The richness-to-mass relation of CAMIRA galaxy clusters from weak-lensing magnification in the Subaru Hyper Suprime-Cam survey

We present a statistical weak-lensing magnification analysis on an optically selected sample of 3029 CAMIRA (Cluster finding Algorithm based on Multiband Identification of Red-sequence gAlaxies) galaxy clusters with richness N > 15 at redshift 0.2 ≤ z < 1.1 in the Subaru Hyper Suprime-Cam survey. We use two distinct populations of colour-selected, flux-limited background galaxies, namely the low-z and high-z samples at mean redshifts of ≈1.1 and ≈1.4, respectively, from which to measure the weak-lensing magnification signal by accounting for cluster contamination as well as masking effects. Our magnification bias measurements are found to be uncontaminated according to validation tests against the ‘null-test’ samples for which the net magnification bias is expected to vanish. The magnification bias for the full CAMIRA sample is detected at a significance level of 9.51σ, which is dominated by the high-z background. We forward-model the observed magnification data to constrain the normalization of the richness-to-mass (N–M) relation for the CAMIRA sample with informative priors on other parameters. The resulting scaling relation is N∝ (M 500 ) ^{0.92 ± 0.13} (1 + z) ^{−0.48 ± 0.69} , with a characteristic richness of N = 17.72 ± 2.60 and intrinsic lognormal scatter of 0.15 ± 0.07 at M 500 = 10 ¹⁴  h ⁻¹  M ⊙ . With the derived N–M relation, we provide magnification-calibrated mass estimates of individual CAMIRA clusters, with the typical uncertainty of ≈39 and ≈32  per cent at richness of ≈20 and ≈40, respectively. We further compare our magnification-inferred N–M relation with those from the shear-based results in the literature, finding good agreement.