` Millions of z>0.6 Luminous Red Galaxies from BigBOSS and WISE Jeffrey A. Newman (U. Pittsburgh), Timothy C. Licquia (U. Pittsburgh), Nick Mostek (LBNL), Kyle Barbary (LBNL), Adam Stanford (UC Davis), Arjun Dey (NOAO), Jean-Paul Kneib (OAMP), Michael Levi (LBNL), David Schlegel (LBNL), and the BigBOSS Team The most luminous red galaxies (LRGs) provide an excellent tracer of large scale structure in the Universe due to their strong clustering, but they are difficult to select at z>0.6 from optical imaging alone. We demonstrate here that simple color cuts based upon SDSS r/i-band and WISE 3.4 micron photometry allow us to efficiently select 0.6 < z < 1 LRGs while strongly rejecting stellar contaminants. This technique exploits the fact that the "1.6 micron bump" (i.e., the bolometric peak of red galaxy SEDs at 1.6 microns) lies near the center of the WISE 3.4 micron passband at z~1. We test these techniques using WISE data in fields covered by the zCOSMOS and DEEP2 survey, investigating the nature of resulting samples. Our methods select 400 candidate LRGs per square degree, yielding more than 5 million targets over the BigBOSS footprint; 97% of the objects targeted are galaxies at z>0.55, 96% of which are on the red sequence. We also explore selection using only Palomar Transient Factory r-band imaging and WISE 3.4 micron data in the COSMOS field, finding that we can still effectively select LRGs, at the cost of a contamination with bright Emission Line Galaxies which are also of interest for BigBOSS. Selecting z>0.6 LRGs with SDSS + WISE We have tested selection techniques for BigBOSS using both data from the WISE Preliminary Data Release and proprietary catalogs provided by the WISE team. As seen in Figure 2, a simple cut in (r−i/r−[3.4μm]) color-color space efficiently selects galaxies which are both red in rest-frame color and located at z > 0.6. Simple methods will select 400 higher-z LRGs per square degree (more than 5 million over the BigBOSS area) to limiting magnitudes of r < 22.35 and i < 21.3; i.e., using only imaging to Sloan Digital Sky Survey depth. Only 2% of the objects targeted are stars, with the remainder being galaxies that are indeed luminous, red in rest-frame color, and at high redshift. 94% of the galaxies selected are at z > 0.6 (cf. Figure 3), while 96% of the galaxies targeted have red- sequence colors. The Challenge of Targeting z>0.6 LRGs The largest-volume surveys of large-scale structure have targeted the highest-mass galaxies in the z < 1 universe, a population commonly known as luminous red galaxies or LRGs (e.g., Eisenstein et al., 2001). They are associated with massive dark matter halos and hence are strongly clustered, making them highly efficient tracers of the large-scale structure. Taking advantage of their strong 4000Å breaks, LRGs at z < 0.6 can be selected efficiently and their redshifts estimated based on optical photometry (Padmanabhan et al., 2007), while the strong absorption features around the break allow redshifts to be identified definitively in spectra of modest signal-to-noise. They have therefore formed the cornerstone of the BOSS spectroscopic redshift survey. Surveying LRGs at higher redshifts will be beneficial for BigBOSS as their strong clustering enhances BAO measurements. However, LRGs are increasingly difficult to select with conventional techniques at z>~0.6; the 4000 A break passes into the i band by z ∼ 0.75, requiring time-expensive longer-wavelength imaging for selection. Instead, we need a new technique for BigBOSS. A new method: 1.6μm bump selection The spectral energy distributions of cool stars exhibit a local maximum at a wavelength of roughly 1.6μm, corresponding to a local minimum in the opacity of H − ions (John, 1988). This feature, commonly referred to as the ”1.6μm bump,” dominates the near-infrared spectra of stellar populations with ages above ∼10 Myr (Wright, Eisenhardt, and Fazio 1994). Since they possess few young stars, luminous red galaxies at z ∼ 0.5−1 will exhibit relatively large near-infrared to optical flux ratios at wavelengths of ∼ 2 − 4μm. The lowest-wavelength channel in WISE satellite imaging is centered at 3.4μm, nearly optimal for selecting these objects as it overlaps the bump at redshift z∼1. The infrared-to-optical flux ratio of LRGs rises monotonically with redshift as z approaches 1, then declines past z ∼ 1.1 (where optically bright LRGs are rare). A simple cut in r - [3.4μm] color can select high-z LRGs effectively; adding information from more optical bands helps in rejecting non-LRGs. Figure 1 illustrates the enhanced brightness of LRGs in WISE imaging. Figure 2: An optical/near- infrared color-color diagram for galaxies observed by both WISE and the DEEP2 Galaxy Redshift Survey. Blue (in rest- frame color) galaxies are indicated by small dark points; large colored squares indicate red sequence galaxies at z > 0.6; stars are indicated by the cyan diamonds. A simple cut in this color-color space can efficiently select LRGs at the redshifts of interest; the selection region used in Figure 3 is shaded in red here. Figure 3: Redshift distributions for z > 0.6 LRG samples selected as shown in Figure 1. The histograms are based on photometric redshifts from COSMOS and spectroscopic redshifts from DEEP2 for objects with r < 22.35 and i < 21.3 in the selection region. For the combined redshift distribution (red curve), we utilize only COSMOS photometric redshifts at z < 0.75, but average COSMOS and DEEP2 redshift distributions above z = 0.75 (where the DEEP2 z>0.7 sample selection has minimal impact) to mitigate the effects of cosmic variance. We find that simple techniques will select 400 objects per square degree, almost all red galaxies at z > 0.6. University of Pittsburgh Selecting z>0.6 LRGs with PTF + WISE SDSS ri imaging is not available for ~25% of the planned BigBOSS footprint (WISE data covers the full sky). As can be seen Figure 2, a sample could still be constructed without i-band imaging by taking all objects above some threshold in r − [3.4μm] color. We have tested this using PTF R-band data in the COSMOS field; PTF imaging will cover the entire BigBOSS area. A simple cut in R − [3.4μm] yields a sample which is 5% stars, and 85% galaxies with z > 0.6. However, 66% of the galaxies targeted are red in rest-frame color, as opposed to 96% with i. As seen in Figure 4, the interloper galaxies are overwhelmingly bright, somewhat less red, massive galaxies at z > 0.7; they are the largest, most strongly-biased emission line galaxies (ELGs). We would likely wish to target these objects for the BigBOSS ELG sample. In total, a PTF + WISE selection would include 1000 objects per square degree down to R = 22.5. Even in regions without SDSS data, PTF R (together with the existing WISE imaging) would be sufficient for us to select LRGs. SDSS gri WISE 3.4µm 1' (60") Figure 1: An example of a z=1.0 Luminous Red Galaxy as seen in SDSS and WISE imaging. This object is at the SDSS 5σ limit in r & i, but is easily detected by the 40cm WISE telescope; in fact, it is the brightest object in this 1' x 1' region. This demonstrates just how much brighter LRGs are at 1.6μm (restframe) than in restframe-blue light. Figure 4: Redshift distributions for LRG samples selected using PTF R + WISE. The histograms are based on photometric redshifts from COSMOS for objects with R < 22.5 and R -[3.4μm] > 3.5. The red-dashed curve shows the histogram only for objects which are intrinsically red in color (restframe U-B > 1, an effective red-sequence cut). This single-color cut will select 1000 objects per square degree, with the vast majority being either red galaxies or massive emission-line galaxies at z > 0.6. z>0.6 LRGs Stars Blue galaxies x r-i r-[3.4] Acknowledgements: This work was supported by a US Department of Energy Early Career grant. We wish to thank the PTF and WISE teams for early access to proprietary data. References: Eisenstein, D. J., et al. 2001, Astron. J., 122, 2267 John, T. L. 1988, A&A, 193, 189 Padmanabhan, N., et al. 2007, Mon. Not. R. Astron. Soc., 378, 852 Wright, Eisenhardt & Fazio 1994, BAAS, 184.2503 BigBOSS