Avalanche susceptibility map using GIS and Google Earth Engine

This preprint presents a simplified and fully open-source workflow for regional-scale avalanche susceptibility mapping, implemented using Google Earth Engine and a Random Forest classification approach. The study replicates and validates an existing Probable Release Areas (PRA)–based methodology by adapting it to freely available datasets and cloud-based tools, with the aim of improving accessibility, reproducibility, and ease of application. The proposed framework was tested in the Aosta Valley (NW Italy) using a regional avalanche inventory and topographic and snow-related predictors derived from open satellite and DEM data. Rather than replacing operational avalanche forecasting or detailed hazard zoning, this work is intended as a low-cost, easily transferable approach for preliminary susceptibility assessment, educational purposes, and exploratory analyses in data-limited regions. The methodology and results are provided as a reproducible research example, highlighting both the potential and limitations of simplified machine-learning–based susceptibility mapping workflows.