Survey Design

Figure 1: The M-z distribution of DEVILS galaxies in comparison to GAMA and zCOSMOS. We will probe Mz=0 to z∼ 1 and 0.1Mz=0 to z∼0.4. Grey shaded region shows space occupied by major merger (<3:1) pairs for Mz=0 galaxies. Our mass limits from TAO also agree with empirical estimates from existing surveys. Note, histograms are in log(number).

Depth & Target Selection

To characterise sub-Mpc structure to z < 0.8, DEVILS will reach the lowest possible stellar mass galaxies, while remaining within the constraints of the AAT+AAOmega. To meet our science goal the DEVILS selection depth is designed to: i) identify multiplicity >3, >1013M groups to z = 0.7 , ii) measure the galaxy stellar mass function to Mz=0 at z = 1 (Fig 1), and iii) measure the major merger rate for M∗z=0 galaxies to z = 0.7 by identifying all systems at < 3:1 mass ratio (> 0.3 Mz=0, grey band Fig 1). A Y < 21.2 selection will meet these constraints. DEVILS will use the Y -band as we wish to i) be close to a single band measurement of stellar mass (NIR), ii) select in flux at rest-frame > 4000 ̊A to z ∼ 1 (> 8000 ̊A observed), iii) select from a single instrument/facility (VISTA has the only contiguous observations across our fields) and iv) meeting previous conditions, be as close to the observing wavelength of AAOmega as possible. The VISTA Y -band at ∼ 1 μm meets all of these criteria, and is the deepest band in VIDEO (5σ point source depth is Y ∼ 25). Observations with the AAT will cover ∼ 3700-8800 ̊A at R ∼ 1000-1600, allowing identification of key spectral features across our redshift range; [OII] and Ca H & K to z ∼ 1.

Area & Field Selection

The robustness of any evolution and structure survey increases with volume. Thus, we wish to maximise area, while once again remaining within observing constraints and retaining maximum legacy value. We require robust number statistics for both structure tracers (groups/pairs) and individual galaxies. Given our depth and key science goals we require ~6deg2 to both parametrise the evolution of the high mass end of the halo mass function and determine the effect of environment on galaxy evolution at intermediate redshifts. To maximise legacy science with auxiliary data, DEVILS will target deep, well studied extragalactic fields.  There are a number of optimal 2dF (∼ 2-3 deg2) deep fields with diverse multi-wavelength data accessible from the AAT. We have selected three (Fig 2, Table 1) which have the most extensive existing legacy data and are soon to be targeted with the next generation of instruments: each field has/will have: deep x-ray (XMM, Chandra), optical (HSC, DES, LSST deep-drill), NIR (VIDEO, UltraVISTA), MIR (Spitzer), FIR (Herschel) imaging, high resolution imaging (Euclid, HST), and radio imaging/spectra (ATCA, MeerKAT, JVLA), as well as existing spectroscopy (e.g. OzDES, zCOSMOS…). As such, these are key legacy fields which would benefit immensely from a highly complete spectroscopic survey – a unique database for the astronomy community. We opt to target a 2 × 1.5 deg2 contiguous region in XMM-LSS to maximise legacy science with x-ray observations of DEVILS groups and large-scale structure. We will however, prioritise a single 1.5deg2 region in XMM-LSS before observing the second, to ensure we obtain the completeness for our science goals.

Figure 2: Left – DEVILS field positions. DEVILS will target three deep extragalactic fields: D02 (XMM-LSS, ~3deg2), D03 (ECDFS, ~1.5deg2) and D10 (COSMOS, ~1.5deg2). Figure made using AstroMap. Right – The predicted light cone distribution of DEVILS sources. We will probe to much greater look-back times than GAMA and SDSS, which retaining a high level of completeness.

Table 1: Target Fields

Common Name
Observable (AAT)
Input Band
Input Depth
Input Survey
D02 XMM-LSS 35.4975 -4.825 02:21:59 -04:49:30 Sept-Dec VISTA-Y Y<21.2 VIDEO
D03 ECDFS 53.65 -28.1 03:34:36 -28:06:00 Sept-Jan VISTA-Y Y<21.2 VIDEO
D10 COSMOS 150.1 2.1 10:00:24 02:06:00 Jan-April VISTA-Y Y<21.2 UltraVISTA

Comparison To Other Surveys

In order to probe the evolution of galaxies, their environment and the dark matter halos in which they reside, we require highly complete samples. While this has been achieved in the relatively local Universe (SDSS, GAMA, etc) until now, deep higher redshift surveys have suffered from low completeness which are inadequate for galaxy group/pair science.  Fig 3 displays a comparison between DEVILS and existing surveys. The only surveys at comparable redshifts and number densities have complex selection functions and/or very low completeness (VIPERS, DEEP2, VVDS). The current state of the art survey for galaxy evolution at these epochs (zCOSMOS- bright) covers much less area, has under half an order of magnitude fewer sources and only ∼ 50 % completeness. None of the key science goals of DEVILS can be undertaken with these surveys.

Figure 3: Comparison of DEVILS with other surveys. Point size = log(area). We show surveys with a simple magnitude selection (circles) and more complex colour/zphoto selections (diamonds). DEVILS is comparable in completeness and number to large area surveys (2dFGRS, SDSS, GAMA) but has much higher source density. The most directly comparable survey is zCOSMOS – DEVILS has over half an order of magnitude more sources, ∼ 6 times larger area and much higher completeness.