2026 Project Description

Making Precision Cosmology Possible: Analytic Covariance for Galaxy Clusters

Supervisors

Tomomi Sunayama
Find out more about supervisors on ASIAA website

Task Description and Goals

Modern cluster cosmology aims to extract percent-level constraints on cosmology from cluster abundance, clustering, and weak lensing. At this level of precision, the limiting factor is often not the signal itself, but our understanding of the covariance matrix—the object that encodes statistical uncertainties and correlations between observables.

In principle, covariances can be estimated from large ensembles of numerical simulations. In practice, this approach is prohibitively expensive: accurate covariance estimation requires thousands of realizations, far exceeding what is feasible for high-resolution simulations and complex cluster observables.

This project tackles this fundamental bottleneck by developing analytic covariance models for cluster cosmology. Using theoretical tools from large-scale structure and the halo model, the student will build covariance matrices for key cluster observables and test their accuracy against simulations.

During the project, the student will:

• Learn how covariance enters likelihood-based cosmological inference
• Derive analytic covariance contributions from shot noise, sample variance, and super-sample covariance
• Apply these models to cluster observables such as cluster abundance, cluster clustering, and cluster–galaxy lensing
• Validate analytic predictions using existing simulations and mock catalogs
• Explore how covariance modeling impacts cosmological parameter constraints

This project sits at the interface of theory, statistics, and data analysis, and directly enables cosmological analyses of current and next-generation surveys such as HSC, LSST, Euclid, and Roman.

Required Background

Strong programming skills, especially in Python and C/C++