Identification of Superclusters and Their Properties in the Sloan Digital Sky Survey
Using the WHL Cluster Catalog
Shishir Sankhyayan1 , Joydeep Bagchi2 , Elmo Tempel1,3 , Surhud More4,5 , Maret Einasto1 , Pratik Dabhade6,7 ,
Somak Raychaudhury4,8 , Ramana Athreya9 , and Pekka Heinämäki10
1 Tartu Observatory, University of Tartu, Observatooriumi 1, 61602 Tõravere, Estonia; shishir.sankhyayan@ut.ee, shishir9898@gmail.com
2 Department of Physics and Electronics, CHRIST (Deemed to be University), Bengaluru-560029, India
3 Estonian Academy of Sciences, Kohtu 6, 10130 Tallinn, Estonia
4 The Inter-University Centre for Astronomy and Astrophysics (IUCAA), S.P. Pune University Campus, Post Bag 4, Pune 411007, India
5 Kavli Institute for the Physics and Mathematics of the Universe (IPMU), 5-1-5 Kashiwanoha, Kashiwashi, Chiba 277-8583, Japan
6 Instituto de Astrofísica de Canarias, Calle Vía Láctea, s/n, E-38205, La Laguna, Tenerife, Spain
7 Universidad de La Laguna (ULL), Departamento de Astrofisica, E-38206, Tenerife, Spain
8 Department of Physics, Ashoka University, Sonipat, Haryana 131029, India
9 Indian Institute of Science Education and Research (IISER), Dr. Homi Bhabha Road, Pashan, Pune 411008, India
10 Tuorla Observatory, Department of Physics and Astronomy, University of Turku, Vesilinnantie 5, 20014 Turku, Finland
Received 2022 December 13; revised 2023 September 11; accepted 2023 September 11; published 2023 November 13
Abstract
Superclusters are the largest massive structures in the cosmic web, on tens to hundreds of megaparsec scales. They
are the largest assembly of galaxy clusters in the Universe. Apart from a few detailed studies of such structures,
their evolutionary mechanism is still an open question. In order to address and answer the relevant questions, a
statistically significant, large catalog of superclusters covering a wide range of redshifts and sky areas is essential.
Here, we present a large catalog of 662 superclusters identified using a modified friends-of-friends algorithm
applied on the WHL (Wen–Han–Liu) cluster catalog within a redshift range of 0.05„ z„ 0.42. We name the most
massive supercluster at z∼ 0.25 as the Einasto Supercluster. We find that the median mass of superclusters is
∼5.8× 1015 Me and the median size ∼65Mpc. We find that the supercluster environment slightly affects the
growth of clusters. We compare the properties of the observed superclusters with the mock superclusters extracted
from the Horizon Run 4 cosmological simulation. The properties of the superclusters in the mocks and
observations are in broad agreement. We find that the density contrast of a supercluster is correlated with its
maximum extent with a power-law index, α∼−2. The phase-space distribution of mock superclusters shows that,
on average, ∼90% of part of a supercluster has a gravitational influence on its constituents. We also show the mock
halos’ average number density and peculiar velocity profiles in and around the superclusters.
Unified Astronomy Thesaurus concepts: Superclusters (1657); Galaxy clusters (584); Catalogs (205); Large-scale
structure of the universe (902)
Supporting material: machine-readable tables
1. Introduction
The intricate network of the distribution of galaxies and
matter in the Universe is called the cosmic web (Bond et al.
1996). The details of the cosmic web help in our understanding
of the cosmological models governing the evolution of
structures in the Universe. The main components of the cosmic
web are clusters, two-dimensional walls, one-dimensional
filaments, and underdense regions called voids.
In the cosmic web, there exist large coherent regions
(∼10–100 Mpc) that are larger than the dimensions of galaxy
clusters, which span about a few megaparsecs. Moreover, these
regions are extremely massive and contain several groups and
clusters of galaxies apart from the galaxies and intergalactic dark
and baryonic matter in between the clusters. Cumulatively, these
regions are called superclusters. Superclusters are not as
abundant as galaxies or even groups and clusters, but they are
known to affect the evolution of the galaxies within them (e.g.,
Einasto et al. 2007b, 2011c, 2014; Lietzen & Einasto 2016). It is
still an open question as to how these giant structures form and
evolve in the cosmic web, and answering it requires detailed
studies involving observations and simulations. Although this is
a growing field of research, which is getting enriched with the
availability of more data, a single widely accepted definition of a
supercluster does not exist as yet. Superclusters are coherent
regions in the cosmic web that have been defined in different
ways in the literature. The two extreme definitions of a
supercluster are based on the largest bound structures in the
Universe and the largest regions with converging peculiar
velocity field flows. The superclusters in the literature can be
broadly divided into three categories: (1) the gravitationally
bound regions; (2) the unbound overdense regions in the
Universe; and (3) the converging peculiar velocity field regions.
According to definition (1), they have been defined as the largest
overdense regions, with sufficiently high matter density to
overcome the global expansion of the Universe. They would
eventually collapse and form gravitationally bound systems
(Dünner et al. 2006; Araya-Melo et al. 2009; Luparello et al.
2011; Chon et al. 2015). According to definition (3), using the
peculiar velocity field, they have been defined as regions in
which, on average, the peculiar velocities of galaxies converge
(Tully et al. 2014; Pomarède et al. 2015; Dupuy et al. 2019;
Pomarède et al. 2020). These superclusters form regions of
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 https://doi.org/10.3847/1538-4357/acfaeb
© 2023. The Author(s). Published by the American Astronomical Society.
Original content from this work may be used under the terms
of the Creative Commons Attribution 4.0 licence. Any further
distribution of this work must maintain attribution to the author(s) and the title
of the work, journal citation and DOI.
1
dynamical influence where they act as great attractors that grow
by the inflow of matter from lower-density regions (Tully et al.
2014). Einasto et al. (2019) and Dupuy et al. (2019) showed that
the whole cosmic web can be divided into regions of dynamical
influence or the basins of attraction. This means that type (3)
superclusters contain some parts of the underdense void regions
surrounding them. Observationally, Laniakea—our home super-
cluster (Tully et al. 2014)—has been identified using the rich
observational data of the peculiar velocities of the nearby
galaxies. But such data are still not available for distant galaxies;
therefore, the velocity field flow method of estimating regions of
convergence cannot be derived for them. However, these
converging peculiar velocity field regions (or the “basins of
attraction”) can be identified in simulations. It is often seen that
the supercluster definition (1) of gravitationally bound systems
picks up the central regions of superclusters defined with other
criteria (definitions (2) and (3)), for example, overdensities in the
luminosity density field (e.g., Einasto et al. 2022). Definition (2)
picks up the overdensities present in the matter density field.
These regions do not contain any parts of underdense void
regions. The overdensity threshold to consider in defining these
types of superclusters is not set, and depending on the focus of
the study, it varies in the literature (see for example, Einasto
et al. 2007a; Liivamägi et al. 2012; Lietzen et al. 2016; Einasto
et al. 2019).
Superclusters are traditionally identified using the friends-of-
friends (FoF) algorithm. FoF is also used to find bound halos in
simulations (Knebe et al. 2011). FoF has been used in several
supercluster-finding works (e.g., Einasto et al. 1997; Chow-
Martínez et al. 2014; Bagchi et al. 2017) in various forms,
using clusters as well as galaxies as their input data. Apart from
FoF, there are other methods as well to extract superclusters.
For example, by applying different density threshold cuts on
the luminosity density field of the galaxy distribution (Einasto
et al. 2007a; Liivamägi et al. 2012; Lietzen et al. 2016) and the
watershed method, by applying it on the number density field
constructed by Voronoi tessellation (Neyrinck 2008; Nadathur
& Crittenden 2016).
Recent studies show that a supercluster’s size, mass, and
luminosity show some evolution with redshift. In this context,
luminosity represents the total luminosity of all galaxies within
the supercluster. Size refers to the maximum comoving extent
of the supercluster, and the mass indicates the total amount of
baryonic and dark matter contained within the supercluster. As
the luminosity and mass increase with time, the overall size
decreases (Araya-Melo et al. 2009; Einasto et al. 2019, 2021a).
The evolution of galaxies inside a supercluster is affected by its
host supercluster’s environment (Lietzen et al. 2012; Seth &
Raychaudhury 2020; Alfaro et al. 2022). Also, the evolution of
clusters and groups could be governed by the environment of
the superclusters in which they reside. Extremely large
superclusters may have a linear growth on average, but on a
megaparsec scale, galaxies and clusters have a nonlinear
growth within them. Overall, massive and large superclusters
are excellent targets to study the diversity of environments they
offer to their resident galaxies. Multiwavelength observations
of such objects will aid our understanding of their growth and
evolution. Ongoing and upcoming deep weak-lensing surveys
will immensely help in estimating the masses of these huge
overdensities.
In the Λ Cold Dark Matter (ΛCDM) model of the Universe,
the overall growth of structures slows down below redshift
z≈ 0.5 due to the influence of dark energy (Frieman et al.
2008; Einasto et al. 2021b). However, the highly overdense
structures are still able to grow. In a spherical collapse model,
structures that are at the turnaround stage at the present epoch,
with a density contrast of ∼13.1, form bound systems.
Structures with a present-day density contrast of ∼8.73 will
reach the turnaround stage in the future (Einasto et al. 2020).
Einasto et al. (2021c), using the spherical collapse model,
suggested that structures with a present-day density contrast of
≈30 (rich clusters and their regions of influence in supercluster
cores) have passed turnaround and started to collapse at
redshifts z≈ 0.3–0.4. Therefore, a supercluster may or may not
collapse, depending on the average density contrast of the
entire supercluster. For example, Einasto et al. (2022) show
that the BOSS Great Wall fragments into different structures in
physical space, and these fragmented structures will collapse
individually in the future.
In order to enhance our understanding of superclusters and
their properties, it is important to identify these structures over
a broad range of redshifts and a wide area of the sky. Moreover,
their dynamical behavior can be explored using simulations.
This paper presents 662 superclusters identified within a
redshift range of 0.05„ z„ 0.42, covering a sky area of
∼14,000 deg2. The properties of these superclusters are
presented and compared with simulations. The layout of the
paper is as follows. In Section 2, we present the observational
data and mocks created through simulation. In Section 3, we
present our analysis methods. In Section 4, we define the
properties of the superclusters. Section 5 presents the results of
our analysis. The properties of the identified superclusters
highlighting the most massive superclusters (e.g., the Corona
Borealis (CB) supercluster: Einasto et al. 2021c; the Sloan
Great Wall (SGW): Gott et al. 2005; and the Saraswati
supercluster: Bagchi et al. 2017), a power-law relation between
the density contrast and the size of the supercluster, and the
effect of the supercluster environment on the cluster properties
are presented in Section 5.1. A comparison of the properties of
superclusters extracted from observations and simulations and
the phase-space distribution of the mock superclusters is
presented in Section 5.2. We conclude and discuss the future
prospects in Section 6, and the Appendix highlights the five
most massive superclusters.
Throughout this paper, we have adopted the following
cosmological parameters: H0= 72 km s
−1 Mpc−1, Ωm= 0.26,
and ΩΛ= 0.74. RΔc is defined as the radius of a spherical
region within which the matter density is Δ times the critical
density (ρc) of the Universe. MΔc is the mass within this
spherical region of radius RΔc. So M500c denotes the mass
within a spherical region of radius R500c where the mass density
is 500 times the critical density of the Universe.
2. Data
Here, we describe the observational data and simulation used
in our analyses.
2.1. Observational Data
Bagchi et al. (2017) used the spectroscopic sample of
galaxies in the Sloan Digital Sky Survey (SDSS; York et al.
2000) to identify the massive supercluster named Saraswati.
We have been motivated by the large excess of clusters
and groups (Wen et al. 2012) found in the Saraswati
2
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
supercluster region (Bagchi et al. 2017). As a result, we
decided to search for other superclusters in SDSS using the
same cluster catalog.
To achieve this, we use the group and cluster catalog of Wen
& Han (2015; hereafter, WH15),11 which is an updated version
of Wen et al. (2012; hereafter, WHL12).12 The WHL12 catalog
consists of 132,684 groups and clusters identified from the
photometric redshift data of galaxies within the redshift range
0.05„ z„ 0.8 from SDSS-III.13 Galaxies with deblending
problems and saturated objects were excluded. With the release
of SDSS Data Release 12 (DR12; Alam et al. 2015), new
spectroscopic redshifts of galaxies became available, leading to
the addition of 25,419 rich clusters (primarily at high redshifts,
z> 0.4) in the WH15 catalog.
In this paper, we utilize the spectroscopic redshifts
available in DR12 for 89% of the groups and clusters listed
in the WH15 catalog. For the remaining 11% of groups and
clusters, we rely on photometric redshifts. Figure 1 shows the
redshift distribution of WH15 groups and clusters. The small
peak observed at z∼ 0.22 and the dip at z∼ 0.28 are attributed
to the redshift distributions of the two major samples
(the luminous red galaxy (LRG) and LOWZ galaxy samples)
comprising SDSS. The LRG sample exhibits an increase
beyond z∼ 0.2, while the LOWZ sample peaks at z∼ 0.35.
The data set covers a sky area of ∼14,000 deg2, as shown in
Figure 2.
Figure 1. The blue histogram represents the redshift distribution of the entire sample of 158,103 WHL clusters, and the red line represents the selected sample of
85,686 clusters within the redshift range of 0.05 „ z „ 0.42.
Figure 2. Sky distribution of the WHL clusters and HR4 halos within 0.05 „ z „ 0.366. The WHL clusters and HR4 halos are the observed and mock samples,
respectively.
11 https://cdsarc.cds.unistra.fr/viz-bin/cat/J/ApJ/807/178
12 https://cdsarc.cds.unistra.fr/viz-bin/cat/J/ApJS/199/34
13 https://www.sdss3.org
3
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
The completeness of the data, defined as the detection rate of
the injected mock clusters using the cluster detection algorithm,
depends on the virial mass M200c. For clusters with mass
M200c 0.6× 1014Me, the completeness is ∼80% (WHL12).
For clusters with M200c> 10
14Me, it exceeds 95%, and it
reaches 100% for clusters with M200c> 2× 10
14Me within the
redshift range 0.05„ z„ 0.42. Consequently, we focus on the
data within this redshift range, selecting 85,686 groups and
clusters to extract superclusters (Figure 1). The false detection
rate of these clusters is 6% (WHL12). For detailed group and
cluster catalog construction information, refer to WHL12 and
WH15. Throughout the rest of the paper, we will refer to the
group and cluster catalog of WH15 (the updated WHL12
catalog) as the WHL cluster catalog. Table 1 shows the sample
of selected 85,686 groups and clusters within the redshift range
0.05„ z„ 0.42.
The halo masses in WH15 are derived from the optical
richness, which was cross-calibrated using clusters having
masses from X-ray and Sunyaev–Zeldovich measurements.
The optical richness is not expected to have a one-to-one
correlation with the halo mass, but will instead be a useful
proxy, rather than being entirely perfect and a proxy that will
involve an intrinsic scatter. Therefore, a selection of clusters
that are volume-limited by mass cannot be obtained in practice.
In addition, the various flux limits of spectroscopic and
photometric surveys will imply that the members of the clusters
at higher redshift will be limited to brighter galaxies, which, on
average, will be hosted by more massive clusters. In principle,
such selection effects can be thoroughly included in the mocks,
but such an implementation will require a cluster catalog that
calibrates both the mean and the scatter of the richness–mass
relation, accounting for the flux limit effects mentioned above.
In the absence of such calibration, including the intrinsic scatter
is beyond the scope of this paper, but is something that we plan
to address in the future. In the current work, we treat
incompleteness to be independent of mass and compensate
for the incompleteness of low-mass clusters at higher redshift
by increasing the linking length of the existing clusters. This
allows the clusters to be linked at larger distances in inverse
proportion to the cube root of their density (see Section 3) and
ensures that the identification of superclusters is not affected by
the incompleteness of the chosen sample. Additionally, we
preferred to use all the data available in the WHL catalog, as
applying a higher mass cut means that we have to discard some
observational data. Instead, we use all the available observa-
tional data and model the incompletenesses in the data. This is
the approach we have adopted here while constructing the
mock catalog.
2.2. Simulation and Mock Clusters
We use the Horizon Run 4 (HR4) cosmological N-body
simulation (Kim et al. 2015) to create a mock WHL cluster
catalog. It has 63003 dark matter particles in a cubic box of
size 3150 h−1 Mpc. The particle mass is mp; 9× 10
9 Me,
and the minimum mass of halos with 30 member particles is
MS; 2.7× 10
11 Me. The HR4 simulation adopted the
Wilkinson Microwave Anisotropy Probe 5 yr (Dunkley
et al. 2009) ΛCDM cosmology: Ωm,0= 0.26, ΩΛ,0= 0.74,
and H0= 100 h km s
−1 Mpc−1, where h= 0.72.
In order to create a mock catalog, we require information on
the position, mass, and peculiar velocity of halos. However, if
we consider the snapshot of the simulation at redshift z= 0,
it does not correspond to the observations that capture
the states of objects at different cosmological times (red-
shifts). Essentially, a halo (or a cluster in observations)
becomes visible to the observer when it enters the observer’s
past light cone (or the center of the simulation box).
Therefore, we need to consider the state of a halo (its
comoving position, mass, and peculiar velocity) at the time
(redshift) when it enters the observer’s past light cone. The
HR4 simulation output products provide the necessary past
light cone information on halos. We use the past light cone
space dark matter halos in the redshift range 0< z„ 0.4. We
displace the halos according to their peculiar velocities to get
the positions of halos in the redshift space (the same as the
observations).
The mock WHL cluster catalog is made by selecting the
halos according to the selection functions of the WHL cluster
data. We matched the sky distribution, redshift distribution, and
mass distribution of the WHL clusters for the extraction of
halos from the HR4 simulation (see Figures 2 and 3). To make
a sky mask for the WHL clusters’ sky distribution, we use the
Hierarchical Equal Area isoLatitude Pixelization (HealPix)
scheme by Górski et al. (2005). HEALPix divides the sky plane
into equal-area and shape-preserving pixels. We use Nside= 64
and create the sky mask of the WHL catalog. A pixel is given
an angular weight of 1 if it has at least one cluster in it and 0 if
it does not contain any cluster. We then apply this mask to the
HR4 halos and extract all the halos on the sky pixels with a
weight equal to 1. Next, we divide the redshift distribution into
equal comoving distance bins. In each bin, halos are extracted
Table 1
Properties of 85,686 Groups and Clusters Extracted from the WHL Cluster Catalog within the Redshift Range 0.05 „ z „ 0.42
ID RA Decl. z R200c M200c SCl
(deg) (deg) (Mpc) (1014 Me)
(1) (2) (3) (4) (5) (6) (7)
J000000.6+321233 0.00236 32.20925 0.1274 1.72 5.9478 0
J000002.3+051718 0.00957 5.28827 0.1694 0.94 1.9627 0
J000003.5+314708 0.01475 31.78564 0.0916 0.94 1.6766 0
J000004.7+022826 0.01945 2.47386 0.4179 0.95 0.4769 0
J000006.0+152548 0.02482 15.42990 0.1731 1.13 1.5391 0
Note. Columns (1)–(4) are taken from the WHL catalog and columns (5)–(7) have been computed in the current paper (see Sections 3 and 4). Column (1): cluster
identifier. Column (2): R.A. of the brightest cluster galaxy (BCG). Column (3): decl. of the BCG. Column (4): redshift of the BCG. Column (5): R200c of the cluster.
Column (6): M200c of the cluster. Column (7): supercluster number to which the cluster belongs. SCl = 0 means the cluster is not a part of any supercluster.
(This table is available in its entirety in machine-readable form.)
4
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
to match the mass distribution and the number of WHL clusters
in the corresponding bin.
Last, we find a small deficiency of large-mass clusters in the
HR4 simulation compared to the WHL clusters. A similar
deficiency with high-luminosity galaxies (the high-mass end of
galactic halos) is also seen in other studies (see Trayford et al.
2015; Tuominen et al. 2021). To address this issue, when we
encounter a deficiency of large-mass clusters within a specific
distance bin, we randomly select halos from the lower-mass
bin. However, we only choose halos that have not already been
selected for the previous mass bin within that specific distance
bin. This ensures that the number of mock clusters in each
distance bin approaches the number of WHL clusters present in
that particular distance bin. As a result, there is a small
difference in the mass distribution between the WHL and mock
clusters.
Additionally, to avoid the artificial change in the number
density of halos at z∼ 0.4 due to the peculiar velocities of the
galaxies in the halos, we have limited the halos to
0.05„ z„ 0.366. The redshift limit of 0.366 is chosen to
avoid any contamination due to redshift space distortion by
assuming a maximum peculiar velocity of a galaxy in a
massive cluster as 5000 km s−1. To err on the side of caution
and ensure no contamination from the Fingers-of-God effect, a
conservative value of 10,000 km s−1 (corresponding to
Δz∼ 0.034) has been chosen. Figure 3 shows the mass and
redshift distributions of the mock and WHL clusters.
3. Methods: Modified Friends of Friends
To identify the superclusters and create a supercluster
catalog, a modified FoF (mFoF) algorithm is used. The FoF
algorithm (Huchra & Geller 1982; Martínez & Saar 2002) finds
overdensities in the distribution of galaxies or clusters. In this
algorithm, a linking length l (threshold distance) is chosen.
Two clusters are considered linked (part of the same
supercluster) if the distance between them is less than or equal
to l. This gives a system of clusters (a supercluster) with all the
member clusters having distances to their natural neighbors „l.
We modify the FoF algorithm to account for the selection
biases in the survey. The mFoF algorithm makes use of
Delaunay triangulation for distance calculations and weighted
linking length to find the member clusters. The procedure is
described below in detail.
3.1. Distances Using Delaunay Triangulation
A simple way is to calculate O(N2) distances for a
distribution of N clusters and then find l for which the number
of superclusters is maximum (see Section 3.4). This method is
computationally very expensive for large values of N. To
increase the computational efficiency, Delaunay triangulation
(Delaunay 1934) is used to link a cluster to its nearest
neighbors, and only the distances between the natural
neighbors are considered. This reduces the calculation of
O(N2) distances to O(NlogN).
3.2. Radial Selection Weights
As the number density of WHL clusters varies with redshift,
the distribution of WHL clusters is not uniform in the radial
direction. The number density f(z) of WHL clusters is shown
in Figure 4. The radial selection weight wr,i of the ith cluster at
redshift z is defined as
w
z
, 1r i,
f
f=
¯
( )
( )
where, f¯ is the mean number density of the survey.
3.3. Weighted Linking Length
The number density of clusters varies with redshift (Figure
4), necessitating the adjustment of the linking length according
to this variation. Consequently, the linking lengths will be
different for different pairs of clusters. The radial selection
weights are computed through three-dimensional density
estimations. Therefore, the cube root of the weights is to be
taken to apply these weights on the one-dimensional linking
length:
w w . 2f i r i, ,
1
3= [ ] ( )
The final weights wf,i of all the clusters as a function of redshift
are shown in Figure 4.
Now, if lo is the selected linking length (as defined in Section
3.4), then the weighted linking length lij between the ith and jth
Figure 3. (a) Mass functions of WHL clusters and mock clusters (HR4 halos) within 0.05 „ z „ 0.366. n is the number density of the clusters/halos. (b) Number
density as a function of the redshift of the WHL and mock clusters (HR4 halos) within 0.05 „ z „ 0.366.
5
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
clusters will be
l
w w
l
2
. 3ij
f i f j
o
, ,= + ( )
3.4. Selecting the Linking Length
The lo is chosen to obtain the maximum number of
superclusters for a given minimum number Nmin of clusters
in a supercluster (Chow-Martínez et al. 2014; Bagchi et al.
2017). In the literature, different FoF algorithms choose
different Nmin for clustering analysis, according to the data
and the aim of the study. There are no fixed criteria for
choosing Nmin, and the choice changes the value of the linking
length and, thus, the clustering properties. In our analysis, we
notice that selecting higher Nmin values and subsequently
calculating the linking length leads to unstable results, with the
linking lengths varying significantly. We are limited by the
number density of WHL clusters, and choosing a higher Nmin
value increases the statistical noise. Moreover, a higher Nmin
gives a higher linking length value, which leads to the
identification of some spurious superclusters exceeding
600Mpc in size, comprised of up to 200 member clusters
and displaying negative density contrast values (indicating
underdense regions). This issue persists even when considering
only the nonstriped, contiguous survey region (discussed
below). We mitigate this issue by choosing a lower value of
N 2min = , which anchors the linking length estimation toward
lower values, thereby preventing it from reaching percolation
thresholds. For our analysis, the chosen linking length (lo) for
the sample is 20.65Mpc (as shown in Figure 5), representing
the maximum number of superclusters detected with N 2min = .
The above method gives a range of masses of superclusters
with at least two clusters in a supercluster, and hence these can
Figure 4. Number density f(z) of WHL clusters as a function of redshift within 0.05 „ z „ 0.42. Inset: corresponding weights wf,i of the clusters as a function of
redshift.
Figure 5. Number of candidate superclusters as a function of unweighted comoving linking length lo (in blue). The red line shows the position lo = 20.65 Mpc where
the number of candidate superclusters is maximum.
6
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
be regarded as only candidate superclusters. The candidate
superclusters with at least 10 member clusters are defined as
superclusters in our catalog. The majority (∼80%) of clusters in
the WHL catalog have masses M200c 1014Me. With a
minimum of 10 member clusters, the lower mass limit of a
supercluster corresponds to ∼1015Me. Consequently, all the
superclusters included in our catalog have masses 1015Me,
which is consistent with the findings regarding supercluster
masses presented in Araya-Melo et al. (2009).
To verify the robustness of our analysis results against the
complex geometry of our sample’s sky footprint, we conducted
an additional check by examining only the contiguous regions
of our sample footprint, excluding the striped regions. We then
performed the mFoF algorithm to detect superclusters. We
found no significant deviations in the properties of the detected
superclusters when comparing them to the findings presented in
Section 5.1 for the entire survey footprint (as illustrated in
Figure 2). This suggests that the inclusion of the striped regions
in our analysis is justified. Furthermore, the detected super-
clusters in the striped regions will be useful for individual
studies. Only 10 of our catalog’s 662 superclusters (Section
5.1) are found within the striped regions. We also note that the
very massive Saraswati supercluster was found in the narrow
Stripe 82 of SDSS (Bagchi et al. 2017).
4. Properties of the Superclusters
4.1. Mass Estimation
In order to calculate the total mass of a supercluster, we use
the following method. First, we calculate the bound mass of the
individual member clusters of a supercluster. Then, we add the
bound masses of all the member clusters to get the total mass of
the supercluster.
WH15 provides information about clusters’ richness in their
catalog, denoted as RL*,500. This richness is the optical
luminosity, L500, measured in units of L
*, the evolved
characteristic luminosity of galaxies in the r band, within
R500c. WH15 also gives the correlation between the mass
M500c (in units of 10
14 Me) and richness RL*,500 of the
clusters as
*M Rlog 1.08 0.02 log 1.37 0.02 . 4c L500 ,500=  - ( ) ( ) ( )
The mass uncertainty estimated by RL*,500 is 0.14Mlog c500s = .
To estimate the mass M500c of a cluster, we utilize RL*,500 in
conjunction with Equation (4).
The relation between R200c and R500c is R500c≈ 0.65× R200c
for a given Navarro–Frenk–White mass profile with a
concentration parameter in the range 4–8 (Ettori & Balestra
2009). And assuming a spherically symmetric distribution
(M Rc c c
4
3
3r= D pD D ) of halo density, M200c can be expressed as
⎜ ⎟
⎛
⎝
⎞
⎠
M
R
R
M
200
500
. 5c
c
c
c200
200
500
3
500= ( )
Using Equations (4) and (5), we get M200c for each cluster. To
account for the mass beyond the virial radius of a cluster, the
virial mass (M200c) of each member cluster is further scaled up
by a factor of 2.2 to get its bound halo mass Mhalo≈M5.6c. In
earlier studies, it is found that the bound halo mass of a cluster
is close to M5.6c∼ 2.2×M200c (Busha et al. 2005; Rines &
Diaferio 2006; Rines et al. 2013). The sum of the bound halo
masses of all the member clusters gives the total mass of the
supercluster. Figure 6 shows the mass distribution of the
superclusters.
4.2. Size and Position
The comoving linear size of a supercluster is calculated by
measuring the maximum distance between pairs of member
clusters of a supercluster. The comoving position X of the
supercluster is taken as the virial mass-weighted average of the
positions of member clusters,
X
XM
M
, 6i c i i
i c i
200 ,
200 ,
= åå ( )
where M200c,i and Xi are the virial mass and comoving position
of the ith member cluster. The distributions of the super-
clusters’ size, mass, and redshift are shown in Figure 6.
4.3. Density Contrast
To estimate the average density contrast δ of a supercluster,
we now need to estimate the volume occupied by a
supercluster. Since these superclusters do not have any specific
shape, we estimate the volume by fitting a convex hull to the
member clusters. A convex hull of a set of points P in three-
dimensional Euclidean space is the convex surface (envelope)
of the minimum possible volume on P. The volume of a
supercluster is much larger than the volume of a single cluster.
Hence, treating clusters as points or spherical objects with fixed
radii has negligible impact on the estimated supercluster
volumes. We use the Qhull algorithm (Barber et al. 1996) to
construct the convex hull and to calculate its volume. Once we
have the mass and volume, we can calculate the density of the
supercluster. The corresponding matter density contrast,
therefore, will be given by
1, 7
m
SCd rr= - ( )
where ρSC is the mass density of the supercluster and ρm is the
background matter density at the redshift (z) of the super-
cluster:
H z
G
3 1
8
. 8m
m0
2 3
r p=
W +( ) ( )
Figure 6 shows the density contrast of the superclusters.
5. Results
5.1. Observation
5.1.1. WHL Superclusters
The mFoF algorithm with at least 10 member clusters gives a
total of 662 superclusters in the redshift range of
0.05„ z„ 0.42. To the best of our knowledge, this is the
largest supercluster catalog constructed from the data of
clusters in this redshift range. Around 12% of clusters (9895
clusters) from the WHL catalog considered for our work reside
in a supercluster environment, and around 28% (183 super-
clusters) of the superclusters have at least one Abell cluster
(Abell et al. 1989) in them. The presence of Abell clusters in
many superclusters (for example, the Shapley supercluster:
Shapley 1930; Raychaudhury 1989; the SGW; the Saraswati
supercluster; and the CB supercluster) demonstrates the
7
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
robustness of our supercluster catalog. Figure 7 shows the
distribution of the superclusters in the redshift slice
0.26„ z„ 0.29. All the WHL clusters within the redshift slice
are represented by dots. Gray dots represent the clusters that are
not part of any supercluster. The other colored dots represent
member clusters of 93 superclusters within the redshift slice.
The supercluster in yellow at 350°„ R. A.„ 359° and
decl.∼ 0° is the Saraswati supercluster.
The sky coordinates, redshifts, and properties of the
superclusters, arranged in decreasing order of their masses, in
our catalog are listed in Table 2. The distributions of the
superclusters’ mass, size, and density contrast are shown in
Figure 6.
Apart from a few superclusters, which are rediscoveries of
previously known superclusters, most of the superclusters
reported here are newly discovered. For example, the Saraswati
supercluster (SCl 4 and SCl 231) and the SGW (SCl 3 and SCl
22) are identified by our algorithm (see Section 5.1.2).
The median size and mass of the superclusters in our catalog
are ∼65Mpc and ∼6× 1015Me, respectively. It is to be noted
that the superclusters listed in our catalog might even be larger
and more massive than currently indicated. We estimate the
Figure 6. (a) Distribution of the superclusters’ mass and size. (b) Distribution of the superclusters’ mass and redshift. (c) Density contrast of the superclusters as a
function of size. The blue dots represent the superclusters, the solid black line is a linear fit to the data points on the log–log plot, and the shaded black region denotes
the associated errors in the fit. The slope of the fitted line is given at the top right corner of the plot.
Figure 7. Sky plane distribution of the superclusters within 0.26 „ z „ 0.29. The dots represent the WHL clusters within the redshift slice. The gray dots are not a part
of any supercluster, and the other colored dots represent the member clusters of 93 superclusters within the redshift slice.
Table 2
Properties of the Sample of 662 Superclusters in Decreasing Order of Their Masses
SCl R.A. Decl. RA_BCG Decl_BCG z Size Mass Nmem NAbells
(deg) (deg) (deg) (deg) (Mpc) (1014 Me)
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
1 209.6316 3.3618 210.2586 2.8785 0.2493 111.10 257.106 54 2
2 235.1322 28.2855 239.5833 27.2334 0.0792 123.59 255.164 57 7
3 200.3919 0.2291 202.7959 −1.7273 0.0828 167.40 210.629 41 10
4 356.0263 −0.0101 355.9481 0.2567 0.2719 108.50 196.421 38 1
5 3.2246 8.3751 3.1893 8.3964 0.3625 123.10 191.222 44 0
Note. BCG stands for brightest cluster galaxy. Column (1): supercluster identification number in our catalog. Column (2): R.A. Column (3): decl. Column (4): R.A. of
the most massive member cluster’s BCG. Column (5): decl. of the most massive member cluster’s BCG. Column (6): redshift. Column (7): linear comoving size.
Column (8): total mass. Column (9): number of member clusters. Column (10): number of Abell clusters present.
(This table is available in its entirety in machine-readable form.)
8
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
masses of these superclusters by adding the bound masses of
their member clusters. However, matter also exists between
these clusters within a supercluster. Currently, we do not have
an estimation of the mass attributed to this intercluster matter.
Consequently, the actual mass of the superclusters may be
higher than what is indicated by our calculations. In this
context, Bagchi et al. (2017) found that if one assumes that the
average density of diffuse matter (baryonic plus dark matter)
dispersed between component (member) clusters is minimally
at the cosmological matter density, then there is an increase in
the total mass of the Saraswati supercluster at least by a factor
of 2. Furthermore, it is worth considering that the superclusters
situated at the edges of the SDSS survey footprint and those
near the redshift limits of our sample might only be partially
identified in terms of their true extent. Therefore, factors such
as survey boundaries and limitations imposed by redshift
constraints could potentially impact our ability to capture the
entirety of these superclusters fully. A very deep and dedicated
spectroscopic survey is needed to unravel their true extent and
content. This sample will also be useful for comparing similar
studies using numerical simulations. Also, since the catalog has
a relatively wide range of redshift coverage, spanning ∼4 Gyr,
the study of the evolution of the properties of galaxies in
supercluster environments can be explored.
Our supercluster catalog differs in some ways from two other
catalogs: the one by Chow-Martínez et al. (2014) and the other
by Liivamägi et al. (2012). In the catalog by Chow-Martínez
et al. (2014), the redshift coverage is less, up to z„ 0.15.
However, their sky coverage is significantly larger than ours,
encompassing almost the entire sky, except for the Galactic
disk region. To identify the superclusters, they implemented a
tunable FoF algorithm based on the selection functions, applied
to the Abell/Abell-Corwin-Olowin clusters (Abell 1958; Abell
et al. 1989). The number of member clusters in their catalog
varies from two to 42, and the sizes of the identified
superclusters range from < 1Mpc to ∼185Mpc. On the other
hand, the supercluster catalog of Liivamägi et al. (2012) has a
redshift coverage of 0.02„ z„ 0.5, but its sky coverage is less
than 50% compared to our supercluster catalog. They took a
different approach, identifying supercluster regions using the
luminosity density field generated from the SDSS DR7
(Abazajian et al. 2009) spectroscopic galaxy data. Applying
an adaptive density threshold on this field, they found
supercluster sizes ranging from ∼22 to ∼260Mpc. In contrast
to these catalogs, our supercluster catalog provides additional
information on mass estimates for the identified superclusters.
Along with the wide sky coverage and comparatively large
redshift range of our catalog, this additional information makes
it a useful resource for further research and understanding of
superclusters.
We also find a correlation between the density contrast and
the size of the superclusters. It is close to a power law with an
index α∼−2. The correlation (Figure 6) is
Llog 2.14 0.07 log 4.95 0.13 , 9d = -  + ( ) ( ) ( )
where δ is the density contrast and L is the size of the
supercluster. In our catalog, we notice that this correlation
arises from the dependency of mass (M) and volume (V ) of the
superclusters on its size (M∝ L0.73 and V∝ L2.64). It may arise
due to the morphology distribution of the superclusters or due
to the evolution of the superclusters/structures with time
(Teerikorpi et al. 2015). We leave this aspect to explore in
future work.
5.1.2. Notes on the Five Most Massive Superclusters
Here, we highlight the five most massive superclusters in our
supercluster catalog. Figure A1 shows the sky distribution of
these five superclusters. Figure A2 shows the optical images of
the most massive member clusters of the top five superclusters
in Table 2 from the DESI Legacy Imaging Surveys (DR9).14
These are some of the richest clusters in the WHL cluster
catalog. The WHL cluster catalog places the cluster ZwCl
2341.1+0000 (Figure A2—SCl 4) as the most massive cluster
of the SCl 4 (Saraswati) supercluster. However, A2631 (also
shown in Figure A2) is the most massive cluster of Saraswati
(Bagchi et al. 2017; Monteiro-Oliveira et al. 2021). Zwcl
2341.1+0000 is a multiple-merger system (Bagchi et al. 2002;
van Weeren et al. 2009; Boschin et al. 2013; Parekh et al.
2022), which is why the mass estimation of this cluster is
slightly higher. Individual notes on the top five superclusters
are given below:
SCl 1—new discovery. The most massive supercluster (SCl
1) is found at redshift ∼0.25 with mass ∼2.57× 1016Me and
size ∼111Mpc. This supercluster contains 54 member clusters,
including A1835 and A1801. A1835 is the most massive
member cluster of SCl 1 (see Figure A2). A1835 is a massive
cool core cluster (Ueda et al. 2017) with a very high mass of
∼ 1015M☉ (Schmidt et al. 2001; LaRoque et al. 2006) and a
radio minihalo (Kale et al. 2015). SCl 1 is a newly discovered
supercluster and is the most massive at this redshift. We
propose naming SCl 1 the Einasto Supercluster in honor of
Prof. Jaan Einasto.15
CB Supercluster—rediscovery. In our catalog, the richest
(maximum number of member clusters) supercluster is SCl 2,
with 57 member clusters, including seven Abell clusters, at
redshift z≈ 0.08, mass M≈ 2.55× 1016M☉, and size
≈124Mpc. It contains most of the CB supercluster (Einasto
et al. 2021a). Figure 8 shows the three-dimensional distribution
of the 57 clusters of SCl 2 in comoving coordinate space.
Einasto et al. (2021a) showed that the CB supercluster consists
of two parts, weakly connected by a chain of galaxies and poor
groups. They found that these parts will separate during future
evolution in physical space and form two separate structures.
One of them, which contains Abell clusters A2065, A2061/
A2067, and A2089, will be one of the most massive bound
systems in the local Universe (Einasto et al. 2021a). Our
algorithm combines the CB supercluster with the supercluster
A2142 (see Einasto et al. 2015, 2018, 2020 for details and
references). As a result, the total mass of this double
supercluster system in our catalog is twice as high as the
sum of masses of the CB and A2142 superclusters according to
the estimates in Einasto et al. (2015, 2021c). Also, A2142 is the
most massive cluster in the CB supercluster in our catalog. It is
worth mentioning that the inner structure of the CB super-
cluster is complicated. It consists of multiple clusters connected
by low-density filaments, resembling a group of spiders (called
a multispider in a morphological classification of superclusters
in Einasto et al. 2011b). Overall, the CB supercluster has a
14 https://www.legacysurvey.org
15 Prof. Jaan Einasto is one of the pioneers of the cosmic web. He was one of
the first to start systematic studies of superclusters and still actively contributes
to the field.
9
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
highly elongated shape, resembling a horseshoe (Einasto et al.
2011b).
SGW—rediscovery. The richest supercluster complex in the
local Universe, the SGW is represented by two superclusters in
our catalog, SCl 3 and SCl 22 (Figure A1). In Park et al.
(2012), the spectroscopic galaxy data from SDSS is utilized to
identify nearby structures. Due to the higher density of these
data than our sparser cluster data, these superclusters have been
identified as a single system (Park et al. 2012). The
morphology of SCl 3, the richest SGW supercluster, resembles
a multibranching filament with rich galaxy clusters connected
by a small number of filaments. The second richest supercluster
in SGW, SCl 22 in our catalog, is a multispider, in which
clusters and groups are connected by a large number of
filaments. During future evolution in physical space, both these
superclusters will fall apart and form smaller superclusters
(Einasto et al. 2016). SGW superclusters differ in galaxy
content and morphology, suggesting that they have had
different evolution (Einasto et al. 2010, 2011c).
Saraswati Supercluster—rediscovery. The Saraswati super-
cluster (Bagchi et al. 2017) is a massive supercluster
surrounded by large voids at a redshift ∼0.28. In our catalog,
it is identified with SCl 4 and SCl 231 (Figure A1). The total
number of member clusters is 51 (38 clusters in SCl 4 and 13 in
SCl 231). In Bagchi et al. (2017), a high-density spectroscopic
galaxy sample was used to identify the Saraswati supercluster.
This sample allowed them to observe a continuous structure
encompassing SCl 4 and SCl 231. The combined mass of SCl 4
and SCl 231 is ∼2.6× 1016 Me, slightly higher than, but still
within the same order of magnitude as, the value reported in
Bagchi et al. (2017). The increased mass is attributed to the
inclusion of more member clusters, made possible by the wider
sky area coverage of the WHL cluster catalog compared to the
spectroscopic galaxy sample used in Bagchi et al. (2017). The
most massive cluster of Saraswati is A2631, and the second
most massive cluster is ZwCl 2341+0000. Bagchi et al. (2017)
showed that the central core region within a radius ∼20Mpc of
this supercluster, including A2631, is gravitationally bound.
The total size of this supercluster is ∼200Mpc, which is the
largest supercluster at this redshift.
SCl 5—new discovery. SCl 5 is our catalog’s fifth most
massive supercluster, containing 44 clusters at a redshift ∼0.36.
Its mass and size are ∼1.9× 1016 Me and ∼123Mpc,
respectively. It does not contain any Abell cluster. It is also a
newly discovered supercluster.
5.1.3. Effect of Supercluster Environment on the Clusters
The cosmic environment affects the growth and evolution of
galaxies and clusters. To explore whether the supercluster
environment plays any role in the evolution of clusters, we
study the mass distribution of clusters in superclusters and in
the field (clusters that are not members of any supercluster).
The left panel of Figure 9 shows the normalized mass
distributions of the member clusters of superclusters and the
field clusters. It shows a slight mass bias of finding massive
clusters in a supercluster environment rather than in the field.
Similarly, there is a slight deficiency of less massive clusters
within superclusters. In other words, the probability of
randomly picking a massive cluster is slightly higher within a
supercluster environment than in the field. And the probability
of randomly picking a less massive cluster is slightly higher in
the field than in a supercluster region. To see the difference
between the two distributions, we perform a two-sample
Kolmogorov–Smirnov test (Kolmogorov 1933; Smirnov 1948).
Figure 8. The three-dimensional distribution of clusters in the richest supercluster of our catalog (SCl 2), the CB supercluster. The colors represent the mass of the
clusters, with blue representing the less massive clusters and yellow the most massive ones. The red spheres represent the Abell clusters; their Abell IDs are written
next to them. There are a total of 57 clusters in SCl 2.
10
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
The null hypothesis that the two samples are drawn from the
same population is rejected with a p-value of 3.7× 10−52. This
shows that the supercluster environment affects the evolution
of clusters.
We also explore the mass distribution of the member clusters
within each supercluster. The right panel of Figure 9 shows the
median values (red dots) of the mass of clusters in a
supercluster as a function of the total mass of the supercluster.
Here, despite a large scatter, we see a trend indicating that the
low-mass superclusters (masses 5× 1015 Me) host low-mass
member clusters. This may suggest that the growth of clusters
is higher in high-mass superclusters. However, this growth in
mass is not very high: the median value goes from
∼1× 1014Me to ∼1.8× 10
14 Me.
5.2. Simulation
5.2.1. Comparison with Simulation
We compare our results with the mock superclusters
extracted from the mock WHL clusters of the HR4 simulation
(as described in Section 2.2). To compare the redshift coverage
of the mock clusters, we restrict our observational data of the
clusters to 0.05„ z„ 0.366. We then apply the same mFoF
algorithm on the observed and mock clusters to identify
superclusters. The results are summarized in Figures 10 and 11
and Table 3.
The distributions of the size, density contrast, and the
number of member clusters match fairly well, but there is a
slight difference in the distribution of masses of superclusters
(Figure 10). The difference in the mass distributions is because
of the deficiency of massive clusters in the mock WHL clusters
made from the HR4 simulation (see Section 2.2).
The linking lengths for WHL and the mock clusters are
19.22Mpc and 19.88Mpc, respectively (Figure 11). This
difference of 0.66Mpc is acceptable, as it is much less than the
mean cluster separation (∼40 Mpc) and less than the mean
diameter of a cluster (∼2 Mpc).
We also see the correlation between the density contrast and
the size of a supercluster in the mock superclusters (Figure 11).
The value of the slope here is ∼−2.04± 0.10.
5.2.2. Peculiar Velocities of Members in the Mock Superclusters
In Figure 12, we show the radial peculiar velocities as a
function of their comoving distance from the centers of all the
superclusters stacked in the reference frame of the center of
mass of the supercluster. The radial peculiar velocity vpec of a
member cluster is
v v rv , 10SCpec mem= -( ) · ˆ ( )
where vmem is the peculiar velocity vector (as given in the HR4
simulation) of the member cluster, vSC is the peculiar velocity
vector of the center of mass of the supercluster, and rˆ is the
comoving position unit vector of the member cluster in the
frame of reference of the center of mass of the supercluster.
Figure 12 shows the phase-space distribution of member
clusters, where the y-axis shows the radial peculiar velocities
(the peculiar velocity component of a member cluster in the
reference frame of the center of mass and along the line from
the center of mass to the member clusters) of the member
clusters of all superclusters and the x-axis is the comoving
distance of a member cluster from the center of mass of a
supercluster. The black line represents a radial peculiar velocity
of 0 km s−1 and the red line represents all the radial peculiar
Figure 9. (a) Normalized histograms of the masses of clusters within superclusters (red histogram) and in the field (blue histogram). (b) Median masses (red dots) of
the member clusters of superclusters as a function of their total mass. The blue error bars denote the quartile values of the member cluster masses above and below the
median values, and the solid black line is the running mean of the red dots.
11
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
velocities equal and opposite to the Hubble flow as seen from
the center of mass frame. The three regions divided by red and
black lines tell us the kinematical state of the member clusters.
Hubble Decoupled: member clusters that have decoupled from
the Hubble flow and are moving toward the supercluster center;
Supercluster Influence: member clusters that have not
Figure 10. Distributions of the size, mass, density contrast, and the number of member clusters of the superclusters in WHL and HR4 within 0.05 „ z „ 0.366.
Figure 11. (a) Number of candidate superclusters for different linking lengths l0 for WHL and the mocks. The curves peak at 19.22 Mpc and 19.88 Mpc for WHL and
the mocks, respectively, and are represented by dotted lines. (b) Density contrast of the mock superclusters as a function of size. The blue dots represent the mock
superclusters, the red line is a linear fit to the data points on the log–log plot, and the shaded red region denotes the associated errors in the fit. The slope of the fitted
line is given at the top right corner of the plot. The black line and the shaded black region denote the corresponding fit and the associated errors of WHL superclusters
presented in Figure 6(c).
12
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
decoupled from the Hubble flow, but have slowed down in
their Hubble expansion; and Outside Influence: member
clusters that have their velocities greater than the Hubble flow
and therefore have gravitational influence from regions outside
of the supercluster. On average, 89%, 9%, and 2% of the
member clusters are in the regions of “Supercluster Influence,”
“Outside Influence,” and “Hubble Decoupled,” respectively.
For example, Figure 13 shows the supercluster-centric spatial
Figure 12. Radial peculiar velocities of the member clusters of all (stacked) superclusters as a function of their comoving distance from the supercluster center. The
black line represents a peculiar velocity of 0 km s−1, and the red line represents all the peculiar velocities equal and opposite to the Hubble flow as seen from the
center. The three regions, divided by red and black lines, are Hubble Decoupled: members that have decoupled from the Hubble flow; Supercluster Influence:
members that have not decoupled from the Hubble flow, but have slowed down in their Hubble expansion; and Outside Influence: member clusters that have their
velocities greater than the Hubble flow and therefore have gravitational influence from regions outside of the supercluster.
Figure 13. The most massive supercluster in the mock. The “∗” symbol indicates the center of mass of the supercluster. Member clusters are represented by circles and
their colors indicate the masses M200c. The red arrows show the supercluster-centric peculiar velocity components (radial peculiar velocities) of the member clusters,
with arrow lengths proportional to the radial peculiar velocities. Here, an arrow with a length of 1 Mpc corresponds to 60 km s−1.
Table 3
Summary of the Properties of the Observed and Mock Superclusters
WHL WHL Mock
(0.05 „z „0.42) (0.05 „z „0.366) (0.05 „z „0.366)
Linking Length lo (Mpc) 20.65 19.22 19.88
Number of Superclusters 662 456 451
Median Mass (1014 Me) 57.80 55.97 46.43
Median Size (Mpc) 64.87 60.49 60.83
Median Density Contrast 11.64 15.40 10.88
Median No. of Members 13 13 12
13
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
distribution of the member clusters of the most massive
supercluster in the mock. The arrows indicate the supercluster-
centric peculiar velocity components (radial peculiar velo-
cities), almost all pointing toward the supercluster center
indicated by the “∗” symbol.
5.2.3. Phase-space Distribution of Mock Halos in and around
Superclusters
As shown in Figure 12, most supercluster member halos
(91%) have a negative supercluster-centric peculiar velocity
component. The member halos are influenced by the
gravitational potential of their host supercluster. This influence
should be negligible on the halos well outside the supercluster
region. The phase-space distribution of halos within the
supercluster region should be distinguishable from those
outside the supercluster region. If we take the supercluster-
centric peculiar velocity components of all halos from the
supercluster center up to some distance outside the super-
cluster, we should see the overall negative components going to
zero values at higher distances outside the supercluster. To
explore this effect, we present the stacked supercluster-centric
phase-space distribution in and around the supercluster region
in Figure 14. The x-axis shows the scaled distance from the
supercluster center. If S is the size of a supercluster, then we
define a quantity, rs= S/2, of a supercluster. The supercluster-
centric comoving distances r of the halos are then scaled to rs.
This gives a unit-scaled distance as the rough extent of a
supercluster. The left panel shows the two-dimensional
histogram of the halos in and around the superclusters (stacked)
up to a scaled distance, r/rs= 8. In this panel, we can see that
there is a detached distribution (r/rs 1 and vpec,h 0) from
the main distribution (r/rs 1), where vpec,h is the supercluster-
centric peculiar velocity component of the halos. This detached
distribution is the supercluster region where most halos have
negative radial peculiar velocity components. Beyond the
supercluster region (r/rs 1), when it starts to encounter
surrounding voids and other cosmic web components/
structures, this distribution of the negative component of the
peculiar velocity gradually and slowly moves toward zero.
The right panel of Figure 14 shows the mean values (blue
line) of the radial peculiar velocities of halos at a certain scaled
distance from the supercluster center. The shaded blue region
shows the standard deviation (1σ) of the supercluster-centric
peculiar velocity distribution above and below the mean values.
It shows that at a distance approximately four to five times rs,
the value goes to the global average of zero, which is expected
on larger scales. The red line shows the relative number density
of halos as a function of the scaled distance. It is calculated by
dividing the average number density ρN (of all stacked
superclusters) at a scaled radius by the total number density
〈ρN〉 of halos within a sphere of radius ∼8× rs. Both the
velocity and density profiles show a characteristic scale at a
distance of ∼1.5× rs.
It should be noted that the superclusters in our catalog are
not spherically symmetric, and the supercluster center, which is
the barycenter of the member clusters, may not coincide with
the geometric center of the supercluster. Because of this
asymmetry, the characteristic scale is not r/rs= 1 and comes
out to be r/rs∼ 1.5. After this characteristic scale, both profiles
gradually approach their global values.
Figure 14. Stacked radial peculiar velocities and relative number density in and around superclusters. The x-axis denotes the distance from the supercluster center in
units of rs (half the size of a supercluster). (a) Two-dimensional histogram of the mock halos in and around superclusters. N denotes the number of halos in each grid.
The region (r/rs  1 and vpec,h  0) contains most of the member halos of a supercluster. (b) Mean (blue line) values of the radial peculiar velocity component as a
function of scaled distance. The shaded blue region shows the standard deviation (1σ) of the radial peculiar velocity component. The red line shows the relative
number density of halos as a function of scaled radius. This relative number density is calculated by dividing the average number density ρN (of all stacked
superclusters) at a scaled radius by the total number density 〈ρN〉 of halos within a sphere of radius ∼8 × rs.
14
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
Figures 12 and 14 support the robustness of the superclusters
identified in our catalog. The density contrast of all the
superclusters in our catalog has values greater than zero. The
relative density shows a knee around 1.5 times the scaled
distance, also where the shape of the radial peculiar velocity
profile changes. This suggests that these superclusters are well
within the basin of attraction. The overall densities of the
superclusters are not sufficiently high enough to classify them
as gravitationally bound structures within the framework of a
spherical collapse model. Instead, they are more likely to
fragment into smaller structures, similar to the observed cases
of the Saraswati supercluster (Bagchi et al. 2017) and the BGW
supercluster (Einasto et al. 2022). We therefore conclude that
our superclusters fall in the supercluster definition (2)—the
unbound overdense regions in the Universe (see Section 1).
6. Discussion and Conclusion
Superclusters, being the largest structures in the Universe,
need more focused multiwavelength studies than their current
status of study. Whether they grow with a bottom-up process, a
top-down process, or both needs to be addressed observation-
ally as well as theoretically. Identifying superclusters and
creating a statistically significant sample will increase our
understanding of their properties and their environmental
effects on the galaxies, groups, and clusters in them.
As the galaxy redshift surveys have become deeper and
larger over time, not only more superclusters, but also bigger
superclusters, have been discovered. Hence, it will not be
surprising if even bigger superclusters are discovered in the
upcoming redshift surveys like DESI (DESI Collaboration et al.
2016), 4MOST (de Jong et al. 2019; Richard et al. 2019),
Euclid (Laureijs et al. 2011), etc. And this may pose a
challenge to the ΛCDM model of the Universe (Park
et al. 2012).
In this work, we have identified 662 superclusters in the
redshift range 0.05„ z„ 0.42 using the spatial distribution of
WHL clusters. We have applied an mFoF algorithm to
overcome the survey/catalog selection effects. These super-
clusters have typical masses >1015 Me and sizes >10 Mpc.
Einasto Supercluster, a new discovery at z∼ 0.25, is found to
be the most massive supercluster in our catalog.
In our process of identifying superclusters, we have
rediscovered many previously known superclusters, as well
as discovered many new ones. We found a power-law relation
between a supercluster’s density contrast and size with an index
∼−2. Whether this suggests a relation with the morphology of
the supercluster or some other relation is a matter of further
study. The topological analysis of the shape of superclusters
may explain this phenomenon more clearly. Recently, Bag
et al. (2023) found that large superclusters with volumes
104 Mpc3 tend to be more filamentary, and Heinämäki et al.
(2022) found that the low-luminosity, small, poor, and low-
mass end of superclusters has pancake-type shapes, while only
a handful exhibit exceptionally spherical shapes. The super-
cluster environment weakly affects the evolution of clusters. It
is slightly more likely to find a massive cluster in a supercluster
environment than in a nonsupercluster environment.
The simulation gives slightly fewer high-mass clusters than
those found in the observations. This gives a difference in the
mass distribution of halos/clusters in the simulation and
observational data and affects our comparison of the properties
of superclusters found in the simulations and observations.
Nevertheless, we found almost comparable properties in the
simulated and SDSS data. The peculiar velocity field of the
mock superclusters in the simulation shows that the most part
of a supercluster points toward the center of mass of the
supercluster. That is, the most part of a supercluster is
influenced by its mass and overdensity. This leads to a slowing
down of the expansion of the supercluster with respect to the
Hubble flow. The phase-space distribution of halos in and
around the superclusters shows a characteristic length scale
when a change in radial peculiar velocity and density profiles is
seen. This length scale lies approximately 1.5 times the scaled
distance from the supercluster center and arises due to the
nonspherical shapes of the superclusters.
Superclusters host some of the most massive galaxy clusters
formed through mergers. Under the right conditions, upon a
merger, the resulting propagating shock manifests as mega-
parsec-scale radio relics (e.g., Bagchi et al. 2002, 2006; van
Weeren et al. 2009) and radio halos (for a review, see van
Weeren et al. 2019). Recently, using the eROSITA Final
Equatorial Depth Survey, Ghirardini et al. (2021) reported a
new supercluster at a redshift of z∼ 0.36 with eight associated
galaxy cluster members and spanning ∼ 27Mpc. They have
also carried out detailed X-ray, optical, and radio studies of the
supercluster, where they found two new radio relics and a radio
halo. The Shapley supercluster has been the subject of
numerous multiwavelength studies in the past few decades.
Its central part has recently been studied at radio wavelengths
using sensitive observations from uGMRT, MeerKAT, and
ASKAP (Venturi et al. 2022). In addition to finding radio halos
and very low surface brightness radio emissions connecting
groups and clusters (∼ 1 Mpc), such deep multifrequency
observations also enable estimation of the equipartition
magnetic field (∼ 0.76 μG) permeating across a large region.
Hence, the study by Venturi et al. (2022) provides compelling
evidence of nonthermal signatures detected from minor
mergers (e.g., megaparsec-scale radio emission from bridges
connecting clusters and groups).
Simulations show that supercluster embryos form at very
early cosmological epochs. Even in observations, superclusters
have been detected at higher redshifts, extending as far back as
z∼ 2.45 (Kim et al. 2016; Lietzen et al. 2016; Cucciati et al.
2018; Shimakawa et al. 2023). The location of superclusters
does not change much during evolution, and the essential
evolutionary changes occur within the supercluster cocoons
(basins of attraction or the regions of dynamical influence;
Einasto et al. 2019, 2021a). On the question of the formation of
the superclusters, Einasto et al. (2011a) and Suhhonenko et al.
(2011), using wavelet analysis of the cosmic web, show that
rich clusters and superclusters form where density waves of
medium (;32 h−1Mpc) and large scales (…64 h−1Mpc)
combine in similar phases to generate high-density regions.
They show that the synchronization (or coupling) of differently
scaled density waves plays an important role in structure
formation. Their results show that the largest structures to form
have sizes of approximately 100 h−1 Mpc, which agrees with
the size of the largest superclusters in our study. Density waves
with larger scales only modify the properties of structures.
Underdense regions (voids) that surround superclusters form
where different phases of the density waves are combined.
Matter within the voids flows toward the inner regions of
superclusters due to gravitational instability. Observationally,
15
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
this has been shown by analyzing local velocity fields (Tully
et al. 2014; Hoffman et al. 2017). The voids may be expanding
more rapidly than the global expansion rate and thus helping in
the growth of the supercluster (Sheth & van de Weygaert 2004;
van de Weygaert 2016; Bagchi et al. 2017; Hoffman et al.
2017). The expansion of large voids can facilitate shaping
surroundings and help in the formation of long coherent
structures (superclusters) along its boundary or periphery.
Superclusters are important to study the formation of
structures on large scales. For further advances in this subject,
a detailed spectroscopic survey of a massive supercluster is
needed (e.g., the Saraswati supercluster), which can inform us
more about the dynamical state of the supercluster. Such
surveys will provide the crucial information needed to answer
open questions in supercluster physics, provide a wealth of data
for individual galaxies, and pave the path for more synergistic
studies. Future deep, wide-sky galaxy surveys will provide the
opportunity to compare the abundance and properties of
superclusters at higher redshifts with the ones in simulations.
Therefore, it is vital to identify and characterize superclusters.
It is important to compare the properties of galaxies in large,
high-density regions like superclusters and in underdense
regions like voids to understand the factors affecting their
growth and evolution. The current era of deep multiwavelength
large sky surveys provides us with the perfect opportunity.
Acknowledgments
We thank the referee for the valuable and constructive
comments, which have enhanced the paper’s clarity and
readability. S.S. acknowledges the support of the European
Regional Development Fund, the Mobilitas Pluss postdoctoral
research grant MOBJD660, and the ETAg grant PRG1006.
E.T. and M.E. acknowledge the support from ETAg grant
PRG1006 and EU through the ERDF CoE TK133. J.B.
acknowledges the help and full support from Christ (Deemed to
be) University in facilitating the research. We kindly thank
Changbom Park, Sung-Wook Kim, and Juhan Kim for
providing the HR4 simulation products on request.
Appendix
Five Most Massive Superclusters
Figure A1 shows the sky plane distribution of the five most
massive superclusters. The member clusters are shown with
their sizes proportional to their R200c, and the color scale
indicates their masses (M200c). The different marker types
denote different superclusters. The black horizontal line shows
a comoving scale of 50Mpc. Figure A2 shows the optical
images of the most massive member clusters of the top five
superclusters in Table 2. For more details, see Section 5.1.2.
16
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
Figure A1. Sky plane distribution of the five most massive superclusters identified in our supercluster catalog.
17
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
Figure A2. Optical images of the most massive member clusters in each of the top five superclusters in Table 2. Abell 2631 is actually the most massive cluster of SCl
4 (the Saraswati supercluster). The images are extracted from the DESI Legacy Imaging Surveys (DR9).
18
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.
ORCID iDs
Shishir Sankhyayan https://orcid.org/0000-0003-2601-2707
Joydeep Bagchi https://orcid.org/0000-0002-2922-2884
Elmo Tempel https://orcid.org/0000-0002-5249-7018
Surhud More https://orcid.org/0000-0002-2986-2371
Maret Einasto https://orcid.org/0000-0003-3722-8239
Pratik Dabhade https://orcid.org/0000-0001-9212-3574
Somak Raychaudhury https://orcid.org/0000-0002-
4864-4046
Ramana Athreya https://orcid.org/0000-0001-7141-7311
Pekka Heinämäki https://orcid.org/0000-0002-1568-0227
References
Abazajian, K. N., Adelman-McCarthy, J. K., Agüeros, M. A., et al. 2009,
ApJS, 182, 543
Abell, G. O. 1958, ApJS, 3, 211
Abell, G. O., Corwin, H. G., Jr., & Olowin, R. P. 1989, ApJS, 70, 1
Alam, S., Albareti, F. D., Allende Prieto, C., et al. 2015, ApJS, 219, 12
Alfaro, I. G., Rodriguez, F., Ruiz, A. N., Luparello, H. E., & Lambas, D. G.
2022, A&A, 665, A44
Araya-Melo, P. A., Reisenegger, A., Meza, A., et al. 2009, MNRAS, 399, 97
Bag, S., Liivamägi, L. J., & Einasto, M. 2023, MNRAS, 521, 4712
Bagchi, J., Durret, F., Neto, G. B. L., & Paul, S. 2006, Sci, 314, 791
Bagchi, J., Enßlin, T. A., Miniati, F., et al. 2002, NewA, 7, 249
Bagchi, J., Sankhyayan, S., Sarkar, P., et al. 2017, ApJ, 844, 25
Barber, C. B., Dobkin, D. P., Dobkin, D. P., & Huhdanpaa, H. 1996, ACM
Trans. Math. Softw., 22, 469
Bond, J. R., Kofman, L., & Pogosyan, D. 1996, Natur, 380, 603
Boschin, W., Girardi, M., & Barrena, R. 2013, MNRAS, 434, 772
Busha, M. T., Evrard, A. E., Adams, F. C., & Wechsler, R. H. 2005, MNRAS,
363, L11
Chon, G., Böhringer, H., & Zaroubi, S. 2015, A&A, 575, L14
Chow-Martínez, M., Andernach, H., Caretta, C. A., & Trejo-Alonso, J. J. 2014,
MNRAS, 445, 4073
Cucciati, O., Lemaux, B. C., Zamorani, G., et al. 2018, A&A, 619, A49
de Jong, R. S., Agertz, O., Berbel, A. A., et al. 2019, Msngr, 175, 3
Delaunay, B. 1934, Sur la sphére vide, Izvestia Akad Nauk SSSR Otdel. Mat.
Sov. Nauk, 7, 793
DESI Collaboration, Aghamousa, A., & Aguilar, J. 2016, arXiv:1611.00036
Dunkley, J., Komatsu, E., Nolta, M. R., et al. 2009, ApJS, 180, 306
Dünner, R., Araya, P. A., Meza, A., & Reisenegger, A. 2006, MNRAS,
366, 803
Dupuy, A., Courtois, H. M., Dupont, F., et al. 2019, MNRAS, 489, L1
Einasto, J., Einasto, M., Tago, E., et al. 2007a, A&A, 462, 811
Einasto, J., Hütsi, G., Saar, E., et al. 2011a, A&A, 531, A75
Einasto, J., Hütsi, G., Suhhonenko, I., Liivamägi, L. J., & Einasto, M. 2021a,
A&A, 647, A17
Einasto, J., Klypin, A., Hütsi, G., Liivamägi, L. J., & Einasto, M. 2021b, A&A,
652, A94
Einasto, J., Suhhonenko, I., Liivamägi, L. J., & Einasto, M. 2019, A&A,
623, A97
Einasto, M., Deshev, B., Lietzen, H., et al. 2018, A&A, 610, A82
Einasto, M., Deshev, B., Tenjes, P., et al. 2020, A&A, 641, A172
Einasto, M., Einasto, J., Tago, E., et al. 2007b, A&A, 464, 815
Einasto, M., Gramann, M., Saar, E., et al. 2015, A&A, 580, A69
Einasto, M., Kipper, R., Tenjes, P., et al. 2021c, A&A, 649, A51
Einasto, M., Lietzen, H., Gramann, M., et al. 2016, A&A, 595, A70
Einasto, M., Lietzen, H., Tempel, E., et al. 2014, A&A, 562, A87
Einasto, M., Liivamägi, L. J., Tago, E., et al. 2011b, A&A, 532, A5
Einasto, M., Liivamägi, L. J., Tempel, E., et al. 2011c, ApJ, 736, 51
Einasto, M., Tago, E., Jaaniste, J., Einasto, J., & Andernach, H. 1997, A&AS,
123, 119
Einasto, M., Tago, E., Saar, E., et al. 2010, A&A, 522, A92
Einasto, M., Tenjes, P., Gramann, M., et al. 2022, A&A, 666, A52
Ettori, S., & Balestra, I. 2009, A&A, 496, 343
Frieman, J. A., Turner, M. S., & Huterer, D. 2008, ARA&A, 46, 385
Ghirardini, V., Bulbul, E., Hoang, D. N., et al. 2021, A&A, 647, A4
Górski, K. M., Hivon, E., Banday, A. J., et al. 2005, ApJ, 622, 759
Gott, J. R., III, Jurić, M., Schlegel, D., et al. 2005, ApJ, 624, 463
Heinämäki, P., Teerikorpi, P., Douspis, M., et al. 2022, A&A, 668, A37
Hoffman, Y., Pomarède, D., Tully, R. B., & Courtois, H. M. 2017, NatAs,
1, 0036
Huchra, J. P., & Geller, M. J. 1982, ApJ, 257, 423
Kale, R., Venturi, T., Giacintucci, S., et al. 2015, A&A, 579, A92
Kim, J., Park, C., L’Huillier, B., & Hong, S. E. 2015, JKAS, 48, 213
Kim, J. W., Im, M., Lee, S. K., et al. 2016, ApJL, 821, L10
Knebe, A., Knollmann, S. R., Muldrew, S. I., et al. 2011, MNRAS, 415, 2293
Kolmogorov, A. L. 1933, G. Ist. Ital. Attuari, 4, 83
LaRoque, S. J., Bonamente, M., Carlstrom, J. E., et al. 2006, ApJ, 652, 917
Laureijs, R., Amiaux, J., Arduini, S., et al. 2011, arXiv:1110.3193
Lietzen, H., & Einasto, M. 2016, in IAU Symp. 308, The Zeldovich Universe:
Genesis and Growth of the Cosmic Web, ed. R. van de Weygaert et al.
(Cambridge: Cambridge Univ. Press), 412
Lietzen, H., Tempel, E., Heinämäki, P., et al. 2012, A&A, 545, A104
Lietzen, H., Tempel, E., Liivamägi, L. J., et al. 2016, A&A, 588, L4
Liivamägi, L. J., Tempel, E., & Saar, E. 2012, A&A, 539, A80
Luparello, H., Lares, M., Lambas, D. G., & Padilla, N. 2011, MNRAS,
415, 964
Martínez, V. J., & Saar, E. 2002, Statistics of the Galaxy Distribution (Boca
Raton, FL: Chapman & Hall/CRC)
Monteiro-Oliveira, R., Soja, A. C., Ribeiro, A. L. B., et al. 2021, MNRAS,
501, 756
Nadathur, S., & Crittenden, R. 2016, ApJL, 830, L19
Neyrinck, M. C. 2008, MNRAS, 386, 2101
Parekh, V., Kincaid, R., Thorat, K., et al. 2022, MNRAS, 509, 3086
Park, C., Choi, Y. Y., Kim, J., et al. 2012, ApJL, 759, L7
Pomarède, D., Tully, R. B., Graziani, R., et al. 2020, ApJ, 897, 133
Pomarède, D., Tully, R. B., Hoffman, Y., & Courtois, H. M. 2015, ApJ,
812, 17
Raychaudhury, S. 1989, Natur, 342, 251
Richard, J., Kneib, J. P., Blake, C., et al. 2019, Msngr, 175, 50
Rines, K., & Diaferio, A. 2006, AJ, 132, 1275
Rines, K., Geller, M. J., Diaferio, A., & Kurtz, M. J. 2013, ApJ, 767, 15
Schmidt, R. W., Allen, S. W., & Fabian, A. C. 2001, MNRAS, 327, 1057
Seth, R., & Raychaudhury, S. 2020, MNRAS, 497, 466
Shapley, H. 1930, BHarO, 874, 9
Sheth, R. K., & van de Weygaert, R. 2004, MNRAS, 350, 517
Shimakawa, R., Okabe, N., Shirasaki, M., & Tanaka, M. 2023, MNRAS,
519, L45
Smirnov, N. 1948, Ann. Math. Stat., 19, 279
Suhhonenko, I., Einasto, J., Liivamägi, L. J., et al. 2011, A&A, 531, A149
Teerikorpi, P., Heinämäki, P., Nurmi, P., et al. 2015, A&A, 577, A144
Trayford, J. W., Theuns, T., Bower, R. G., et al. 2015, MNRAS, 452, 2879
Tully, R. B., Courtois, H., Hoffman, Y., & Pomarède, D. 2014, Natur, 513, 71
Tuominen, T., Nevalainen, J., Tempel, E., et al. 2021, A&A, 646, A156
Ueda, S., Kitayama, T., & Dotani, T. 2017, ApJ, 837, 34
van de Weygaert, R. 2016, in IAU Symp. 308, The Zeldovich Universe:
Genesis and Growth of the Cosmic Web, ed. R. van de Weygaert et al.
(Cambridge: Cambridge Univ. Press), 493
van Weeren, R. J., de Gasperin, F., Akamatsu, H., et al. 2019, SSRv, 215, 16
van Weeren, R. J., Röttgering, H. J. A., Bagchi, J., et al. 2009, A&A, 506, 1083
Venturi, T., Giacintucci, S., Merluzzi, P., et al. 2022, A&A, 660, A81
Wen, Z. L., & Han, J. L. 2015, ApJ, 807, 178
Wen, Z. L., Han, J. L., & Liu, F. S. 2012, ApJS, 199, 34
York, D. G., Adelman, J., Anderson, J. E., Jr., et al. 2000, AJ, 120, 1579
19
The Astrophysical Journal, 958:62 (19pp), 2023 November 20 Sankhyayan et al.