Shear Test Experiments
The goal of this S15 project is to build a framework for testing shear measurement algorithms. Primarily, we want to characterize the error and bias bias in the measurement of galaxies with known profiles, as well as how that changes depending on measurement conditions.
GalSim Galaxy Simulations:
To do this characterization, we use the GalSim program to create a sample of galaxy images with constant shear and a different Psf for each galaxy. For each run, the Psfs are sampled from a library of Psf images. The great3sims package had most of the capability we need to create the input yaml files for GalSim, so we adapted this set of scripts. As it was implemented for Great3, great3sims can create a variety of simulations with known galaxy profiles, atmospheric conditions, and generated Psf using a variety of Psf models. However, it does not allow Psfs to be introduced in a parameter free form, such as a kernel image. Our primary modification to great3sims scripts was to create a new PsfBuilder class which allows the Psf for each galaxy to be introduced as a .fits file.
We have chosen to do our first set of tests with the “control/ground/constant” branch, which creates a galaxy set with applied Psf and Shear. The shear is constant for each subfield, and is sampled from the range [1E-2, 5E-2] with a range of rotation angle from [0, 2π].
PhoSim Psf Simulations:
The stellar images used to create the Psf .fits files is created by PhoSim, using different combinations of
seeing (.5, .7, and .9 arcsec raw seeing values)
position on the focal plane (randomly distributed over 189 sensors)
filter (R or I)