Galaxy Clusters (A: photo-z calibration)

One of the great advantages of using galaxy clusters as cosmological probes with a photometric survey is the incredibly precise photometric redshifts that are achievable by using a large number of similarly colored cluster galaxies. However, achieving this precision to z~1 requires a precise calibration of the galaxy cluster red sequence at high redshift, including spectroscopic redshifts for a large sample of clusters. Fortunately, this calibration only requires spectroscopic redshifts for the central galaxies of each cluster, and as these are some of the brightest galaxies at these redshifts these targets are well within range of the AAOmega instrument with reasonable exposure times.

In order to select appropriate targets, we begin with a rough estimate of the red sequence model. We then run a cluster finder to find suitable central galaxies. The photometric redshifts thus computed in this preliminary step will be significantly biased and have increased scatter, but even this approximate first pass is sufficient to identify central cluster galaxies at high redshift. In this way we construct a candidate list of ~20-50 targets per square degree.

Galaxy Clusters (B: velocity dispersion measurements)

Before we can use clusters for cosmology, we require mass proxies, in addition to securing accurate cluster redshifts. Calibration of cluster mass proxies for DES will be done using a small, but representative subset of DES clusters, and will involve a variety of techniques, including X-ray observables and velocity dispersions. Simulations have shown that with roughly 20 galaxy redshifts per clusters, one can measure velocity dispersions with sufficient precision to provide useful mass calibrations via scaling laws (in order to measure masses for individual clusters, many more redshifts are required). Even with the minimum fibre distances set by the AAOmega instrument, it is still possible to gather several galaxy redshifts per cluster in one mask. In the case of multiple repeat observations (as expected by the OzDES project), it should be feasible to gather the 20 or so needed for mass calibrations.

In order to select appropriate targets, we begin with clusters detected using the XMM X-ray telescope, as these have well defined centroids and additional mass proxy information (X-ray temperatures or luminosities). Initially we may not know the cluster redshift, but can take a guess using either the RedMapper results, and/or individual galaxy photo-z's (using the Arbor-z technique) to select against background and foreground galaxies. We will also identify candidate AGN near the X-ray cluster centroid, using DES colours (if those AGN turn out to at the same redshift as the cluster, then the X-ray observables will suffer from contamination). We will construct a candidate list of ~20-50 targets per cluster. These targets will be densely packed, so will be prioritised carefully, to ensure the fibre positioning software is not overly taxed. Once the cluster redshift has been unambiguously determined, the target selection will be refined appropriately. We note that from the December 2012, successful redshifts up to z=0.9 were obtained using this technique.

White Dwarfs

The DES’s science requirements drive its requirements for high-precision (i.e., relative) and high-accuracy (i.e., absolute) photometry. The DES has a relative photometry requirement – a requirement that the photometry must be internally consistent to 2% rms over the full survey area in each of the five filters – of 2% (and a goal of 1%). To achieve this requirement, the DES will make use of a strategy of multiple tilings (with large offsets compared with the size of the DECam focal plane) of the survey footprint in each filter band. The DES also has absolute photometry requirements – requirements on the conversion of count rates in ADU/sec to a physical specific flux measure (in, say, ergs/sec/cm^2/Hz) – of 0.5% in both color (bandpass-to-bandpass calibration) and flux (overall intensity).

Due to their relatively simple spectra, pure hydrogen atmosphere (“DA”) white dwarfs are the current standard for absolute calibration (see, e.g., Holberg & Bergeron 2006; Holberg, Bergeron, & Gianninas 2008). The process requires calculating the synthetic grizY magnitudes from the well-modeled spectra of a good set of DA white dwarfs and comparing their synthetic magnitudes with their observed grizY magnitudes. The measured differences between the synthetic and observed grizY magnitudes are the offsets required to perform the absolute calibration of the DES photometry by placing it firmly onto an AB magnitude system (a magnitude system tied to units of ergs/sec/cm^2/Hz). In order to do this, we need to ensure that we have a good sample of DA white dwarfs observable by DECam (preferably within the DES footprint) that have well-modeled spectrophotometry. Unfortunately, there is a dearth of confirmed DA white dwarfs in the Southern skies, especially ones with well-characterized spectra. The hydrogen Balmer lines can be measured to extract parameters, Teff and log g, that can be used to predict the natural DA white dwarf spectra. High precision and control of systematics in this entire procedure is important to be relevant to the required precision of DES.

For the OzDES AAT proposal, which focuses primarily on covering areas associated with the DES Supervnova Fields, we are drawing our candidate DA white dwarfs from several sources, in particular, from the SDSS (where there is overlap) and from Rowell & Hambly’s (2011) all-sky catalog of proper-motion-identified white dwarf candidates. Color selection from Dark Energy Survey data using its preliminary calibration will also be used. In the end, we hope to have several well-measured DA white dwarfs in each DES Supernova Field. By the end of the DES, combining data from the OzDES AAT program with other candidate DA White Dwarf programs, we hope to have a “Golden Sample” of several tens to a couple hundred well-characterized DA white dwarfs for the calibration of the DES (as well as other Southern surveys that happen to overlap the DES footprint).

Active Transient Follow-up

We obtain a spectrum of all transient sources with a differential magnitude brighter than r=22.5 detected by DES in the supernova fields. The past run we also partnered with La Silla/QUEST because we were uncertain whether the DES search would be running in time and to provide transients of the interesting SNLS S2 field that is not surveyed by DES.

There are several science goals of this survey. The AAT plays a major role in the DES plan to get spectroscopic typing of a major fraction of its low-redshift Type Ia supernova discoveries. The magnitude completeness and spectroscopic observations of non-Ia transients provide control for the false-positive rate in photometric classification. Spectroscopic typing allows for calculation of rates for the range of transient sources. There is the potential for the discovery of novel types of transients.

Supernova Host Galaxies

The DES Supernova (SN) Survey is expected to discover nearly 4000 SNe Ia that will have light-curves suitable for use in cosmological analyses. Spectroscopic confirmation of the SN type for this quantity of transient objects, however, is not feasible. To classify a SN based solely on its light-curve is possible, but to achieve a high accuracy a redshift prior is required for the fit. For this task, the accuracy provided by spectroscopic redshifts is required.

Following the precedent of Lidman et al. (2012), we propose to use AAOmega/2dF instrument, with its near-perfect match in size to the DES footprint, to obtain redshifts from the host galaxies of SNe in the 10 DES SN fields. In addition to the large field-of-view and the multitude of fibers, obtaining redshift measurements from the SN host eliminates the time-sensitivity of standard SN observations, allowing observations to be made for a large number of objects at once. Fainter galaxies can have fibers placed on them repeatedly throughout the duration of the survey, resulting in high-confidence spectroscopic measurements for even very faint hosts.

The data obtained from ~10 observations of each DES SN field to their nominal depth in each filter during the extended SV period (2012-2013) will allow us to have a deep galaxy catalog on hand for the start of DES operations. Using this catalog, real-time decisions can be made about whether a SN host is bright enough to rely on future AAT observations, or if the SN must be directly observed to obtain type & redshift information. We anticipate this cutoff to be in the vicinity of r~24 for passive galaxies and r~26 for starforming galaxies. As we will be observing the hosts of all types of SNe, not just SNe Ia, observations of host-galaxies with AAT will allow us to type > 5000 SNe.

Photo-z Main Sample

The main photo-z sample is defined as a simple magnitude-limited, i < 21 galaxy sample and is intended to improve the calibration of DES photo-z's. This sample of thousands of galaxy redshifts that will be obtained provides a valuable training set to derive the DES photo-z solution at these magnitudes. More importantly, this AAOmega galaxy redshift sample will be distributed over the 30 sq. deg. area of the DES supernova fields, providing about 10 times larger sky coverage compared to existing deep training sets (e.g., VVDS and DEEP2) that will also be used for DES photo-z calibration. The much larger total area of the AAOmega sample will be needed to help alleviate the impact of large scale structure fluctuations in the training set data on the redshift distributions derived from DES photo-z's. This so-called "sample variance" effect on DES photo-z's has recently been identified as an important systematic (Cunha et al. 2012, MNRAS, 423, 909) to be overcome to optimize DES weak lensing cosmology constraints. Moreover, we will also measure angular cross correlations (Matthews & Newman 2010, ApJ, 721, 456) that will let us use the AAOmega sample to calibrate the redshift distribution of fainter galaxies, down to the DES photometric limit i ~ 24.


The repeat imaging and AAT spectroscopy of the SN fields will produce an extremely valuable dataset for several QSO science projects. These include over an order of magnitude increase in the number of luminous QSOs with black hole mass estimates from reverberation mapping, measurement of the faint end of the QSO luminosity function at high redshift, enable cross-correlation studies of galaxies and QSOs, and provide a high surface density of QSOs suitable for tomography of the IGM. Here are brief science cases for the first two of these projects:

Reverberation Mapping of Luminous QSOs

Reverberation mapping provides the only direct method to estimate of the masses of the supermassive black holes that power QSOs. This method combines a geometric estimate of the size of the broad line region, obtained with multi-epoch photometry and spectroscopy, with the characteristic velocity of the broad emission lines, to estimate the virial mass of the black hole. To date, reverberation mapping has measured black hole masses in over 50 AGN, yet all but three are low-luminosity and at low-redshift, rather than the luminous QSOs that represent the bulk of supermassive black hole growth over cosmic time. This is because the size of the broad line region in luminous QSOs can be on order a light year across, and consequently several years of data are needed for a single measurement. We propose to spectroscopically monitor approximately ten bright QSOs per SN field at z=0.5 to 2 to increase the number of QSOs with direct black hole mass measurements by an order of magnitude over the five year survey.

Going Below the Knee of the QSO LF at High Redshift

Accurate measurement of the knee of the QSO LF before the peak of their space density at z~2 is an important constraint on the relationship between QSOs, galaxies, and halos early in cosmic history. This is because the knee of the QSO LF is related to the knee of the galaxy mass function and halo mass function, and the sharpness of the break constrains the scatter between QSOs, galaxies, and halos. The DES SN fields extend over a magnitude deeper and over nearly an order of magnitude more area than the most comparable, existing surveys. This combination of depth and area is a powerful combination to measure the knee of the QSO LF at high redshift. The extensive variability data that will be available for the SN fields, combined with the substantial ancillary data that already exist, will enable very efficient QSO selection for spectroscopic follow-up observations. We propose to observe several hundred QSOs per SN field to better measure the knee of the QSO LF over the important redshift 2 < z < 4. This estimate is based on present data and models, which suggest that to g<23 mag (AB) there are about 60 QSOs per deg^2 at 2<z<3 and about 15 per deg^2 at 3<z<4. Measurement of on order a hundred QSOs per SN field with single observations, combined with a comparable number of fainter QSOs with repeat, stacked exposures will produce a substantial step forward in understanding the coevolution of galaxies and QSOs.

Radio galaxies

We propose an AAOmega spare fibres program, as part of the OzDES survey, to conduct a spectroscopic survey of the radio-detected galaxies in the Australia Telescope Large Area Survey (ATLAS). Combining a sensitive radio survey with the detailed environmental metrics provided by the DES survey will allow measurement of the role of AGN feedback in galaxy formation/evolution and the growth of structure. We propose to use ~30 spare AAOmega fibres per OzDES observation, yielding redshifts and spectroscopic typing to i<22 for 300-600 radio sources each campaign year.