Astronomy & Astrophysics manuscript no.

main ©ESO 2025

April 4, 2025

Galaxy Transformation Across the Cosmic Web: Evolution of

Stellar Colours and Star Formation Rates in Filaments
S. Zarattini1, 2, 3 & J. A. L. Aguerri2, 3

1
Centro de Estudios de Física del Cosmos de Aragón (CEFCA), Unidad Asociada al CSIC, Plaza San Juan 1, 44001 Teruel, Spain
2
Instituto de Astrofísica de Canarias, calle Vía Láctea s/n, E-38205 La Laguna, Tenerife, Spain
3
Departamento de Astrofísica, Universidad de La Laguna, Avenida Astrofísico Francisco Sánchez s/n, E-38206 La Laguna, Spain
[email protected]

April 4, 2025
arXiv:2504.02026v1 [astro-ph.GA] 2 Apr 2025

ABSTRACT

Context. Galaxies form and evolve within the diverse environments of the cosmic web, which shape their properties in unique ways.
However, disentangling the effects of internal properties, local environments, and large-scale structures on galaxy evolution involves
significant complexities, since these effects are frequently overlapping.
Aims. The aim of this work is to provide evidence of the imprints left by large-scale structure on galaxy properties by selecting
extensive samples of galaxies from different environments, while matching their intrinsic and local environmental characteristics.
Methods. We investigate the effects of the large-scale structure on the g − r stellar colour and star formation rate (SFR) in galaxies
with stellar mass M∗ > 1010 M⊙ , within the redshift range 0.05 ≤ z ≤ 0.1, and selected from the main galaxy sample of the Sloan
Digital Sky Survey Data Release 16. We use a spectroscopic catalogue available in the literature to define samples of galaxies located
in field and filament environments.
Results. Galaxies located in the field tend to exhibit bluer g − r stellar colours and higher SFR than those in filaments. These
differences persist even in samples of galaxies matched by mass and galaxy local overdensity, indicating that they are not produced
by internal or local environmental processes. These differences cannot be attributed to variations in morphology between the two
matched samples. There is also a variation in stellar colour and SFR with distance to the filaments (D f ila ). The SFR of galaxies
becomes statistically smaller than that of field galaxies for objects located at D f ila < 5 Mpc, while changes in stellar colour occur at
smaller distances (D f ila < 1 Mpc). This could indicates that the typical filament width is about 2.5 - 5 Mpc.
Conclusions. The variations in colour and SFR between galaxy samples in the field and in filaments, with matched masses and local
overdensities, indicate that large-scale environmental factors drive these transformations rather than local or internal galaxy properties.
These transformations are not strong enough to produce changes in galaxy morphology. They could be explained by a cosmic web
starvation process, in which galaxies moving from the field into filaments become detached from their gas supply, leading to a gradual
decline in SFR and a subsequent reddening of their stellar populations.
Key words. large-scale structure of Universe – galaxies: evolution

1. Introduction have observationally confirmed the theoretical predictions re-

garding the structure of the cosmic web.
The Cold Dark Matter (CDM) paradigm states that the large- Galaxy evolution is strongly driven by gravity through grav-
scale distribution of matter in the Universe on megaparsec scales itational interactions, both local and global. Local gravitational
is not homogeneous. Instead, it forms a cosmic web consisting interactions occur in typical scales of smaller than 1 Mpc due
of galaxy filaments, walls, voids, and nodes. Galaxies form and to nearby companions or within the local galactic environment.
evolve within these different environments, which directly influ- These interactions can significantly impact a galaxy’s morphol-
ence their properties throughout their lifetimes (see e.g. Kraljic ogy, dynamics, and stellar/gas content. Several types of local
et al. 2018; Xu et al. 2020). gravitational interactions can affect galaxies, including galaxy
mergers, tidal interactions, satellite accretion, and interactions
The primary factor shaping this cosmic web is the influence with the hot intracluster medium (Blanton & Moustakas 2009;
on the matter of the large-scale tidal field. Subsequent gravita- Schaye et al. 2015)
tional collapse amplifies anisotropies in the primordial matter Galaxy mergers are crucial for understanding galaxy evo-
distribution, resulting in a today Universe with highly asymmet- lution. The current cosmological paradigm of galaxy formation
ric structures (Zel’dovich 1970). The cosmic structure, predicted establishes that galaxies are assembled in a bottom-up scenario,
by theory, has also been validated by N-body numerical simu- through the mergers of smaller halos (White & Rees 1978; White
lations (see e.g. White et al. 1987; Springel et al. 2005; Dolag & Frenk 1991). Mergers between galaxies of similar mass typ-
et al. 2006). Moreover, extensive spectroscopic surveys, such ically result in the formation of massive spheroidal galaxies,
as the Sloan Digital Sky Survey (SDSS, York et al. 2000), the leading to significant transformations in the morphologies of the
2dF Galaxy Redshift Survey (2dFGRS, Colless et al. 2001), the parent galaxies (see e.g. Toomre & Toomre 1972; Somerville
Galaxy And Mass Assembly survey (GAMA, Driver et al. 2009), & Davé 2015). These interactions can also trigger intense star
and the 2MASS redshift survey (2MASS, Huchra et al. 2012), formation in the resulting galaxy due to the rapid consumption
Article number, page 1 of 11
 A&A proofs: manuscript no. main

of gas from the pre-merger galaxies (Barnes & Hernquist 1996; forming than their counterparts far from filaments. This result
Kim et al. 2009). However, other studies such as Knapen et al. would be associated with streams of cold gas moving together
(2015) or Pearson et al. (2019) did not find this star formation with galaxies in the filaments, that are able to prevent strangula-
triggering. The merger processes have often been invoked to ex- tion by shielding the effect of the hot intra-cluster medium. Das
plain the observed dependence of galaxy morphology (Dressler et al. (2023) studied the effect of filaments on galaxy pairs, find-
1980), stellar colour (Baldry et al. 2006; Bamford et al. 2009), ing that close pairs are bluer and more star forming in filaments,
and star formation (Kauffmann et al. 2004) on the local environ- whereas the trend inverts for large pairs, with separation > 50
ment. Conversely, mergers between galaxies of vastly different kpc.
masses—often referred to as satellite accretion—slightly affect The large-scale structure can not only influence intrinsic
the structure of the more massive galaxy while mixing the gas galaxy properties but also produce significant effects on the effi-
and stars of the less massive galaxy into the larger one (Aguerri ciency of galaxy formation. Specifically, Guo et al. (2015) found
et al. 2001; Eliche-Moral et al. 2006, 2011). that the satellite luminosity function (LF) of galaxies in filaments
When galaxies pass near one another, they experience tidal is higher (i.e. more satellites) than that of galaxies not located in
gravitational fields that can alter their morphology, resulting in filaments.
long tidal tails or even inducing structures such as bars (Hib- For years, there has been ongoing debate about whether the
bard & Mihos 1995; Łokas et al. 2016; Martinez-Valpuesta et al. observed variations in galaxy properties are primarily driven by
2017) or lopsided discs (Łokas 2021, 2022). Repeated fast tidal internal factors or external environmental influences. Some stud-
encounters produce what is called galaxy harassment which can ies suggested that internal properties, like stellar mass and mor-
also shape the morphology of the galaxies (Moore et al. 1996, phology, play a dominant role in shaping galaxy characteristics
1998). Additionally, interactions among galaxies in local envi- (Alpaslan et al. 2015; Wegner et al. 2019). On the other hand,
ronments, such as galaxy clusters, can lead to variations in their research by Eardley et al. (2015) argued that environmental fac-
cold gas content (through processes like ram pressure stripping, tors, such as local galaxy density and position within the cosmic
Gunn & Gott 1972; Quilis et al. 2000) or their hot gas con- web, are significant drivers of these variations. Recently, O’Kane
tent (due to starvation or strangulation Bekki et al. 2002; Fujita et al. (2024) analyse the morphology and star formation of galax-
2004), ultimately reducing their star formation rates. ies of the nearby Universe in different environments from voids
On scales of several megaparsecs, galaxies are embedded in to cluster cores. They found that galaxies in filaments tend to be
the cosmic web, and its gravitational field can also influence their less star forming and favour early-type morphologies. However,
properties. About 35-40 % of the total galaxy luminosity is lo- these differences observed between the population of galaxies
cated in filaments (Tempel et al. 2014). These galaxies may ex- in filaments and the field vanish when they compare samples of
perience morphological and stellar content transformations due galaxies matching their mass and local density. This points to-
to interactions with nearby galaxies, the gas present within the ward the difficulty to split between different processes that si-
filaments and/or their orientation respect to the large-scale struc- multaneously influence galaxies. Although internal and exter-
ture (Dubois et al. 2014). Additionally, the presence of filaments nal processes act at the same time, they could have very dif-
can affect the accretion of gas into galaxies, slowing down their ferent time scales. The results presented in O’Kane et al. (2024)
star formation. These processes can lead to a reduction or sup- highlight the crucial role that sample selection plays in these re-
pression of star formation in galaxies within filaments (Aragon search outcomes. This finding also emphasizes the importance of
Calvo et al. 2019; Bulichi et al. 2024). analysing galaxy samples with matched properties to effectively
Several observational studies have found the imprints left by distinguish between the influences on galaxy characteristics re-
large-scale structure on the properties of galaxies. In particular, sulting from local versus large-scale processes.
the fraction of red galaxies, for a given stellar mass, turned out The aim of this paper is to analyse the influence of the cos-
to be higher in filaments than in the field (Kraljic et al. 2018; mic web environment on various galaxy properties. In particu-
Laigle et al. 2018; Pandey & Sarkar 2020; Hoosain et al. 2024), lar, we study the stellar colour and star formation rate (SFR) of a
which contributes to galaxies displaying redder stellar colours large number of galaxies (≈ 120 000) with stellar masses greater
near filaments (Rojas et al. 2004; Kuutma et al. 2017; Luber et al. than 1010 M⊙ within the redshift range of 0.05 < z < 0.1, lo-
2019). Moreover, Kuutma et al. (2017) found that the Elliptical- cated in different cosmic web environments. Special attention
to-Spiral ratio varies with the distant to filaments, being larger at will be given to the selection of samples with similar mass and
smaller distances to filaments. Chen et al. (2017) discover that local density properties. Matching in mass help us avoid differ-
galaxies closer to filaments are redder, larger and more massive ences arising from internal processes within the galaxies, while
than those located at large distances. Some of this galaxy proper- matching overdensities eliminate variations in galaxy properties
ties observed in the nearby Universe are kept at higher redshift. due to different local environments.
In this sense, Malavasi et al. (2017) found that the most massive This work is organized as follows: the sample and the selec-
and quiescent galaxies are closer to the filaments in the VIPERS tion of the different environments are presented in Sect. 2. The
survey at z ∼ 0.7. These observational findings have recently impact of the large-scale environments on galaxy properties are
been confirmed analysing several galaxy properties in SIMBA shown in Sect. 3 . The discussion and conclusions are given in
simulation (Bulichi et al. 2024). In addition to these morpholog- Sect. 4 and 5, respectively.
ical properties, it has been also observed that there is a correla- Throughout this paper we have used the ΛCDM cosmology
tion of the spin and principal axes of the galaxies with nearby with Ωm = 0.3, ΩΛ = 0.7, and H0 = 70 km s−1 Mpc−1 .
filaments. This could be related with the acquisition of the angu-
lar momentum by the baryons in galaxies (Tempel & Libeskind
2013; Zhang et al. 2013). 2. Sample of galaxies and their cosmic web
However, Kotecha et al. (2022) found an opposite trend in the location
central regions of galaxy clusters. In fact, these authors claimed
that, within one virial radius from clusters of The Three Hundred In this section we present the mother galaxy sample and the cos-
project simulation, galaxies in filaments are bluer and more star mic web catalogue that we use to define the large-scale structure.
Article number, page 2 of 11
 S. Zarattini, J. A. L. Aguerri: Galaxy Transformation Through the Cosmic Web

Fig. 1: The SDSS footprint represented with black dots. In yel- Fig. 2: The dependence of the stellar mass with redshift for the
low, we show the galaxies with spectroscopic redshift. In green, SDSS sample with r p < 17.77. In red, our final sample of galax-
we over plotted the region analysed by Chen et al. (2016). Fi- ies with 0.05 ≤ z ≤ 0.1 and M∗ ≥ 1010 M⊙ .
nally, the area delimited by the red thick solid lines is the one
used in this work and the corresponding α J2000 and δ J2000 values
are highlighted in red in the axes. Please note that we only ren- We also applied a redshift cut, specifically excluding all
dered a random subsample of points for each colour. This was galaxies with z < 0.05, as the filaments and intersections cat-
done for display porpoises only. alogue is defined within the range 0.05 ≤ z ≤ 0.7. We then
restricted our sample to z ≤ 0.1 to ensure mass completeness.
Figure 2 illustrates the dependence of mass on redshift for our
Moreover, we discuss how we associate galaxies with different galaxy sample, with mass estimates derived from colours fol-
environments (filaments, field, and intersections) and how we es- lowing Roediger & Courteau (2015). The adopted redshift cuts
timate the local density of galaxies. provide a galaxy sample that is complete in stellar mass down
to M∗ = 1010 M⊙ . These galaxies, highlighted in red in Fig. 2,
represent our final sample, totalling 120 927 galaxies.
2.1. Mother galaxy sample

We use data from the SDSS Data Release 16 (DR16, Ahumada 2.2. The cosmic web catalogue
et al. 2020) and, in particular, we use all objects catalogued as
As we already did in previous works (Zarattini et al. 2022, 2023)
galaxies brighter than r p = 17.77, where r p is the r−band Pet-
we used the catalogue of filaments and intersections presented
rosian magnitude. This is the same magnitude limit used to se-
in Chen et al. (2016) as the reference for computing distances
lect the SDSS spectroscopic sample called main galaxy sample
between our galaxies and such large-scale structures.
(Strauss et al. 2002). The separation between galaxies and stars
in the Sloan Survey is based on several photometric properties Chen et al. (2016) constructed the catalogue using the Sub-
of the objects. This galaxy/star classification is not free of mis- space Constrained Mean Shift (SCMS) method, which consists
classifications. For this reason, we filter the objects classified as of multiple two-dimensional maps representing filaments and in-
galaxies in the SDSS database by removing those that, based on tersections, beginning at a redshift of 0.05. Each map is gener-
their position in the r − < µ(r50 ) > plane, are inconsistent with ated by grouping galaxies within redshift slices of ∆z = 0.005
being galaxies (see e.g., Sánchez-Janssen et al. 2005; Aguerri up to a redshift limit of z = 0.7. This slicing was performed for
et al. 2020). Here, < µ(r50 ) > represents the mean surface bright- several reasons, including the reduction of the so-called finger-
ness of the object within its effective radius (r50 ). of-god effect (see Chen et al. 2015, for details). Filaments are
defined as a series of points in α J2000 and δ J2000 , outlining the
The final selected galaxies span a large area in right ascen- filament spine. For intersections, however, a single position in
sion (α J2000 ) and declination (δ J2000 ). This area is not regular, α J2000 and δ J2000 is provided. In this catalogue, intersections are
as shown with black dots in Fig. 1. In the same figure we high- defined as points where filaments cross. Structure formation the-
lighted in yellow galaxies with spectroscopic redshift in SDSS. ory suggests that galaxy clusters are often located at these fil-
Finally, green dots represent the area cover by the Chen et al. ament intersections. Chen et al. (2016) show that the distance
(2016) catalogue of filaments and intersections. This catalogue from galaxy clusters to intersections is generally smaller than
has been used to identify galaxies within the cosmic web, and is that from galaxies to intersections; however, not all intersections
briefly described in Sect. 2.2. host a galaxy cluster (see Chen et al. 2016). In Sect. 2.4.2, we
Since our goal is to study the dependence of the stellar colour will examine the relationship between galaxy clusters and inter-
and star formation rate of galaxies as a function of environment sections in terms of local galaxy overdensity.
(local density and distance to filament, in particular), we select The cosmic web catalogue is divided into thin slices of red-
our final sample in a rectangular region that is included within shift, covering the range 0.05 ≤ z ≤ 0.7, and with a thickness
the red thick lines in Fig. 1. This selection ensures us to cover a of 0.005 (Chen et al. 2016). As a first step, we identify the slice
uniform area in the survey and avoid border effects in the com- corresponding to the redshift of a specific galaxy. The filaments
putation of the local galaxy density and the distance to filaments. are described with a series of points with right ascension and
Article number, page 3 of 11
 A&A proofs: manuscript no. main

redshift slice of ∆z = 0.005, as defined in the cosmic web cat-

alogue (Chen et al. 2016). This overdensity will be used in the
present work as tracer of the local galaxy environment.

2.4. Selection of different environments

In this section we explain how we proceed in defining the differ-
ent large-scale environment used in our study.

2.4.1. Intersections, filaments, and field

We divide our sample of galaxies according to its position rela-
tive to the cosmic web. Figure 3 shows the location of the galax-
ies in the D f ila − Dint plane. We have also overploted the position
in this plane of the clusters from the eROSITA groups and clus-
ter catalogue (Bulbul et al. 2024). This catalogue is built from
12 247 groups and clusters detected in the X-ray band 0.3 − 2.3
Fig. 3: Distribution of the galaxies in the D f ila − Dint plane. keV. These systems are optically confirmed and span an area of
Galaxies selected in filaments, intersections and field are rep- 13 116 deg2 in the Western galactic half of the sky. The red-
resented in violet, red and blue points, respectively. The green shift range is 0.003 < z < 1.32 and the mass range is be-
stars represent eROSITA clusters within 0.05 ≤ z ≤ 0.1 and tween 5 × 1012 and 2 × 1015 M⊙ . About two third (68%) of the
M500 > 1014 M⊙ . eROSITA clusters and groups are new discoveries. For each clus-
ter and group the catalogue provides the mass M500 , that can be
converted to the M200 mass (Navarro et al. 1996) according to
declination. These points are not contiguous, but they are usu- M200 = 1.516 × M500 (Arnaud et al. 2005). We can also obtain
ally dense enough to identify the spine of each filament. Since the R200 radius by using
the slices are very thin, we assume that galaxies and filaments
are exactly at the same redshift and we define the distance be- 4
tween them as the minimum projected mutual distance (D f ila ). M200 = π(200ρc )R3200 (1)
This methodology is the same that we adopted in Zarattini et al. 3
(2022, 2023) and distances are measured in Mpc. We calculate where ρc = 3H02 /8πG is the critical density of the Universe.
the distance to intersections (Dint ) using the same methodology,
There are a total of 54 clusters (M500 > 1014 M⊙ ) in our foot-
with the only difference that, in Chen et al. (2016) catalogue, an
print and with redshift between 0.05 ≤ z ≤ 0.1. These are the
intersection is defined as a single point (thus, its distance corre-
clusters from eROSITA plotted in Fig. 3. Notice that eROSITA
spond, by definition, to the minimum distance). In other works,
clusters show a wide range of Dint values indicating that there
D f ila is referred to as D skel , and Dint is denoted as Dnode (see
is no a clear match between the location of these clusters and
Kraljic et al. 2018).
the intersections of filaments provided by Chen et al. (2016).
Following Fig. 3, we define a population of galaxies that lives
2.3. Local density of galaxies close to intersections (Dint < 2 Mpc), one that reside close to
filaments (Dint > 20 Mpc and D f ila < 1 Mpc), and a field pop-
One of the main drivers of galaxy evolution is the local envi- ulation (D f ila > 10 Mpc). Taking these definitions into account,
ronment, which can be traced by the projected local density of we identified 2 963 galaxies in intersections, 7 042 galaxies in fil-
galaxies (Σg ). We compute this value for each galaxy in our sam- aments, and 14 657 galaxies in the field. These definitions were
ple by selecting the five nearest neighbours within ±3000 km established to avoid at maximum the possible cluster contamina-
s−1 . This cut was used to properly manage the high-density en- tion in the filaments and field populations (see Fig. 3).
vironments of galaxy clusters, since it corresponds to the typical This division is well representative of the regions that we
cluster escape velocity (Abell 1962; Pignataro et al. 2021; Zarat- want to map. In Fig. 4 we show one of our redshift slices colour-
tini et al. 2022; Boschin et al. 2023). The spectroscopic sample coded to highlight the position of the three populations, that are
of galaxies from SDSS-DR16 is incomplete, which means that clearly separated.
the local galaxy density calculated using only spectroscopic data The cumulative distribution of the galaxy overdensity in the
will be a lower limit of the real one. To account for this incom- different environments is shown in the left panel of Fig. 5 us-
pleteness and avoid underestimating the local galaxy density, we ing solid lines. As expected, galaxies in the field are found at
apply a photometric correction to Σg by counting the number of the lowest overdensities with 77% of the galaxies found in den-
galaxies without redshift in the area defined by the fifth spectro- sities smaller than the mean one (δg ≤ 0), and only 5% in
scopic neighbour. We assume that the fraction of galaxies with- δg > 3. For the filaments, these fractions are 32% and 28%,
out redshift contributing to the local density is the same as that respectively, whereas in intersections we find 30% and 35% in
for galaxies with spectroscopic redshift, and we add this number the same overdensity ranges. These numbers indicate a trend of
to the local density calculated previously. increasing overdensity from the field population to the intersec-
The global mean density of galaxies in our sample is Σg = tion regions. However, within each of the three environments,
25.14 galaxies per square degree, obtained as the 50% of the galaxies are found in distinctly different local settings. Previ-
cumulative distribution of local density of the entire sample ous works found that filaments are regions of mean overdensity
(0.05 ≤ z ≤ 0.1). We compute the local overdensity as δg = where galaxy-galaxy interactions or mergers can take place (Guo
(Σg − Σz,g )/Σz,g , being Σz,g the mean density of galaxies in each et al. 2015; Kuutma et al. 2017; Aragon Calvo et al. 2019). In
Article number, page 4 of 11
 S. Zarattini, J. A. L. Aguerri: Galaxy Transformation Through the Cosmic Web

Fig. 4: Plot of the entire field of view between redshift 0.075 < z < 0.08. Red galaxies are those selected to be close to intersections
(Dint < 2 Mpc), violet ones are close to filaments only (Dint > 20 Mpc and D f ila < 1 Mpc), and blue galaxies are selected to be
in the field (D f ila > 10 Mpc). Turquoise circles are intersections and pink stripes are filaments from Chen et al. (2016). In yellow
we show the position of eROSITA clusters that we used in this region. Please note that symbol’ sizes of filaments, intersections,
and clusters are for display purposes only and have no connection with the physical sizes of these objects. The size in Mpc of the
separation between two ticks (5 degrees) is indicated in the upper left corner of the image.

particular, Guo et al. (2015) found that only a higher relative ve- 2.4.2. Intersections and cluster environments
locity between galaxies in the filament or a higher local density
can justify the merging-rate increase observed in filaments with The aim of this subsection is to compare the local overdensity
respect to the field. Figure 5 also shows the galaxy overdensity of galaxies at intersections with those at various distances from
for a set of galaxies located in clusters (dashed lines). These ob- the centres of galaxy clusters. This comparison allows us to as-
jects will be discussed in Sect. 2.4.2 sess whether the galaxy population at intersections is represen-
tative of that within clusters. As we mentioned before, there are
The right panel of Fig. 5 further illustrates the cumulative a total of 54 clusters (M500 > 1014 M⊙ ) from the eROSITA sam-
mass distribution functions of galaxies in the three environments. ple in our footprint and with redshift between 0.05 ≤ z ≤ 0.1.
A clear trend in masses is visible, with smaller mass galaxies We use the positions of these objects to define a galaxy popu-
predominantly located in the field population, whereas galax- lation within the clusters and compare their local overdensities,
ies near filaments tend to be more massive. This segregation of up to 5 × R200 , with those of the galaxies located in the inter-
galaxy mass based on position within the cosmic web has also sections of filaments. To achieve this result, we exclude clusters
been observed by other authors (Chen et al. 2017; Luber et al. located within less than 5 × R200 from the outskirts of another
2019; Hoosain et al. 2024). system. This choice reduces the number of eROSITA clusters in
Figure 5 clearly indicates that the galaxies located in field, our footprint and redshift range to 28. The remaining objects are
filament and intersection environments have different internal still sufficient, and this filtering ensures that the galaxies found
properties and are located in different local environments. This in the outer regions of one cluster are not contaminated by those
suggests that a direct comparison of galaxy properties between in the core of a nearby cluster (e.g., galaxies that would be in
samples in these environments may be biased, as internal prop- higher overdensities than expected in the outskirts of clusters).
erties and local environment can become the dominant mecha- We add the positions of the eROSITA groups and clusters to Fig.
nisms in their evolution. 4 in yellow.
Article number, page 5 of 11
 A&A proofs: manuscript no. main

Fig. 5: Left panel: the cumulative distribution of the overdensity of galaxies in the different environments. Solid lines represent the
field (blue), filaments (violet), and intersections (red). The dashed lines are the eROSITA clusters and, in particular, galaxies within
2 − 5 × R200 (blue), 1 − 2 × R200 (green),< 1 × R200 (black), < 0.5 × R200 (red). Right panel: the cumulative distribution of the stellar
mass of galaxies in field (blue) filament (violet) and intersections (red).

For this sample, we classify galaxies based on their distance 3. Results

from the center of the cluster (defined by the position of the
eROSITA detection). The cumulative distribution of overden- In this section, we discuss our results by focusing on two main
sities in each environment is shown in the left panel of Fig. 5 topics: the evolution of stellar g − r galaxy colour and the SFR
with dashed lines. We divide the galaxies into four cluster envi- from field to filament environments.
ronments: one representing the core of the clusters (distances
smaller than 0.5 × R200 , 392 galaxies), galaxies within R200 3.1. The impact of the environment on galaxy colours and
(746), galaxies between 1−2 × R200 (659), and galaxies between SFR
2 − 5 × R200 (1 775). Galaxies at the largest radii (2 − 5 × R200 )
are shown with a blue dashed line, and their cumulative pro- In the left panel of Fig. 6 we show the cumulative distribution
file is very similar to that of galaxies in filaments (as defined in function of the stellar galaxy colour (g − r) as a function of the
the previous section). The overdensity steadily increases, but it local environment. Solid lines are galaxies in the field (blue), and
is notable that the significant separation occurs between galax- filaments (violet). We can see that the stellar colour of galaxies
ies in the 1 − 2 × R200 region and those within R200 . It is also in the field are bluer than in filaments. This change can be quan-
important to note that galaxies located at intersections exhibit tified using the median values of the distributions, that is 0.74
similar overdensities to those found beyond the R200 radius of for the field and 0.81 for the filaments. This result is similar to
eROSITA clusters. This suggests that this population of galax- other previously found in the literature (Kuutma et al. 2017; Lu-
ies does not represent typical virialized cluster environments but ber et al. 2019; O’Kane et al. 2024).
rather galaxies residing in local densities comparable to the out- To compute the SFR of the galaxies we have used the expres-
skirts of clusters. This could be expected, since this population sion:
was selected within a radius of 2 Mpc from the position of the
intersection. In order to test this scenario, we select a galaxy
population in a smaller radius, namely within 1 Mpc from the S FR (M⊙ yr−1 ) = 7.9 × 10−42 LHα (erg s−1 ) (2)
intersection (e.g. similar to the typical cluster’s virial radius).
The median value of the overdensity slightly grows from 1.3 to This expression assumes a Salpeter Initial Mass Function with
1.4, whereas the median value of the core of eROSITA cluster masses between 0.1 and 100 M⊙ and solar metallicities (see Ken-
is larger than 10. We thus confirm that galaxies in intersections nicutt 1998). The Hα luminosity were obtained from the contin-
are not representative of galaxy clusters. This is not surprising, uum subtracted Hα flux measured for their spectra by MPA-JHU
1
since Chen et al. (2015) applied the SCMS algorithm to labeled emission line analysis for the galaxies of the SDSS-DR7. The
simulated data generated via the Voronoi model of van de Wey- Hα luminosity was corrected for dust absorption by using the
gaert (1994) to show that it preferentially detects structures la- ratio between this line and the Hβ line and the Cardelli et al.
beled as filaments. As a consequence, we will exclude this pop- (1989) extinction law. In addition, the SFR obtained from the
ulation from further analysis and instead focus on the properties Hα luminosity from the MPA-JHU catalog represents the fiber
of galaxies located in field and filament environments. SFR. To obtain the total SFR an aperture correction must be ap-
plied. This aperture correction was done similarly to Salim et al.
1
https://wwwmpa.mpa-garching.mpg.de/SDSS/DR7/

Article number, page 6 of 11

 S. Zarattini, J. A. L. Aguerri: Galaxy Transformation Through the Cosmic Web

(2007). The right panel of Fig. 6 presents the cumulative distri-

bution function of the SFR for galaxies located in the field and
filament environments. A Kolmogorov-Smirnov (KS) test con-
firms that the SFR distributions of galaxies in filaments and the
field are statistically distinct, with galaxies in filaments exhibit-
ing lower SFR compared to those in the field. This difference in
SFR may explain the variations observed in their stellar g − r
colour. The result obtained here is in agreement with Hoosain
et al. (2024). They found that galaxies nearer to filaments pos-
sess less gas, which directly results in lower SFRs. Additionally,
numerical simulations show a dependence of the SFR on the dis-
tance from the filaments (Aragon Calvo et al. 2019; Hasan et al. Fig. 6: Left panel: cumulative distribution of the g − r stellar
2023). There are also some results in the literature which found colours of galaxies in the field (blues line), and filament (violet
evidences suggesting that interactions between galaxies within line). Right panel: cumulative function of the star formation rate.
filaments can enhance their SFR (Darvish et al. 2014; Vulcani The colour code is the same as in left panel.
et al. 2019).
The differences observed in the stellar g−r colour and SFR of
samples, indicating that the two distributions are statistically dif-
galaxies in filaments versus those in the field may be influenced
ferent.
by the disparities in mass and local galaxy density between the
In summary, the analysis of stellar colour and SFR for galax-
two samples (see Fig. 5). More massive galaxies, like those near
ies in mass and overdensity-matched samples indicates that fila-
filaments, tend to exhibit redder colours and may deplete their
ments have a measurable impact on these two properties. Specif-
gas reservoirs more rapidly. Additionally, galaxies located in re-
ically, galaxies in filaments are redder and exhibit lower SFR
gions of higher local density, such as filaments, may experience
compared to their counterparts in the field.
gas suppression due to higher fraction of interactions, which
would lead to redder colours and reduced SFR. To disentangle
the effects of large-scale structures from local environmental in- 3.3. Change of galaxy properties with the distance to
fluences, it is essential to compare galaxy samples in both field filaments
and filament environments with similar stellar masses and local
galaxy densities. We have found that the SFR and g − r stellar colour of galax-
ies in filaments and in the field are statistically different, even
when matching their stellar mass and local galaxy overdensity.
3.2. Stellar mass and overdensity-matched galaxy However, we defined filament galaxies as those with D f ila < 1
samples Mpc and field galaxies as those with D f ila > 10 Mpc. A key
question is whether there is a specific value of D f ila at which
We construct stellar mass and local overdensity-matched sam- these changes in galaxy properties occur. To investigate this, we
ples for galaxies in filaments and the field. To achieve this, we defined three intermediate populations: 1 < D f ila < 2.5 Mpc
take each galaxy from the filament environment2 and find its (Inter1), 2.5 < D f ila < 5 Mpc (Inter2), and 5 < D f ila < 10 Mpc
counterpart in the field with the closest stellar mass and local (Inter3). All galaxies in these populations also follow Dint > 20
overdensity. Both galaxies, from the field and filament environ- Mpc. These definitions give us a total of 8 005, 8 395 and 9 589
ments, are then added to their respective matched samples. This galaxies in the Inter1, Inter2 and Inter3 populations, respectively.
process is repeated for all filament galaxies, ensuring no replace- Table 1 presents a statistical comparison, based on the KS
ment of the field counterparts once selected. This procedure en- probability parameter pKS , of the cumulative distribution func-
sure us to have two samples of galaxies in filaments and field tions of SFR and g − r stellar colour between galaxies in the field
with similar stellar masses and overdensities. The KS test gives and the other galaxy populations defined earlier. Notably, there is
probabilities of 0.72 and 0.99 for the cumulative distributions a statistically significant difference in the cumulative SFR distri-
of stellar mass and overdensity in the matched samples, respec- bution between galaxies in the field and those at D f ila < 5 Mpc.
tively, indicating a good match between the two populations. At larger radii the difference in SFR between the field and inter-
The results for the stellar colour and SFR of the matched mediate population is not statistically different at 95% c.l., but
galaxy samples are presented in Fig. 7. This figure demonstrates with PKS close to the threshold of 0.05 (see Table 1). In contrast,
that the differences in g − r stellar colour between galaxies in the differences in g − r colour begin for galaxies with D f ila < 1
filaments and the field are smaller than those seen in Fig. 6. Mpc. This suggests that, as galaxies transition from the field into
Specifically, both distributions show similar median g − r colour filaments, they undergo a change in their star formation rate be-
values. However, the KS test indicates that the two colour cumu- tween approximately 2.5 - 5 Mpc. Meanwhile, the shift in stellar
lative distributions are still statistically different (pKS = 0.002), colour occurs at smaller distances from the filament (D f ila < 1
with the most significant difference occurring in the 50% bluest Mpc). This variation of the stellar colour with the distance to the
galaxies (see Fig. 7). Moreover, galaxies in filaments exhibit filaments have also been observed in other galaxy samples (e.g.
lower SFR compared to those in the field, even in the matched Kuutma et al. 2017).
samples (see Fig. 7). Similar to the stellar colours, the matched These findings may also indicate that the typical width of a
samples show smaller differences in SFR than the full galaxy filament could be around 2.5 - 5 Mpc, as this is the distance from
samples. However, the KS test gives a probability of 2.13×10−4 the filament spine where galaxies experience changes in proper-
for the cumulative SFR distribution functions of the matched ties such as SFR. Castignani et al. (2022) modelled the galaxy
density in the Virgo filaments using an exponential profile. Their
2
We select galaxies from the filament sample as it has a smaller num- analysis revealed a considerable diversity in the scale widths of
ber of galaxies than the field sample. the filaments. Notably, the widest filaments had scales ranging
Article number, page 7 of 11
 A&A proofs: manuscript no. main

late-type. Using this classification, 62% of the galaxies in fila-

ments and 64% in the field were classified as late-type. This in-
dicates that both samples are dominated by late-type galaxies, as
expected in low-overdensity environments (Dressler 1980). It is
also clear that the two subsamples do not present significant mor-
phological variation. Therefore, the differences in stellar colour
and SFR between galaxies in the field and in filaments cannot
be explained by differences in morphology in our case. What-
ever process is producing the observed differences should not
be strong enough to produce significant morphological and/or
structural changes in the galaxies.
Fig. 7: Cumulative distribution functions of the g − r stellar Differences in the gas mass of the galaxies in filaments and
colours (left panel) and SFR (right panel) of galaxies in the field could also explain the observed differences in stellar colour
field (blue lines), and filaments (violet lines) for stellar mass and and SFR. Luber et al. (2019) found a variation of the MHI /M∗ as
overdensity-matched samples. a function of the distance to filaments. In particular, this ratio is
smaller for galaxies closer to filaments. However, they also ob-
served an increase of the stellar mass for those galaxies closer
from 2.2 to 2.7 Mpc, likely corresponding to the types of struc- to filaments. This implies that the variation of the MHI /M∗ ra-
tures identified in the Chen et al. (2016) catalogue. tio could be explained by the observed variation of the stellar
In the framework where galaxies transition from the field mass alone. We can compute an estimation of their gas content
population into the cosmic web via filaments, it seems to be a from the SFR of galaxies. In particular, assuming a typical de-
delay between the changes in their star formation rate and stel- pletion time τ = 2 × 109 yr for galaxies with moderate SFR and
lar colour. This result suggests that the SFR of galaxies changes high mass as those studied here, the ratio between the mean val-
first, likely due to a decrease in the galaxy gas content due to the ues of the mass gas for galaxies in filament (Mgas, f ila ) and field
lack of external gas supply, followed by a subsequent change in (Mgas, f ield ) is: Mgas, f ila /Mgas, f ield = 0.86. Therefore, the mass of
the stellar colour. This indicates a phased transformation where the gas in field galaxies in about 10% larger than galaxies with
the suppression of star formation occurs before the galaxy’s stel- similar mass in filaments. This smaller amount of gas for galax-
lar population reddens. ies in filaments could produce a decrement of their SFR as ob-
served in the present work.
The decrease in SFR and gas content of galaxies from field
4. Discussion to filaments presented in this work can be explained by sev-
The results presented in this paper show differences in stellar eral physical mechanisms. Some, such as AGN and supernova
colour and SFR between samples of galaxies located in the field feedback processes (Silk & Rees 1998; Di Matteo et al. 2005),
and in filaments, despite having similar stellar masses and local are linked to internal galaxy processes. Other mechanisms, in-
densities. This indicates that the cosmic web produce an imprint cluding ram pressure stripping (Gunn & Gott 1972; Quilis et al.
in these properties of the galaxies. 2000), harassment (Moore et al. 1996), strangulation (Bekki
Differences in galaxy properties between filaments and field et al. 2002; Fujita 2004), or preprocessing (Treu et al. 2003; Goto
have also been reported in the literature. For instance, several et al. 2003), are driven by external factors. These external mech-
early studies have found that the fraction of red galaxies depends anisms are highly dependent on both local galaxy density and
on their environment (Balogh et al. 2004; Bamford et al. 2009). mass. Therefore, in order to differentiate between the effects of
Furthermore, this fraction of red galaxies has been shown to vary the different mechanisms, it is crucial to compare samples of
with distance from filaments (see Kraljic et al. 2018; Pandey & galaxies with matched masses and local densities as we did in
Sarkar 2020; Hoosain et al. 2024). Additionally, Kuutma et al. the present work. Aragon Calvo et al. (2019) proposed the so-
(2017) found that bright galaxies (Mr ≤ −20.0) exhibit redder called Cosmic Web Detachment (CWD) model, which unifies
g−i colour and lower SFR as they moves from voids to filaments. several starvation mechanisms that halt star formation in galax-
The differences in stellar colour and SFR between galaxies ies due to various external processes. This model suggests that
in the field and in filaments, as found in this paper, could re- a galaxy can reduce or stop its star formation by being sepa-
flect variations in the fraction of galaxy types between both sam- rated from its supply of star-forming gas. This can be achieved
ples. In this sense, Kuutma et al. (2017) observed that early-type via galaxy-galaxy interactions, galaxy accretions to other halos
galaxies were more abundant near filaments, which could ac- or to filaments (see Aragon Calvo et al. 2019). The fact that the
count for the colour differences observed in their sample. To de- morphology of galaxies in filaments and the field remains simi-
termine whether differences in morphology between field and fil- lar for our samples suggests that the processes affecting galaxies
ament environments could explain our findings, we analysed the discard strong tidal interactions or galaxy mergers. In contrast,
morphology of the galaxies in the two studied environments. We our results indicate that, as galaxies enter in filaments, their gas
obtained the morphological classification for our galaxies from reservoirs are slowly depleted and this cosmic web depletion of
Huertas-Company et al. (2011), who used a machine-learning the gas reservoirs in the galaxies could explain the observed dif-
technique to classify the galaxies in the SDSS-DR7 into four ferences in the stellar colour and SFR.
morphological classes: Ellipticals (E), Lenticulars (S0), Early In a recent study, Bulichi et al. (2024) analysed the proper-
Spirals (Sab), and Late Spirals (Scd). Each galaxy in the sur- ties of galaxies from the SIMBA simulation as a function of their
vey is assigned a probability of belonging to each of these mor- distance to cosmic filaments throughout the Hubble time. They
phological classes. We considered a galaxy to be early-type if identified significant variations in stellar mass, specific star for-
the combined probabilities of being elliptical (PE) and lenticu- mation rate (sSFR), gas fraction, and stellar metallicity depend-
lar (PS0) exceeded 0.5. Otherwise, the galaxy was classified as ing on how close the galaxies are to the filaments. Specifically,
Article number, page 8 of 11
 S. Zarattini, J. A. L. Aguerri: Galaxy Transformation Through the Cosmic Web

Table 1: KS probabilities of the cumulative distribution functions

of the field and the other galaxy populations.

Filament Inter1 Inter2 Inter3

D f ila [Mpc] <1 1-2.5 2.5-5 5-10
PKS (g − r) 0.002 0.085 0.24 0.26
PKS (S FR) 4.68×10−6 0.003 0.006 0.053

galaxies located near the filament spines tend to be more mas-

sive, quenched, gas-poor, and exhibit higher metallicity.
Fig. 8: Left panel: cumulative distribution of the g − r stellar
Additionally, the depletion of gas reservoirs could also ex- colours of galaxies in the field (blues line), and filament (violet
plain why the SFR of galaxies differs from that of the field sam- line). Right panel: cumulative function of the star formation rate.
ple at greater distances from the filament compared to their stel- The colour code is the same as in left panel. In both panels, only
lar colours. In this scenario, as galaxies approach the filaments, galaxies with 109 < M < 1010 M⊙ are considered.
the infall of external gas would cease, allowing them to sustain
their SFR using internal gas, but with a gradual decline. The re-
maining gas would be converted into stars within a closed-box less gas than their counterparts in the field, in agreement with
framework, leading to an increase in gas metallicity due to the our main sample (M > 1010 M⊙ ).
lack of external gas supply. Indeed, Domínguez-Gómez et al. To better analyse the impact of filaments on galaxy proper-
(2023) found a slight increase of the stellar metallicity of galax- ties as a function of mass, we examined the stellar colour and
ies in filaments with respect to the field ones. Moreover, signifi- star formation rate (SFR) of galaxies with 109 < M < 1010 M⊙
cant changes in stellar colours would not occur until star forma- in both filament and field environments. These galaxies were in-
tion is nearly suppressed, which would happen as galaxies move cluded in the analysis done in the previous paragraph but, since
closer to the filament spine. In this quenching process, driven by they are few in number, their impact on the results could have
a cosmic web starvation mechanism, it is expected that the gas been diluted. With this new test, we aim at highlight the proper-
in field galaxies would be less metal-rich than that in galaxies ties of this population. As in our previous analysis, we matched
within filaments. the mass and overdensity distributions of galaxies in filaments
Peng et al. (2010) demonstrated that the effects of mass and and the field.
environment on the quenching process of galaxies can be distin- Figure 8 presents the cumulative distribution functions of the
guished, identifying two concurrent mechanisms: mass quench- g − r stellar colour and SFR for the matched samples of galax-
ing and environmental quenching. The results of this work show ies in filaments and the field with 109 < M < 1010 M⊙ . The
that there are small but statistically significant differences in their KS test shows that, as for our main sample of galaxies with
stellar colour and SFR properties when comparing samples of M > 1010 M⊙ , those located in filaments exhibit redder g − r
galaxies with similar mass and located in similar local environ- colours and lower SFRs compared to their counterparts in the
ments (i.e., experiencing similar mass quenching and environ- field. These differences cannot be explained by variations in
mental quenching). The only distinction between the two galaxy morphological fractions, as the fraction of late-type galaxies is
samples is their cosmic web location. Therefore, the results of 0.97 ± 0.02 in filaments and 0.99 ± 0.01 in the field.
this study suggest that the cosmic web environment leaves a dis- The only relevant difference that we found is that low-mass
tinct imprint on the quenching process of galaxies. galaxies in filaments contain approximately 20% less gas than
those in the field, suggesting that they may be more affected by
4.1. Galaxies with masses smaller than 1010 M⊙ the large-scale structures, leading to larger gas depletion. A simi-
lar result was found in Singh et al. (2020) using the EAGLE cos-
We applied a mass cut at M = 1010 M⊙ , as Fig. 2 indicates that mological simulation. In fact, these authors claimed that galaxies
the sample is mass-complete at that mass within the analysed are redder and less star forming when they are closer to the spine
redshift range (0.05 < z < 0.1). Additionally, Fig. 2 shows that of the filaments. These trends are stronger for galaxies less mas-
only a small number of objects with M < 1010 M⊙ can be in- sive than 1010 M⊙ . In addition, Galárraga-Espinosa et al. (2023),
cluded in our samples. Specifically, 814 and 2230 galaxies with using data from the TNG50 simulation, showed that the stream
M < 1010 M⊙ are found in the filament and field samples, respec- connectivity of low-mass galaxies (M ∼ 109.5 M⊙ ) depends on
tively. However, recent studies such as Galárraga-Espinosa et al. their large-scale environment. Galaxies in higher-density envi-
(2023) suggested that these galaxies can have enhanced SFR in ronments are connected to fewer streams compared to those in
filaments, due to the so-called cold accretion mode (e.g. the gas the field. This highlights the dependence of galaxy gas accretion
flows directly to the centre because small galaxies are not able on the surrounding environment specially for low mass galaxies.
to support shocks, see also Kereš et al. 2005). Nonetheless, with the limited number of low-mass galaxies
To assess the impact of including these lower-mass galaxies, in our sample it is challenging to draw a definitive conclusion
we tested our results with an extended sample incorporating all about the dependence of the large-scale environment on galaxy
galaxies in the mass range 109 M⊙ < M < 1012 M⊙ . We found no properties as a function of their stellar mass.
significant changes in the observed trends for the properties of
galaxies in the filament and field samples. In both cases, galax- 4.2. Misidentified galaxies across the cosmic web
ies in filaments exhibit redder g − r colours and lower SFRs than
those in the field. These differences cannot be explained by vari- The cosmic web is a complex network composed of various
ations in morphological fractions. Furthermore, in this extended large-scale structures, including voids, walls/sheets, filaments,
sample, galaxies in filaments also contain approximately 10% and clusters (see e.g., Cautun et al. 2014). Catalogs, such as the
Article number, page 9 of 11
 A&A proofs: manuscript no. main

one used in this study, aim to accurately reproduce this intricate SDSS-DR16 with M∗ ≥ 1010 M⊙ and within the redshift range
structure but are not free from uncertainties. In this work, we an- 0.05 ≤ z ≤ 0.1. The catalogue from Chen et al. (2016) was used
alyzed the properties of galaxies in two distinct environments: to classify the galaxies into three large-scale environments: field,
the field and filaments. However, these two populations may be filaments, and intersections. Additionally, we computed the lo-
affected by contamination from misidentified galaxies across the cal overdensity of the galaxies to account for their local environ-
cosmic web. ment.
One potential source of contamination in the filament and There is a clear segregation of mass and overdensity among
field galaxy populations could be the presence of galaxy clus- galaxies located in the field, filaments, and intersections. In par-
ters that were not identified as nodes or intersections in the Chen ticular, galaxies in filaments tend to have higher stellar masses
et al. (2016) catalog. As discussed in Sect. 2.4.2, the galaxy pop- compared to those in the field, which are generally less massive.
ulations located at the intersections defined by Chen et al. (2016) Additionally, a gradient in local overdensity is observed, with
do not exhibit the characteristics of typical virialized cluster en- galaxies in the field showing the lowest overdensities, followed
vironments. Chen et al. (2016) also measured the average dis- by those in filaments, and the highest overdensities found in in-
tance between RedMapper clusters and their identified intersec- tersections. Galaxies in filaments inhabit environments spanning
tions. From their Fig. 11, this distance appears nearly constant a wide range of overdensities, from those typical of field re-
with redshift, with the best agreement found in the lowest ana- gions to others with overdensities akin to intersections. The in-
lyzed redshift range (0.10 < z < 0.15), where the difference is tersections of filaments, as defined by Chen et al. (2016), exhibit
less than one degree (closer to 0.5, though no exact value is pro- galaxy overdensities that differ significantly from those found in
vided in the text). This corresponds to less than 7 Mpc at their the virialized regions of clusters, indicating that these regions do
lowest limit, z=0.1. It is worth noting that no study was done by not trace typical galaxy cluster environments.
Chen et al. (2016) in the redshift bin used in our paper. How- Galaxies in the field and filaments show distinct g − r stel-
ever, we were able to repeat a similar analysis using eROSITA lar colours, with galaxies in filaments generally appearing red-
clusters: we found a median separation of 9 Mpc between these der than those in the field. Additionally, galaxies in filaments
clusters and intersections. For this reason, we think that our se- exhibit lower SFR compared to their field counterparts. These
lection of filament galaxies (which include a distance to inter- conclusions hold even for samples of galaxies in filaments and
section larger than 20 Mpc) is quite robust versus this issue. To the field that are matched for stellar mass and local overdensity,
further assess our results, we compute the total number of galax- indicating that filaments imprint on galaxy properties by effec-
ies that are within the virial radius of an eROSITA cluster and tively suppressing their star formation.
also accomplish our definition of filament galaxy. We found that No differences in morphologies were found between the
only 0.01% of the galaxies accomplish this definition. This frac- mass- and overdensity-matched samples of galaxies in the field
tion rises to about 2% if we extend the radius to 2r200 . Moreover, and filaments. This indicates that whatever process producing
we selected the field galaxy population as those galaxies with the observed differences between galaxies in field and filaments
distance to the filaments larger than 10 Mpc. Only 1 out of 54 is not strong enough to significantly change the morphology
clusters from eROSITA is located in this region, for a fraction and/or the structure of the galaxies. In contrast, a slight variation
of galaxies smaller than 1%. Therefore, we also believe that our in the gas content was observed, with galaxies in filaments con-
selection of field galaxies is largely free from significant cluster taining about 10% less gas compared to those in the field. This
contamination. difference rises to about 20% for galaxies smaller than 1010 M⊙ .
On the other hand, the field galaxy population may also be These differences in gas content may contribute to the lower star
contaminated by galaxies residing in wall/sheet structures. These formation rates seen in galaxies within filaments, as gas is a key
structures are challenging to detect, and a fraction of our field fuel for star formation.
galaxies could indeed belong to sheets. To test this possibility, There is a variation of the g−r and the SFR of the galaxies as
we created a subsample of field galaxies with δg < 0, which a function of the distance to the filaments. The SFR of the galax-
should be less affected by sheet galaxies, that are expected to be ies located at D f ila < 5 Mpc is statistically lower than that for
found in slightly larger overdensities. This refined field sample field galaxies. The change in the g − r stellar colour is observed
contains 11 159 objects. The median and minimum distances to at closer distances to the filament (D f ila < 1 Mpc).
filaments of the galaxies in the new field sample do not change These results indicate that filaments induce moderate
with respect to the original field ones. This implies that both changes in galaxies, such as gas suppression and the quenching
galaxy samples are similarly distributed in the sky. We then com- of star formation, but do not lead to substantial transformations,
pared their stellar colour and SFR properties with those of the such as major changes in galaxy morphology. This suggests that
filament population, finding no significant differences with the while environmental effects from the large-scale structure im-
results presented in previous sections. Specifically, galaxies in pact star formation and gas content, they may not be strong
filaments still exhibit redder g − r stellar colours and lower SFRs enough to drive significant structural changes in galaxies. Fur-
than those in the new field sample with δg < 0. Additionally, thermore, as galaxies approach filaments, the infall of external
the fraction of late-type galaxies remains similar between fila- gas would cease at distances of about 2–5 Mpc from the filament
ments and the field and the gas mass ratio between filament and spine, gradually slowing their SFR. Significant changes in stellar
field galaxies slightly changes to 0.92, indicating that filament colours would occur only as galaxies move closer to the filament
galaxies contain about 10% less gas than those in the new field spine (D f ila < 1 Mpc) and their internal gas is depleted, ceasing
sample. These results suggest that our filament galaxy selection the SFR and representing a galaxy quenching process driven by
is robust against contamination from sheet galaxies. cosmic web starvation. In this context, differences in gas metal-
licity between galaxies in the field and those in filaments would
also be expected. This aspect will be explored in a future study
5. Conclusions using the galaxy sample analysed in the present work.
We analysed the variation in the g − r stellar colour and the SFR Acknowledgements. We would like to thank to Dra. C. Muñoz-Tuñón for her
as a function of environment for a large sample of galaxies from valuable comments during the preparation of this work. SZ and JALA acknowl-

Article number, page 10 of 11

 S. Zarattini, J. A. L. Aguerri: Galaxy Transformation Through the Cosmic Web

edge financial support provided by the Spanish Ministerio de Ciencia, Inno- Driver, S. P., Norberg, P., Baldry, I. K., et al. 2009, Astronomy and Geophysics,
vación y Universidades (MICIU) through the projects PID2020-119342GB-I00 50, 5.12
and PID2022-136598NB-C31. JALA and SZ also acknowledge support from Dubois, Y., Pichon, C., Welker, C., et al. 2014, MNRAS, 444, 1453
the Agencia Estatal de Investigación del Ministerio de Ciencia, Innovación y Eardley, E., Peacock, J. A., McNaught-Roberts, T., et al. 2015, MNRAS, 448,
Universidades (MCIU/AEI) under grant “WEAVE: EXPLORING THE COS- 3665
MIC ORIGINAL SYMPHONY, FROM STARS TO GALAXY CLUSTERS” Eliche-Moral, M. C., Balcells, M., Aguerri, J. A. L., & González-García, A. C.
and the European Regional Development Fund (ERDF) with reference PID2023- 2006, A&A, 457, 91
153342NB-I00 / 10.13039/501100011033. Eliche-Moral, M. C., González-García, A. C., Balcells, M., et al. 2011, A&A,
SZ acknowledges the financial support provided by the Governments of Spain 533, A104
and Aragón through their general budgets and the Fondo de Inversiones de Fujita, Y. 2004, PASJ, 56, 29
Teruel, the Aragonese Government through the Research Group E16_23R, Galárraga-Espinosa, D., Garaldi, E., & Kauffmann, G. 2023, A&A, 671, A160
the Spanish Ministry of Science and Innovation and the European Union - Goto, T., Yamauchi, C., Fujita, Y., et al. 2003, MNRAS, 346, 601
NextGenerationEU through the Recovery and Resilience Facility project ICTS- Gunn, J. E. & Gott, J. Richard, I. 1972, ApJ, 176, 1
MRR-2021-03-CEFCA, and the Spanish Ministry of Science and Innovation Guo, Q., Tempel, E., & Libeskind, N. I. 2015, ApJ, 800, 112
(MCIN/AEI/10.13039/501100011033 y FEDER, Una manera de hacer Europa) Hasan, F., Burchett, J. N., Abeyta, A., et al. 2023, ApJ, 950, 114
with grant PID2021-124918NB-C44. Hibbard, J. E. & Mihos, J. C. 1995, AJ, 110, 140
Funding for the Sloan Digital Sky Survey IV has been provided by the Al- Hoosain, M., Blyth, S.-L., Skelton, R. E., et al. 2024, MNRAS, 528, 4139
fred P. Sloan Foundation, the U.S. Department of Energy Office of Science, Huchra, J. P., Macri, L. M., Masters, K. L., et al. 2012, ApJS, 199, 26
and the Participating Institutions. SDSS acknowledges support and resources Huertas-Company, M., Aguerri, J. A. L., Bernardi, M., Mei, S., & Sánchez
from the Center for High-Performance Computing at the University of Utah. Almeida, J. 2011, A&A, 525, A157
The SDSS web site is www.sdss4.org. SDSS is managed by the Astrophysical Kauffmann, G., White, S. D. M., Heckman, T. M., et al. 2004, MNRAS, 353,
Research Consortium for the Participating Institutions of the SDSS Collabo- 713
ration including the Brazilian Participation Group, the Carnegie Institution for Kennicutt, Robert C., J. 1998, ARA&A, 36, 189
Science, Carnegie Mellon University, Center for Astrophysics Harvard & Smith- Kereš, D., Katz, N., Weinberg, D. H., & Davé, R. 2005, MNRAS, 363, 2
sonian (CfA), the Chilean Participation Group, the French Participation Group, Kim, J.-h., Wise, J. H., & Abel, T. 2009, ApJ, 694, L123
Instituto de Astrofísica de Canarias, The Johns Hopkins University, Kavli In- Knapen, J. H., Cisternas, M., & Querejeta, M. 2015, MNRAS, 454, 1742
stitute for the Physics and Mathematics of the Universe (IPMU) / University Kotecha, S., Welker, C., Zhou, Z., et al. 2022, MNRAS, 512, 926
of Tokyo, the Korean Participation Group, Lawrence Berkeley National Lab- Kraljic, K., Arnouts, S., Pichon, C., et al. 2018, MNRAS, 474, 547
oratory, Leibniz Institut für Astrophysik Potsdam (AIP), Max-Planck-Institut Kuutma, T., Tamm, A., & Tempel, E. 2017, A&A, 600, L6
für Astronomie (MPIA Heidelberg), Max-Planck-Institut für Astrophysik (MPA Laigle, C., Pichon, C., Arnouts, S., et al. 2018, MNRAS, 474, 5437
Garching), Max-Planck-Institut für Extraterrestrische Physik (MPE), National Łokas, E. L. 2021, A&A, 655, A97
Astronomical Observatories of China, New Mexico State University, New York Łokas, E. L. 2022, A&A, 662, A53
University, University of Notre Dame, Observatório Nacional / MCTI, The Ohio Łokas, E. L., Ebrová, I., del Pino, A., et al. 2016, ApJ, 826, 227
State University, Pennsylvania State University, Shanghai Astronomical Obser- Luber, N., van Gorkom, J. H., Hess, K. M., et al. 2019, AJ, 157, 254
vatory, United Kingdom Participation Group, Universidad Nacional Autónoma Malavasi, N., Arnouts, S., Vibert, D., et al. 2017, MNRAS, 465, 3817
de México, University of Arizona, University of Colorado Boulder, University Martinez-Valpuesta, I., Aguerri, J. A. L., González-García, A. C., Dalla Vecchia,
of Oxford, University of Portsmouth, University of Utah, University of Virginia, C., & Stringer, M. 2017, MNRAS, 464, 1502
University of Washington, University of Wisconsin, Vanderbilt University, and Moore, B., Katz, N., Lake, G., Dressler, A., & Oemler, A. 1996, Nature, 379,
Yale University. 613
Moore, B., Lake, G., & Katz, N. 1998, ApJ, 495, 139
Navarro, J. F., Frenk, C. S., & White, S. D. M. 1996, ApJ, 462, 563
O’Kane, C. J., Kuchner, U., Gray, M. E., & Aragón-Salamanca, A. 2024, MN-
References RAS[arXiv:2409.09028]
Pandey, B. & Sarkar, S. 2020, MNRAS, 498, 6069
Abell, G. O. 1962, in IAU Symposium, Vol. 15, Problems of Extra-Galactic Re- Pearson, W. J., Wang, L., Alpaslan, M., et al. 2019, A&A, 631, A51
search, ed. G. C. McVittie, 213 Peng, C. Y., Ho, L. C., Impey, C. D., & Rix, H.-W. 2010, AJ, 139, 2097
Aguerri, J. A. L., Balcells, M., & Peletier, R. F. 2001, A&A, 367, 428 Pignataro, G. V., Bergamini, P., Meneghetti, M., et al. 2021, A&A, 655, A81
Aguerri, J. A. L., Girardi, M., Agulli, I., et al. 2020, MNRAS, 494, 1681 Quilis, V., Moore, B., & Bower, R. 2000, Science, 288, 1617
Ahumada, R., Prieto, C. A., Almeida, A., et al. 2020, ApJS, 249, 3 Roediger, J. C. & Courteau, S. 2015, MNRAS, 452, 3209
Alpaslan, M., Driver, S., Robotham, A. S. G., et al. 2015, MNRAS, 451, 3249 Rojas, R. R., Vogeley, M. S., Hoyle, F., & Brinkmann, J. 2004, ApJ, 617, 50
Aragon Calvo, M. A., Neyrinck, M. C., & Silk, J. 2019, The Open Journal of Salim, S., Rich, R. M., Charlot, S., et al. 2007, ApJS, 173, 267
Astrophysics, 2, 7 Sánchez-Janssen, R., Iglesias-Páramo, J., Muñoz-Tuñón, C., Aguerri, J. A. L., &
Arnaud, M., Pointecouteau, E., & Pratt, G. W. 2005, A&A, 441, 893
Vílchez, J. M. 2005, A&A, 434, 521
Baldry, I. K., Balogh, M. L., Bower, R. G., et al. 2006, MNRAS, 373, 469
Schaye, J., Crain, R. A., Bower, R. G., et al. 2015, MNRAS, 446, 521
Balogh, M. L., Baldry, I. K., Nichol, R., et al. 2004, ApJ, 615, L101
Silk, J. & Rees, M. J. 1998, A&A, 331, L1
Bamford, S. P., Nichol, R. C., Baldry, I. K., et al. 2009, MNRAS, 393, 1324
Singh, A., Mahajan, S., & Bagla, J. S. 2020, MNRAS, 497, 2265
Barnes, J. E. & Hernquist, L. 1996, ApJ, 471, 115
Somerville, R. S. & Davé, R. 2015, ARA&A, 53, 51
Bekki, K., Couch, W. J., & Shioya, Y. 2002, ApJ, 577, 651
Springel, V., White, S. D. M., Jenkins, A., et al. 2005, Nature, 435, 629
Blanton, M. R. & Moustakas, J. 2009, ARA&A, 47, 159
Strauss, M. A., Weinberg, D. H., Lupton, R. H., et al. 2002, AJ, 124, 1810
Boschin, W., Girardi, M., De Grandi, S., et al. 2023, A&A, 672, A199
Bulbul, E., Liu, A., Kluge, M., et al. 2024, A&A, 685, A106 Tempel, E., Kipper, R., Saar, E., et al. 2014, A&A, 572, A8
Bulichi, T.-E., Davé, R., & Kraljic, K. 2024, MNRAS, 529, 2595 Tempel, E. & Libeskind, N. I. 2013, ApJ, 775, L42
Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 Toomre, A. & Toomre, J. 1972, ApJ, 178, 623
Castignani, G., Vulcani, B., Finn, R. A., et al. 2022, ApJS, 259, 43 Treu, T., Ellis, R. S., Kneib, J.-P., et al. 2003, ApJ, 591, 53
Cautun, M., van de Weygaert, R., Jones, B. J. T., & Frenk, C. S. 2014, MNRAS, van de Weygaert, R. 1994, A&A, 283, 361
441, 2923 Vulcani, B., Poggianti, B. M., Moretti, A., et al. 2019, MNRAS, 487, 2278
Chen, Y.-C., Ho, S., Brinkmann, J., et al. 2016, MNRAS, 461, 3896 Wegner, G. A., Salzer, J. J., Taylor, J. M., & Hirschauer, A. S. 2019, ApJ, 883,
Chen, Y.-C., Ho, S., Freeman, P. E., Genovese, C. R., & Wasserman, L. 2015, 29
MNRAS, 454, 1140 White, S. D. M. & Frenk, C. S. 1991, ApJ, 379, 52
Chen, Y.-C., Ho, S., Mandelbaum, R., et al. 2017, MNRAS, 466, 1880 White, S. D. M., Frenk, C. S., Davis, M., & Efstathiou, G. 1987, ApJ, 313, 505
Colless, M., Dalton, G., Maddox, S., et al. 2001, MNRAS, 328, 1039 White, S. D. M. & Rees, M. J. 1978, MNRAS, 183, 341
Darvish, B., Sobral, D., Mobasher, B., et al. 2014, ApJ, 796, 51 Xu, W., Guo, Q., Zheng, H., et al. 2020, MNRAS, 498, 1839
Das, A., Pandey, B., & Sarkar, S. 2023, Research in Astronomy and Astro- York, D. G., Adelman, J., Anderson, Jr., J. E., et al. 2000, AJ, 120, 1579
physics, 23, 115018 Zarattini, S., Aguerri, J. A. L., Calvi, R., & Girardi, M. 2022, A&A, 668, A38
Di Matteo, T., Springel, V., & Hernquist, L. 2005, Nature, 433, 604 Zarattini, S., Aguerri, J. A. L., Tarrío, P., & Corsini, E. M. 2023, A&A, 676,
Dolag, K., Meneghetti, M., Moscardini, L., Rasia, E., & Bonaldi, A. 2006, MN- A133
RAS, 370, 656 Zel’dovich, Y. B. 1970, A&A, 5, 84
Domínguez-Gómez, J., Pérez, I., Ruiz-Lara, T., et al. 2023, A&A, 680, A111 Zhang, Y., Yang, X., Wang, H., et al. 2013, ApJ, 779, 160
Dressler, A. 1980, ApJ, 236, 351

Article number, page 11 of 11