Code for "Natural variability can mask forced permafrost response to stratospheric aerosol injection in the ARISE-SAI-1.5 simulations"

We assess how natural variability may mask the forced response of permafrost to a hypothetical stratospheric aerosol injection (SAI) scenario. We train two logistic regression models to predict whether maps of active layer depth and permafrost soil temperature came from the SAI or SSP2-4.5 (control) simulations, and find that it may take 10-30 years of SAI to accurately detect its influence on permafrost.