Residual Smoothing: Using Mocks to Correct Model Covariance Matrices

TL;DR

This paper presents a data-driven residual smoothing method that enhances model covariance matrices by incorporating missing features identified from mocks, improving accuracy at non-Gaussian scales in cosmological surveys.

Contribution

It introduces a novel residual smoothing technique that corrects model covariance matrices using mock data, extending their validity to smaller, non-Gaussian scales.

Findings

Improved covariance matrices for non-Gaussian scales (8-40 Mpc/h).

Method maintains model precision while enhancing accuracy.

Applicable to BOSS-like survey data.

Abstract

Abstract Covariance matrix estimation is a challenging problem in cosmology. Recent work has shown that model covariance matrices can be precise, and that at relatively large scales they can also be accurate. We introduce a data-driven method that can identify features from a mock covariance matrix that are missing from a corresponding model, then incorporate them into the model without significantly degrading the model's precision. We apply this method to a BOSS-like survey and extend a model covariance to be valid at scales relevant for measurements of Redshift Space Distortions (8-40 Mpc/h), where the galaxy field is significantly non-Gaussian.