Photometric Redshift Requirements for Self-Calibration of Cluster Dark Energy Studies

The ability to constrain dark energy from the evolution of galaxy cluster counts is limited by the imperfect knowledge of cluster redshifts. Ongoing and upcoming surveys will mostly rely on redshifts estimated from broad-band photometry (photo-z's). For a Gaussian distribution for the cluster photo-z errors and a high cluster yield cosmology defined by the WMAP 1 year results, the photo-z bias and scatter needs to be known better than 0.003 and 0.03, respectively, in order not to degrade dark energy constrains by more than 10% for a survey with specifications similar to the South Pole Telescope. Smaller surveys and cosmologies with lower cluster yields produce weaker photo-z requirements, though relative to worse baseline constraints. Comparable photo-z requirements are necessary in order to employ self-calibration techniques when solving for dark energy and observable-mass parameters simultaneously. On the other hand, self-calibration in combination with external mass inferences helps reduce photo-z requirements and provides important consistency checks for future cluster surveys. In our fiducial model, training sets with spectroscopic redshifts for ~5%-15% of the detected clusters are required in order to keep degradations in the dark energy equation of state lower than 20%.