Spurious Shear in Weak Lensing with LSST

TL;DR

This paper quantitatively estimates the systematic errors in weak lensing shear measurements for LSST, focusing on atmospheric PSF effects, and finds that these errors are manageable with LSST's large dataset.

Contribution

First to analyze and quantify the absolute level of additive systematic errors in LSST weak lensing shear measurements using high fidelity simulations.

Findings

Atmospheric PSF variation causes spurious shear correlations at 10^{-4}–10^{-3} level.

Large LSST dataset averages out stochastic errors, reducing systematic impact.

Systematic errors are unlikely to severely limit LSST cosmological constraints.

Abstract

The complete 10-year survey from the Large Synoptic Survey Telescope (LSST) will image $\sim$ 20,000 square degrees of sky in six filter bands every few nights, bringing the final survey depth to $r\sim27.5$, with over 4 billion well measured galaxies. To take full advantage of this unprecedented statistical power, the systematic errors associated with weak lensing measurements need to be controlled to a level similar to the statistical errors. This work is the first attempt to quantitatively estimate the absolute level and statistical properties of the systematic errors on weak lensing shear measurements due to the most important physical effects in the LSST system via high fidelity ray-tracing simulations. We identify and isolate the different sources of algorithm-independent, \textit{additive} systematic errors on shear measurements for LSST and predict their impact on the final cosmic shear measurements using conventional weak lensing analysis techniques. We find that the main source of the errors comes from an inability to adequately characterise the atmospheric point spread function (PSF) due to its high frequency spatial variation on angular scales smaller than $\sim10'$ in the single short exposures, which propagates into a spurious shear correlation function at the $10^{-4}$--$10^{-3}$ level on these scales. With the large multi-epoch dataset that will be acquired by LSST, the stochastic errors average out, bringing the final spurious shear correlation function to a level very close to the statistical errors. Our results imply that the cosmological constraints from LSST will not be severely limited by these algorithm-independent, additive systematic effects.