UDS DR11 Catalogue (2019 Edition)

Overview

This page presents the UDS DR11 12-band matched catalogue (2019 Edition), created by Omar Almaini, David Maltby, Will Hartley and Chris Simpson. Further details will be provided in two forthcoming data papers (Almaini et al., Hartley et al., in preparation). An updated set of photometric redshifts is currently in preparation.

The catalogue above contains photometry in 12 bands, mask information, columns to identify cross-talk artefacts, star-galaxy classification, photometric redshifts, absolute magnitudes, and stellar masses. Note that coordinates are given in the WFCAM astrometric frame, calibrated to 2MASS.

A simplified description of the catalogue columns may be found here. See below for further details. Further information on the UDS project can be obtained from the UKIDSS UDS web page.

WARNING: Please do not use this catalogue without first applying some basic cuts (see notes below). For a quick galaxy catalogue please select the "good galaxy" subset identified by the corresponding Boolean column; this excludes secure stars, applies a basic optical/IR mask and removes cross-talk objects and candidates stars. Depending on your goals, you may also wish to apply a chi2 limit for the photo-zs or apply a K-band magnitude cut. Use the "best" subset to select objects covered by the deepest IRAC photometry.


Photometry

The catalogue is K-band selected, based on the UDS DR11 release. The 95% completeness limit is estimated to be K=25.0 (AB). A completeness curve based on simulated galaxies can be found here. Photometry is provided in 12 bands (U, B, V, R, i', z', Y, J, H, K, 3.6, 4.5), where available, in AB magnitudes, measured from 2-arcsec diameter apertures. Note that the fluxes/magnitudes provided have been aperture-corrected to total magnitudes. Photometry from a wider of range of apertures can also be provided on request. Raw photometry, with no aperture corrections, can also be provided if required. Fluxes in all bands can be converted to AB magnitudes as follows: m_AB = -2.5 log (f) +30.

The JHK photometry is from the DR11 release of UKIDSS UDS. Optical imaging (B,V,R,i',z') is from Subaru Suprimecam (Furusawa et al. 2008), with improved background subtraction and zero-point scaling (images available here). Complementary U-band imaging is from CFHT (PI: Almaini), while Y-band imaging is obtained from VISTA VIDEO. IRAC imaging at 3.6 and 4.5 microns is based on the SpUDS Spitzer Legacy Survey (PI: Dunlop), combined with deeper data from SEDS. In the outer regions of the field shallower data is used from SWIRE (see subsets described below).

Masks, stars, and cross-talk

The "maskflag" column ($86) provides information on the multiwavelength masks. Masked regions correspond to image boundaries, artefacts, and bright stars. The flagging system uses a 2N system to determine whether an object is flagged in each of the input images. See here for details. To simplify the use of these flags, we have pre-defined subsets that will be suitable for most purposes (see below). The multiwavelength mask images are also provided below.

Candidate stars are identified using the Boolean column "Stars-final". Brighter stars (K<20) are identified using JHK colours and profiles only (Almaini et al., in preparation). Fainter stars (K>20) are identified using 12-band photometric fits combined with the K-band image profile, using the method described in Simpson et al. (2013). The list has been extended to fainter magnitudes compared to previous catalogues, and now contains 8275 objects (formerly ~5500).

Cross-talk are identified with a Boolean column (XTALK_DR11), which identifies false cross-talk images, or real objects within 1-arcsec of a likely cross-talk feature, for which the JHK photometry is likely to be compromised. The number of X-TALK candidates is considerably expanded compared to previous versions of this catalogue, and now contains 5910 objects (formerly ~3000).

Pre-defined subsets

We have combined the mask information, cross-talk, and stellar classification to create four subsets, which ought to be useful for most purposes. The "best" subset includes only data with optimal 12-band photometry, including deeper IRAC coverage from SEDS/SpUDS. The "good" subsets also have 12-band photometry, but extend the areal coverage by ~10% by including shallower IRAC data from SWIRE.

1. "Best": this Boolean column identifies unmasked objects within the deepest 12-band multiwavelength coverage, with cross-talk removed. Number of objects: 208,190.

2. "Best galaxies": as above, with stars excluded. Number of objects: 202,282.

3. "Good": this Boolean column identifies unmasked objects with 12-band multiwavelength coverage, including shallower IRAC data from SWIRE. Note that "Best" objects are a subset of "Good". Cross-talk have been removed. Number of objects: 223,766.

4. "Good galaxies": as above, with stars excluded. Number of objects: 217,429.

Photo-zs

The photometric redshifts were obtained using EAzY (Brammer et al.), using a method similar to that described in Simpson et al. (2013). We note that Chris identified a bug in EAzY in early 2018, which was found to be applying the K-band prior incorrectly. This issue has now been fixed.

Please refer to the EAzY manual for a description of the photo-z columns, but for most purposes we recommend using the maximum posterior redshift, z_p. Note that the stellar masses and absolute magnitudes are evaluated at z_p. For certain applications you may prefer the mean posterior redshift (z_m2), or redshifts that do not make use of the K-band prior, e.g. z_a. High values of χ 2 (chi2p or chi2a) or low values of "odds" suggest low reliability. We have also generated p(z) distributions (see below) if greater nuance is required. Note that photometric redshifts are assigned z=-99 in regions where there are fewer than 5 photometric data points, or if only one (or no) filters show the minimum required signal-to-noise (flux/err>1).

Photo-zs were calibrated using ~8000 spectroscopic redshifts, including data from UDSz, VANDELS, VIPERS, and 3DHST. The resulting dispersion in dz/(1+z) is σNMAD=0.0187, using the "best galaxies" subset, after the exclusion of known AGN (0.0197 including AGN). Further details of the spectroscopy available in the UDS field can be obtained from the UDS Data Page.

Stellar masses

Stellar masses and absolute magnitudes are provided, evaluated at the max. likelihood redshift, z_p. Upper and lower bounds are also provided. Stellar masses have been evaluated for a range of metallicities, as provided in BC03 (m22, m32, m42, m52, m62, m72, corresponding to Z= 0.0004, 0.001, 0.004, 0.008, 0.02 [solar], and 0.03 respectively). A marginalised stellar mass (Mstar_z_p) is also provided.


Image masks

The full multiwavelength mask allows you to separate masks from the individual input images, with values matching the "maskflag" parameter for objects in the catalogue. The flagging system uses a 2N system to determine whether an object is flagged in each of the input images. See here for details. The "good" regions of the mask correspond to maskflag=0, 1, 32, 33 within this full mask file, while the "best" correspond to maskflag=0, 1. For simplicity, we have also created simpler binary masks corresponding to the "good" and "best" regions. Please note that stars and cross-talk are identified separately using the catalogues themselves, and are not identified using the masks. Therefore the "good galaxies" within your catalogue correspond to objects satisfying three criteria: (a) they lie within the "good" mask region identified below, (b) they are not X-talk, and (c) they are not classified as stars.

Please note, these images are provided in the original 2MASS astrometric frame.

  NB: The raw UDS pixel scale is 0.1342 arcsec/pixel. Note that the effective areas corresponding to the "good" and "best" catalogues are 0.62774 and 0.58202 sq deg respectively. These are slightly smaller than the unmasked pixel regions given above, given the requirement to obtain unmasked 2-arcsec diameter photometry.


P(z) distributions

The P(z) distributions for all 296,007 K-band detected sources within the UDS field are provided below. The peaks in these distributions should correspond to the redshift z_p in the catalogue. For ease of use, we provide these data in two formats: ascii and binary. Note that the ascii files are considerably larger. A Python script is provided to convert the binary files to ascii, if required.

i) Ascii files

  The P(z) distribution for each source is contained within a single row of the P(z) distribution file (DR11pz.dat). The source ID for each row is identical to those in the DR11 catalogue provided above.

ii) Binary files



Last update May 4 2022 by Omar Almaini