Modeling Soil Burn Severity Prediction for Planning Measures to Mitigate Post Wildfire Soil Erosion in NW Spain

Abstract Post-fire soil erosion is of major concern because of the effects it has on both soil and water resources. Rapid assessment of post-fire soil burn severity is essential for planning mitigation actions aimed at reducing post-fire flood and erosion risk. Soil burn severity is spatially heterogeneous and mainly depends on the type of vegetation, topography, and weather attributes. It is a descriptor of the magnitude of the fire-induced changes and is the result of the interactions between the above factors and soil characteristics, including soil moisture content. In this study, we used an approach combining pre- and post-fire digitalized georeferenced information to explore the capacity of an ensemble learning method to predict the spatial pattern of soil burn severity. The mean accuracy of both approaches was approximately 75%. The importance of the vegetation variables suggests that pre-fire vegetation structure may play an important role in determining the spatial distribution of soil burn severity.