1 Introduction

Clusters of galaxies are the largest virialized structures in the Universe, and accurate knowledge of their global properties is needed to constrain models of galaxy formation and evolution. The first step in this direction requires the construction of a statistically well-defined sample of clusters in the nearby universe to be used as a "local template". Unfortunately, due to historical and observational reasons, and in spite of much effort, existing samples cannot be considered ideal. Existing cluster catalogs, in fact, fall into three main categories: i) large catalogs derived from photographic surveys (POSS-I, UKST) by visual inspection and covering wide portions of the sky (Abell 1958; Abell et al. 1989; Zwicky et al. 1961-68) but missing the needed depth, homogeneity and completeness (Postman et al. 1986; Sutherland 1988); ii) catalogs machine extracted with objective criteria from photographic plates (cf. Dodd & MacGillivray 1986; Dalton et al. 1992; Lumsen et al. 1992), reaching, in some cases, limiting magnitudes fainter than (i), but not covering equally wide areas of the sky and so far available only for the Southern hemisphere (only UKST plates); iii) accurate and deeper catalogs usually derived from CCD data and selected on the basis of objective criteria but covering much smaller regions of the sky and containing only a small number of objects (cf. Postman et al. 1996; Olsen et al. 1999). Excluding the ones, which will be derived from the Sloan Digital Sky Survey (SDSS) (Kim et al. 2000; Kepner et al. 1999), for the Northern sky no automatically extracted catalogs of galaxies (nor of clusters) could be produced until the recent completion of the Digital Palomar Sky Survey (DPOSS). DPOSS, which covers three bands (J,F,N), is characterized by deeper limiting magnitudes than other catalogs extracted from photographic surveys and therefore offers an important opportunity to investigate galaxy clusters in the nearby (i.e. z < 0.4) Universe. Recently, DPOSS photometric calibration (Weir et al. 1995) and extraction of the Palomar Norris Sky Catalog (PNSC) were completed at Caltech in collaboration with the observatories of Rome, Naples and Rio de Janeiro, partners in the CRoNaRio project (Djorgovski et al. 1998a; Andreon et al. 1997). This catalog contains astrometric, photometric and morphological information for all objects detected down to limiting magnitudes of $g_J \sim 21.5$, $r_F \sim 20.5$ and $i_N \sim 20.0$ in the Gunn & Thuan photometric system. With the availability of these new data, various methods to build galaxy cluster catalogs based on color information have been proposed (Gal et al. 1999; Gal et al. 2000a).

Figure 1: Detected overdensities in the density map in a $5^{\circ } \times 5^{\circ }$ sky region centered at RA = 1 h and Dec = +15$^{\circ }$. The smoothing on this map is performed by a $3 \times 3$ Gaussian filter. Squares mark known Abell and Zwicky clusters. Circles mark putative - previously unknown - clusters.

In order to exploit the scientific potential of the DPOSS, we developed a model independent procedure to search for cluster candidates. The procedure is detailed elsewhere and therefore we briefly summarize its main characteristics here (Puddu et al. 2000; Puddu et al., in preparation). First, for a given region of the sky, we extract the individual catalogs obtained from the corresponding J, F and N DPOSS plates and calibrate them to the g, r and i Gunn-Thuan System using the procedure described in Weir et al. (1995). Then, after correcting for misclassified objects in at least one of the three filters (see Puddu et al. 2000), we produce a matched (in the three bands) catalog complete to ${\it r} \sim 19.75$ and ${\it g} \sim 20.2$ and, in order to ensure completeness, we disregard all objects fainter then ${\it r} = 19.5$. The spatial galaxy distribution in the matched catalog is then binned into equal area square bins of 1.2 arcmin to produce a density map. S-Extractor (Bertin & Arnouts 1996) is run on the resulting image in order to identify and extract all the overdensities having number density 2 $\sigma$ above the mean background and covering a minimum detection area of 4 pixels (equivalent to 4.8 arcmin²) on the map convolved with a $3 \times 3$ Gaussian filter. We wish to stress that, as in the Schectman (1985) approach, we are not assuming any a priori cluster model. The procedure is run on a given plate catalog to produce a preliminary candidate cluster catalog. All previously known Abell and Zwicky clusters are recovered together with many new cluster candidates (Fig. 1 shows, as an example, DPOSS field 610) which need to be confirmed. The best way to validate cluster finding algorithms would be to apply them to a field for which redshifts are available for a fairly deep magnitude limited sample of galaxies. Since such a sample does not exist, we were obliged to follow an alternative route. It is well known that early-type galaxies are preferentially located in the cores of rich clusters and that they present a small scatter in the color-magnitude diagram (Dressler 1980; Dressler & Gunn 1992; Stanford et al. 1998). These properties turn the color-magnitude diagram into a powerful tool to disentangle true clusters from overdensities of galaxies caused by random object alignment along the line of sight. In this paper we use the early type sequences detected in the color-magnitude diagrams obtained from multiband optical photometry to confirm a sample of 12 candidates. These candidate clusters were selected from some fully reduced DPOSS plates available at the time the test was performed. The selection was performed randomly in order to include overdensities covering a wide range of S/N ratios in the detection maps. We also use a sample of 8 X-ray selected clusters at nearby and intermediate redshift as templates to calibrate the photometric redshift estimate. The paper is structured as follows: in Sect. 2 we describe the data and the data reduction strategy, in Sect. 3 we show the color-magnitude diagrams for the calibration sample and illustrate the procedure implemented to derive the photometric redshift estimate, while in Sect. 4 we apply this method to the candidates. In Sect. 5 conclusions and future developments are discussed.

 

 

Table 1: Selected overdensities. Notes:(1) Zwicky cluster.

OBJ	RA (2000)	Dec (2000)	Richness within	Detection
			the isopleth	S/N
27_694(1)	05 00 07.24	+10 15 52.00	76	8
44_778	08 59 52.68	+04 10 53.20	21	4
17_778	09 08 28.20	+06 03 39.55	63	8
5_778	09 12 11.09	+02 23 10.22	17	4
1_778	09 12 15.34	+02 32 18.11	48	7
64_781	09 57 25.10	+03 39 06.70	27	5
72_781	09 57 53.23	+03 27 10.09	59	7
6_725	15 24 52.60	+11 20 27.10	146	12
1_799(1)	16 03 11.78	+03 14 17.63	132	11
24_694	05 03 38.92	+10 38 08.59	50	7
21_694	05 04 42.34	+10 48 49.00	70	8
26_727	15 57 48.70	+08 52 04.39	8	3

 

 

Table 2: Sample of known clusters. G&L: Gioia & Luppino (1994); S&R: Struble & Rood (1999).

Id	RA (J2000)	Dec (J2000)	z	Ref. (z)
MS0821.5+0337	08 24 07.104	+03 27 45.22	0.347	G&L
Abell 1437	12 00 24.960	+03 20 56.40	0.1339	S&R
MS1253.9+0456	12 56 28.827	+04 40 01.87	0.23	G&L
Abell 1835	14 01 02.399	+02 52 55.20	0.2532	S&R
MS1401.9+0437	14 04 29.378	+04 23 00.33	0.23	G&L
MS1426.4+0158	14 28 58.768	+01 45 11.94	0.32	G&L
Abell 2033	15 11 23.518	+06 19 08.40	0.0818	S&R
MS1532.5+0130	15 35 02.739	+01 20 57.15	0.49	G&L