Diagnosing space telescope misalignment and jitter using stellar images

Abstract

Accurate knowledge of the telescope's point spread function (PSF) is essential for the weak gravitational lensing measurements that hold great promise for cosmological constraints. For space telescopes, the PSF may vary with time due to thermal drifts in the telescope structure, and/or due to jitter in the spacecraft pointing (ground-based telescopes have additional sources of variation). We describe and simulate a procedure for using the images of the stars in each exposure to determine the misalignment and jitter parameters, and reconstruct the PSF at any point in that exposure's field of view. The simulation uses the design of the SNAP (http://snap.lbl.gov) telescope. Stellar-image data in a typical exposure determines secondary-mirror positions as precisely as $20 {\rm nm}$. The PSF ellipticities and size, which are the quantities of interest for weak lensing are determined to $4.0 \times 10^{-4}$ and $2.2 \times 10^{-4}$ accuracies respectively in each exposure, sufficient to meet weak-lensing requirements. We show that, for the case of a space telescope, the PSF estimation errors scale inversely with the square root of the total number of photons collected from all the usable stars in the exposure.