PREDICTING LANDSCAPE-SCALE EROSION AFTER LARGE WILDFIRES

In recent years, the size and severity of wildfires have both increased. One of the major impacts of wildfires is an increased risk of severe erosion and related downstream impacts. Erosion risk can be reduced by treatments such as mulching. These treatments, however, are both expensive and best limited to sensitive areas at greatest risk to erosion. Predictions of landscape erosion risk following wildfire can be a useful aid in targeting treatments to areas where the combination of fire severity, topography, and potential weather may result in severe erosion. We will demonstrate how we used the Water Erosion Prediction Project (WEPP) Geospatial interface and an online database developed with support from NASA to predict erosion distribution following the 2015 Butte Fire in Calaveras County, California. We will compare those predictions with the predicted erosion for a simulated wildfire from a 2014 study of the nearby Mokelumne Basin, which intersects the Northeastern quarter of the Butte Fire. Within the basin predicted erosion rates for the Butte fire ranged from 0-166 Mg/ha/yr; compared to 0-252 Mg/ha/yr for the same area derived for the simulated fire. We found that the simulated burn severity (low 50%, moderate 29% or high 21%) was the same as the satellite-derived map of burn severity (low 57%, moderate 33% or high 11%) for 39 percent of the pixels, and the predicted number of high severity pixels was twice as great as the number of observed high severity pixels. Burn severity distribution for the entire Butte fire was: low 37%, moderate 41% or high 22%. We will discuss modeling challenges such as defining an appropriate climate to drive the erosion model, linking burn severity determined from a satellite image to soil properties, sources of data to support post wildfire erosion prediction, and evaluating our ability to predict fire severity with a fire spread model in order to support fuel treatment planning.