redMaPPer II: X-ray and SZ Performance Benchmarks for the SDSS Catalog

Abstract

We evaluate the performance of the SDSS DR8 redMaPPer photometric cluster catalog by comparing it to overlapping X-ray and SZ selected catalogs from the literature. We confirm the redMaPPer photometric redshifts are nearly unbiased (<Δz> < 0.005), have low scatter (σ_z ~ 0.006-0.02, depending on redshift), and have a low catastrophic failure rate (~ 1%). Both the T_X-λ and Mgas-λ scaling relations are consistent with a mass scatter of σ_{\ln M|λ} ~ 25%, albeit with a ~ 1% outlier rate due to projection effects. This failure rate is somewhat lower than that expected for the full cluster sample, but is consistent with the additional selection effects introduced by our reliance on X-ray and SZ selected reference cluster samples. Where the redMaPPer DR8 catalog is volume limited (z < 0.35), the catalog is 100% complete above T_X > 3.5 keV, and L_X > 2\times 10^{44} erg/s, decreasing to 90% completeness at L_X ~ 10^{43} erg/s. All rich (λ> 100), low redshift (z < 0.25) redMaPPer clusters are X-ray detected in the ROSAT All Sky Survey (RASS), and 86% of the clusters are correctly centered. Compared to other SDSS photometric cluster catalogs, redMaPPer has the highest completeness and purity, and the best photometric redshift performance, though some algorithms do achieve comparable performance to redMaPPer in subsets of the above categories and/or in limited redshift ranges. The redMaPPer richness is clearly the one that best correlates with X-ray temperature and gas mass. Most algorithms (including redMaPPer) have very similar centering performance, with only one exception which performs worse.