Transformationally decoupling clustering and tracer bias

TL;DR

This paper proposes a method using Gaussianizing transformations to enhance cosmological data analysis by increasing Fisher information and decoupling clustering from local tracer bias, improving insights from nonlinear scales.

Contribution

It introduces a strategy combining the 1-point PDF analysis with Gaussianized clustering statistics to improve cosmological parameter estimation and tracer bias modeling.

Findings

Increases Fisher information on nonlinear scales.

Decouples clustering statistics from local bias.

Enhances analysis of tracer sampling in cosmology.

Abstract

Gaussianizing transformations are used statistically in many non-cosmological fields, but in cosmology, we are only starting to apply them. Here I explain a strategy of analyzing the 1-point function (PDF) of a spatial field, together with the 'essential' clustering statistics of the Gaussianized field, which are invariant to a local transformation. In cosmology, if the tracer sampling is sufficient, this achieves two important goals. First, it can greatly multiply the Fisher information, which is negligible on nonlinear scales in the usual $\delta$ statistics. Second, it decouples clustering statistics from a local bias description for tracers such as galaxies.