0 Assigned ID: 2936 ----- Attempting Submission 1 (Thu Mar 24 02:31:32 GMT 2022) ----- HST Phase I Proposal 2936 successfully submitted. Receipt: # 2936-1 ----- Attempting Submission 2 (Fri Mar 25 23:42:03 GMT 2022) ----- Elemental abundance variations in globular clusters are universal but not well understood. HST photometry has been instrumental in detecting these populations, but does not directly provide the underlying elemental abundances, which are the physical property that can be compared to theoretical models; those measurements come from spectroscopic data for which it is very difficult to obtain large samples. We propose to combine literature spectroscopic measurements of O, Na, and Fe abundances of red giants with archival HST photometry of the same stars spanning 57 globular clusters. We will use this combined data set to train a machine learning regression model to predict the abundances for the vast majority of red giants in the HST data without spectroscopic measurements, and use them to build full elemental distribution functions and chemospatial maps of the clusters. J. Bailin constructs models on the formation of globular clusters, especially the effects of supernova self enrichment on the iron abundance distribution. Bailin has analyzed spectroscopic abundances of stars in globular clusters, and has extensive experience with resolved stellar population analysis of HST imaging data as part of the GHOSTS project. He is part of the Machine Learning For Science (ML4SCI; https://ml4sci.org/) collaboration that runs annual machine learning projects as part of the Google Summer of Code, and annual machine learning hackathons. D. Chambers is an undergraduate with a background in astronomy, computing, and aerospace engineering, who performed the proof-of-concept machine learning fits to the NGC 2808 spectroscopic and photometric data in the proposal. Stellar Populations and the Interstellar Medium 0 Regular false false true false false false SUPPORTED