Galaxy Clusters from the DESI Legacy Imaging Surveys. I. Cluster Detection
Hu Zou, Jinghua Gao, Xin Xu, Xu Zhou, Jun Ma, Zhimin Zhou, Tianmeng Zhang, Jundan Nie, Jiali Wang, Suijian Xue
aa r X i v : . [ a s t r o - ph . GA ] F e b Draft version February 24, 2021
Typeset using L A TEX twocolumn style in AASTeX62
Galaxy Clusters from the DESI Legacy Imaging Surveys. I. Cluster Detection
Hu Zou, Jinghua Gao, Xin Xu,
Xu Zhou, Jun Ma,
1, 2
Zhimin Zhou, Tianmeng Zhang, Jundan Nie, Jiali Wang, and Suijian Xue Key Laboratory of Optical Astronomy, National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100012, China School of Astronomy and Space Science, University of Chinese Academy of Sciences, Beijing 101408, China (Received —; Revised —; Accepted —)
Submitted to ApJSABSTRACTBased on the photometric redshift catalog of Zou et al. (2019), we apply a fast clustering algorithmto identify 540,432 galaxy clusters at z . . Monte-Carlo simulations indicate that the false detection rate of our detectingmethod is about 3.1%. The total masses of galaxy clusters are derived using a calibrated richness–massrelation that are based on the observations of X-ray emission and Sunyaev & Zel’dovich effect. Themedian redshift and mass of our detected clusters are about 0.53 and 1 . × M ⊙ , respectively.Comparing with previous clusters identified using the data of the Sloan Digital Sky Survey (SDSS), wecan recognize most of them, especially those with high richness. Our catalog will be used for furtherstatistical studies on galaxy clusters and environmental effects on the galaxy evolution, etc. Keywords: galaxies: clusters: general — galaxies: distances and redshifts — galaxies: photometry INTRODUCTIONGalaxy clusters are the most massive gravitation-ally bound systems in the universe. They are formedon the cosmic web, tracing the large-scale structure.Galaxy clusters are powerful probes for the struc-ture growth and provide important constraints oncosmological parameters (Kravtsov & Borgani 2012;Planck Collaboration et al. 2016a). Galaxy clustersare also excellent astrophysical laboratories for study-ing the galaxy evolution, galaxy collision, gravitationallense, dark matter, and extreme physics in dense en-vironments (Dressler 1980; Butcher & Oemler 1984;Moore et al. 1996; Lin & Mohr 2004; Sand et al. 2004;Kneib & Natarajan 2011; Kravtsov & Borgani 2012;Ebeling et al. 2014).Galaxy clusters are a collection of galaxies that arebonded by the gravity. In addition to stars constitut-ing the luminous matter of galaxies in a cluster, thereare multiple observable compositions, including cold gasand dust in galaxies, hot ionized intra-cluster medium
Corresponding author: Hu [email protected] (ICM), and dark matter. The multi-component natureof galaxy clusters make them visible across the elec-tromagnetic spectrum. The X-ray emission and Sun-yaev & Zel’dovich effect in microwave band originatingfrom the hot ICM (Sunyaev & Zeldovich 1972; Voit2005; Bleem et al. 2015) can be used to detect redshift-independent samples of galaxy clusters (Piffaretti et al.2011; Reichardt et al. 2013; Planck Collaboration et al.2016b; Tarr´ıo et al. 2019). These samples usuallypresent relatively high purity and completeness buttend to be high-mass systems. The most effective detec-tions of galaxy clusters are based on large-scale opticaland near-infrared imaging data. A variety of clusterfinding techniques have been developed attributing toongoing and upcoming wide and deep large-scale pho-tometric surveys. Generally, there are two kinds ofcluster finding methods: one is to find the overdensityof galaxies in the three-dimensional space (Szabo et al.2011; Wen et al. 2012; Gao et al. 2020) and the other isbased on the red-sequence feature (Koester et al. 2007;Hao et al. 2010; Rykoff et al. 2014), which describes thenarrow color distribution of red member galaxies in acluster (also known as the E/S0 redgeline). The red-sequence method utilizes intrinsic color properties ofmember galaxies, while the overdensity finding method
Hu Zou et al. can identify clusters not presenting the red-sequencefeature, especially at high redshift.In Gao et al. (2020), we introduced a new fast clus-tering algorithm, called as Clustering by Fast Searchand Find of Density Peaks (CFSFDP), to identifygalaxy clusters from a photometric redshift (photo-z) catalog, which is based on the 7-band photometryof South Galactic Cap U-band Sky Survey (SCUSS;Zhou et al. 2016; Zou et al. 2016), SDSS, and unWISE(Lang et al. 2016) surveys. A total of about 20,000clusters were discovered over a sky area of 3700 deg in the south Galactic cap. The latest data release of the DESI (Dark Energy Spectroscopic Instrument;DESI Collaboration et al. 2016) legacy imaging surveysprovide 5-band photometry over a sky area of about20,000 deg in both northern and southern Galacticcaps (Dey et al. 2019). They are composed of threeoptical components in grz bands and one near-infraredcomponent from the Wide-field Infrared Survey Explorer (WISE) satellite. We have obtained a catalog of accu-rate photo-zs and stellar masses for more than 0.3 billionmorphologically classified galaxies with r <
23 mag andin the redshift range of z < m = 0 .
3, Ω Λ = 0 .
7, and H = 70 km s − Mpc − . IMAGING DATA AND PHOTO-ZThe DESI is planned to conduct a large-scale spec-troscopic survey with 5000-fiber robots installed on thefocal plane of the 4m Mayall telescope at Kitt Peak,Arizona (DESI Collaboration et al. 2016). It will mea-sure the redshifts of about 35 million galaxies andquasars over a 5-year period, aiming to explore the structure growth and expansion history of the universe.The DESI legacy imaging surveys are designed to pro-vide spectroscopic targets for DESI. They are composeof three public optical surveys including the Beijing-Arizona Sky Survey (Zou et al. 2017), the Dark EnergyCamera Legacy Survey (Blum et al. 2016), and the May-all z -band legacy survey (Silva et al. 2016), which jointlyimage a sky area of ∼ in g , r , and z bands.The legacy surveys also integrate the latest WISE obser-vations and provide deep force photometry in two WISEW1 and W2 bands (Lang et al. 2016; Meisner et al.2019). The latest data release (DR8) covers an area ofabout 20,000 deg in both northern and southern Galac-tic caps, which includes additional data from Dark En-ergy Survey (Dark Energy Survey Collaboration et al.2016). The optical imaging depths are 1.5–2 mag deeperthan the SDSS and the infrared W W which contains about 0.3 bil-lion galaxies with r <
23 mag. The redshift and massranges are about z < . < log( M ∗ ) < . < − . , PHOTO ZERR < . , (1)where MAG ABS R is the r -band absolute magnitudeand PHOTO ZERR is the photo-z error. The cut of thephoto-z error is used to eliminate galaxies with largephoto-z uncertainty. The cut of the absolute magnitudeis set to keep more luminous galaxies and hence improvethe completeness. Finally, a total of 0.18 billion galaxiesare left for identifying clusters in the rest of this paper. CLUSTER IDENTIFICATION3.1.
Detecting method http://batc.bao.ac.cn/ ∼ zouhu/doku.php?id=projects:desi photoz:start alaxy clusters We adopt the resolutionparameter of nside = 64, giving each HEALPixpixel area of about 0.89 deg . The above pixe-lation is to facilitate the parallelization of clusterdetecting.2. For each galaxy at a given redshift z in a spec-ified HEALPix pixel, we calculate the local den-sity (Φ) and background density (Φ bkg ). Thesetwo quantities are computed from the galaxies inthis pixel and its surrounding 8 neighbor pixels,which have a total area of about 8 deg . The areais large enough to estimate a proper local back-ground. Φ is computed as the number of galaxieswith distance R < . z ± . z ). Φ bkg is computed as the numberof galaxies with R > D in Mpc is defined as theshortest distance of the specified galaxy to othergalaxies with higher Φ.3. The cluster centers should be located on the over-density peaks with large enough local density andadequately far away from other peaks. They are http://healpix.jpl.nasa.gov/ identified as the galaxies with Φ > bkg and D > r band asthe cluster center. This cluster center is not alwaysthe brightest cluster galaxy (BCG) that is usuallyclose to the densest region. We simply regard the r -band brightest galaxy within 0.5 Mpc aroundthe density peak as the BCG. Since each galaxy ispointed to the nearest neighbor with higher Φ, thegalaxies in the redshift slice can be traced and clus-tered to the cluster center. The clustered galaxiesare roughly considered as the member candidates.With these members, we can calculate the averageposition and redshift for each cluster.4. The number of member galaxies within 1 Mpcaround the cluster center are calculated, which issubtracted the local background and denoted as N . N is regarded as the first-order ap-proximate of the cluster richness. We require thatclusters should have N >
10. The total r -band luminosity of member galaxies around thecluster center ( L ) is calculated in unit of L ∗ ,where L ∗ is the characteristic luminosity. L will be used as a proxy of the cluster mass andrichness later. When calculating L , we alsosubtract the background luminosity that is esti-mated in the same way as Φ bkg . Note that therichness and luminosity in this paper are not cor-rected for the variation of the imaging depth. Suchcorrections are quite complicated due to the mix-ture nature of our data from three different opti-cal surveys plus the WISE infrared survey and theselection of the photo-z sample. Actually, morethan 90% of the area has the 5 σ r -band magni-tude limit larger than 23.4 mag. As presented inZou et al. (2019), the completeness of galaxies at r <
23, which is the magnitude cut for the photo-z catalog used in this paper, is higher than 90%.As a preliminary estimate, the correction of therichness or luminosity would be less than 10%.The above process is performed over the photo-z cat-alog. Finally, we obtain a catalog of 540,432 galaxyclusters at z .
1. The catalog content is described inAppendix A. Figure 1 presents four examples of our de-tected clusters at different redshifts. As can be seenfrom this figure, the prominent overdensity features ofgalaxies make them easily identified from our imagingdata by using the above cluster finding method.3.2.
Reliability analysis
Hu Zou et al.
Figure 1.
Four examples of galaxy clusters at different redshifts: (a) z = 0 .
05, (b) z = 0 .
34, (c) z = 0 .
54, and (d) z = 0 . − N (N 1MPC), richness (RICHNESS), total mass (M 500), and characteristic radius (R 500) (see Section 3.3 and Table 2). Northis up and east is left. The cluster detecting process is executed in relativelylarge redshift slices due to the photo-z uncertainty. Itis inevitable to introduce the project effect and hencecause false cluster detections. In addition, the localbackground fluctuation will also bring contaminationsto the low-richness clusters. We perform a Monte-Carlosimulation to evaluate the detecting approach and esti-mate the false detection rate following a similar methodof Wen & Han (2011). The simulation is based on theactual photometric data and the specific steps are de-scribed as follows. (1) An arbitrary region with an area of 400 deg is selected in our photo-z catalog. (2) Galax-ies in this region are redistributed on the sky with ran-dom walks from original locations in the range of 1–2.5Mpc and then their redshifts are shuffled. This step cangenerate new random galaxy backgrounds and mean-while retain the project effect in some degree. (3) Thesame clustering detecting process used in this paper isapplied to these mock galaxies and to finding fictitiousclusters.We make a total of 10 simulations and identify thefalse detections in these simulations. The false rate is de- alaxy clusters N . From this figure, we can see that the falserate generally increases with redshift and decreases with N . In other words, the contamination of our detec-tions is more serious for distant or low-richness clusters.The overall false rate is about 3.1% and it can go up toabout 7-8% at high redshift and lowest richness.It should be noticed that the above simulations canestimate the false detections inducing by chance associ-ations and part of false detections caused by the projec-tion effect of uncorrelated structures, because we gen-erate random galaxy backgrounds with limited positionchanges. Nevertheless, the projection effects include theimpacts from both correlated and uncorrelated large-scale structures. It is very important to estimate theireffect on basic cluster properties (e.g. richness and halomass) especially when the cluster catalog is used toconstrain the cosmological parameters. For example,Costanzi et al. (2019) combined both real data and nu-merical simulations to characterize the impact of projec-tion effects on the redMaPPer cluster catalog and theyfound that the projection effect obviously biases the cos-mological parameter measurements. Principally, we canuse the mock catalogs from N -body simulations to as-sess the reliability of our cluster finder. However, thesynthetic data are model-dependent and are probablybiased from realistic observations, such as the halo oc-cupation distribution of cluster galaxies, galaxy colordistributions, photometric uncertainty and correspond-ing photo-z uncertainty. Considering the complexity ofthe projection effects, we expect an investigation of theirimpact on the reliability of our detecting method and onthe following mass estimation in future work.3.3. Mass and richness estimation
The X-ray and SZ surveys provide large samples ofgalaxy clusters at different redshifts, which has reason-ably estimates of the cluster mass with relatively smallintrinsic scatter. The observables, such as luminosityand temperature in X ray and the integrated Compton-isation parameter (or SZ parameter), are tightly linkingto the cluster mass via scaling relations.Wen & Han (2015) collected a sample of 1,192 clustersfrom the literature and scaled the cluster masses derivedby different authors to a unified calibration. These sam-ples are limited to match the galaxy clusters identifiedfrom SDSS with the maximum redshift is 0.75. The clus-ters at high redshift are insufficient, so we supplementthe sample of Wen & Han (2015) with other clusters from X-ray and SZ observations (Piffaretti et al. 2011;Takey et al. 2013, 2014; Wen & Han 2015; Takey et al.2016; Planck Collaboration et al. 2016b). The clustermasses in different catalogs were estimated using differ-ent mass proxies and scaling relations, so it is necessaryto recalibrate the masses. So we use the rescaling re-lations derived by Wen & Han (2015) to calibrate thecluster mass for these additional data. All the clustermasses are rescaled to the standard of Vikhlinin et al.(2009). We should note that the published Planck cata-log has not applied the Eddington bias correction, whichcan systematically bias the SZ masses (Battaglia et al.2016; Medezinski et al. 2018). The above recalibrationprocess can reduce this kind of systematic effects. Here,the cluster mass is denoted as M , which is the masswithin a characteristic radius ( R ). R is definedas the radius within which the mean density of a clus-ter is 500 times the critical density of the universe ( ρ c ).According to the definition, M and R satisfy thefollowing relation: M = 4 π R × ρ c . (2)We collect 3,157 clusters with reliable estimations of M and R , which spreads over the whole celestialsphere. They are listed in a separate table of Table 3 inAppendix B, which we called as the calibration catalog.As described in Gao et al. (2020), L is used as anoptical proxy of the cluster mass. We also use L to estimate the total mass of our clusters. The galaxyclusters detected in this work are cross-matched with theclusters in the calibration catalog of Table 3, followinga similar procedure in Bleem et al. (2020): 1) rankingthe clusters in the calibration catalog by the decreas-ing mass (equivalent to the richness); 2) matching eachcluster in the calibration catalog to the richest clusterin our catalog with a redshift error of ∆ z = 0 . z )and a projected separation error of 1 Mpc; 3) removingeach matched cluster from our catalog and repeating theabove matching process for all clusters in the calibrationcatalog. As a result, we find 1,797 matched clusters thatcan be used to determine the relation between L and M . The left panel of Figure 3 shows M asfunction of L . We can see that there is a tight cor-relation between these two quantities. The correlationcan be described by the following equation:log( M ) = a log( L ) + b log(1 + z ) + c, or M = 10 c L a (1 + z ) b , (3)where a , b , and c are constants to be determined andthe term of (1 + z ) is regarded as the correction term forthe redshift evolution and incompleteness. Hu Zou et al. F a l s e r a t e
10 20 30 40 50N
Figure 2.
Left: false detection rate as function of redshift. Right: false detection rate as function of N . The error barshows the standard deviation of 10 simulations. The above luminosity–mass relation suffers from theMalmquist bias as the cluster sample is constructed withflux-limited X-ray and SZ observations. The Malmquistbias makes the sample preferentially contain brighteror massive clusters. It is especially serious for clusterswith low richness at high redshift. The compilations ofour clusters are from X-ray and SZ observations withdifferent depths, which may mitigate the effect of theMalmquist bias, but also induces the complication ofthe bias correction. We hope a further investigation ofthis problem in the future. Considering the bias, we as-sign the data with weight of L z when fitting the datawith Equation (3). This will put more weight to the clus-ters with larger richness and at lower redshift. We getthe best-fit parameters via weighted linear regression: a = 0 . ± . b = 0 . ± .
14, and c = 12 . ± . M and the fitted one asfunctions of L and redshift. The overall RMS of theresidual is about 0.2 dex. The residual RMS changeswith L , ranging from 0.28 to 0.12 dex. The residualbiases at high redshift, partly reflecting the Malmquistbias. We apply the Equation (3) to estimate M andEquation (2) to estimate R for our detected clusters.We also use the Equation (17) in Wen & Han (2015) toestimate the richness. These parameters are included inTable 2.3.4. Cluster statistics and photo-z accuracy
Figure 9 shows the parameter distributions for thewhole 540,432 galaxy clusters detected in this paper.The median redshift is about 0.53 and the number den-sity decreases at z > ∼ .
6, where the cluster detectionshould be more incomplete. The median richness isabout 22.5. The range of the total mass log( M ) is13.5–14.8 and the median is about 14.1 (equivalent to1 . × M ⊙ ). The characteristic radius of R rangesfrom about 0.4 Mpc to 1.0 Mpc and the median value isabout 0.63 Mpc.Figure 5 displays both the distributions of r -banddepth and number density of our clusters over the DESIimaging footprint. Note that the depth is referred to 5 σr -band magnitude limit for extended sources with thecorrection of the Galactic extinction. The depth mapin Figure 5a shows three distinct regions with differentimaging depths, which reflects that the observing datawere obtained by different facilities. The general depthdifferences among these three components are about 0.6mag. Although the depth is inhomogeneous across thesurvey footprint, we still obtain uniformly distributedcluster samples attributing to our relatively conserva-tive magnitude cut of our photo-z catalog (see Figure5b). In order to more clearly show the effect of theimaging depth on our cluster detection, we present thenumber densities as function of redshift for three sub-samples of our clusters under different r -band depths inFigure 6. From this figure, we can see that the imagingdepth only affects the cluster detection at z > .
4. More alaxy clusters Figure 3.
Left: correlation between log( M ) and log( L ). The error bar shows the error of log( M ). Middle: thedifference (∆ log( M )) between log( M ) and fitted values as function of L . The red error bar shows the median andstandard deviation in each luminosity bin. Right: ∆ log( M ) as function of redshift. The red error bar shows the median andstandard deviation in each redshift bin. The horizontal line displays ∆ log( M ) = 0. N (a) 0 25 50 75Richness050000100000150000 N (b)13.5 14.0 14.5log(M ) (logM ⊙ )0250005000075000100000125000150000 N (c) 0.4 0.6 0.8 1.0r ⊙(Mpc)0250005000075000100000125000 N (d) Figure 4.
Distributions of redshift (a), richness (b),log( M ) (c), and R (d) for our clusters. clusters are detected in the deeper area and it becomesmore pronounced as the redshift increases.The photo-z accuracy of our galaxy clusters are es-timated by comparing the photometric redshifts of theBCGs in our catalog with spectroscopic redshifts col-lected by Zou et al. (2019). There are a total of 12,2390BCGs with spectroscopic redshifts. The top panel ofFigure 7 shows the comparison between the photomet-ric redshift ( z phot ) and spectroscopic redshift ( z spec ).The bottom panel presents the photo-z accuracy of∆ z norm = ( z phot − z spec ) / (1 + z spec ) as function of z spec .We can see that the photo-z accuracy changes from (a)
23 25 r-band depth (mag) (b)
10 40
Number density (deg −2 ) Figure 5. (a) The r -band depth map in Mollview projec-tion. (b) Number density distribution of our detected clus-ters. about 0.01 at z ∼ . z ∼ .
9. Theredshift histogram in Figure 8 shows the distribution of σ ∆ z norm of the total sample. The overall photo-z accu-racy is σ ∆ z norm = 0 . Hu Zou et al. N u m b e r d e n s i t y ( d e g − ) Figure 6.
Number density as function of redshift for threesubsamples of our clusters under different r -band imagingdepths. The redshift bin is 0.1. which is shown in the blue histogram of Figure 8. Sucha comparison also gives a similar photo-z accuracy of σ ∆ z norm = 0 . Figure 7. (a) Comparison between photometric and spec-troscopic redshifts of the BCGs in our cluster catalog. Thediagonal line denotes z phot = z spec . (b) σ ∆ z norm as func-tion of z spec . The red diamonds and blue crosses display themedians and standard deviations in different redshift bins.4. MATCHING WITH OTHER CLUSTERCATALOGS −0.10 −0.05 0.00 0.05 0.10ΔZ norm d e n s i t y Δ d i s t r i b u t i o n Figure 8.
The σ ∆ z norm distribution (red) of the BCGs in ourclusters, which have spectroscopic redshift measurements.The blue histogram is the σ ∆ z norm for confirmed galaxy clus-ters with reliable redshift measurements, which are detectedin X-ray and SZ-effect observations. The most famous cluster catalog is the Abell cat-alog, which was constructed by the visual identifica-tion on the photographic plates. There are also sev-eral catalogs of galaxy clusters identified from the SDSSphotometric data. These catalogs are either based onthe red-sequence feature or the overdensity feature inthe photo-z space. Through matching our catalog withthese catalogs, we can check the detecting repeatabil-ity of the clusters from both shallower photometric dataand deeper DESI imaging data. Such comparisons canalso indicate the reliability of our cluster detection. TheAbell clusters were visually identified as galaxy over-densities (Abell et al. 1989). Among other catalogsthat are based on SDSS data, the clusters of MaxBCG(Koester et al. 2007), GMBCG (Hao et al. 2010), andredMaPPer (Rykoff et al. 2014) were identified throughthe red-sequence feature, while the clusters of AMF(Szabo et al. 2011) and WHL15 (Wen & Han 2015) wereobtained using the overdensity feature. We summarizethe redshift range, adopted method and number of clus-ters in the DESI footprint of these catalogs in Table 1.If the clusters in the Abell catalog have spectroscopicredshifts, we use their spectroscopic redshifts as the dis-tance indicator. For those clusters without redshifts inthe Abell catalog, we only use our photo-z as the dis-tance tracer to match the clusters. The projected sep-aration is limited to 1 Mpc, which is set to be a lit- alaxy clusters R of mostclusters as shown in Figure 9). For the SDSS clustercatalogs, we match them with our cluster catalog usinga redshift tolerance of ∆ z < . z ) and the sameseparation tolerance. The larger the separation toler-ance is set, the more real clusters would be matched inconsideration of the uncertainty of the cluster center,while more possible false association would occur andvice versa. The matching results with our catalog arelisted in Table 1.As checked, most of the mismatching clusters in thosecatalogs based on the SDSS data are the clusters withlower richness or at higher redshift, which are possiblefalse detections. We select two representative catalogs,redMaPPer and WHL15, to show the matching rates asfunction of redshift and richness and present the com-parison of richness and redshift with those in our cata-log in Figure 9. Generally, the matching rate increasesas the richness increases or the redshift decreases. Therichnesses among these catalogs show good correlations,although there are somewhat large scatters. The photo-zalso shows pretty correlations. Compared to the WHL15photo-z, the photo-z of the redMaPPer clusters is moreconsistent with that of our clusters. SUMMARYGalaxy clusters are excellent probes for studyinggalaxy formation and evolution in the dense environ-ments and they can be also used to constrain cosmolog-ical parameters. Using the photometric redshift catalogof Zou et al. (2019), we identify a catalog of galaxyclusters over the 20000-deg sky footprint of the DESIlegacy imaging surveys. This sample of clusters will beused to statistically study the galaxy clusters at z < z .
1. These clusters are uniformly distributed over theimaging footprint. Monte-Carlo simulations show thatthe false detection rate is about 3.1% and it becomesworse as the redshift increases or the richness decreases.The false rate at high redshift or low richness can reachup to about 8%. We utilize the mass measurements fromX-ray and radio observations to calibrate the total massand richness of the detected clusters by using the opti-cal luminosity L . The median mass and richness areabout 1 . × M ⊙ and 22.5, respectively. Compar-ing our catalog with the Abell catalog, we can recoveralmost all Abell clusters. We also compare our detectedclusters with those based on the SDSS data that areabout 2 magnitude shallower than the DESI imaging surveys. It is found that we can recognize most of theclusters identified from SDSS, especially those with highrichness.We thank the anonymous referee for his/her thought-ful comments and insightful suggestions that improveour paper greatly. This work is supported by Major Pro-gram of National Natural Science Foundation of China(No. 11890691, 11890693). It is also supported by theNational Natural Science Foundation of China (NSFC;grant Nos. 11733007, 11673027, 11873053, 12073035)and the National Key R&D Program of China No.2019YFA0405501.The Legacy Surveys consist of three individual andcomplementary projects: the Dark Energy CameraLegacy Survey (DECaLS; NOAO Proposal ID Hu Zou et al.
Table 1.
Matching our catalog with other cluster catalogsCatalog Redshift Method Number Matched Percent(1) (2) (3) (4) (5) (6)Abell · · · visual overdensity 2,298 2,143 93.3%MaxBCG 0 . < z < . . < z < .
55 red sequence 55,317 29,980 54.2%AMF 0 . < z < .
78 overdensity 44,692 33,731 75.5%redMaPPer 0 . < z < .
55 red sequence 25,835 24,645, 95.4%WHL15 0 . < z < . Note —(1) Cluster catalogs: Abell (Abell et al. 1989), MaxBCG (Koester et al. 2007), GMBCG (Hao et al. 2010), AMF(Banerjee et al. 2018), redMaPPer (Rykoff et al. 2014), and WHL15 (Wen & Han 2015). (2) Redshift range. (3) Adoptedmethod for identifying clusters. (4) Number of clusters in the DESI imaging footprint. (5) Number of matched clusters usinga matching separation tolerance of 2 Mpc. (6) Matching rate using a matching separation tolerance of 2 Mpc. (7) Number ofmatched clusters using a matching separation tolerance of 1 Mpc. (8) Matching rate using a matching separation tolerance of1 Mpc.
Figure 9. (a) The matching rate between the redMaPPer catalog and our catalog as function of redshift and richness. (b)The correlation between the richness of our clusters and that of the redMaPPer clusters ( λ ). (c) The correlation between thephoto-z of our clusters and that of the redMaPPer clusters. (d) The matching rate between the WHL15 catalog and our catalogas function of redshift and richness. (e) The correlation between the richness of our clusters and that of the WHL15 clusters(RL 500). (f) The correlation between the photo-z of our clusters and that of the WHL15 clusters. tado do Rio de Janeiro, Conselho Nacional de Desen-volvimento Cientifico e Tecnologico and the Ministe-rio da Ciencia, Tecnologia e Inovacao, the DeutscheForschungsgemeinschaft and the Collaborating Institu-tions in the Dark Energy Survey. The Collaborat-ing Institutions are Argonne National Laboratory, theUniversity of California at Santa Cruz, the University of Cambridge, Centro de Investigaciones Energeticas,Medioambientales y Tecnologicas-Madrid, the Univer-sity of Chicago, University College London, the DES-Brazil Consortium, the University of Edinburgh, theEidgenossische Technische Hochschule (ETH) Zurich,Fermi National Accelerator Laboratory, the Universityof Illinois at Urbana-Champaign, the Institut de Cien- alaxy clusters A. INFORMATION ABOUT OUR CLUSTER CATALOGIn total, we obtain 540,432 galaxy cluster with a false detection rate of about 3.1%. Table 2 lists the contentcontained in our catalog. It includes the position and redshift of the cluster center, which is the galaxy at thelocal density peak. The catalog also includes the basic information of the BCG including the position, redshift,observed magnitudes, absolute magnitudes, and stellar mass, which are inherited from the photo-z catalog. Thecluster properties estimated in this paper are also contained in this catalog. This catalog will be uploaded to CDS andit is also available at http://batc.bao.ac.cn/ ∼ zouhu/doku.php?id=projects:desi clusters:start. B. GALAXY CLUSTERS WITH THE TOTAL MASS MEASUREMENTS FROM X-RAY AND RADIOOBSERVATIONSTable 3 lists the galaxy clusters detected by X-ray and SZ-effect observations and their total masses are effectivelyestimated and rescaled by us to a unified calibration. The characteristic radius of R is calculated according toEquation (2). REFERENCES Abell, G. O., Corwin, H. G., Jr., & Olowin, R. P. 1989,ApJS, 70, 1Banerjee, P., Szabo, T., Pierpaoli, E., et al. 2018, NewA,58, 61Battaglia, N., Leauthaud, A., Miyatake, H., et al. 2016,JCAP, 2016, 013. doi:10.1088/1475-7516/2016/08/013Bleem, L. E., Stalder, B., de Haan, T., et al. 2015, ApJS,216, 27Bleem, L. E., Bocquet, S., Stalder, B., et al. 2020, ApJS,247, 25. doi:10.3847/1538-4365/ab6993 http://batc.bao.ac.cn/ ∼ zouhu/doku.php?id=projects:desi photoz:start Blum, R. D., Burleigh, K., Dey, A., et al. 2016, AmericanAstronomical Society Meeting Abstracts Hu Zou et al.
Table 2.
Column description for our cluster catalog
Column Unit Format DescriptionCLUSTER ID Long Cluster IDRA PEAK degree Double R.A. for the density peak (J2000)DEC PEAK degree Double decl. for the density peak (J2000)PZ PEAK Float Photometric redshift for the density peakSZ PEAK Float Spectroscopic redshift for the density peak if existingDEN PEAK Integer Local density Φ for the density peakBKG PEAK Float Local background density Φbkg for the density peakRA MEAN degree Double Mean R. A. of possible membersDEC MEAN degree Double Mean decl. of possible membersN 1MPC Integer Number of member galaxies within 1 Mpc from the cluster centerL 1MPC L ∗ Float Total luminosity of member galaxies within 1 Mpc from the cluster centerM 500 log10( M ⊙ ) Float Total mass of the cluster M R g -band magnitude for the BCGMAG R BCG mag Float r -band magnitude for the BCGMAG Z BCG mag Float z -band magnitude for the BCGMAG W1 BCG mag Float W W g -band magnitude error for the BCGMAGERR R BCG mag Float r -band magnitude error for the BCGMAGERR Z BCG mag Float z -band magnitude error for the BCGMAGERR W1 BCG mag Float W W g -band absolute magnitude for the BCGMAG ABS R BCG mag Float r -band absolute magnitude for the BCGMAG ABS Z BCG mag Float z -band absolute magnitude for the BCGMAG ABS W1 BCG mag Float W W M ⊙ ) Float Logarithmic stellar mass for the BCGMASS INF BCG log10( M ⊙ ) Float Lower limit of logarithmic stellar mass with 68% confidence level for the BCGMASS SUP BCG log10( M ⊙ ) Float Upper limit of logarithmic stellar mass with 68% confidence level for the BCG Table 3.
Galaxy clusters with the total mass measurements from X-ray and SZ-effect observationsID RA DEC Z M500 E M500 R500 E R500 REF(1) (2) (3) (4) (5) (6) (7) (8) (9)1 4.57107 16.29432 0.55 4.11 0.77 0.93 0.06 W152 8.32687 -21.41319 0.19 0.99 0.20 0.66 0.04 T133 9.82501 0.69981 0.28 0.66 0.13 0.56 0.04 W154 9.84351 0.80273 0.41 1.62 0.31 0.72 0.05 T13, T165 9.92591 0.75922 0.42 1.27 0.25 0.66 0.04 T13, T166 10.16344 25.51840 0.15 1.24 0.24 0.72 0.05 T137 10.48690 25.53105 0.13 0.93 0.19 0.66 0.04 T138 10.62969 0.85368 0.16 0.69 0.15 0.59 0.04 T13, T169 10.71922 0.71671 0.27 1.16 0.23 0.68 0.04 T13, T1610 10.72397 -9.57311 0.41 1.49 0.29 0.70 0.05 T13
Note —(1) sequence number. (2) right ascension in degree (J2000). (3) declination in degree (J2000). (4) redshift. (5) totalmass M in 10 M ⊙ . (6) error of M in 10 M ⊙ . (7) R in Mpc. (8) error of R in Mpc. (9) references: P11 forPiffaretti et al. (2011), P15 for Planck Collaboration et al. (2016b), W15 for Wen & Han (2015), T13 for Takey et al. (2013),T14 for Takey et al. (2014), and T16 for Takey et al. (2016).(This table is available in its entirety in machine-readable form and it can be also available athttp://batc.bao.ac.cn/ ∼ zouhu/doku.php?id=projects:desi clusters:start.)Ebeling, H., Stephenson, L. N., & Edge, A. C. 2014, ApJL,781, L40Gao, J., Zou, H., Zhou, X., et al. 2020, PASP, 132, 024101 Hao, J., McKay, T. A., Koester, B. P., et al. 2010, ApJS,191, 254Kneib, J.-P. & Natarajan, P. 2011, A&A Rv, 19, 47 alaxy clusters13