Abstract
Molecular abundances in protoplanetary disks are highly sensitive to thelocal physical conditions, including gas temperature, gas density, radiationfield, and dust properties. Often multiple factors are intertwined, impactingthe abundances of both simple and complex species. We present a new approach tounderstanding these chemical and physical interdependencies using machinelearning. Specifically we explore the case of CO modeled under the conditionsof a generic disk and build an explanatory regression model to study thedependence of CO spatial density on the gas density, gas temperature, cosmicray ionization rate, X-ray ionization rate, and UV flux. Our findings indicatethat combinations of parameters play a surprisingly powerful role in regulatingCO compared to any singular physical parameter. Moreover, in general, we findthe conditions in the disk are destructive toward CO. CO depletion is furtherenhanced in an increased cosmic ray environment and in disks with higherinitial C/O ratios. These dependencies uncovered by our new approach areconsistent with previous studies, which are more modeling intensive andcomputationally expensive. Our work thus shows that machine learning can be apowerful tool not only for creating efficient predictive models, but also forenabling a deeper understanding of complex chemical processes.