Characteristic functions for cosmological cross-correlations

Abstract

ABSTRACT We introduce a novel unbiased, cross-correlation estimator for the one-point statistics of cosmological random fields. One-point statistics are a useful tool for analysis of highly non-Gaussian density fields, while cross-correlations provide a powerful method for combining information from pairs of fields and separating them from noise and systematics. We derive a new Deconvolved Distribution Estimator that combines the useful properties of these two methods into one statistic. Using two example models of a toy Gaussian random field and a line intensity mapping survey, we demonstrate these properties quantitatively and show that the deconvolved distribution estimator can be used for inference. This new estimator can be applied to any pair of overlapping, non-Gaussian cosmological observations, including large-scale structure, the Sunyaev–Zeldovich effect, weak lensing, and many others.