Decomposing Regional Nonstationarity in U.S. Extreme Precipitation with a Bayesian Regime-Mixture Framework

Data and Code Availability This repository contains the supplementary dataset and Stan model code used for the Bayesian two-component Gumbel mixture analysis of annual-maximum daily precipitation (AMDP) across U.S. regions. The dataset provides station information, posterior summaries and diagnostics, annual parameter and 100-year return-level trajectories, counterfactual decomposition results, external diagnostic summaries, sensitivity analyses, and numerical data used to generate study figures. Raw precipitation, tropical-cyclone track, and ERA5 reanalysis data are not redistributed and should be obtained from their original data providers.