Methodological refinement of the submillimeter galaxy magnification bias. I. Cosmological analysis with a single redshift bin

Abstract

The main goal of this work is to test the results of a methodological improvement in the measurement of the magnification bias signal on a sample of submillimeter galaxies. In particular, we investigate the constraining power of cosmological parameters within the Lambda CDM model. We also discuss important points that can affect the results. We measured the angular cross-correlation function between a sample of foreground GAMA II galaxies in a single wide spectroscopic redshift bin of $0.2<z<0.8$ and a sample of background submillimeter galaxies from Herschel-ATLAS. We focused on the photometric redshift range of $1.2<z<4.0,$ with an improved methodological framework. Interpreting the weak lensing signal within the halo model formalism and performing a Markov chain Monte Carlo (MCMC) algorithm, we obtained the posterior distribution of both the halo occupation distribution and cosmological parameters within a flat Lambda CDM model. Our analysis was also performed with additional galaxy clustering information via a foreground angular auto-correlation function. We observed an overall remarkable improvement in terms of uncertainties in both the halo occupation distribution and cosmological parameters with respect to previous results. A priori knowledge about beta , the logarithmic slope of the background integral number counts, is found to be paramount to derive constraints on $ when using the cross-correlation data alone. Assuming a physically motivated prior distribution for beta , we obtain mean values of $ $ and $ $ and an unconstrained distribution for the Hubble constant. These results are likely to suffer from sampling variance, since one of the fields, G15, appears to have an anomalous behavior with a systematically higher cross-correlation. We find that removing it from the sample yields mean values of $ $ and $ $ and, for the first time, a (loose) restriction of the Hubble constant is obtained via this observable: $h=0.79^ $. The addition of the angular auto-correlation of the foreground sample in a joint analysis tightens the constraints, but also reveals a discrepancy between both observables that might be an aggravated consequence of sampling variance or due to the presence of unmodeled aspects on small and intermediate scales.