Automated Wildfire Hazard Modeling Using Google Earth Engine

Abstract

Timely wildfire hazard assessments are critical for disaster preparedness and response. Current approaches to susceptibility modeling are tedious, labor and data-intensive, and often reliant on proprietary software. These limitations can lead to inconsistent workflows, reduced reproducibility, and delays in delivering actionable information for mitigation strategies. This project delivers an automated, open-source workflow designed to estimate wildfire hazard for any study area within the United States. Using Google Earth Engine, the model integrates meteorological, vegetation, and terrain variables obtained from publicly accessible geospatial data. Users can modify the area of interest, define start and end dates for the analysis period, and adjust coefficient weights for model variables to better reflect environmental and temporal dynamics surrounding wildfire seasons. The workflow is tested using the 2024 Line Fire in the San Bernardino National Forest. The primary result of the project is a comprehensive script that generates a Wildfire Hazard Index raster. Built-in helper functions provide flexibility and adaptation for different regions and timeframes, supporting resource managers and emergency planners. By automating laborious tasks of data acquisition and pre-processing, the workflow allows producing consistent and repeatable hazard assessments that can support better decision-making.