Sparsely Sampling the Sky: A Bayesian Experimental Design Approach

TL;DR

This paper explores how sparse sky sampling, guided by Bayesian Experimental Design, can optimize galaxy surveys by balancing survey area and observational time to improve cosmological parameter constraints.

Contribution

It introduces a Bayesian framework to evaluate sparse sampling strategies, demonstrating potential efficiency gains for large-scale galaxy surveys.

Findings

Sparse sampling increases parameter errors by at most 0.45% when observing the same area.

Sampling a larger area with the same time reduces parameter errors by 28%.

Bayesian design effectively guides optimal survey strategies.

Abstract

The next generation of galaxy surveys will observe millions of galaxies over large volumes of the universe. These surveys are expensive both in time and cost, raising questions regarding the optimal investment of this time and money. In this work we investigate criteria for selecting amongst observing strategies for constraining the galaxy power spectrum and a set of cosmological parameters. Depending on the parameters of interest, it may be more efficient to observe a larger, but sparsely sampled, area of sky instead of a smaller contiguous area. In this work, by making use of the principles of Bayesian Experimental Design, we will investigate the advantages and disadvantages of the sparse sampling of the sky and discuss the circumstances in which a sparse survey is indeed the most efficient strategy. For the Dark Energy Survey (DES), we find that by sparsely observing the same area in a smaller amount of time, we only increase the errors on the parameters by a maximum of 0.45%. Conversely, investing the same amount of time as the original DES to observe a sparser but larger area of sky we can in fact constrain the parameters with errors reduced by 28%.