Survey depth and filter choice. The depth required is discussed in several sections above. Most importantly, to detect galaxy clusters to z=1.5 (section 5.2) requires K=21. The depth K=21 will also ensure virtually complete identification of all SWIRE sources (section 5.4.1). In the redshift range of interest the reddest galaxies besides EROs are passively evolving ellipticals which are no redder than J-K=1.5. Therefore a J limit 1.5 mag deeper than the K limit can identify unusually red objects by J-K colour (or colour upper limits). The photo-z study of Bolzonella et al. (2000) demonstrates that where optical data exist, the J and K bands provide most of the SED information in the near-ir, unsurprising since K lies at the long wavelength end, and J lies midway between K and the z band.
To provide colour discrimination within the ERO population it is desirable to reach even deeper. For this sub-population alone this is most efficiently achieved in the H band, since the short wavelength difference to K provides the maximum contrast. For this reason we propose to cover a small part of the DXS, 5 sq. degs, to a depth H=22.
Survey Area. The chosen survey area 35sq. degs is set by the goals of determining omegam to an accuracy of 0.03 (section 5.2), and the requirement of section 5.3 to survey a volume as large as the volume of the 2dF galaxy redshift survey.
Choice of fields. We have made a preliminary choice of the XMM-LSS, Lockman Hole, and Elais N1 fields as our targets. Table 5.1 summarises the coordinates of the fields and the XMM coverage, as well as coverage at other wavelengths. The XMM-LSS field is chosen because of the wide-field deep XMM pointings, while the two northern fields are well placed for SZ coverage, given the northern latitudes of the SZ observatories. These two fields also already have some XMM coverage.
| Field | RA, Dec | multiwavelength coverage | ||||
|---|---|---|---|---|---|---|
| DXS | XMM | GALEX | SIRTF | optical | ||
| J, K area | area/texp | |||||
| XMM-LSS | 02 21, -05 00 | 10 sq. degs | 10 sq. degs / 10+ ks | Y | Y | CFHLS |
| Lockman Hole | 10 45, +58 00 | 15 sq.degs | 0.2 sq. degs / 200 ks | Y | Y | |
| Elais N1 | 16 13, +55 16 | 10 sq. degs | 1.1 sq. degs / 30 ks | Y | Y | INT |
Table 5.1. Summary of provisionally selected target fields. The columns list: 1. field name, 2. the field coordinates, 3. the DXS area to be imaged in J and K, 4. the current completed or scheduled XMM areal coverage and the XMM integration time in kilosec, 5. whether or not GALEX (uv) and 6. SIRTF (mid to far ir) observations are planned, and 7. the existing optical coverage of the field (CFHLS=Canada France Hawaii Legacy Survey, INT=Wide Angle Survey with the INT Wide Field Camera).
Optical coverage. Deep ugriz optical data is required for most of the scientific goals set out in this proposal. This would be feasible with the INT wide-field camera. Alternatively it may be attractive to target VIRMOS fields, or fields from the CFH Legacy Survey. One can anticipate substantial convergence of target fields between groups surveying at the several different wavelengths. Although coverage of 35 sq. degs to the required depths to match the proposed J and K limits is a large undertaking, compared to the DXS itself this would require a relatively modest allocation of telescope time with contemporary optical wide-field cameras such as SuprimeCam on Subaru, and MegaCam on CFHT, and we do not foresee this as a limiting factor.
Follow-up spectroscopy with FMOS. FMOS is a 400 fibre R=600 multiobject spectrograph being built for Subaru, with expected completion in 2005. FMOS has two characteristic features which make it an excellent counterpart to the DXS. Firstly the wavelength coverage is Z through H, which therefore includes H-alpha (the strongest useful galaxy emission line) in the target redshift range 1<z<1.5. The second advantage is the wide field of view 0.5o. With OH suppression FMOS will provide the fastest survey speed of any spectrograph for the measurement of redshifts of early-type galaxies at 1<z<1.5, for scales of order 1o. At J=22, for an early-type galaxy, FMOS achieves S/N=15 per resolution element in 4 hours, over the full wavelength range. The speed for measuring redshifts for emission line galaxies will be much faster.
FMOS provides the capability for measuring velocity dispersions of detected galaxy clusters (section 5.2). As mentioned in section 5.3 preliminary planning for a sparse-sampled redshift survey of 105 z>1 galaxies, covering the entire DXS region, has begun. The scope of this survey is similar to the 2dF galaxy redshift survey. This will provide the opportunity for a statistical analysis of the galaxy spectra, which can be compared against the recent results at low redshift of Baldry et al. (2001).