bernatj/ai-based-attribution-cyclones: Code release accompanying the manuscript "Forecast-based Attribution of Extratropical Cyclones Using AI Weather Models."

Code release accompanying the manuscript "Forecast-based Attribution of Extratropical Cyclones Using AI Weather Models." This release provides the scripts and notebooks used to generate the factual and pseudo-global-warming (PGW) simulation data and to perform the subsequent forecast-skill and attribution analyses for the four AI weather prediction models used in the study: Pangu-Weather, FourCastNet-v2, Aurora, and AIFS. Included in this release: scripts to download and prepare initial conditions for the AIWP models; scripts to apply CMIP6-derived PGW thermodynamic perturbations; forecast runners for factual and counterfactual/PGW simulations; optional scripts to download and convert ECMWF IFS Open Data forecasts for benchmarking; notebooks for multimodel forecast-skill and attribution diagnostics; an AIFS precipitation-attribution notebook; a minimal CMIP6 multi-model-mean delta bundle required by the supported PGW workflow; starter conda environment files for each model and for the analysis notebooks. This release corresponds to the version of the code used for the revised manuscript submission.