Maestra 3

While the original MUNICS catalogue contained photometry in the K0, J, I, R and V bands, we decided to do complementary observations in the B filter for all MUNICS mosaic fields. Adding more information about the blue part of the spectrum has two advantages: Firstly, the photometric redshifts (Baum 1962, Koo 1985, Bender et al. 2001) which are used for the bulk of the objects in the survey catalogue reach higher accuracy with more available filters. This is especially true for objects at low redshifts, where the V filter is insufficient in tracing prominent breaks in the spectral continuum. The use of photometric redshifts in MUNICS is described in detail in Drory (2002) and Drory et al. (2003, hereafter MUNICS II). Secondly, at redshifts z>_∼0.5 or so the flux in the B band traces the ultraviolet part of the objects’ spectral energy distribution and thus can be used to estimate the star formation rate of MUNICS galaxies.

B-band observations of the MUNICS mosaic fields were carried out 9 and 11 Jan-

uary 2002 and 8-10 June 2002 with the Calar Alto Faint Object Spectrograph (CAFOS) focal reducer at the 2.2-m telescope at Calar Alto Observatory (Spain).

The data were reduced in a fairly standard manner usingIRAF2(Tody 1986, 1993), except for cosmic ray cleaning. The B-band data reduction is very similar to the anal- ysis of the other optical data in MUNICS, described in Feulner (2000) and Drory et al. (2001b). The frames were bias/over-scan corrected and then flat-fielded using a com- bination of dome flats and sky flats.

Cosmic ray events were identified by searching for narrow local maxima in the image and fitting a bivariate rotated Gaussian to each maximum. A locally deviant pixel is then replaced by the mean value of the surrounding pixels if the Gaussian obeys appropriate flux ratio and sharpness criteria Gössl & Riffeser (2002). Such a procedure is much more expensive in terms of computing time (roughly 10 CPU minutes per frame) compared to standard median filtering techniques, but is much more reliable in finding cosmic ray events in the wings of objects and in cleaning long cosmic ray trails.

The re-imaging system of CAFOS causes substantial radial distortion of the image which had to be dealt with before co-adding the offset images. Therefore the frames 2 IRAF, the Image Reduction and Analysis Facility, is distributed by the National Optical Astronomy Observatories, which are operated by the Association of Universities for Research in Astronomy, Inc, under cooperative agreement with the National Science Foundation.

2.3. MUNICS B-BANDIMAGING 33

were rectified using the known distortion equation, a polynomial of fourth order in the distance from the optical axis (K. Meisenheimer, private communication).

10

15

20

25

30

0

2

4

6

8

Maddox et al. (1990) APM Arnouts et al. (1999) Arnouts et al. (2001) Bertin & Dennefeld (1997) Couch & Newell (1984) Hedon−Dumbleton (1989) Infante et al. (1997) Jarvis & Tyson (1981) Jones et al. (1991) Koo (1986) Kron (1978) Lilly et al. (1991) Maddox et al. (1990) Metcalfe et al. (1991) Metcalfe et al. (1995a) Metcalfe et al. (1995b) Metcalfe et al. (1998) Metcalfe et al. (2000) WHDF Metcalfe et al. (2000) HDFN Metcalfe et al. (2000) HDFS Peterson et al. (1986) Stevenson et al. (1986) Tyson (1988) Yasuda et al. (2001)

B [mag]

log N (B) [mag

−1

deg

−2

]

Figure 2.2: Galaxy number counts in the MUNICS B-band images compared to litera- ture data. No correction for incompleteness has been applied. The published data were taken from a compilation of N. Metcalfe (private communication).

If necessary, variations in the background intensity across the frames caused by scattered light were fitted and subtracted in each individual frame. The images were then corrected for atmospheric extinction and scaled to a common photometric zero- point before finally being added using the positions of∼15 bright objects for determi- nation of the offsets between the individual frames.

During photometric nights, photometric standard stars were observed (Landolt 1992). For each field, a photometric zero point and the atmospheric extinction were determined. The extinction coefficients were usually consistent with a Rayleigh atmo- sphere, with a few nights showing higher extinction, albeit within the variations typical for Calar Alto.

The B images of each field were registered against the K0-band image by matching the positions of_∼200 bright homogeneously distributed objects in the frames and de- termining the coordinate transform from the K0-band system to each image in the other

34 CHAPTER2. THEMUNICHNEAR-INFRARED CLUSTER SURVEY

four pass-bands using the tasksXYXYMATCHand GEOMAPwithin IRAF. The scatter in the determined solutions is less than 0.2 pixels RMS. Note that the frames them- selves are not transformed. We only determine accurate transformations and apply these later to the apertures in the photometry process.

To show the quality of the B-band data for the MUNICS project we compare the galaxy number counts, i.e. the surface density of galaxies per magnitude bin, to pub- lished number counts (see Figure 2.2). Stars have been excluded based on their spec- tral energy distributions (see 7.3 for details). Note that the number counts shown in the Figure are not corrected for incompleteness at fainter magnitudes.

Clearly, the MUNICS B-band observations are reasonably complete to B_∼24. It also nicely illustrates the scope of the survey: While there are surveys with much larger field which are able to find large numbers of bright (low redshift) galaxies, and the classic deep fields which probe the very faint (and very high redshift) galaxy popu- lation, MUNICS constitutes an important probe of the intermediate-redshift universe.

Furthermore, in Figure 7.1 we show the completeness functions for the B-band images of the ten MUNICS mosaic fields, based on Monte-Carlo simulations of artifi- cial point-sources placed at random in the images. After running the object detection algorithm on the images, the fraction of re-detected artificial objects as a function of magnitude is computed. This is the quantity shown in the Figure. It is obvious that the data quality is very good and homogeneous over all ten MUNICS fields.