On the "Lensing is Low" of BOSS Galaxies

Recently, Leauthaud et al discovered that the small-scale lensing signal of Baryon Oscillation Spectroscopic Survey (BOSS) galaxies is up to 40% lower than predicted by the standard models of the galaxy-halo connections that reproduced the observed galaxy stellar mass function (SMF) and clustering. We revisit such "lensing is low" discrepancy by performing a comprehensive Halo Occupation Distribution (HOD) modelling of the SMF, clustering, and lensing of BOSS LOWZ and CMASS samples at Planck cosmology. We allow the selection function of satellite galaxies to vary as a function of stellar mass as well as halo mass. For centrals we assume their selection to depend only on stellar mass, as informed by the directly measured detection fraction of the redMaPPer central galaxies. The best-fitting HOD successfully describes all three observables without over-predicting the small-scale lensing signal. This indicates that the model places BOSS galaxies into dark matter halos of the correct halo masses, thereby eliminating the discrepancy in the one-halo regime where the signal-to-noise of lensing is the highest. Despite the large uncertainties, the observed lensing amplitude above 1 Mpc/h remains inconsistent with the prediction, which is however firmly anchored by the large-scale galaxy bias measured by clustering at Planck cosmology. Therefore, we demonstrate that the "lensing is low" discrepancy on scales below 1 Mpc/h can be fully resolved by accounting for the halo mass dependence of the selection function. Lensing measurements with improved accuracy is required on large scales to distinguish between deviations from Planck and non-linear effects from galaxy-halo connections.