Orca: The Overdense Red-sequence Cluster Algorithm

Abstract

We present a new cluster detection algorithm designed for the Panoramic Survey Telescope and Rapid Response System (Pan-STARRS) survey but with generic application to any multiband data. The method makes no prior assumptions about the properties of clusters other than (a) the similarity in colour of cluster galaxies (the "red sequence") and (b) an enhanced projected surface density. The detector has three main steps: (i) it identifies cluster members by photometrically filtering the input catalogue to isolate galaxies in colour-magnitude space, (ii) a Voronoi diagram identifies regions of high surface density, (iii) galaxies are grouped into clusters with a Friends-of-Friends technique. Where multiple colours are available, we require systems to exhibit sequences in two colours. In this paper we present the algorithm and demonstrate it on two datasets. The first is a 7 square degree sample of the deep Sloan Digital Sky Survey equatorial stripe (Stripe 82), from which we detect 97 clusters with z<=0.6. Benefiting from deeper data, we are 100% complete in the maxBCG optically-selected cluster catalogue (based on shallower single epoch SDSS data) and find an additional 78 previously unidentified clusters. The second dataset is a mock Medium Deep Survey (MDS) Pan-STARRS catalogue, based on the Lambda-CDM model and a semi-analytic galaxy formation recipe. Knowledge of galaxy-halo memberships in the mock allows a quantification of algorithm performance. We detect 305 mock clusters in haloes with mass >10^13 solar masses at z<=0.6 and determine a spurious detection rate of <1%, consistent with tests on the Stripe 82 catalogue. The detector performs well in the recovery of model Lambda-CDM clusters. (abridged)