Principal Investigator: Meredith Durbin
PI Institution: University of California - Berkeley
Investigators
(xml)
Title: Modeling Spatiotemporal Systematics in Multiwavelength Stellar Photometry Catalogs
Cycle: 29
Abstract
The value of precision in unlocking new science has been repeatedly demonstrated across multiple fields of astrophysics (e.g., CMB cosmology and Milky Way dynamics with Gaia). Stellar photometry with the Hubble Space Telescope (HST) has progressed to the point that a similar push towards precision would reap immediate scientific benefits. Detailed and thorough characterization of HST's optical performance has resulted in unparalleled PSF photometry, but residual systematic errors currently form a barrier in achievable precision. Time-dependent changes to the telescope focus and visible sky background conspire to limit photometric precision to the ~5% level. This barrier currently prevents time-critical science including high-resolution dust mapping in nearby galaxies, and red giant branch distance measurements which are urgently needed to shed independent light on the uncomfortable tension between Cepheid- and CMB-based estimates of the local Hubble parameter.
We propose to remove this barrier by deriving corrections at the level of cataloged HST photometry, using machine learning techniques. Specifically, we will leverage archival wide-area multiwavelength data, including the PHAT survey, to model photometric biases induced by variations in the HST PSF, sky background, and other hidden effects. This model will correct resolved stellar photometry as a function of detector position, exposure date, and other telescope parameters. This methodology will be broadly applicable to photometry and cross-calibration of all existing and future space-based stellar catalogs, including HST and Roman, and will allow a new generation of critical precision stellar population studies.