An interpretable remote sensing–machine learning framework for wildfire soil burn severity prediction in the northwestern United States
Authors:
- Subhankar DasSDSubhankar Das
- Mariana DobreMDMariana Dobre
- Roger LewRLRoger Lew
- Sarah A. LewisSLSarah A. Lewis
- Erin S. BrooksEBErin S. Brooks
- Peter R. RobichaudPRPeter R. Robichaud
- Anurag SrivastavaASAnurag Srivastava
- Mary Ellen MillerMMMary Ellen Miller
- Alex M. WatanabeAWAlex M. Watanabe
- Marta BassoMBMarta Basso
Open Access
- 4 August 2026
- journal article
- Published by CSIRO Publishing in International Journal of Wildland Fire
- Vol. 35 (8)
Abstract
Background: Although soil burn severity (SBS) is traditionally mapped post-fire, pre-fire predictive approaches remain limited despite their potential to identify areas at risk of soil degradation, runoff and erosion. Aim: To develop a large-scale model for predicting post-fire SBS from pre-fire conditions at ~30 m spatial resolution. Methods: A harmonized dataset of 62 wildfires (2018–2024) was used to build machine learning (ML) frameworks based on Random Forest, eXtreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM), using pre-fire vegetation, climate, soil–terrain and anthropogenic predictors. Performance was evaluated using five-fold cross-validation (CV), leave-one-fire-out (LOFO) validation and independent testing on 2025 fires. Key Results: All models showed similar CV overall accuracy (OA ~0.58), but LOFO performance varied. LightGBM maintained more consistent OA (0.51–0.58) across fire size classes, whereas XGBoost and Random Forest exhibited comparable performance. Key predictors included vegetation biomass (Normalized Difference Moisture Index, Enhanced Vegetation Index), fuel stress (Evaporative Stress Index, Evapotranspiration), elevation and soil moisture. High SBS was associated with dry conditions and high biomass availability, with region-specific controls modulating these patterns. Conclusion: ML models using pre-fire environmental data enable proactive SBS prediction. Implications: This framework supports fuel management planning, potential post-fire hydrological impact assessment and the prediction of SBS from pre-fire conditions.Keywords
Funding Information
- National Aeronautics and Space Administration (80NSSC22K0920P00006)
This publication has 100 references indexed in Scilit:
- Application of the BNDVI index as proxy for burn severity assessmentEarth Science Informatics, 2025
- Improved fire severity prediction using pre-fire remote sensing and meteorological time series: Application to the French Mediterranean areaAgricultural and Forest Meteorology, 2025
- Learning from wildfires: A scalable framework to evaluate treatment effects on burn severityEcosphere, 2024
- Moderating effects of past wildfire on reburn severity depend on climate and initial severity in Western US forestsEcological Applications, 2024
- Digital soil mapping of soil burn severitySoil Science Society of America Journal, 2024
- Designing tools to predict and mitigate impacts on water quality following the Australian 2019/2020 wildfires: Insights from Sydney's largest water supply catchmentIntegrated Environmental Assessment and Management, 2021
- Multi-scaled drivers of severity patterns vary across land ownerships for the 2013 Rim Fire, CaliforniaLandscape Ecology, 2019
- Large wildfire trends in the western United States, 1984–2011Geophysical Research Letters, 2014
- Current research issues related to post-wildfire runoff and erosion processesEarth-Science Reviews, 2013
- Development of gridded surface meteorological data for ecological applications and modellingInternational Journal of Climatology, 2011