimentation or landslide risk on the downslope. Under- standing this multi-level aspect (where ES come from and flow to and at what point in time) is crucial for an effective management of ES flows in rural areas. In this chapter, we reflect on the role of spa- tial information on ES for the sustainable management of agricultural areas. The use and selection of ES to consider and their mapping approaches depend on: i) the strength of the relationship between agricul- tural production systems and ES supply and ii) the spatial extent of the supply, flow and management level of the ES. Ecosystem services and agricultural production links In 2014, The Economics of Ecosystems and Biodiversity initiative (TEEB) initiated a specific study on the value of ES and bio- diversity across agricultural systems: TEEB for Agriculture and Food (TEEBAgFood). TEEBAgFood has identified the positive (provisioning and regulating services) and Mapping Ecosystem Services318 negative (environmental impacts) flows to and from agricultural systems. The quanti- fication of these services helps to assess the dependence and impact of production sys- tems on ES supply. However, not all ES have equal relevance for all farming systems. In Figure 1, we show the assumed and simplified link for high to low input farming systems to relevant ES based on their supply, ES dependence and ES impact. The figure also shows on which spatial level these interactions take place and therefore need to be managed. “Input” refers here to pesticides, fertilisers and water (not to labour or machinery). The white arrows in this figure indicate the farming systems for which the specific ES (and thus information on this ES) is relevant. The general assump- tion is that low input farms are more depen- dent and have less impact on ES compared to conventional high input farming. For ex- ample, the supply of the ES ‘nutrient cycling’ is particularly relevant for low input farming systems. In contrast, closely managing nutri- ent cycling via an ES based approach is not as relevant on farms where this is provided by synthetic fertilisers. In Figure 1, this is shown by the arrow indicating the lower input farm- ing systems only for this ES. Some ES are rel- evant for all farming systems: all farms will produce food, fodder or fuel crops, they all rely on specific water and climate conditions and all conversions of land to agriculture will impact the natural habitat. Figure 1 could be used as a general guide for selecting the specific ES to be mapped, in ad- dition to the location-specific ES information needs and focus. Maps of ES play an import- ant role in land management for: the assess- ment of the current state of ES in rural ar- eas, impact analyses of agriculture on ES and Figure 1. Linkages between ES and agricultural management types for ES production, ES dependence and ES impact per spatial level. The white arrows indicate to which farming type the ES relate, from low to high input. Chapter 7 319 the monitoring of ES to support sustainable management of agricultural areas. Land man- agement, as well as the generation of spatial information, has so far mostly focused on the ES supply (agricultural goods) and ES impact (e.g. environmental impact assessments) and less so on the enabling of common public goods on which ES depend (central blue bar of Figure 1). The TEEBAgFood project calls these the ‘invisible’ positive flows. Maps can make these invisible flows ‘visible’, facilitating their inclusion in decision-making. Ecosystem service maps for farms and beyond Decisions on agricultural practices are typi- cally made at farm level. However, most ES on which agriculture depends and impacts often have a spatial level exceeding the farm. Figure 1 shows that difference: few ES are purely linked to field level, while many ES are related to the ‘full eco-agri-system’ which can cover landscapes, watersheds or even the global system depending on the ES in ques- tion. Thus, when mapping ES to support decision-making in agricultural manage- ment, farm and field level maps alone are insufficient, as agriculture mostly supplies, impacts and depends on ES from larger spa- tial extents. The spatial extent of ES and the related mapping requirements (data resolu- tion, accuracy) are described in Chapter 5.2. Applications of ES mapping in agricultural areas Current work demonstrates that ES maps and the process of generating maps can address important land management ques- tions in agricultural areas across the globe. Studies have shown that the process of map- ping ES as well as the maps themselves can be used to: i) visualise the scales at which different services operate; ii) assess locations of ES supply and beneficiaries highlighting dependencies; iii) visualise impacts which are often considered invisible externalities of agriculture, both positive and negative; iv) facilitate negotiations amongst stakeholders, including payment schemes and v) target intervention locations required to ensure or improve ES supply. An example of this type of ES mapping study is presented in Box 1. Box 1 . Managing reservoir catchments to secure transboundary ES delivery in the Volta basin The Volta River flows through six West African countries, draining a 407,000 km2 area that is home to over 20 million people. The Volta basin is subject to highly variable rainfall, yet timely supply of a sufficient quantity of quality water is essential for the rural households that rely on crop, fish or livestock production for their livelihood. Over 1000 small and several large dams have been constructed in the basin since the 1950s to help maintain a year-round supply of agricultural water. Ecosystem processes in the reservoir catchments provide a service for reservoir-users by regulating the quality, quantity and timing of reservoir water supplies, making the network of land-users, reservoir systems and water beneficiaries tightly inter- connected. Bioversity International and its partners are working with smallholder farmers and local and regional government in the Volta basin to facilitate evidence-based ES management decisions. Many of these stakeholders identify soil erosion and associated sedimentation as a key threat to reservoir water sup- plies and water management authorities are seeking to minimise erosion through improved management of land adjacent to the stream network. The ES model WaterWorld, is used here to investigate the effect on water supply and the control of soil erosion rates by ensuring: 1) 100 % herbaceous plant cover and 2) Mapping Ecosystem Services320 100 % tree cover, on land within 100 m of waterways in dam catchments across the Volta basin. Results indicate that targeting herbaceous vegetation cover in riparian zones (Scenario 1) would be more effective than targeting tree cover (Scenario 2) for improving water availability, although benefits are unevenly dis- tributed across the region and generally higher in the south. Local variations in annual water balance are expected particularly under the tree cover scenario, with the annual water supply falling to less than half of its baseline level (a decrease of more than 100 %) in several dispersed locations across the region. The area, highlighted in the annual water supply inset maps below, illustrates that water supplies are generally expected to decrease on the Burkinabé side of the border under both scenarios while, on the Ghanaian side, water balance is expected to increase by up to 10 % or more in most places under herbaceous cover (Scenario 1), but continue to fall under tree cover (Scenario 2). The difference in water supply results between the scenarios can be largely explained by a difference in evapo-transpiration losses which will be higher from tree cover than herbaceous cover. In contrast, both vegetation types appear to be effective at controlling sediment. Both scenarios indicate erosion control rates adjacent to waterways will increase across the basin where there is perennial vegetation cover, with the largest erosion prevention impacts occurring near the headwaters of the stream network where slopes are steepest. The erosion control inset maps below illustrate that reduced erosion rates may be up to 100 % compared to baseline levels in some areas. The model outputs show that ensuring year-round vegetation cover on land adjacent to waterways, particularly with herbaceous plants and near stream headwaters, could be an effective strategy to control sedimentation rates and improve regional water supplies. Much of this riparian land is currently used for crop and livestock production and restricting agriculture on this land would negatively impact on thousands of smallholder farmers. Careful management of vegetation cover on existing agricultural land combined with protection and restoration of natural vegetation in adjacent areas could represent a viable option for implementing a riparian management scheme with minimal losses to food production. This would mean agricultural land in riparian zones is selectively managed to ensure year-round plant cover by, for example, using perennial species such as bananas, perennial rice and cover crops, while natural vegeta- tion is restored and protected on adjacent non-agricultural land. Mapping relative changes in ecosystem servces across the Volta basin under two riparian buffer management scenarios. Scenario 1: Herbaceous plant cover (natural, crops, cover crops) in 100 m buffer along waterways in dam watersheds. Change from baseline (%) -1000% - -100% -99% - -11% -19% - -1% 0% (no change) 1% - 10% 11% - 100% 101% - 1,000% Main map scale: 1:17,000,000. Minor map scale: 1:5,000,000. Data sources: GAUL (admin bounderies); GRUMP (settlements) WaterWourld V2 - KCL/AmbioTEK (all other data) Chapter 7 321 Further reading Fremier AK, Declerck FAJ, Bosque-Pérez NA, Carmona NE, Hill R, Joyal T, Keesecker L, Klos PZ, Martínez-Salinas A, Niemey- er R, Sanfiorenzo A, Welsh K, Wulfhorst JD (2013) Understanding Spatiotemporal Lags in Ecosystem Services to Improve In- centives. BioScience 63: 472-482. Mulligan M (2013) WaterWorld1: a self-pa- rameterising, physically based model for application in data-poor but problem-rich environments globally. Hydrology Re- search 44(5): 748. TEEB (2015) TEEB for Agriculture & Food: an interim report. United Nations Environ- ment Programme, Geneva, Switzerland. 1 www..org/waterworld Poppy GM, Chiotha S, Eigenbrod F, Harvey CA, Honzák M, Hudson MD, Jarvis A, Madise NJ, Schreckenberg K, Shackleton CM, Villa F, Dawson TP (2014) Food secu- rity in a perfect storm: using the ecosystem services framework to increase understand- ing. Philosophical Transactions of the Royal Society B: Biological Sciences: 369. Power AG (2010) Ecosystem services and agriculture: tradeoffs and synergies. Phil- osophical Transactions of the Royal Soci- ety of London B: Biological Sciences 365: 2959-2971. Scenario 2: Tree cover (natural, orchards, plantations) in 100 m buffer along waterways in dam watersheds. Change from baseline (%) -1000% - -100% -99% - -11% -19% - -1% 0% (no change) 1% - 10% 11% - 100% 101% - 1,000% Main map scale: 1:17,000,000. Minor map scale: 1:5,000,000. Data sources: GAUL (admin bounderies); GRUMP (settlements) WaterWourld V2 - KCL/AmbioTEK (all other data) Mapping Ecosystem Services322 7.3.3. Mapping forest ecosystem services Sandra Luque, Julen Gonzalez-Redin & Christine Fürst Introduction Forests are a crucial element not only of landscapes but of human living conditions. Forests have supported people’s livelihoods throughout history, particularly when crops failed. Covering nearly a third of the earth’s land surface, they provide multiple ecosystem services (ES) and habitats for a multitude of species. They hold the majority of the world’s terrestrial species. However, these biological- ly-rich systems are increasingly threatened, largely as a result of human activity, such as land-use and climate change, deforestation, afforestation, wildfires, storms, insects and pathogen outbreaks. Timber production has often dominated the way in which forests were managed until the 20th century. New challenges and increasing pressures in the 21st century have stimulated a multi-functional approach, involving the delivery of multiple goods and services in- cluding regulating ES (e.g. climate regulation and mitigation, erosion control, hydrological regulation). Nowadays, in most regions of the world, forests, trees on farms and agro-forestry systems play important roles in the livelihoods of people by providing employment, ener- gy, nutritious foods and a wide range of ES. Well-managed forests have a high potential to contribute to sustainable development and to a greener economy. Applications of ES mapping in forest management A successful multifunctional forest man- agement approach needs to consider the interests and needs of a great variety of ac- tors and sectors. In doing so, adequate tools, information and mapping of ES are needed to support policies and decision-making. In Europe, as an example, over 155 mil- lion hectares of forests are under manage- ment plans, representing over 70 % of the forest area in the region. Despite this, data sharing and adequate ES mapping for deci- sion-making is still lacking. The recent decision by European govern- ment leaders to increase the share of renew- able energy in Europe to 20 % by 2020 is expected to result in a much greater demand for forest biomass for bio-energy generation. This higher demand will intensify the com- petition for resources between forest indus- try, the energy sector and nature conserva- tion/other protective functions and services (including biodiversity, protection from nat- ural hazards, landscape aesthetics, recreation and tourism). This competition may lead to more intensive forest management such as plantation of fast-growing tree species, more frequent cuttings, shorter rotations and increasing export of coarse woody debris which has not traditionally been harvested. These increasing economic demands from society and complex relationships between humans and ES drive our actions towards the need for spatially explicit analysis and tools to map both the capacity of the eco- systems to deliver services to society and the societal demand for ES. Chapter 7 323 Data challenges for mapping forest-related ES The first challenges in quantifying forest ES involve having relatively comprehensive data on stand structure and composition (species composition, diameter distribu- tion, spatial distribution of trees) and, if possible, their dynamics (growth, mortali- ty and regeneration). These static and dy- namic data are indeed essential to provide information for ES indicators that may be relevant for producing maps to support management and planning. The acquisition of these data may be based on a dedicated device in situ (e.g. forest inventories, plot data at different levels, botanical surveys, surveys of forest companies, statistics for taxation) but also on remote sensing (RS) data and tools to give spatial form to the information. New RS developments such as very high resolution satellite imagery, LiDAR techniques that support the mea- suring of forest structure amongst other parameters, can really help to speed up the process of mapping at different scales. More satellite imagery is becoming available as open data, such as the imagery from the European SENTINELS. To improve forest ES mapping capabilities, current free and open data policy (i.e. RS data at different resolutions, large species data and open access forest inventory data) will have a dramatic impact on our ability to understand how forests are being affect- ed by anthropogenic pressures. We need to improve our knowledge on the status of forest systems which play key roles in trade- offs between provisioning ES supply and maintenance of, for example, carbon stocks, biodiversity and other related ES. In recent years, advances in working with different sensors (optical and non-optical sensors) at different resolutions are allowing work not just at finer resolutions but also for work on areas where cloud-cover was a problem (e.g. tropical forests, boreal forests). To collect indicator data in relation to forest habitat quality as an example, a number of information sources (besides more conven- tional data sources) exist today that are be- coming very popular such as citizen science (see Chapter 5.6.3); forest pedagogics proj- ects; the use of crowdsource information and social networks amongst others (see Chapter 5.5.3). It is also important to as- sess how changes to ecosystem management might alter the flow of ES either positively or negatively and who will be affected. Forest ES indicators A key aspect in the assessment of forest ES is the consideration of the long-term temporal dynamics of forest ecosystems that strongly determine ES capacities of the system. Con- sequently, indicators that provide informa- tion about forest ES supply need to be related to the ecosystem conditions, including infor- mation on age (ranges), tree species compo- sition and spatial distribution as well as stand density. The research project “RegioPower” (see example in Chapter 5.7.5) developed an approach for a combined assessment of typi- cal forest ES in a landscape context. Referring to case studies undertaken in this context in Finland, Germany, Slovenia and Sweden, making use of the CICES frame- work (see Chapter 2.4) and to approved frameworks for assessing sustainability of forests (MCPFE), we propose the following indicators to be adopted as shown in Table 1. Additional to these suggested indicators which mainly consider either the stand lev- el, or the level of forest areas/districts, the capacities of forest ecosystems to supply ES (see Chapter 5.1) are also greatly depen- Mapping Ecosystem Services324 dent on structural parameters at the (for- est) landscape level (see landscape metrics; Chapter 3.6). Benefits from cultural ES continue to be overlooked in many forest assessments be- cause of the many difficulties associated with measuring and mapping them. How- ever, cultural ES and other types of social values are often fundamentally important to understand how people use and value na- ture. Forests and woodlands play an import- ant cultural role and a number of spatially explicit methodologies have been developed which attempt to explore the values of their cultural ES. However, the current indicators of well-being linked to cultural and social values that are used in mapping approaches, if present at all, tend to be the more generic and easily quantifiable values. These include ES such as recreation, tourism and some aesthetic values. There is very limited repre- sentation of non-market ES such as spiritual connections with woodlands or emotional attachment to local places. This presents a significant barrier to understanding the wid- er societal benefits associated with wood- lands and similar green spaces. The result is that cultural and social values of woodlands are underestimated (see Chapter 5.5.3 for an alternative approach to assess cultural ES). Forest biodiversity and ES Forest biodiversity contributes to ecosystem functioning by maintaining a sustainable production of related forest ES. Therefore, losses of biodiversity can impose substantial costs at local and national scale, but many of the costs of changes in forest biodiver- sity have not been accounted for in deci- sion-making. Recognising the links between forest biodiversity and ES would help stake- holders to avoid biodiversity losses which lead to unacceptable ES shortfalls. Setting aside forest stands from commercial use reduces wood harvest possibilities and in- creases timber prices which affect forestry and forest industries. Employment opportunities in the forest sector are important for the rural population and the export income from the forest industry products are important for national economies. Although there are often not sufficient economic resources for the pro- tection of biodiversity yet, difficult choices on how to prioritise conservation need to be made. In order to support decision-making, integrative tools and analyses that simultane- ously consider the goals and economic im- pacts of conservation are needed. Section Division Class Indicators Provisioning Materials Biomass stand level / tree species level: stocking volume (m3 / ha); growth (m3 / ha x a); yield (m3 / ha x a) Regulation & Maintenance Maintenance of physical, chemical and biological conditions Global climate regulation GHG emissions / ha x a; above and belowground sequestered carbon; humus forms Cultural Spiritual, symbolic Symbolic abundance of rare species; number of above-average aged / thick single trees / breeding burrow trees, dead-wood stock (m3 / ha) Table 1: Examples for forest ES indicators according to the CICES scheme. Chapter 7 325 An integrated methodology, based on link- ing Bayesian Belief Networks (BBN; Chap- ter 4.5) with GIS is proposed in Box 1, for combining available evidence to help forest managers evaluate implications and trade- offs between forest production and conser- vation measures in order to preserve biodi- versity in forested habitats. Final considerations Future efforts should aim at improving mea- sures on the importance of forests for society at large. Therefore we need to improve our understanding of the people who live in and around forests – in many cases depending directly on forest ES for their livelihoods. BOX 1 . An integrated approach for forest ES mapping The approach for forest ES mapping, incorporates GIS-based data with expert knowledge to consider trade-offs between the biodiversity value for conservation and timber production potential with the focus on a complex mountain landscape in the French Alps. Figure 1. Forest trade-offs management: Target areas with high potential for intensification of for- estry practices (in brown, left side) as opposed to areas with conservation suitability potential (in green, right side). The final map in red represents areas showing conflicts (darkest red) in terms of trade-offs needed to balance interests of potential forest production and forest biodiversity conser- vation targets (from Gonzalez-Redin et al. 2016). Total Suitability High_Suitability_Production 5% Trade_off 55% High_Suitability_Conservation 37% No_Suitable 3% Timber Production Suitability Very Low (<0.10) Very Low (<0.21) Very Low (<0.19) Low (0.10 - 0.17) Low (0.21 - 0.39) Low (0.19 - 0.32) Moderate (0.17 - 0.28) Moderate (0.39 - 0.52) Moderate (0.32 - 0.42) High (0.28 - 0.42) High (0.52 - 0.66) High (0.42 - 0.56) Vert High (0.42<) Vert High (0.66<) Vert High (0.56<) Trade-offs Conservation Suitability Mapping Ecosystem Services326 Well-managed forests have a high potential to contribute to sustainable development and to promote food security. We need then stronger collaborative efforts to collect data and monitor trends, to raise awareness and monitor progress towards sustainable forest management. We need operational inte- grative methods to ensure spatially explicit mapping of complex forest ES to facilitate communication and planning adequate for- est management. Further reading Arnold FE, van der Werf N, Rametsteiner E (2014) Strengthening evidence-based forest policy-making: linking forest mon- itoring with national forest programmes. Forestry Policy and Institutions Working Paper 33. Rome, FAO. Fürst C, Frank S, Witt A, Koschke L, Make- schin F (2013) Assessment of the effects of forest land use strategies on the provision of Ecosystem Services at regional scale. Journal of Environmental Management, 127: 96-116. García-Nieto AP, García-Llorente M, Ini- esta-Arandia I, Martín-López B (2013) Mapping forest ecosystem services: from providing units to beneficiaries. Ecosystem Services 4: 126-138. Gonzalez-Redin J, Luque S, Poggio L, Smith R, Gimona A (2016) Spatial Bayesian be- lief networks as a planning decision tool for mapping ecosystem services trade-offs on forested landscapes. Environmental Research 144 Part B: 15-26. Luque S, Vainikainen N (2008) Habitat Qual- ity Assessment and Modelling for Biodi- versity Sustainability at the Forest Land- scape Level. pp. 241-264. In: Lafortezza R, Chen J, Sanesi G, Crow T (Eds) Patterns and Processes in Forest landscapes: Mul- tiple Use and Sustainable management – Part III Landscape-scale indicators and projection models. Springer publications. 370 pp. Kallio M, Hänninen R, Vainikainen N, Luque  S (2008) Biodiversity value and the optimal location of forest conservation sites in Southern Finland. Ecological Eco- nomics 67: 232-243. State of the World’s Forests (SOFO) (2014) Enhancing the socioeconomic benefits from forests. FAO, Rome. E-ISBN 978- 92-5-108270. Chapter 7 327 7.3.4. Nature protection Petteri Vihervaara, Joachim Maes, Laura Mononen, Constantin Cazacu & Mihai Adamescu Biodiversity, ecosystem functioning and ecosystem services Biodiversity, i.e. genetic, species and eco- systems diversity, is at the core of ecosystem services (ES). Their relationship is two-di- rectional. On the one hand, it is commonly stated that biodiversity underpins the deliv- ery of ES. Increasing species diversity is asso- ciated with enhanced ecosystem stability and productivity which, in turn, supports the delivery of multiple ES at higher production levels (Chapter 2.2). This is evident, for in- stance, in grasslands where processes such as ecosystem productivity or recycling of nutri- ents achieve higher rates if more species are present. The more species which are present in an ecosystem, the higher the probability that one of the species is very productive in delivering particular functions and particu- lar services. Similar observations are report- ed for forests or rivers where higher species richness is associated with higher potential and actual service delivery (Chapter 7.3.3). Knowing the relationship between species diversity and ES is useful for mapping ES. If certain habitats or species are key service providers, it is usually sufficient to map the distribution or presence of these species for mapping ES. This concept is also known as service providing areas (SPA; Chapter 5.2). It links habitats and species to the spatially explicit supply of ES by assigning different roles to service providers depending on their contribution in the delivery of ES. On the other hand, nature management targeted at maintaining or enhancing the delivery of ES may also improve the state of biodiversity. Thus, the assumption is that measures which increase the extent of ecosystems through land conversion or development of green infrastructure or measures improving the quality or the condition of ecosystems with the particu- lar aim of increasing ES, have a spill-over effect on biodiversity. More species would be able to profit from restored ecosystems or from new green infrastructure and this has a positive effect on overall biodiversity. There is indeed much scientific support for the positive relationships observed between biodiversity and ES. However, not all evidence points in the same direction. Some studies report nega- tive, no or weak correlations between bio- diversity and ES. Many species are rare and most species are very rare. This log-normal distribution of the relative abundance of species is used to describe biodiversity across different levels of taxonomic organisation, biomes, ecosystems or bio-geographical re- gions. Only few species dominate ecosys- tems or ecological communities. As a con- sequence, most flows of matter and energy are processed by a relatively small number of dominating species. This is very evident in croplands which farmers maintain in a particular state to maximise production by a single species, but it is also the case in nat- ural systems where few species deliver most of the services. Mapping Ecosystem Services328 Conservation as a management strategy Global targets for nature protection come from the Aichi targets of the Convention on Biological Diversity. The Aichi target 11 states that: “By 2020, at least 17 per cent of terrestrial and inland water and 10 per cent of coastal and marine areas, especially areas of particular importance for biodiver- sity and ecosystem services, are conserved through effectively and equitably managed, ecologically representative and well con- nected systems of protected areas and other effective area-based conservation measures and integrated into the wider landscapes and seascapes.” The Natura 2000 network in the European Union is one of the most comprehensive nature protection networks in the world – it protects around 18 % of land in the EU countries. It aims to protect valuable, endangered habitats and species all over the EU. It follows that often conservation approach- es, which usually target rare habitats or species, exist next to approaches based on adaptive ecosystem and ES management. Conservation as an approach to preserve biodiversity remains an important instru- ment in environmental policies and legis- lation (e.g. the Convention of Biological Diversity, the EU Biodiversity Strategy to 2020, national legislation). Conservation through the delineation and management of protected areas remains crucially important since most of the available evidence suggests that biodiversity continues to decline de- spite global efforts to stop biodiversity loss. Conservation targets vulnerable species and habitats and protects their status by protect- ing land from development such as urban- isation and agriculture. Climate change is shifting distributions of many species and new conservation tools are needed to adapt to these changes. Conservation is based on intrinsic values and humans have a moral obligation to share the planet with other species. Consequently, conservation mapping is based on mapping protected areas and nature reserves. Species distribution mapping and habitat mapping are however important tools for support- ing conservation. Species distribution and habitat mapping are usually based on field observations which are then up-scaled for instance through niche modelling using en- vironmental and climate data sets. The as- sumption is that species which are observed under a particular set of environmental con- ditions will also occur in places which are not monitored but which are characterised by the same conditions. Some well known software packages to model species distribu- tions include MAXENT and DIVA-GIS. Ecosystem service approach as a management strategy An ES-based management approach is fre- quently based on instrumental and social val- ues. It aims to conserve ES and restore natu- ral resources while, at the same time, meeting socio-economic and cultural targets. Often it complements conservation approaches, since the aim is not to protect vulnerable species but rather to ensure human well-being by enhancing ES. Evidently, this requires other mapping methods which are described in detail in this book (Chapter 5). The impor- tance of restored areas to support ES has also increased their socio-economic significance. Outside nature protection areas, this means, for instance, raising of concepts (e.g. green infrastructure) in land use planning to im- prove the state of biodiversity and increased ecosystem quality of the connection corridors between the more strictly protected areas. Ac- tually, there is a need to move from the lim- ited conceptual framework for nature pro- tection which only relies on protected areas Chapter 7 329 and move beyond this, towards a connected network of sites and, even more, facilitate the capacity of ecosystems to support ES outside the protected areas. There is cross-fertilisation between the two approaches of nature protection and ES management but sometimes they are also in conflict with each other. It is not always possible to use limited resources to preserve protected species while improving the ca- pacity of ecosystems to, for instance, store more carbon and contribute to climate change mitigation. An example of land use management in northern Finland demon- strates this dichotomy of forestry versus conservation (Figure 1). Synergies between different land uses can be improved, but first we need information about the effects of different management strategies on vari- ous ES. Mapping can help to reconcile con- servation values with instrumental or social values or at least help to understand where conflicting cases may occur so that appro- priate solutions can be found and proposed for policy-making and management. Overlaying maps used for conservation with maps of potential and actual ES is usually a first good approach to provide information for nature conservation managers. Besides protecting habitats and species, nature reserves usually have high capacity to provide a whole range of ES. In particular, regulating and cul- tural ES reach high levels in conservation ar- eas. In practice, ES maps can be overlaid on a map with nature reserves and zonal statistics are then used to derive values for ES which can be compared for selected places outside protected areas. Such information is usually of value for park managers as it can help make a business case for funding proposals. Figure 1. Forestry (upper part of maps) and conservation (below) are main land use types in northern Finland having opposite impacts on biodiversity and ES, for instance carbon storage which needs to be considered in management plans. Above-ground carbon stocks Tons of C/ha Biotope classes No Data Shallow lakes Sparsley wooded alpine heaths Underground brooks Regenerating scots pine-dominated heath forest Small streams and brooks Densley wooded deciduous forest Large rivers Densley wooded mixed forest ©Metsähallitus © LUKE Open hummock-level mires Mature sparsley wooded stocs pine-dominated heath forest Alpine deflation basins Border of protection area 0,01 - 10,00 10,01 - 20,00 20,01 - 30,00 30,01 - 40,00 40,01 - 50,00 Mapping Ecosystem Services330 Many researchers have also tried to compare maps of biodiversity with maps of ES in or- der to find synergies and trade-offs. Obvious- ly, there are always trade-offs between biodi- versity protection and, for instance, delivery of especially provisioning ES, such as food and timber (Figure 2). Due to ES trade-offs in land use and nature management, there can, at times, even be conflicts between the two. Nature protection areas provide an im- portant basis for developing ES maps due to readily available data sets which can support methodological improvements of mapping techniques. Areas where there is spatial con- gruence between biodiversity and ES could receive higher priority in management plans. Areas, where both biodiversity and ES are low, can be considered for development of more nature through green infrastructure projects. Areas, where ES are not or nega- tively related to biodiversity, could show that other approaches are necessary for sustainable management. For example, inclusive conser- vation (i.e. where priorities are directed to protecting biodiversity with the acceptance of low level disturbance), profit from regu- lating or cultural services such as recreation. Many of the tools described in other chapters such as social (Chapter 4.2) and participatory (Chapter 5.6.2) mapping techniques are used in such cases. Besides simple overlaying different maps to guide policy and management, optimisation software such as MARXAN and ZONA- TION are quite useful tools to assist land planning and for managers responsible for conservation, biodiversity or natural re- sources, such as forests or watersheds. These tools allow the choice from the best of both worlds and specifically target or select areas where win-win situations can be achieved. Provisioning S uppor ting, R egulatin g, Cultura l Ag ric ul tu ra l s ys te m s Q ua nt ity O ld g ro w n fo re st s W et la nd s Fo re st p la nt at io n System complexity services goods Energy subsidies (energy input from economic system) High production Low consumption inside system High transfer for human consumption Low service providers Maximisation of the transfer of high quality energy into services Energy needed for transformation Figure 2. Nature protection areas offer valuable sites to study trade-offs between different ES and natural conditions. Chapter 7 331 From policy to practice Bridging the gap between different approach- es of nature conservation and adaptive man- agement of ecosystems to enhance their ser- vice provision is key to biodiversity policy. A comprehensive mapping of all ES and better use of spatially explicit biodiversity data, sup- plementing species richness indicators with abundance and functional traits, will support biodiversity policy. However, it is of equal importance to mobilise financing to continue support for conservation while investing in ecosystem restoration and green infrastruc- ture. This requires using the best available spatial data to help investments in identified priorities so that they deliver multiple bene- fits in terms of biodiversity gains, ES, human well-being and climate adaptation. Intrinsic values and instrumental values to protect biodiversity and ecosystems need not be in opposition, although they do re- flect the hard choices that conservation of- ten faces. They can, instead, be matched to contexts in which each one best aligns with the values of the many audiences that we need to engage. Mapping these values by mapping biodiversity and ES can show what works and what fails in conservation and ecosystem management and thus reconcile different stakeholders. Further reading Sandifer P, Sutton-Grier A, Ward B (2015) Exploring connections among nature, bio- diversity, ecosystem services and human health and well-being: Opportunities to enhance health and biodiversity conserva- tion. Ecosystem Services 12: 1-15. Tallis H, Lubchenco J (2014) Working togeth- er: A call for inclusive conservation. Na- ture 515: 27-28. doi:10.1038/515027a. Egoh B, Reyers B, Rouget M, Bode M, Rich- ardson DM (2009) Spatial congruence be- tween biodiversity and ecosystem services in South Africa. Biological Conservation 14: 553-562. Maes J, Paracchini ML, Zulian G, Dunbar MB, Alkemade R (2012) Synergies and trade-offs between ecosystem service sup- ply, biodiversity and habitat conservation status in Europe. Biological Conservation 155: 1-12. Mapping Ecosystem Services332 7.4. Applying ecosystem service mapping in marine areas Nicola Beaumont, Katie Arkema, Evangelia G. Drakou, Charly Griffiths, Tara Hooper, Camino Liquete, Lida Teneva, Anda Ruskule & Anna-Stiina Heiskanen Introduction Accessibility and availability of spatially explicit information on marine ecosys- tem functions and ecosystem services (ES) are key components for successful marine management. As the uses and users of the marine environment increase in number and variety, there is a growing need for detailed Marine Spatial Planning (MSP), delineating spatial and temporal extents of different resource uses and the likely inter- actions of these uses, as well as impacts on the ecosystem and associated ES. In Europe, despite the new interest fostered by the Ma- rine Spatial Planning Directive or the Bio- diversity Strategy 2020, there are still very few initiatives for mapping marine ES at national or regional scales. Marine ecosys- tem service mapping is crucial for enabling sustainable marine resource use and is also equally important for ensuring successful marine protection through, for example, the designation of marine protected areas. In ac- cordance with the EU legal framework for marine protection and planning of sea uses (Marine Strategy framework Directive and MSP Directive), MSP can enable the imple- mentation of the ecosystem-based approach in management of human activities. This means that the collective pressure of hu- man activities should be kept within levels compatible with the achievement of good environmental status and that the capaci- ty of marine ecosystems to respond to hu- man-induced changes is not compromised, while enabling the sustainable use of marine goods and services by present and future generations. Mapping can provide informa- tion on integrated sustainable development and conservation with positive outcomes for ecosystems as well as people. Marine and coastal ES (MCES) mapping is still in its infancy (see Chapter 5.7.4) al- though several mapping studies have recently been undertaken. In most cases, these studies focus on mapping ES stocks and potential supply. However, in a few cases, it is has been attempted to associate marine ecosystems with the flow of benefits or the demand for them. This chapter explores the methods and data required to undertake a mapping exer- cise and how these vary depending upon the drivers of the mapping exercise, the scale of the study, the data available and the final use of the mapping by stakeholders. Drivers of mapping Mapping exercises may be driven by local communities (Box 1), local/regional policy and governance regimes (Box 2) or national/ international policy (Box 3). The aim of ES mapping may simply be to understand and highlight current ES provision and to pro- vide a baseline for future management strat- egies (Boxes 1 and 2), or an alternative aim Chapter 7 333 may be to produce Marine Spatial Plans to enable trade-offs between different uses and users, ensuring the balanced and sustainable use of the coastal and marine environment for human benefit both nationally and across the world (Box 3). In deriving the approach to mapping, it is essential to maintain clarity in the drivers and aims of the exercise and to ensure regular communication with the end users to ensure the final product is both fit for purpose and readily understood. As such, it is recommended that the aim and methods are clearly defined from the outset with expectations managed accordingly. Scale of mapping Mapping exercises can vary in scale from lo- cal (Box 2) to regional (Box 1) to national (Box 3). In some cases, a mapping exercise may be designed to explore a single ecosys- tem service whereas others may explore a host of ES (Box 2 and 3). The scope of the ES analysis will influence methods and data requirements. Thus, the objectives, scale and constraints of the analysis should be clear- ly defined at the outset. ES mapping on a larger scale may yield results of greater un- certainty than mapping on a smaller scale. Thus, when deciding the scale of the map- ping exercise, the end-user should be aware of this trade-off. Data availability In some cases, existing data may be suffi- cient for a particular mapping exercise (Box 2); however, in other cases, new data (Box 1) or a combination of primary and second- ary data (Box 3) may be necessary. In da- ta-limited contexts, practitioners often use habitat type as a proxy for ES supply (Boxes 2 and 3), especially in the case of regulat- ing services. There is however a high level of uncertainty associated with this approach and innovative methods for modelling ES are becoming more common. Surveys tend to be used to access additional information on provisioning and cultural services (Boxes 1 and 3). If surveys are undertaken, it is ad- visable that approaches which are used are participatory, emphasising the design and implementation by community members who are also resource-users. Data gaps and uncertainty The lack of empirical assessment of ES and their supporting habitats and attributes, remains a key challenge. Low resolution habitat data continues to be an issue at all levels, generating generalised service provi- sion maps at best (Box 3). The use of un- certain underlying information reduces the confidence in mapped outputs. As such, the communication of uncertainty and confi- dence is important in mapping ES (Chap- ter 6.3), to aid interpretation of the outputs by end-users (Box 2) and to ensure decisions are made with the full knowledge of poten- tial uncertainty in the underlying data. Stakeholder engagement Stakeholder engagement is essential for suc- cessful marine ES mapping, from defining the aim and parameters of the exercise, to providing data, context, ownership and val- idation. As explored in Box 1, the combina- tion of a participatory approach along with the mapping approach of provisioning and cultural ES allows for novel, informative and management-relevant maps of flow of benefits that help communities, especially those in collaborative management settings. To ensure stakeholders are engaged effec- Mapping Ecosystem Services334 tively, it is important to establish a two way dialogue throughout the process. Conclusions Under the present regulatory frameworks and the pressure to foster sustainable Blue Growth, it is crucial to undertake more ac- curate, policy-driven mapping of marine ecosystems and their services. Competing uses of marine resources should be analysed from a holistic perspective. ES maps should reveal the supply and demand of essential services across sectors and scales and should be co-developed and validated through iter- ative engagement with decision-makers, key stakeholders and the general public. A com- bination of methods is required to carry out MCES mapping, ranging from participatory mapping, stakeholder surveys, field measure- ments, to models. Care should be taken to ensure that the mapping exercise is well-de- fined at the outset with the aims, scope and scale agreed upon and the methods developed accordingly. The use of proxies and models can help to fill the data gaps until primary data can be attained, but uncertainty associ- ated with such data tends to be high. Key rec- ommendations should include the following: • Be fully aware of the reasons for the mapping exercise and active encourage- ment of stakeholder engagement at the start of the mapping process, including the use of local champions, to ensure that: i) the ES mapping is designed to meet stakeholder, policy-maker and practitioner needs; ii) the best available data is collected; iii) the outputs are us- able; iv) stakeholders can take owner- ship of the outputs. • Clearly define the scale of mapping at the outset and design the approach ac- cordingly. • Collect and share more spatially-explic- it data, ideally including low resolution data and with higher confidence levels. Data availability is still a limiting factor at all stages of marine ES assessments, from our understanding of the ecosys- tems and how they provide the ES, to the final social benefits and location of demand. Therefore national policy is recommended to actively promote the research on marine ecosystems in order to obtain more credible data on distri- bution of ES. • Improve accessibility to modelled infor- mation which is often highly technical. • Find ways of measuring and communi- cating uncertainty to stakeholders and end-users, as this is likely to be a signif- icant factor in all marine ES mapping. Chapter 7 335 Box 1 . From reef to table: Seafood security from community fisheries, Main Hawaiian Islands . A small Hawaiian community was interested in understanding the total biomass of their fisheries as well as the community dependency on the ecosystem as a food source in order to promote better local sustainable fishing practices and community management initiatives. Methods included field expert surveys, participatory mapping and data quantification. The reason for mapping the seafood catch benefit from Kīholo Bay across the island was to understand how this bay feeds the rest of the island and the magnitude of the food provisioning ES it provides. This study, involving collaboration between Conservation International, University of Hawaii and the community organisation, Hui Aloha Kīholo, mapped how seafood caught in Kīholo Bay travelled across the island and fed communities near and far. The location of people’s fishing activities was not discretely mapped, as fishing ground locations remain local knowledge and confidential. The ES which was mapped was essentially the seafood benefit in equivalent number of meals which were generated and also exported from Kīholo Bay. The methods used included fishermen’s surveys upon returning to shore and collecting data on species catch and size. Interviews with the fishermen revealed information on the end-users of the catch in order to assess the food miles (distance between the landing area and the place of consumption). The survey also investigated if catches were handled by the commercial sector or through non-commercial or not-for-profit activities. This single small-scale coastal fishery can provide more than 30,000 meals per year per square mile (2.6 km2) and represents nearly $80,000 in landed value (Figure 1). Approximately 90 % of the catch is consumed at home or given away as part of cultural practice. These fisheries provide a significant source of food security and economic security. The results from this study are likely to be used by the community to propose local legislation that would ensure a sustainable local subsistence fishery. Figure 1. Mapping the transport of a small reef fishery harvest in Kīholo Bay, Hawiian Islands, from the land zone to place of consumption. Quantities (kg) are depicted by the size of the pie charts which also indicate the type of transaction. Mapping Ecosystem Services336 Box 2 . Mapping ES provision and associated uncertainty in the Plymouth Sound to Fowey region, UK In the Plymouth Sound to Fowey region, UK, local marine managers requested maps of ES to enable understanding to be gained and communication about the current level of service provision, to provide a baseline against which future changes could be measured and to provide information for local policies and plans which include the Cornwall Maritime Strategy. This area comprises a range of marine habitats, supports diverse human uses and covers 934 km2, extending 22 km offshore. A variety of ES were mapped including carbon sequestration, water purification, fish nursery habitat, nutrient cycling, pollution immo- bilisation and sea defence. The mapping exercise combined local knowledge, expert knowledge, habitat data and published literature, into a series of maps using ESRI ArcGIS v10.2. As empirical assessments of ES within the case study were lacking, the habitat type was used as a proxy for service delivery using pub- lished literature to determine these relationships. In most cases, this resulted in a three-point qualitative scale (low, medium, high) representing the level of each service provided by each habitat. The fish nursery service was, however, considered in terms of the number of commercially important species utilising the habitat in their early life stages. A confidence scale was also provided for each service, based on the quality and quantity of the available data. Habitat data from a number of sources was used to produce habitat maps. These maps were then combined with the ES data and confidence information, allowing the map- ping of the level of service provision and confidence for each service. Figure 2. A map of carbon sequestration in the Plymouth Sound to Fowey region, UK. Chapter 7 337 Box 3 . Maritime Spatial Planning (MSP) for the Latvian territorial waters and the Exclusive Economic Zone Marine ES were mapped as an input for the Latvian national MSP. Areas significant for supply of provi- sioning, regulating and cultural services were mapped to avoid their deterioration when allocating space for new developments in the sea. Depending on data availability, different methodological approaches were used. Empirical assessments and spatial data on ES supply were available only for two provisioning services – wild animals and plants, including the catch of commercially important fish species (sprat, herring, cod, flounder) and red algae beds. The areas important for the fishery were mapped using data from fishery logbooks and visualised by calculation of the total value of fish catch and fishing acts within grid cells with a spatial resolution of 2.8 × 3 km2. The area covered by red algae beds was calcu- lated as a percentage of area unit based on actual field data from benthic habitat surveys. The potential supply of regulating services was mapped using benthic habitat data, expert judgement and indicators from literature. The habitat distribution map was used as a proxy for ES supply, including regulation of eutrophication processes, accumulation of pollutants in sediments, filtration by mussels, maintenance of nursery habitats and carbon storage. The ES distribution was presented in both individual maps and a summary map (Figure 3). The supply of a cultural service (tourism and recreation) was mapped using data on recreational options and their accessibility. The maps were a useful tool in assessing possible impacts of alternative development scenarios and deciding on optimum locations of new uses - offshore wind farms and marine aquaculture farms. The main limitation of the mapping approach was a lack of empirical survey data on habitat distribution, resulting in a low certainty level of the maps on regulating ES. Figure 3. Diversity of benthic habitat-related ES in Latvian marine waters. Legend 0-5 indicates the sum of services identified within each grid cell. Mapping Ecosystem Services338 Further reading Arkema K, Verutes G, Wood S, Clarke C, Canto M, Rosado S, Rosenthal A, Ruck- elshaus M, Guannele G, Toft J, Faries J, Silver JM, Griffin R, Guerry AD (2015) Improved returns on nature’s benefits to people from using ecosystem service mod- els in marine and coastal planning in Be- lize. Proceedings of the National Academy of Sciences 112 (24): 7390-7395. doi: 10.1073/pnas.1406483112. Kittinger JN, Teneva L, Koike H, Stamoulis KA, Kittinger DS, Oleson KLL, Conklin E, Gomes M, Wilcox B, Friedlander AM. From reef to table: social and ecological factors affecting coral reef fisheries, arti- sanal seafood supply chains and seafood security. PLoS One 10(8): e0123856. doi: 10.1371/journal.pone.0123856. Mandle L, Tallis H, Sotomayor L, Vogl AL (2015) Who loses? Tracking ecosystem ser- vice redistribution from road development and mitigation in the Peruvian Amazon. Frontiers in Ecology and the Environment 13(6): 309-315. Tallis H, Wolny S, Lozano JS, Benitez S, Saenz S, Ramos A (2013) “Service sheds” Enable Mitigation of Development Im- pacts on Ecosystem Services. Natural Capital Project. Potts T, Burdon D, Jackson E, Atkins J, Saun- ders J, Hastings E, Langmead O (2014) Do marine protected areas deliver flows of eco- system services to support human welfare?. Marine Policy 44: 139-148. Chapter 7 339 7.5. Business and industry Léa Tardieu & Neville D. Crossman Introduction The private sector has strong relationships with ecosystem services (ES). Business and industries receive benefits from ES but they can also have major impacts on ecosystems and ES delivery. ES degradation can have a significant impact on a company’s perfor- mance in sectors such as food production, construction, hydropower, tourism or bio- technology. There are very few examples of ES accounting used to support business management and decision-making. It is uncommon for firms to make the link between ecosystem manage- ment and financial performance and there is a general lack of understanding of the extent of firms’ dependence and impact on ecosystems. In some cases the exclusion is due more to a lack of guidance on how a company conducts such an analysis than to a lack of knowledge. A further complication is the public-good nature of ES and the absence of markets for many ES. As a consequence, many ES ben- efits/impacts are not represented in market prices. Land-use decisions by the private sec- tor tend to maximise only single objectives which may lead to a decline in other ES. There are several arguments for ES consider- ation in company decision-making, partic- ularly given the strong interactions between industry and ES and increasing consumer awareness of the contribution of ecosystems to well-being. Table 1 lists advantages of ac- counting for ES in business decisions. In this chapter we show how the inclusion of ES in business decision-making can im- prove company management and perfor- mance. We also show how ES mapping leads to more optimal land management decisions. We then highlight particular chal- lenges faced in mapping ES in the private sector and we present some examples from existing applications and case studies. Potential advantages Potential disadvantages Greening the company’s image Improving ES management Adaptation to novel techniques Respond to consumer demand for green products Produce life cycle assessment or environmental impact assessment accounting for ES Consideration by different investors and for bank loans grants Helps in demonstrating corporate sustainability. Determining more cost-effective investments Identifying new opportunities/risks Answer to legal regulations and eventually reduce taxes or become eligible for other financial incentives Develop leadership in considering ES New complementary tool for project design, enhancing project acceptability by strengthening existing approaches. - Cost and time consuming - Adaptation of ES analysis to existing tools - Availability of data - Uncertainty on the results - May need the collaboration with research partners - May reveal commercially sensitive information. Table 1. Potential advantages and disadvantages in accounting for ES in business and industry. Mapping Ecosystem Services340 Box 1 . Mapping ES for a transport infrastructure construction project in France ES maps have been used to assess ES loss caused by infrastructure construction in order to account for it in the project evaluation tools. The analysis proved to be a powerful complementary means of com- paring implementation options at different stages of environmental impact assessment (see Figure 1). It allows for the consideration of impacts otherwise overlooked, but also better targeting of mitigating measures. Further, since ES loss is expressed in monetary terms, the loss induced by the final selected route can be integrated as a standard social cost in the cost-benefit analysis, allowing a more efficient control of natural capital loss. ES mapping for business and industry By providing spatially explicit descriptions of ES, mapping can be used to evaluate business opportunities and to reduce risks for companies whose operations rely on nat- ural resources and ES. Mapping ES can improve decision support and evaluation tools commonly used in the private sector, such as environmental impact assessments (Box 1), lifecycle assessments, risk assessments, cost-benefit analyses (Box 2), land-use plans, or off-site mitiga- tion plans. Maps can be used to assess the impacts of alternative business decisions or courses of action on the location, quantity and value of ES. A company can also use ES maps to assess the direct, indirect and cu- mulative impact of their operations on ES, as well as how activities from other indus- tries affect their operations and profits. Modelling and mapping ES supply, in both biophysical and monetary terms, assists pri- vate sector decision-makers to locate ES de- livery hotspots or cold-spots. These types of maps allow a company to identify and then take advantage of ES benefits. By modelling scenarios of change, land use alternatives and the synergies and trade-offs between delivery of ES can be assessed in order to enhance the provision or the use of multiple ES. Maps and modelled ES scenarios are useful for monitoring consequences of different business investment strategies, improving resource management and/or determining and locating new opportunities for business investment (e.g. identifying best locations to offset carbon emissions or offset biodiversity impacts from infrastructure developments). Mapping can help reduce risks for companies Figure 1. ES mapping for infrastructure construction projects (Source: Egis, AULNES ©, based on Tardieu et al. 2015). Map of ES loss in preliminary studies (local climate change regulation service here) Overlay of multiple ES losses in preliminary studies ES loss analysis during implementation option comparison Chapter 7 341 that depend on ES (e.g. mapping flood dam- age risks for the construction sector). Mapping ES supply can identify potential foregone benefits (opportunity costs) in- curred by a business decision (e.g. foregone agricultural production). Opportunity cost maps can be used to spatially target locations for investment which are most cost effective (i.e. provide greatest returns for least cost). Locations of comparative advantage in ES supply can be identified and investment de- cisions can be made based on whether it is better to jointly generate multiple ES in a region or to specialise in one ES. This will help companies manage trade-offs in opera- tions, investments and management. Mapping ES values derived from beneficia- ries (in monetary or non-monetary terms), such as through a participatory GIS process (Chapter 5.6.2), can be used to identify ar- eas with ES benefits specific to economic sectors (e.g. tourism sector). By assessing and mapping the variation of these benefits ac- cording to different land uses, companies can estimate losses or gains from their operations (See Box 2 for an illustration) and they can target cost-effective risk adaptation or miti- gation measures (e.g. determining where to implement a fauna passageway at a new road infrastructure development). Table 2 lists ex- amples of the use of ES maps in business. Particular challenges in ES mapping for business and industry Spatially-explicit ES valuation is not sim- ple. The process requires multi-disciplinary expertise: environmental and ecological sci- ence, geographic information systems and socio-economics. However there are tools that companies can access to help map ES such as InVEST1 (Chapter 4.4), but these tools can be difficult to implement or adapt to private sector activities. Partnerships between companies and researchers are becoming more common for developing brand-friendly toolkits (e.g. AULNES2©, EarthGenome3) or platforms for advice, tools and techniques (e.g. Oppla4). A grow- ing number of initiatives to help the private sector in realising ES benefits are available, such as the Corporate Ecosystem Services Review Guidelines. 1 http://www.naturalcapitalproject.org/invest/ 2 http://www.climatesolutionsplatform.org/solu tion/aulnes 3 http://www.earthgenome.org/ 4 http://oppla.eu/ Box 2 . Lafarge example in the Presque Isle quarry, Michigan (Natural Capital Project, WRI and WWF) Lafarge is one of the largest construction materials companies in the world. InVEST was used to map and value two ES relevant to Lafarge’s operations on quarry sites: erosion control and water purification. ES mapping located areas where vegetation contributes to sediment retention and evaluated the monetary value of the service provided by avoiding dredging costs. It also identified areas where vegetation could be grown to reduce potential sedimentation of Lake Huron. The assessment of the water purification service by calculating the amount of nitrogen retained by the site has also been analysed. Subsequent economic valuation showed that Lafarge’s efforts to maintain vegetation provided a clear benefit by avoiding water treatment costs. Case study available at: http://www.wri.org/ sites/default/files/esrcasestudylafarge.pdf Mapping Ecosystem Services342 The major challenges can be classified into methodological and operational. The main methodological challenges are: i) defining and prioritising ES; ii) determining the type of impact of operations on ES; iii) modelling and mapping multiple ES in large areas and iv) dealing with the future (e.g. temporal trends, discount rate, evolution of ES pric- es). The main operational challenges are: i) the integration in existing evaluation tools; ii) the cost, time and resources required for such analysis; iii) the need for exhaustive as- sessments and precision of data for trade-offs and iv) the balance between scientific reliabil- ity and reproducibility. Note: Tardieu (2016) (reference below) should be consulted for ex- planation of these major challenges. Further reading Crossman ND, Bryan BA (2009) Identifying cost-effective hotspots for restoring natu- ral capital and enhancing landscape mul- tifunctionality. Ecological Economics 68: 654-668. Mandle L, Bryant BP, Ruckelshaus M, Genel- etti D, Kiesecker JM, Pfaff A (2015) Entry points for considering ecosystem services within infrastructure planning: how to in- tegrate conservation with development in order to aid them both. Conservation Let- ters 9(3): 221–227 Ruijs A, Kortelainen M, Wossink A, Schulp CJE, Alkemade R (2015) Opportunity cost estimation of ecosystem services. Environ- mental and Resource Economics: 1-31. Tardieu L, Roussel S, Thompson JD, Labar- raque D, Salles J-M (2015) Combining direct and indirect impacts to assess eco- system service loss due to infrastructure construction. Journal of Environmental Management 152: 145-157. Tardieu L (2016) Economic evaluation of the impacts of transportation infrastruc- tures on ecosystem services. Chapter 6, In Handbook on biodiversity and ecosystem services in impact assessment. In Genel- etti D (Ed). Edward Elgar, Cheltenham. Forthcoming, 113–139. Business sector Example of ES assessment and mapping potentially useful for the sector Forestry Mapping wood production for forest profitability versus provision of other ES (global climate regulation, recreation, regulation of water flows) to identify areas with comparative advantages Agriculture Mapping pollinators probability of presence and increase potential crop yields and revenues Aquaculture Assess and map different farming practices, location of farms in relation to climate change to determine how it affects harvests Water treatment by beverage producers Map pesticide diffusion and water purification performed by wetlands to minimise contamination of watersheds and identify how to manage upstream land sustainably Hydropower companies Map avoided erosion to identify land areas upstream that are important for erosion control and reduce the costs of removing sediment from reservoirs Transportation Map impacts on ES of alternative routes and identify best location for mitigation measures to increase probability of project approval Tourism Identifying risky areas to avoid when locating businesses or identify areas with particular recreational benefits Table 2. Example of ES maps of practical business relevance in different sectors. Chapter 7 343 TEEB (2012) The Economics of Ecosystems and Biodiversity in Business and Enter- prise. Edited by Joshua Bishop. Earthscan, London and New York. Hanson C, Ranganathan J, Iceland C, Finis- dore J (2012) The corporate ecosystem services review: guidelines for identifying business risks and opportunities arising from ecosystem change. World Resources Institute, Washington, DC. Mapping Ecosystem Services344 7.6. Mapping health outcomes from ecosystem services Hans Keune, Bram Oosterbroek, Marthe Derkzen, Suneetha M Subramanian, Unnikrishnan Payyappalimana, Pim Martens & Maud Huynen Introduction The practice of mapping ecosystem services (ES) in relation to health outcomes is only in its early developing phases. Air purifica- tion by vegetation and the resulting avoided respiratory disease burden is a health-related ES that is currently mapped for several areas in the world (see Figure 1 for an example in the United States). Another example is the attenuation of ocean waves by marine ecosystems and the subsequent reduction in population at risk from flooding. The latter is a health proxy as no connections are made to drowning. Of course, the value of other ES is approximated through maps as well, but map values are often biophysical rather than human health related. Table 1 lists sev- eral examples. ES - health mapping challenges When combining information about human health with information about ecological sys- tems - and with social complexity which is part of social ecological and environmental health systems - we not only combine complex in- formation which is different in nature, but we also combine scientific cultures containing a diversity of methodological approaches, data and evidence. We also need to make choices: we can never fully grasp nor take into account all potentially relevant complexity. This is not only just a matter of choice, it also has im- portant consequences for the quality of our outputs. Especially regarding the links be- tween nature and human health, “the devil is in the detail”: we need to take into account specific characteristics of nature and target groups whose health is affected. Here we in- troduce some specific challenges. First, ES supply and demand often relate to different spatial locations (Chapter 5.2). This is specifically relevant to health-related ES as they often benefit from close to the supply source. Due to the spatial explicit- ness of supply and demand, mapping is also a proper solution for this challenge. High resolution data are needed on, amongst others, the location of vegetation and the location of exposed people (e.g. places with a high population density). We also need to take into account different effects for differ- ences in vulnerability of different groups. Figure 1. Estimations of the annual number of asthma exacerbation cases that may be avoided due to total nitrogen dioxide removed by trees per census block group. (Shown here is Durham, North Carolina.) Adopted from EPA’s “EnviroAtlas Interactive Map”. Chapter 7 345 The second challenge is that health-related ES are often buffered or enhanced by socio-eco- nomic factors. In the case of flood protection, the effect of flooding on human casualties depends strongly on flood response pro- grammes and man-made structures to pre- vent flooding. A third challenge is the pres- ence of health-related ecosystem disservices which are perceived as harmful, unpleasant or unwanted. In several cases, these originate in the same ecosystem types and affect the same health outcomes as their ES counter- parts, but increase health burden. Examples of the latter are emissions of VOC (Volatile Organic Compounds), allergens and locally increasing air pollution concentrations and the potentially dual role of biodiversity in re- lation to infectious diseases. Several other challenges of mapping health-related ES are more ES-specific. For recreation, quantitative epidemiological ex- posure-response models are needed to link to health outcomes such as a reduction in depression. ES supply also depends on the ecosystem structure at micro scale such as vegetation type, height and density; dense shrubbery is effective for lowering noise lev- els, while clean and cool air is mainly pro- vided by trees. Most ES maps do not yet incorporate such spatial and thematic detail. Figure 2 shows a map which was built using high resolution spatial data that differentiate several vegetation types. The result is that the bundle of ES provided can differ substan- tially for districts within the same city, even when they are equal in terms of the surface area occupied by vegetation and water. Thus, to be able to map ES that moderate environ- mental risks to health on a city scale, detailed data of ecosystem types are needed. ES - health mapping design options Health indicators are necessary to make health outcomes spatially explicit and to assess health impacts. The choice of indi- Mapped ecosystem service Example indicator used Prevented health outcomes Air purification Air pollutant uptake (mass per area unit per year) Respiratory diseases, cardiovascular diseases, cancer Flood protection Reduced wave height, shoreline erosion Drowning, infectious diseases, mental disorders, respiratory diseases Biological control of infectious diseases Habitat suitability (index / categorical values, habitat presence likelihood) Infectious and parasitic diseases Noise reduction Reduced noise intensity (per area unit) Hearing loss, cardiovascular diseases Cooling Temperature reduction (per area unit) Heat stroke, heat exhaustion, mental disorders Recreation / provision of aesthetic values Index value, relative value, monetary value, number of visits (per area unit) Mental and behavioural disorders, cardiovascular diseases, obesity Medicinal plants and other medicinal resources Availability, associated traditional knowledge, threat status, volume of trade market value and non- monetary value Several conditions depending on species and associated knowledge Table 1. Examples of direct health-related ES that are currently mapped and provide promising starting points to assess health impacts Mapping Ecosystem Services346 cators and metrics depends on the specific research objective: if focussed on a single ES-related health outcome, then one spe- cific indicator can be used. Maps could then display avoided cases of a specific dis- ease (per area unit per year), avoided in- fectious disease outbreaks or areas where a health threshold value is exceeded (e.g. drinking water quality or noise intensi- ty threshold). However, if the objective is more integrative, for example, to calculate a region’s total (avoided) health burden or to assess an area’s net health effect (positive or negative), then an aggregate health indi- cator or common metric would give more useful insights. Such metrics to express the health effect of several health-related ES in a common unit are for example mortality, life expectancy, the disability adjusted life year (DALY), a monetary value (such as avoided costs of hospital visits) or the num- ber of affected people. Each comes with its own advantages and disadvantages. For example, mortality as an indicator would not include the effects of several non-lethal diseases and conditions with severe effects on well-being, whereas DALYs make use of disability weight factors (reflecting the severity of the disease) which are often dif- ficult to estimate. Additionally, some ar- gue that such integrative health indicators still fail to capture the full breadth of the complex linkages between biodiversity and health (including social determinants and cultural underpinnings) and that therefore a more holistic approach is necessary. Complexity often means making difficult methodological choices on what we need to take into account (and how). Hence, we also need to critically think about the process of methodological decision-making: who is in- volved in making those choices and whose knowledge, information and viewpoints are taken into account? In Western expert cul- ture, expert-driven mapping is still dominant. Mapping can also relate to processes that facilitate assessment of natural and human resources contributing to health and further strengthening them. The next section exem- plifies alternative approaches that include traditional local knowledge and participatory bottom-up mapping techniques relevant to health. The focus is on participatory assess- ment methods and tools that identify health- care delivery issues amongst local commu- nities and how these may be alleviated with resources from the proximate ecosystems. Participatory ES - health mapping The significance of ecosystem specific plants and other resources and related lo- Figure 2. Supply of ES bundles, aggregated to district level in Rotterdam, The Netherlands. Background colours depict total urban green and blue space (UGS) area. ES supply bundle 56 - 70 % 00 .5 12 34 42 - 56 % 28 - 42 % 14 - 28 % 0 - 14 % Carbon storage Noise reduction Cooling Recreation Total UGS area Chapter 7 347 cal traditional knowledge is much more profound for the health and nutritional se- curity of people in marginalised regions of the world in addition to their cultural rel- evance. Identifying local health priorities and supplementing them with ecosystem and community-specific traditional medi- cal knowledge and resources through pri- mary health programmes, is critical both to ensure conservation of biodiversity and health security at the local level. Important dimensions of participatory mapping and prioritisation of healthcare issues at the level of local communities are: 1) ranking of health challenges in a local communi- ty/region; 2) discourse-based mapping of traditional knowledge-based remedies for prioritised health challenges; 3) catalogu- ing medicinal biological resources and their availability in local communities; 4) mapping various other resources such as human-, sociocultural- and economic-pro- duced resources. In India, such rapid validation methodology is applied for determining effective commu- nity-based traditional medical knowledge practices. This is a rapid assessment as it involves no detailed laboratory or clinical studies on the efficacy of selected practices but depends on secondary literature reviews of revealed practices. Following an exhaus- tive documentation and prioritisation of health conditions, data obtained on local medicinal plant resources and associated knowledge in relation to the selected health conditions are matched. Subsequently, a detailed compilation of the global data on safety and efficacy of the selected remedy is done from various phytochemical, phar- macological and clinical literature. It also includes collecting exhaustive data from codified traditional medical systems of the region. Once the dossier has been prepared, a participatory assessment is conducted in the respective communities with involve- ment of various disciplinary experts. Each practice is discussed in detail, based primar- ily on a community’s historical experience of the traditional knowledge practice as well as the secondary literature on their safety and efficacy. These are made into comprehen- sive user manuals that are used to build the capacities of village health workers to pop- ularise the practices. Shortlisted plants are grown in nursery networks to be supplied for establishing home as well as community health gardens. Often participatory clinical cohort studies are conducted to examine efficacy of the selected practices from such local pharma- copeia. Several such participatory mapping and assessment of traditional knowledge programmes have been conducted across India and selected locations in Asia and Africa since 2008. For example, to tackle the onset of malarial infection, community mapping of traditional knowledge practic- es has been performed in endemic regions in India. Applying the above documenta- tion and participatory rapid assessment methodology, several location-specific prophylactic malaria remedies were select- ed for cohort clinical studies in order to explore their efficacy. The programme has demonstrated that significant health im- provements are possible through commu- nity level intervention using local resources and associated knowledge. Further information Interactive maps of health outcomes or health proxies: EPA, Enviroatlas Interactive Map: http://www2.epa.gov/enviroatlas/enviroat- las-interactive-map Coastal Resilience mapping portal: http://maps.coastalresilience.org/network/ Mapping Ecosystem Services348 Further reading Derkzen ML, van Teeffelen AJA, Verburg PH (2015) Quantifying urban ecosys- tem services based on high-resolution data of urban green space: an assessment for Rotterdam, the Netherlands. Jour- nal of Applied Ecology 52: 1020-1032. doi:10.1111/1365-2664.12469. Keune H et al. (2013) Science–policy chal- lenges for biodiversity, public health and urbanization: examples from Belgium, In: Environmental Research Letters, special issue Biodiversity, Human Health and Well-Being. Nagendrappa PB, Naik MP, Payyappallimana U (2013) Ethnobotanical survey of malar- ia prophylactic remedies in Odisha, India, Journal of Ethnopharmacology 146(3): 768-772. Oosterbroek B, De Kraker J, Huynen MMTE (2016) Assessing ecosystem impacts on health: A tool review. Ecosystem Services doi:10.1016/j.ecoser.2015.12.008. Pickard BR, Daniel J, Mehaffey M, Jackson LE, Neale A (2015) EnviroAtlas: A new geospatial tool to foster ecosystem services science and resource management. Ecosys- tem Services 14: 45-55. Raneesh S, Abdul H, Hariramamurthi BA and Unnikrishnan PM (2008) Docu- mentation and Participatory Rapid As- sessment of ethnoveterinary practices, Indian Journal of Traditional Knowledge 7(2): 360-364. http://nopr.niscair.res.in/ bitstream/123456789/1602/1/IJTK%20 7%282%29%20360-364.pdf. WHO – World Health Organisation (2006) Ecosystems and Human well-being: Health Synthesis – A report of the Millen- nium Ecosystem Assessment. WHO, Ge- neva. http://www.maweb.org/documents/ document.357.aspx.pdf. WHO & CBD Secretariat (2015) Connecting Global Priorities: Biodiversity and Human Health: a State of Knowledge Review: World Health Organisation. Chapter 7 349 7.7. Environmental security: Risk analysis and ecosystem services Adam Pártl & David Vačkář Introduction Various environmental drivers impact eco- systems and their capacity to provide ecosys- tem services (ES). The maintenance of this capacity influences the quality of human life and society at large. In a context of envi- ronmental change, environmental security is an important part of human and societal security. For instance, climate or land cover changes in ecosystems impact ecosystems and can lead to a loss of a wide range of ES, thus undermining the environmental secu- rity of human society. The Millennium Project defined environ- mental security as “environmental viability for life support, along with components that: a) prevent or remedy environmental damage; b) prevent or respond to environ- mental conflicts and c) protect the environ- ment due to its inherent moral value”. Socio-economic and ecological sustain- ability including a high quality of life thus depend on protecting ES and maintaining their provision, because they are responsible for the supply of natural resources - includ- ing water, land, energy and minerals. Increasing societal demands has altered the capacity to provide ES rapidly, even at a glob- al scale. This is notably illustrated with food production, for which 38 % of the land is now reserved (which also initiated the idea of the so-called Anthropocene as a new geo- logical era). Whereas agriculture has without doubt improved the quality of life, food pro- duction has resulted in negative externalities leading to the degradation of ecosystems and provision of their services (Chapter 7.3.2). Integrated risk analysis Often a relatively simple model is used for risk assessments with a single hazard focus: Risk = hazard x vulnerability; variations are possible depending on context and focus. In disaster risk science, the original pseu- do-equation has been further reworked and specified by adding the exposure dimension. Hazards are not considered as disasters when they occur on, for example, a deserted island as people nor property are affected. Vulner- ability can be defined as a certain sensitivi- ty or condition of environment, society and ecosystems to hazards which increase their susceptibility to the impacts. Vulnerability is determined by the potential for damage or disruption of ecosystems and human popu- lations through specific sources of risk. Both hazard and vulnerability are required to con- stitute a disaster. Exposure is the last part of the risk which reflects the people, property or ecosystems affected by hazards. We applied the disaster risk approach to as- sess the risk for losing ES in order to map Mapping Ecosystem Services350 the areas where the actual ES provision could be threatened by a combination of important hazards. Hazards are related to the ecosystem and consequently, to the services provision by their ability to impact their functioning, condition and quality. Consequently, the risk function was adjusted by adding the indica- tor of ES as for the exposure - to modify the equation for this specific case: R = H x V x ES. Thus, the risk is a function of hazard, vul- nerability and ES (Figure 1). Figure 1. Conceptual relation of the risk of the ES provision. Some examples of these include: erosion and floods can damage agriculture ecosys- tems and thus the provision of services; high nitrogen deposition hampers forest ecosys- tems; invasive species change the structure and biodiversity and pollution can cause the failure of aquatic ecosystems. All these different hazards can be included within the integrated multi-hazard approach. Clearly, risk drivers and their interac- tions with ecosystems are, in reality, more complex than suggested by this relatively simple approach. On the other hand, the multi-hazard approach can provide a quick overview of places which need more focus and where the combination of different haz- ards can lead to the decline of ES delivery. Risk Further reading Brown I, Ridder B, Alumbaugh P, Barnett C, Brooks A, Duffy L et al. (2011) Climate change risk assessment for the biodiversity and ecosystem services sector. Final Report to Defra - UK Climate Change Risk As- sessment 471: 51-57. Burkhard B, Kroll F, Müller F (2009) Land- scapes‘ Capacities to Provide Ecosystem Services – a Concept for Land-Cover Based Assessments. Landscape Online 15: 1-22. Collins TW, Grineski SE, de Lourdes Romo Aguilar M (2009) Vulnerability to environ- mental hazards in the Ciudad Juárez (Mex- ico)–El Paso (USA) metropolis: A model for spatial risk assessment in transnational context. Applied Geography 29: 448-461. Faber JH, van Wensem J (2012) Elaborations on the use of the ecosystem services con- cept for application in ecological risk as- sessment for soils. Science of the Total En- vironment 415: 3-8. Frélichová J, Vačkář D, Pártl A, Loučková B, Harmáčková ZV, Lorencová E (2014) In- tegrated assessment of ecosystem services in the Czech Republic. Ecosystem Services 8: 110-117. Liu X, Zhang J, Tong Z, Bao Y (2012) GIS- based multi-dimensional risk assessment of the grassland fire in northern China. Natural Hazards 64: 381-395. Xie H, Wang P, Huang H (2013) Ecological risk assessment of land use change in the Poyang Lake Eco-economic Zone, China. International Journal of Environmental Research and Public Health 10: 328-346. Wisner B, Blaikie P, Cannon T, Davis I (2004)  At Risk: Natural Hazards, Peo- ple’s Vulnerability and Disasters (2nd ed.) Routledge, New York. Risk Hazard Ecosystem services Vulnerability Chapter 7 351 Box 1 . Case study: Pilot risk assessment in the Czech Republic The risk approach was used in the Czech Republic at the national level. The study aimed to assess the risk of losing ES based on selected environmental hazards which play important roles in delivery of ES at the national scale within the Czech Republic. The analysis was undertaken in GIS with spatial data representing each risk component: hazards, vulnerability and ES. The hazards included erosion, nitrogen deposition, water pollution, floods, invasive species, urbanisation and contamination (based on mapping of old sites with brownfields, contaminated sites, long-term pollution spills, etc.). Human population density and ecosystem fragmentation together made up the vulnerability part. For example, high population density is linked with the highest demand for the ES providing benefits, for example, derived from regulating services for safety and risk reduction. The other part of vulnerability - sensitiv- ity of ecosystems to impacts from hazards - is represented by ecosystem fragmentation. ES values were based on monetary data (Euro per ha per year) from the pilot national services assessment. All data were standardised, unified to a common grid to enable direct calculations and overlaid to obtain a final risk layer (Figure 2). Figure 2. Final distribution of risk of losing ES in the Czech Republic (projection: S-JTSK / Krovak East North). Generally, over one third of the area was assessed with a low risk of losing ES. On the contrary, the highest risk values were in areas with designated formal nature conservation status (National parks and specially protected areas) showing the most valuable places at highest risk. This finding illustrates the importance of risk mapping to find out which areas need more and integrated focus and priority to mitigate the risk in order to maintain the high services provision. Mapping Ecosystem Services352 7.8. Mapping ecosystem services for impact assessment Davide Geneletti & Lisa Mandle Introduction Impact assessment (IA) processes aim to identify the future consequences of pro- posed actions to provide information for decision-making. Different types of IA exist, focusing on different topics (e.g. Environ- mental IA, Social IA, Health IA) or actions from individual projects to high-level policies (e.g. Regulatory IA, Policy IA, Strategic Envi- ronmental Assessment). The content of IAs is constantly evolving to reflect new perspectives and emerging issues and concerns. A case in point is the treatment of ecosystem services (ES), a cross-cutting theme which is increas- ingly included in different IA types, following the recent progress in literature and the de- velopment of guidance material. This chapter briefly describes the contribution of ES map- ping to IA and presents two illustrative appli- cations related to Strategic Environmental As- sessment of plans and Environmental Impact Assessment of projects, respectively. ES mapping across IA stages Even though IA processes differ widely and cannot be formatted into a standard se- quence of activities, most IA include the fol- lowing stages (not necessarily in this order): – Scoping and baseline analysis – Consultation – Developing alternatives – Assessing impacts of alternatives – Proposing mitigations During the scoping stage, ES mapping can be undertaken to select priority ES, i.e. the services that are most relevant for the ac- tion under analysis and the socio-ecological context. Priority services are of two types: the services upon which the action depends (e.g. tourism development requiring specif- ic cultural services to be profitable) and the services that the action will affect, positively or negatively (e.g. tourism development af- fecting storm regulation provided by coast- al ecosystems). Successful identification of priority ES requires understanding of the spatial relationship between the area affect- ed by the action, the area where the ES are produced and the area where they are used by beneficiaries. Hence, ES maps (even in a qualitative form) represent an essential in- put for this stage. During consultation, ES maps help to focus the debate and engage stakeholders. In addi- tion, participatory mapping exercises can be performed to better characterise key features of the local context and understand how ES are perceived and valued by different benefi- ciary groups (see Chapter 5.6.2). This infor- mation can be used to inform the subsequent development of alternatives, for example, by identifying “no-go” areas for specific activi- ties, suggesting priority locations for facilities or land-use conversions, etc. Concerning the assessment of the impact of different alternatives, spatial analysis allows impacts to be traced to specific beneficiaries. Chapter 7 353 It provides more explicit information that can be incorporated into environmental and social management plans, as compared to qualitative and non-spatial approaches, by illuminating where and how environmen- tal changes are affecting benefits to people. In this way, it also enables identification of more efficient mitigation options by bring- ing together environmental and social as- pects. In addition, by allowing tracking of benefits to specific people or groups of people, spatially explicit analysis provides the opportunity to ensure that development and any associated mitigation actions do not lead to the creation or extension of inequal- ity in service provision. All these aspects suggest that ES mapping can contribute to IA by reducing the like- lihood of plan or project delays due to un- foreseen impacts, reduce reputational risk to public authorities and developers from un- intended social impacts and improve overall outcomes of actions and mitigation. An application in Strategic Environmental Assessment This section exemplifies how spatial analysis of ES can be used to provide information for Strategic Environmental Assessment of urban plans. Particularly, it presents part of a case study related to the Urban Plan of the city of Trento (Italy). Amongst other things, the plan identifies sites for residential area development, mainly located within the existing urban fabric (Figure 1, left side). These sites consist of ninety-one vacant lots, with a surface area ranging from 1,000 to 5,000 m2. The purpose of the analysis is to use ES to support the selection of priority sites. Particularly, the analysis presented here focuses on the climate regulation service provided by green urban infrastructures. The cooling capacity of existing green urban infrastructure was estimated by applying a spatial model tailored to the local climate conditions, based on green areas characteris- tics, such as soil cover, tree canopy and size. Then, for each urban development site, the expected cooling capacity provided by the surrounding green infrastructures was cal- culated and classified into six classes (from A+ to D). This allows the sites to be ranked according to the thermal benefit that they are expected to receive, as shown in Figure 1 (right side). The results show that vacant lots which should be prioritised are, in general, the most peripheral and can be found both in the northern sector part of the city (at the borders of the green wedge that penetrates the built spaces) and in the southern sector (next to the surrounding wooded slopes). However, some vacant lots within the city centre also reach the highest level of ther- mal benefit provided by the surrounding green infrastructure due to the proximity to urban parks and water bodies. This applica- tion shows how ES mapping can be used to compare alternatives and identify priority interventions which represent typical tasks of Strategic Environmental Assessment of spatial and urban plans. An application in Environmental Impact Assessment In this section, we show how spatial analysis of ES can contribute to Environmental Im- pact Assessment for a proposed infrastruc- ture project, using the Peruvian portion of the proposed Pucallpa-Cruziero do Sul road between Peru and Brazil as a case study. We evaluate the likely impacts of the road on sev- eral ES provided to over 100 local communi- ties (Figure 2, centre) and determine where Mapping Ecosystem Services354 restoration has the potential to mitigate these ES losses (Figure 2, right side). We focus on carbon storage for climate regulation and sediment, nitrogen and phosphorous reten- tion for drinking water quality regulation. The combined direct and indirect impacts of the road were estimated by using a spa- tially explicit land use change model. Based on past trends, the model estimates where road construction is likely to spur conversion of forest to agriculture in the surrounding landscape. We then use the InVEST carbon, sediment retention and nutrient retention models (Chapter 4.4) to estimate how these services would change with road develop- ment and associated deforestation, account- ing for factors such as soil, climate and land use/land cover characteristics. We use the ES models to determine which population centres were likely to be affected and which services they would lose (Figure 2, centre). Changes in carbon storage affect climate reg- ulation services for everyone, due to circula- tion and mixing of the Earth’s atmosphere. In contrast, only those population centres that take their drinking water from places situated downstream of the road or its associated de- forestation, will experience a loss in drinking water quality regulation services. Then, to determine where and how restoration might mitigate these losses, we prioritise potential restoration sites in the surrounding area. The prioritisation was based on the ability of res- toration in each location to enhance carbon storage, sediment and nutrient retention and for these functions to benefit the same popu- lations affected by the road (Figure 2, right). The results show that population centres would lose between one and four ES, de- pending on the location of the population centre relative to the road and the projected land use change, as well as the characteristics of the intervening landscape. Potential res- toration sites in the south-western portion of the watershed are expected to return the greatest ES benefits to affected populations, although complete mitigation of ES losses is not possible in this case. This example shows how spatial ES analysis and mapping can be used as part of an Environmental Impact Figure 1. Sites for residential areas development (red dots) identified by the urban plan of Trento (left) and classification of the thermal benefits received by those sites (right). The first quintile include the sites which receive the lowest benefits. Source: Modified after Geneletti et al. 2016. Chapter 7 355 Assessment process, linking environmental change from project impacts and mitigation options to changes in benefits to people. Further reading Geneletti D (2015) A Conceptual Approach to Promote the Integration of Ecosystem Services in Strategic Environmental As- sessment. Journal of Environmental As- sessment Policy and Management 17(4): 1550035. Geneletti D, Zardo L, Cortinovis C (2016) Promoting nature-based solutions for cli- mate adaptation in cities through impact assessment. In Geneletti D (2016) (ed) Handbook on biodiversity and ecosystem services in impact assessment. Edward El- gar (Cheltenham, UK and Northampton, MA, USA), 428-552. Landsberg F, Treweek J, Stickler NM, Venn O (2013) Weaving ecosystem services into impact assessment. Washington, DC World Resource Institute. Mandle L, Tallis H, Sotomayor L, Vogl AL (2015) Who loses? Tracking ecosystem ser- vice redistribution from road development and mitigation in the Peruvian Amazon. Frontiers in Ecology and the Environment 13(6): 309-315. Mandle L, Tallis H (2016) Spatial ecosystem service analysis for environmental impact assessment of projects. In Geneletti D (2016) (ed) Handbook on biodiversity and ecosystem services in impact assess- ment. Edward Elgar (Cheltenham, UK and Northampton, MA, USA), 15-40. Sharp R, Tallis HT, Ricketts T, Guerry AD, Wood SA, Chaplin-Kramer R, Nelson E, Ennaanay D, Wolny S et al. (2014) In- VEST User’s Guide. Stanford, CA: Natu- ral Capital Project. UNEP (2014) Integrating ecosystem services in strategic environmental assessment: a guide for practitioners. A report of Proe- coserv. Geneletti D. United Nations Envi- ronment Programme, Nairobi. Figure 2. In Peru (left), population centres around the proposed Pucallpa-Cruzeiro do Sul road are be expected to lose climate regulation and drinking water quality regulation services (sediment, nitrogen and phosphorous retention services) with road development and associated land use change (centre). Potential ES mitigation areas (right) in surrounding watersheds can be prioritised by accounting for areas where restoration is both possible and would restore ES benefits to those impacted by road development. Source: Based on Mandle et al. 2015. Mapping Ecosystem Services356 7.9. The ecosystem services partnership visualisation tool Evangelia Drakou, Louise Willemen, Neville D. Crossman, Benjamin Burkhard, Ignacio Palomo, Joachim Maes & Michele Conti Introduction Data sharing and open access to informa- tion are key elements for successful spatial ecosystem service (ES) assessments. The development of the Ecosystem Services Partnership Visualisation Tool (ESP-VT) emerged from the aim of the ES community (namely the ESP Thematic Working Groups on Mapping and Modelling1 ES) to system- atically organise and publish ES maps and associated data for ES map users, the scien- tific community and the general public. The effort started in March 2013 and the alpha version was released in September of the same year. The ESP-VT was then tested by ES map-makers and practitioners and, after several modifications, the beta version was released in September 2015. ESP-VT comes as a complement to a range of already available tools and toolkits (see Chapter 3.4) which provide researchers with the possibility of conducting ES as- sessments, generating and sharing ES maps and data. Such tools can be classified into three broad categories: a) the data catalogue tools, allowing users to access catalogues of ES assessments and obtain an overview of previous research in the field (e.g. the MESP database2); b) the mapping and modelling tools, that allow users to enter their own data in an existing platform and conduct 1 http://www.es-partnership.org/ 2 http://marineecosystemservices.org/ their own ES assessments (e.g. the ARIES3 and InVEST4 toolkits that are widely used) and c) the combined tools, that combine functionalities of both (a) and (b), usually focusing on a specific ES (e.g. the Hugin OPENESS tool5 or the BioCarbon Tracker6; see also Chapters 3.4 and 4.4) or a specific ecosystem type (see Chapter 3.5). Within this plurality of tools, the ESP-VT was built to serve as a catalogue for ES maps. Within it, ES map-makers, map users and practitioners can find, access, view and share ES maps. This chapter briefly presents the ESP-VT, its functions, uses and actual and potential users. It describes the contribution of the ESP-VT to the ES mapping commu- nity, highlighting the benefits of data sharing. The Ecosystem Services Partnership Visualisation Tool (ESP-VT) The ESP-VT is an online platform available through esp-mapping.net that systematical- ly organises ES maps and makes them avail- able for the ES community. 3 http://aries.integratedmodelling.org/ 4 http://www.naturalcapitalproject.org/ 5 http://openness.hugin.com/gui 6 http://www.greenergy.com/Environment/biocar bon_tracker.html Chapter 7 357 The ESP-VT consists of: a) a database where all maps and metadata are stored and b) a map and data viewer which is the user in- terface. The database is structured using an adapted version of the ES mapping blueprint, devel- oped in 2013 as a first attempt to create a checklist for ES maps and models. The da- tabase systematically organises the ES maps metadata and the contextual background of the ES maps (e.g. purpose of the study, focal biomes, ES mapped). The ES data are cur- rently organised following the TEEB classi- fication system (see Chapter 2.4). Within the map and data viewer, the users can: i) search the database for available ES maps and data; ii) view and access maps and associated metadata within the viewer and iii) download the maps or data of interest. Registered users can also upload their ES maps and associated metadata. The latter are published online after a quality control check by the system administrator. Within the ESP-VT platform, users also have access to a user guide that allows them to understand the basic functionalities of the platform. More detailed documentation is also provided online. An overview of the tool functionalities is given in Figure 1. ESP-VT uses and users The ESP-VT is currently used by ES re- searchers to publish and share ES maps and associated metadata. ES maps resulting from initiatives and proj- ects such as EU MAES7 or ESMERALDA8 will be published through the platform. ESP-VT is also planned to store and visual- 7 http://biodiversity.europa.eu/maes 8 http://esmeralda-project.eu/ a)  Upload  data  and  metadata b)  View  maps  and  metadata c)  Search  the  database d)  Access  the  database Figure 1. The basic components of the ESP-VT. The central figure is the ESP-VT starting page. On the four corners, the captions of the different interfaces show the ESP-VT web component seen by the users when they: a) upload ES maps and metadata; b) view ES maps and metadata; c) search the database and d) access the database. Mapping Ecosystem Services358 ise maps and data published within the new open access data journal One Ecosystem9. ESP-VT also serves as an ES map repository that allows researchers to search for relevant ES mapping efforts, methodologies and data used. In the future, with more functional- ities added to the ESP-VT, users will be able to perform spatial queries and/or analysis within the maps stored in the database. The ESP-VT is designed to go beyond be- ing a tool just for the scientific community. It can be easily used by practitioners, urban planners and the general public who might require information on how ecosystem ben- efits are distributed in an area of interest. ESP-VT is built using the principles of open access and data sharing, thus allowing local experts (upon registration) to comment and validate the quality and accuracy of the pub- lished information. Lessons learnt and future visions The major challenges faced during the ESP- VT development were: a) the heterogene- ity among ES mapping approaches; b) in- creased complexity of the ESP-VT as new functionalities were included. a. The heterogeneity among ES mapping approaches is an aftermath of a plurali- ty in ES classification systems, tools and methods used to produce ES maps, units and visualisation methods. This is relat- ed to the different purposes for which ES maps were constructed: to answer differ- ent questions; for different users, like ES practitioners, policy makers or the gen- eral public (see Chapter 7). 9 http://oneecosystem.pensoft.net/ b. Increased complexity of the ESP-VT as new functionalities were included. The database of the ESP-VT is populated with ES maps by ES map-makers. Its contribution to information-sharing is therefore based on the willingness of re- searchers to share their outputs with the ES community of practice. So far, data standards on biome types and quantification units are used to organise the heterogeneous data populating the ESP- VT. To structure ES information, ESP-VT follows the TEEB classification. The com- munity of ES researchers and practitioners agrees that there is no “one-size-fits-all” ES classification system and that local or re- gional specificities should be taken into ac- count. The OpenNESS glossary10 can allow ES researchers to “translate” ES to other ES classification systems (see also Chapter 2.4). On the other hand, establishing ES stan- dards, populating the ESP-VT with maps and making these ES maps accessible to all under the open data sharing principles will: a) maximise research efficiency by avoiding replication of errors and duplication of ef- forts; b) allow for “self-correction” within the ES research community; c) open the door to innovation, synthesis work and fu- ture research and d) allow for inter-opera- bility and hence free flow of information among other ES-related tools and toolkits. Lastly, initiatives like ESMERALDA and One Ecosystem should boost the interest of the research community towards sharing in- formation on ES maps through the ESP-VT platform. As more initiatives are added, the development and impact of ES maps will improve. 10 http://openness.hugin.com/example/cices Chapter 7 359 Further reading Crossman ND, Burkhard B, Nedkov S, Wil- lemen L, Petz K, Palomo I, Drakou EG, Martín-Lopez B, McPhearson T, Boyano- va K, Alkemade R, Egoh B, Dunbar MB, Maes J (2013) A blueprint for mapping and modelling ecosystem services. Ecosys- tem Services 4: 4-14. Drakou EG, Crossman ND, Willemen L, Burkhard B, Palomo I, Maes J, Peedell S (2015) A visualisation and data-sharing tool for ecosystem service maps: Lessons learnt, challenges and the way forward. Ecosystem Services 13: 134-140. Pagella TF, Sinclair FL (2014) Development and use of a typology of mapping tools to assess their fitness for supporting man- agement of ecosystem service provision. Landscape Ecology 29: 383-399. Chapter 8 361 Chapter 8. Conclusions Joachim Maes & Benjamin Burkhard Mapping ecosystem services (ES) has de- veloped over the past years into a mature scientific field. That much is clear from this book and other publications, research and ongoing related activities. Many researchers involved in ES mapping projects can count on much attention and sessions on mapping ES during scientific conferences invariantly attract many participants. There are a number of good reasons why mapping ES has come of age. Firstly, different policies and, in particular, global biodiversity policy have embraced the concept of ES in their strategic planning and development. Following the publication of the Millennium Ecosystem Assessment in 2005, different levels of government from local to global scale have then started to use the concept of ES as a bridge between nature and society. However, concepts need to be underpinned by evidence based on sound data and suitable methods in order to be relevant and reliable in the long term. This has, for example, been made clear in the EU Biodiversity Strategy to 2020 which calls ex- plicitly for mapping ES at national scales. ES maps are recognised as tools to help policy and decision-making, to monitor implemen- tation of policy and decisions and to provide baseline information against which change or progress to targets can be assessed. A second important reason for understand- ing the success of ES mapping is the applica- bility of maps for different user groups. De- mand for spatially explicit ES data is spurred by conservation managers, urban and land- scape planners, regional development, busi- ness sectors, marine spatial planners as well as different consulting or executive agencies which help local, regional and national gov- ernments with all aspects of natural resource management. ES maps are not only power- ful tools to communicate messages related to land use trade-offs, but they also simply provide the essential data which are crucial to mainstream biodiversity, ecosystems and ES into policy and decision-making. Of particular relevance is the ability to map ES bundles or to illustrate ES trade-offs which arise between competing sectors such as, for instance, forestry and agriculture. It must be clear that mapping ES is not a demand-driven activity alone. Mapping ES addresses critical scientific questions includ- ing the impact of local or regional policy decisions on biodiversity and ecosystems not only at the actual location but also in other places. Mapping ES supply, flow and demand in a spatially explicit manner can provide essential information to understand the consequences of such decisions. Under- standing ecosystem conditions, including spatial structures, processes and their spa- tio-temporal interactions on different scales, is essential for sustainable management of natural resources. Further degradation of natural capital and the biodiversity base will have significant impacts on ES supply and human well-being for today’s and, especial- ly, for future generations. Mapping ES is founded in geography, land- scape ecology and further related disciplines and it profits from the available knowledge base and the ever-increasing importance of open access spatial data, GIS platforms and multi-dimensional data visualisation in our society. The potential for mapping ES Mapping Ecosystem Services362 to bring different scientific disciplines to- gether in one framework while also reach- ing out to other scientific disciplines such as economy and social sciences, is one of the most appealing but also challenging aspects of the ES concept. Many problems have a spatial nature. Mapping ES offers a framework for combining spatial data and trans-disciplinary knowledge of different sources. More and more quantified ecolog- ical data on species, biodiversity and eco- system processes is combined with expert knowledge through participatory mapping. It demonstrates that mapping ES embraces stakeholders of different backgrounds and that expert- and citizen-based values are not ignored. This is particularly relevant for in- clusion of ES that are difficult to map into, for example, the planning process. The research progress of ES mapping can be inferred from the wide variety of methods, tools and models which have become avail- able. Models and tools for mapping come with different complexity levels, data needs and uncertainties; they are available for dif- ferent spatial and temporal scales and target different user communities. Many of these are illustrated in this book. Often, all these mapping methods, tools and models share their strong dependency on land cover and land use data. These data sets are now readily available, frequently for several points in time and open access and provide a crucial data foundation for mapping ES. They are used throughout this book as an underlying data source to many of the published maps. Nevertheless caution is needed when using land cover and land use data. Errors and un- certainty with respect to land cover and land use data are often unquestioned by research- ers, mainly due to their easy access and ap- plicability. Furthermore, ecosystems are not synonymous with land cover and ES are sup- plied by ecosystems, not by land cover types. The ecology of boreal forests in Sweden is, for instance, quite different from that of a tropical rainforest; yet these differences can fade on land cover maps. Besides land cover and land use, other parameters are essential determinants to control the flow of ES. Soil properties, water availability, local species diversity and climatic variability are import- ant co-variables which should be considered when mapping ecosystems and thus also ES. Clearly, one of the challenges for the next generation of ecosystem (service) map-mak- ers is better mapping of different ecosystem and habitat types. Uncertainty of ES maps has other sources as well. As well illustrated by the ES cascade model, ES flow from nature to society. Map- ping the different components which con- stitute ES introduces errors which may be propagated along the ES cascade. More sci- entific rigour does no harm and may come from natural capital accounting. Several initiatives of a consistent quantification of ES are ongoing. The ultimate goal is to set up a system which is comparable to the sys- tem of economic accounts. This would re- quire a rigorous and validated mapping ap- proach resulting in the regular publication of geo-referenced ES data. Such data need to be accompanied by uncertainty measures giving information about the reliability of each used variable. Even if questions about uncertainty are per- tinent and justified, this does not curtail the wide application of ES maps by different sectors. This book presents a great deal of evidence for this. ES maps are being used, for example, in urban planning, agricul- ture, forestry and nature conservation. The business sector also adopts this approach. A promising avenue for application of ES mapping is related to health issues. Where- as monetary valuation of ES is often con- troversial, human and public health is less Chapter 8 363 so. Maps help demonstrate how ecosystems can reduce exposure to pollutants or envi- ronmental risks such as flood hazards and thereby provide tangible benefits which can be well-understood by policy makers and the public. Using ecosystems and ES to ad- dress important challenges with respect to planning, resource use and public health is now coined as nature-based solutions. They combine innovation with sustainability and are based on a thorough knowledge of eco- system processes, functions and services. It follows that ES mapping will remain an es- sential research activity to support a sustain- able future. The ongoing data revolution, driven by en- hanced earth observation techniques and by the ever-increasing availability of open, large, digital data, will be part of this future. There are enormous opportunities for ES mapping research to profit from this devel- opment. High-resolution data of land, wa- ter, biodiversity and ecosystems, obtained from remote sensing, offer the possibility to map ecosystems in a more accurate way and to assess trends over time. Validation should increasingly depend on the capacity of indi- vidual people to monitor the environment and to share their observations. More work is needed to base ES maps on existing and new sources of data and to integrate these maps in consistent and regularly updated account systems to support decisions at dif- ferent levels, across different sectors and in the long term. In this sense, this book is not only a synthe- sis of the state-of-the-art of ES mapping but it provides a comprehensive overview and guidance for those mapping ES themselves or for those using ES maps. Glossary 365 Abiotic: Referring to the physical (non-living) environment, for example, temperature, moisture and light, or natural mineral substances [Modified from Lincoln et al. (1998: 1)] Agro-ecosystem: An ecosystem, in which usu- ally domesticated plants and animals and other life forms are managed for the pro- duction of food, fibre and other materials that support human life while often also providing non-material benefits. Aquaculture: Breeding and rearing of aquat- ic organisms (fish, molluscs, crustaceans and aquatic plants) in ponds, enclosures, or other forms of confinement in either fresh or marine waters for direct harvest of the product [Adapted from MA (2005), extended by FAO yearbook Fishery and Aquaculture Statistics (2011)]. Assessment: The analyses and review of infor- mation derived from research for the pur- pose of helping someone in a position of responsibility to evaluate possible actions or think about a problem. Assessment means assembling, summarising, organis- ing, interpreting and possibly reconciling pieces of existing knowledge and commu- nicating with an appropriate person so that they are relevant and helpful to the intelli- gent but inexpert decision–maker [Parson (1995), taken from MAES (2014)]. Bayesian [Belief ] Network (BBN): A prob- abilistic graphical model for reasoning under uncertainty, consisting of an acy- clic, directed graph describing a set of dependence and independence properties between the variables of the model repre- sented as nodes and a set of (conditional) probability distributions that quantify the dependence relationship [Adapted from Kjærulff & Madsen (2013)]. Beneficiary: A person or group whose well-be- ing is changed in a positive way by (in this case) an ecosystem service. Benefits (derived from ES): The direct and indirect outputs from ecosystems that have been turned into goods or experiences that are no longer functionally connected to the systems from which they were derived. Ben- efits are things that can be valued either in monetary or social terms [OpenNESS]. Biodiversity: The variability amongst living organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diver- sity within species, between species and of ecosystems. Biodiversity is a contraction of ‘biological diversity’ [CBD]. Bioenergy: Renewable energy made available from materials derived from biological sources. Biomass: The mass of living organisms in a population, ecosystem, or spatial unit de- rived by the fixation of energy though or- ganic processes [Common usage and MA (2005)]. Biome: The largest unit of ecological classifica- tion that is convenient to recognise across the entire globe. Terrestrial biomes are typ- ically based on dominant vegetation struc- ture (e.g. forest, grassland). Ecosystems, Glossary Terms in this Glossary are based on different sources (as indicated); most terms were taken from the OpenNESS project [Potschin M, Haines-Young R, Heink U, Jax K (Eds) (2016) OpenNESS Glossary (V3.0). Grant Agreement No 308428, available from: http://www.openness-project.eu/ glossary ] and the ESMERALDA project [Potschin M, Burkhard B (2015) Glossary for Ecosys- tem Service mapping and assessment terminology. Deliverable D1.4 EU Horizon 2020 ESMER- ALDA Project. Grant agreement No. 642007, http://esmeralda-project.eu/documents/1/ ]. Mapping Ecosystem Services366 within a biome, function in a broadly sim- ilar way, although they may have very dif- ferent species composition. For example, all forests share certain properties regard- ing nutrient cycling, disturbance and bio- mass that are different from the properties of grasslands. Marine biomes are typically based on biogeochemical properties. The WWF biome classification is used in the MA [MA (2005)]. Biophysical Structure: The architecture of an ecosystem that results from the interaction between the abiotic, physical environment and organisms or whole biotic communi- ties [Modified MA (2005)]. Biophysical Valuation: A method that derives values from measurements of the physical costs (e.g. in terms of labour, surface re- quirements, energy and material inputs) of producing given goods or a service [TEEB]. Capacity Building: A process of strengthen- ing or developing human resources, insti- tutions, organisations or networks. Also referred to as capacity development or capacity enhancement [UK NEA (2011)]. Carbon Sequestration: The process of in- creasing the carbon content of a reservoir other than the atmosphere [MA (2005)]. Cartography: The art and science of represent- ing geographic data by geographical means. Classification System [for ES]: An organised structure for identifying and organising ES into a coherent scheme [Common usage]. Choropleth Map: Used to map data collected for areal units, such as states, census areas or eco-regions. Their main purpose is to provide an overview of quantitative spa- tial patterns across the area of interest. To construct a choropleth map, the data for each unit is aggregated into one value. Ac- cording to their values, the areal units are typically grouped into classes and a colour is assigned to each class. Conservation: The protection, improvement and sustainable use of natural resources for present and future generations. Coordinate System: It is used to define the positions of the mapped phenomena in space. Furthermore, it acts as a key to com- bine and integrate different datasets based on their location. Cost-Benefit Analysis: A technique designed to determine the economic feasibility of a project or plan by quantifying its econom- ic costs and benefits [MA (2005)]. Cultural Ecosystem Service (CES): All the non-material and normally non-con- sumptive outputs of ecosystems that affect physical and mental states of people. CES are primarily regarded as the physical set- tings, locations or situations that give rise to changes in the physical or mental states of people and whose characters are funda- mentally dependent on living processes; they can involve individual species, habi- tats and whole ecosystems [CICES]. Decision-maker: A person, group or an organi- sation that has the authority or ability to de- cide about actions of interest [MA (2005)]. Disservice: Negative contributions of ecosys- tems to human well-being; undesired neg- ative effects resulting in the degeneration of ecosystem services [after OpenNESS, modified TEEB]. Ecological Process: An interaction amongst organisms and/or their abiotic environ- ment [shortened from Mace et al. (2012)]. Ecological Status: A classification of an eco- system state amongst several, well-defined value categories. [after Maes et al. (2013)]. Ecosystem: Dynamic complex of plant, an- imal and microorganisms’ communities and their non-living environment interact- ing as a functional unit. Humans may be an integral part of an ecosystem, although the expression ‘socio-ecological system’ is sometimes used to denote situations in which people play a significant role, or where the character of the ecosystem is heavily influenced by human action. [Modified MA (2005)]. Eco-agri-food System: An interacting com- plex of ecosystems, agricultural lands, in- Glossary 367 frastructure and markets playing a role in growing, processing, distributing and con- suming food. Ecosystem Accounting: The process of or- ganising information about natural capital stocks and ecosystem service flows, so that the contributions that ecosystems make to human well-being can be understood by de- cision-makers and any changes tracked over time. Accounts can be organised in either physical or monetary terms [OpenNESS]. Ecosystem Assessment: A social process through which the findings of science con- cerning the causes of ecosystem change, their consequences for human well-be- ing and management and policy options are brought to bear on the needs of deci- sion-makers [UK NEA (2011)]. Ecosystem Capacity: Ecosystem capacity re- fers to the ability of a given ecosystem (or ecosystem asset) to generate a specific (set of ) ecosystem service(s) in a sustainable way for the future [Based on SEEA-EEA]. Ecosystem Condition: The physical, chemi- cal and biological condition of an ecosys- tem at a particular point in time. For the purpose of mapping ES, ecosystem condi- tion is, however, usually used as a synonym for ‘ecosystem state’ [EEA (2016)]. Ecosystem Function: The subset of the in- teractions between biophysical structures and ecosystem processes that underpin the capacity of an ecosystem to provide ecosys- tem services. See ecosystem capacity and ecosystem condition [OpenNESS]. Ecosystem Functioning: The operating of an ecosystem. Very often, there is a normative component involved, insofar as ecosystem functioning not only refers to (any) func- tioning/performance of the system but also to ‘proper functioning’ and thus im- plies a normative choice on what is consid- ered as a properly functioning ecosystem (operating within certain limits) [Based on Jax (2010)]. There are many ways in which this is assessed and conceptualised, for ex- ample, as good ecological status, ecosystem health, ecosystem integrity, or implied by the desired state of ecosystem services de- livered by the systems. When using ecosys- tem functioning, the emphasis should be on the overall performance of the system and not so much on selected processes or purposes. Ecosystem Integrity: This is often defined as an environmental condition that exhibits little or no human influence, maintaining the structure, function and species compo- sition present, prior to, and independent of, human intervention [i.e. integrity is closely associated with ideas of natural conditions, particularly the notion of pristine wilder- ness [after Angermeier and Karr (1994), Callicott et al. (1999), Hull et al. (2003)]. Ecosystem Process: A dynamic ecosystem characteristic that is essential for the eco- system to operate and develop. Examples of ecosystem processes are fluxes of nutri- ents and energy (production and decom- position) and characteristics determining population dynamics, such as seed dis- persal and migration. (See also ecosystem structure and biophysical characteristic) [OpenNESS]. Ecosystem Properties: Attributes which char- acterise an ecosystem, such as its size, bio- diversity, stability, degree of organisation, as well as its functions and processes (i.e. the internal exchanges of materials, energy and information amongst different pools) [MA (2005) and UK NEA (2011)]. Ecosystem Services (ES): These are the con- tributions of ecosystem structure and function – in combination with other inputs – to human well-being [after Bur- khard et al. (2012)]. Ecosystem State: The physical, chemical and biological character of an ecosystem at a particular point in time [OpenNESS]. Ecosystem Structure: A static characteristic of an ecosystem that is measured as a stock or volume of material or energy, or the com- position and distribution of biophysical elements. Examples include standing crop, Mapping Ecosystem Services368 leaf area, percentage ground cover, species composition [OpenNESS]. Environmental Accounting: See term ‘Natu- ral Capital Accounting. ES Bundle (supply side): A set of associated ES that are linked to a given ecosystem and that usually appear together repeatedly in time and/or space [OpenNESS]. ES Bundle (demand side): A set of associated ecosystem services that are demanded by humans from ecosystem(s) [OpenNESS]. ES Mapping: The process of creating a car- tographic representation of (quantified) ecosystem service indicators in geographic space and time. ES Model: A scientific (usually comput- er-based) for quantifying various so- cio-ecological indicators of an ecosystem service. ES Potential: This describes the natural con- tributions to ES generation. It measures the amount of ES that can be provided or used in a sustainable way in a certain region. This potential should be assessed over a sufficiently long period of time. ES Supply: The provision of a service by a par- ticular ecosystem, irrespective of its actual use. It can be determined for a specified period of time (such as a year) in the pres- ent, past or future. ES Flow: A measure for the amount of ES that are actually mobilised in a specific area and time. It includes a dynamic temporal di- mension and conceptually links ES supply with demand. ES Demand: The need for specific ES by society, particular stakeholder groups or individuals. It depends on several factors such as culturally-dependent desires and needs, availability of alternatives, or means to fulfil these needs. It also covers prefer- ences for specific attributes of a service and relates to risk awareness. Forestry: The science, art and practice of man- aging and using trees, forests and their as- sociated resources. Generalisation (map): This aims to represent the ES-information on a level of detail ap- propriate for a given scale, user group and use context. It is necessary in cases where the visual density in maps is increasing too rapidly, symbols overlap or topological conflicts become evident due to graphical scaling. Geographic Information System (GIS): A computer-based system for the Input, Management, Analysis and Presentation (IMAP) of spatially referenced data. Goods: The objects from ecosystems that peo- ple value through experience, use or con- sumption, whether that value is expressed in economic, social or personal terms. Note that the use of this term here goes well beyond a narrow definition of goods simply as physical items bought and sold in markets and includes objects that have no market price (e.g. outdoor recreation). The term is synonymous with benefit (as proposed by the UK NEA) and not with service (as proposed by the MA). Green Infrastructure (GI): A strategically planned network of natural and semi-nat- ural areas with other environmental fea- tures designed and managed to deliver a wide range of ES. It incorporates green spaces (or blue if aquatic ecosystems are concerned) and other physical features in terrestrial (including coastal) and marine areas. On land, GI is present in rural and urban settings [EC (2013)]. Habitat: The physical location or type of envi- ronment in which an organism or biologi- cal population lives or occurs. Terrestrial or aquatic areas distinguished by geographical, abiotic and biotic features, whether entirely natural or semi-natural. Note the Council of Europe definition is more specific: the habitat of a species, or population of a spe- cies, is the sum of the abiotic and biotic fac- tors of the environment, whether natural or modified which are essential to the life and reproduction of the species within its natu- ral geographic range [MA (2005)]. Glossary 369 Health (Human): A state of complete physi- cal, mental and social well-being and not merely the absence of disease or infirmi- ty. The health of a whole community or population is reflected in measurements of disease incidence and prevalence, age-spe- cific death rates and life expectancy [UK NEA (2011)]. Hemeroby: is the degree of the anthropogenic influence on a land use (LU) or land cover (LC) type. Human Inputs: Encompass all anthropogen- ic contributions to ES generation such as land use and management (including sys- tem inputs such as energy, water, fertilis- er, pesticides, labour, technology, knowl- edge), human pressures on the system (e.g. eutrophication, biodiversity loss) and pro- tection measures that modify ecosystems and ES supply. Human Well-Being: A state that is “intrinsi- cally and not just instrumentally valuable” (or good) for a person or a societal group. In the MA, components (or drivers) of hu- man well-being have been classified into: basic material for a good life, freedom and choice, health and bodily well-being, good social relations, security, peace of mind and spiritual experience, not precluding other classifications [Adapted from Alex- androva (2012) and MA (2005)]. Impact: Negative or positive effect on in- dividuals, society and/or environmental resources resulting from environmental change [Modified after Harrington et al. (2010)]. Indicator: An indicator in policy is a metric of a policy-relevant phenomenon used to set environmental goals and evaluate their ful- filment (cf. Heink & Kowarik, 2010). An indicator in science is a quantifiable metric which reflects a phenomenon of interest (the indicandum) [OpenNESS, modified from Heink & Kowarik (2010)]. Intrinsic Value: Intrinsic value is the value something has independent of any inter- ests attached to it by an observer or po- tential user. This does not necessarily mean that such values are independent of a valu- er (i.e. values which exist per se); they may also require a (human) valuer (but this is a matter of disagreement amongst philoso- phers) [OpenNESS, adapted from various sources]. Land Cover (LC): The physical coverage of land, usually expressed in terms of vegeta- tion cover or lack of it. Related to, but not synonymous with, Land Use [UK NEA (2011)]. Landscape: An area, as perceived by people, whose character is the result of the action and interaction of natural and/or human factors. The term “landscape” is thus de- fined as a zone or area as perceived by local people or visitors, whose visual features and character are the result of the action of natural and/or cultural factors. Recog- nition is given to the fact that landscapes evolve through time and are the result of natural and human activities. Landscape should be considered as a whole - natural and cultural components are taken togeth- er, not separately [European Landscape Convention Article 1]. Landscape metrics: Landscape metrics cap- ture composition and configuration of landscape structure in mathematical terms. Not only spatial but also temporal proper- ties of processes can be characterised by a quantifying landscape pattern. Land Use (LU): The human use of a piece of land for a certain purpose such as irrigat- ed agriculture or recreation. Influenced by, but not synonymous with, land cover [UK NEA (2011)]. Map: The main product of cartographic work and is the graphic representation of fea- tures of an area of the Earth or of any other celestial body drawn to scale. Mapping: See term “ES Mapping”. Model (scientific): A simplified representa- tion of a complex system or process includ- ing elements that are considered to be es- sential parts of what is represented. Models Mapping Ecosystem Services370 aim to make it easier to understand and/or quantify by referring to existing and usual- ly commonly accepted knowledge [Open- NESS, based partly on Wikipedia]. Modifiable Areal Unit Problem (MAUP): A cartographic phenomenon associated with the use of data (i.e. statistical data or ob- served data) and their aggregation to geo- graphical areas. The assignment of data to geographical areas and their boundaries do not always make sense, in the context of both scale and aggregation. Monetary Valuation (for ES): The process whereby people express the importance or preference they have for the ES or bene- fits that ecosystems provide in monetary terms. See also ‘Non-monetary valuation’ [OpenNESS, modified from TEEB]. Multifunctionality: The characteristic of eco- systems to simultaneously perform multi- ple functions which may be able to provide a particular ES bundle or bundles [Open- NESS]. Multiple-use Management: Management of land or resources for more than one pur- pose. Natural Asset: A component of Natural Cap- ital [OpenNESS]. Natural Capital: The elements of nature that directly or indirectly produce value for peo- ple, including ecosystems, species, freshwa- ter, land, minerals, air and oceans, as well as natural processes and functions. The term is often used synonymously with natural as- set, but, in general, implies a specific com- ponent [Modified after MA (2005)]. Natural Capital Accounting: A way of organ- ising information about natural capital so that the state and trends in natural assets can be documented and assessed in a systematic way by decision-makers. [OpenNESS]. Non-Monetary Valuation: The process whereby people express the importance or preference they have for the service or benefits that ecosystems provide in terms other than money. See Monetary Valua- tion [OpenNESS]. Policy Maker: A person with the authority to influence or determine policies and prac- tices at an international, national, regional or local level [Modified UK NEA (2011)]. Provisioning Ecosystem Services: Those material and energy outputs from ecosys- tems that contribute to human well-being [Shortened from CICES]. Public Good: A benefit where access to the benefit cannot be restricted [Modified from UK NEA (2011)]. Pragmatics (graphics): Analyse the relation- ships between signs and their users. Projection (of a map): A mathematical rep- resentation of the Earth’s spherical body on a plain surface through mathematical transformations from spherical (latitude, longitude) to Cartesian (x, y) coordinates. Regulating Ecosystem Services: All the ways in which ecosystems and living organisms can mediate or moderate the ambient en- vironment so that human well-being is enhanced. It therefore covers the degra- dation of wastes and toxic substances by exploiting living processes [Modified after CICES]. Rivalry: The degree to which the use of one ES prevents other beneficiaries from using it. Non-rival ES, in return, provide bene- fits to one person and do not reduce the amount of benefits available for others [after Schröter et al. (2014), Kemkes et al. (2010), Costanza (2008), Burkhard et al. (2012)]. Scale (spatial and temporal): The physical di- mensions, in either space or time, of phe- nomena or observations. Regarding tem- poral aspects of ES supply and demand, hot moments are equally as important as spatially relevant hotspots [after Burkhard et al. (2013), Reid et al. (2006)]. Scale (on a map): Represents the ratio of the distance between two points on the map to the corresponding distance on the ground. Scenario: Plausible, but simplified descriptions of how the future may develop, based on a coherent and internally consistent set of Glossary 371 assumptions about key driving forces and relationships. Scenarios are not predictions of what will happen, but are projections of what might happen or could happen giv- en certain assumptions about which there might be great uncertainty [OpenNESS, modified from UK NEA (2011)]. Semantics (graphics): The study of the rela- tionships between signs and symbols and what they are actually representing. Syntactic (graphics): Deals with the formal properties of languages and systems of symbols. Service Benefiting Area (SBA): Spatial unit to which an ecosystem service flow is de- livered to beneficiaries. SBAs spatially de- lineate groups of people who knowingly or unknowingly benefit from the ecosystem service of interest. Service Connecting Area (SCA): Connect- ing space between non-adjacent ecosystem service-providing and service-benefiting areas. The properties of the connecting space influence the transfer of the benefit. Service Providing Area (SPA): Spatial unit within which an ecosystem service is pro- vided. This area can include animal and plant populations, abiotic components as well as human actors. Service Providing Unit (SPU): see Service Providing Area. Social–Ecological System (SES): Interwoven and interdependent ecological and social structures and their associated relation- ships [OpenNESS]. Species: A group of related organisms having common characteristics. Stakeholder: Any group, organisation or indi- vidual who can affect, or is affected by, the ecosystem’s services [OpenNESS]. Sustainability: A characteristic or state whereby the needs of the present and lo- cal population can be met without com- promising the ability of future generations or populations in other locations to meet their needs. Weak sustainability assumes that needs can be met by the substitution of different forms of capital (i.e. through trade-offs); strong sustainability posits that substitution of different forms of capital is seriously limited [UK NEA (2011)]. Synergies: Ecosystem service synergies arise when multiple services are enhanced si- multaneously [Raudsepp-Harne et al. (2010)]. Tiered Approach: A classification of available methods according to level of detail and complexity with the aim of providing ad- vice on method choice. The provision and integration of different tiers enables ES as- sessments to use methods consistent with their needs and resources. Trade-offs: Situations in which one ES in- creases and another one decreases. This may be due to simultaneous response to the same driver or due to actual interac- tions amongst ES [OpenNESS]. Transdisciplinarity: A reflexive, integrative, method-driven scientific principle aim- ing at the solution or transition of socie- tal problems and concurrently of related scientific problems by differentiating and integrating knowledge from various sci- entific and societal bodies including local, place-based knowledge and practitioners’ knowledge [Modified based on Lang et al. (2012) and Turnhout et al. (2012)]. Travel Costs Analysis: Economic valuation techniques that use observed costs to travel to a destination and to derive de- mand functions for that destination [MA (2005)]. Uncertainty: An expression for the degree to which a condition or trend (e.g. of an ecosystem) is unknown. Uncertainty can result from lack of information or from disagreement about what is known or even what can be known. It may have many types of sources, from quantifiable errors in the data to ambiguously defined termi- nology or uncertain projections of human behaviour. Uncertainty can therefore be represented by quantitative measures (e.g. a range of values calculated by various Mapping Ecosystem Services372 models) or by qualitative statements (e.g. reflecting the judgement of a team of ex- perts) [Modified from UK NEA (2011)]. Urban (environment): Environmental con- dition linked to high population density, extent of land transformation, or a large energy flow from surrounding area [Open- NESS, (after McIntyre 2000)]. Value: The worth, usefulness or importance of something. Thus value can be measured by the size of the well-being improvement de- livered to humans through the provision of goods. In economics, value is always asso- ciated with trade-offs, i.e. something only has (economic) value if we are willing to give up something to get or enjoy it [After UK NEA (2011), Mace et al. (2012) and De Groot, (2010)]. Glossary references Alexandrova A (2012) Well-being as an object of science. Philosophy of Science 79: 678- 689. Angermeier PL, Karr JR (1994) Biological in- tegrity versus biological diversity as policy directives. BioScience 44(10): 690-697. Burkhard B, de Groot RS, Costanza R, Sep- pelt R, Jørgensen SE, Potschin M (2012) Solutions for Sustaining Natural Capital and Ecosystem Services. Ecological Indi- cators 21: 1-6. Burkhard B, Crossman N, Nedkov S, Petz K, Alkemade R (2013) Mapping and Model- ling Ecosystem Services for Science, Policy and Practice. Ecosystem Services 4: 1-3. Callicott JB, Crowder LB, Mumford K (1999) Current Normative Concepts in Conser- vation. Conservation Biology 13: 22-35 Costanza R (2008) Ecosystem services: multi- ple classification systems are needed. Bio- logical Conservation 141: 350-352. De Groot RS, Fisher B, Christie M, Aronson J, Braat LC, Haines-Young R, Gowdy J, Maltby E, Neuville A, Polasky S, Portela R, Ring I (2010) Integrating the ecologi- cal and economic dimensions in biodiver- sity and ecosystem service valuation. In: Kumar P (Ed.) TEEB Foundations, The Economics of Ecosystems and Biodiversi- ty: Ecological and Economic Foundations. Earthscan, London (Chapter 1). EC (European Commission) (2013) Green Infrastructure (GI) — Enhancing Europe’s Natural Capital. COM (2013) 249 final, Brussels, 6.5.2013. Harrington R, Dawson TP, de Bello F, Feld CK, Haslett JR, Kluvánkova-Oravská T, Kontogianni A, Lavorel S, Luck GW, Rounsevell MDA, Samways MJ, Skourtos M, Settele J, Spangenberg JH, Vandewalle M, Zobel M, Harrison PA (2010) Ecosys- tem services and biodiversity conservation: concepts and a glossary. Biodiversity Con- servation 19: 2773-2790. Heink U, Kowarik I (2010) What are indi- cators? On the definition of indicators in ecology and environmental planning. Eco- logical Indicators 10(3): 584-593. Hull RB, Richert D, Seekamp E, Robertson D,. Buhyoff GJ (2003) Understandings of environmental quality: Ambiguities and values held by environmental profession- als. Environmental Management 31: 1-13. Jax K (2010) Ecosystem functioning. Cam- bridge University Press, Cambridge. Kemkes RJ, Farley J, Koliba CJ (2010) Deter- mining when payments are an effective pol- icy approach to ecosystem service provision. Ecological Economics 69: 2069-2074. Kjærulff UB, Madsen A (2013) Bayesian Networks and Influence Diagrams: A Guide to Construction and Analysis. In- formation Science and Statistics, Spring- er. Available from: http://link.springer. com/book/10.1007/978-1-4614-5104-4/ page/1. Lang DJ, Wiek A, Bergmann M, Stauffacher M, Martens P, Moll P, Swilling M, Thom- as CJ (2012) Transdisciplinary research in Glossary 373 sustainability science – practice, princi- ples, and challenges. Sustainability Science 7: 25-43. Lincoln R, Boxshall G, Clark P (1998) A dictionary of ecology, evolution and sys- tematics Cambridge, Cambridge Univer- sity Press. MA [Millennium Ecosystem Assessment] (2005) Ecosystems and Human Wellbe- ing: Current State and Trends. Volume 1, Island Press, Washington D.C. Mace GM, Norris K, Fitter AH (2012) Biodi- versity and ecosystem services: a multi-lay- ered relationship. Trends in Ecology and Evolution 27(1): 19-26. Maes J, Teller A, Erhard M et al. (2013) Map- ping and Assessment of Ecosystems and their Services. An analytical framework for ecosystem assessments under action 5 of the EU biodiversity strategy to 2020. Publications office of the European Union, Luxembourg. McIntyre NE, Knowles-Yánez K, Hope D (2000) Urban ecology as an interdisciplin- ary field: differences in the use of “urban” between the social and natural sciences. Urban Ecosystems 4: 5-24. OpenNESS [Potschin M, Haines-Young R, Heink U, Jax K (Eds.)] (2016): OpenNess Glossary (V3.0). Grant Agreement No 308428, available from: http://www.open- ness-project.eu/glossary. Raudsepp-Hearne C, Peterson GD, Bennett EM (2010) Ecosystem service bundles for analysing tradeoffs in diverse landscapes. Proceedings of the National Academy of Sciences. doi/10.1073/pnas.0907284107. Reid WV, Berkes F, Wilbanks T, Capistrano D (Eds.) (2006) Bridging scales and knowl- edge systems: concepts and applications in ecosystem assessment / Millennium Eco- system Assessment. Island Press/World Re- sources Institute, Washington, DC. Schröter M, Barton DN, Remme RP, Hein L (2014) Accounting for capacity and flow of ecosystem services: A conceptual mod- el and a case study for Telemark, Norway. Ecological Indicators 36: 539-551. Turnhout E, Bloomfield B, Hulme M, Wynne B (2012) Listen to the voices of experi- ence. Nature 488: 454-455. UK NEA (2011) The UK National Ecosystem Assessment. Technical Report. UNEP-WC- MC, Cambridge. Available from: http:// uknea.unep-wcmc.org/Default.aspx. Glossary internet sources CICES: http://cices.eu/ ESMERALDA: http://esmeralda-project.eu/ European Landscape Convention: http:// www.coe.int/de/web/landscape FAO Fisheries and Aquaculture Department: http://www.fao.org/fishery/statistics/en MAES: http://biodiversity.europa.eu/maes OpenNESS: http://www.openness-project.eu/ SEEA-EEA: http://unstats.un.org/unsd/en- vaccounting/eea_project/default.asp TEEB: http://www.teebweb.org/ Wikipedia: https://www.wikipedia.org/ Mapping ecosystem services delivers essential insights into the spatial characteristics of various goods’ and services’ flows from nature to human society. It has become a central topic of science, policy, business and society – all belonging on functioning ecosystems. This textbook summarises the current state-of-the-art of ecosystem services mapping, related theory and methods, different ecosystem service quantification and modelling approaches, as well as practical applications. The book is produced by various international experts in the field, in a professional but understandable format to be used by stakeholders, students, teachers, practitioners and scientists involved or interested in ecosystem services mapping. ISBN 978-954-642-852-3 Benjamin Burkhard is a professor for physical geography at Leibniz Universität Hannover, Germany. He is doing research and teaching in geography and landscape ecology and has been involved in ecosystem services mapping since many years. Joachim Maes is a researcher at the European Commission’s Joint Research Centre in Ispra, Italy. He is engaged in the scientific support of EU biodiversity policy and develops maps of ecosystem services at European scale. 9 789546 428295