The SPHEREx View of Galaxy Clusters: A Simulation-based Validation of the Forced Photometry Pipeline for Extended Sources

Abstract

We present a simulation-driven assessment of the performance of the SPHEREx pipeline for galaxy cluster science, focusing on photometry, source blending, survey depth, and photometric redshift accuracy. To do that, we compile a sample of eight galaxy clusters spanning a wide redshift range ($z \approx 0.02$-$1.1$) and develop an end-to-end pipeline. We use the ancillary data from the DESI Legacy Survey and COSMOS survey, and generate realistic mock SPHEREx observations with the SPHEREx Sky Simulator. By performing forced photometry on these images with The Tractor, we quantify the characteristic biases and uncertainties relevant to cluster science. We find that the photometry is generally unbiased, but source blending is the primary driver of catastrophic outliers, particularly when the combined flux of neighbors is comparable to the flux of targets. Measuring the effective survey depth, we find that SPHEREx detects members down to $K_{s}\approx 20$ AB ($5\sigma$), 7-9 mag fainter than the brightest cluster galaxy (BCG) in nearby clusters but only 1-2 mag for clusters at $z \sim 1$, where the BCG itself has faded close to this depth. Despite these challenges, we demonstrate that SPHEREx can achieve a photometric redshift precision of $\sigma_{\mathrm{NMAD}}\approx 0.003$-$0.01$ for cluster galaxies with an appropriate sample selection based on brightness or signal-to-noise. Combining the redshifts of quality-selected members, we recover cluster redshifts with a bias of $|\Delta z|/(1+z) < 0.002$ and a scatter of $\sigma \approx 0.002$ at $z \lesssim 0.5$, meeting the precision required for cluster cosmology.