ABSTRACT
Following on our purpose of developing a unified pipeline for large-scale structure data analysis with angular power spectra, we now include the weak lensing effect of magnification bias on galaxy clustering in a publicly available, modular parameter estimation code. We thus forecast constraints on the parameters of the concordance cosmological model, dark energy, and modified gravity theories from galaxy clustering tomographic angular power spectra. We find that a correct modelling of magnification is crucial not to bias the parameter estimation, especially in the case of deep galaxy surveys. Our case study adopts specifications of the Evolutionary Map of the Universe, which is a full-sky, deep radio-continuum survey, expected to probe the Universe up to redshift z ∼ 6. We assume the Limber approximation, and include magnification bias on top of density fluctuations and redshift-space distortions. By restricting our analysis to the regime where the Limber approximation holds true, we significantly minimize the computational time needed, compared to that of the exact calculation. We also show that there is a trend for more biased parameter estimates from neglecting magnification when the redshift bins are very wide. We conclude that this result implies a strong dependence on the lensing contribution, which is an integrated effect and becomes dominant when wide redshift bins are considered. Finally, we note that instead of being considered a contaminant, magnification bias encodes important cosmological information, and its inclusion leads to an alleviation of its degeneracy between the galaxy bias and the amplitude normalization of the matter fluctuations.
Cosmological parameter estimation from large-scale structure deep learning
