Controls on Post-fire Debris Flow Grain Sizes Across the Intermountain West

Post-fire debris flows represent one of the most destructive and potentially hazardous consequences associated with increasing wildfire severity. While the abundance of literature has explored the initiation processes and generation of post-fire debris flows, investigations into their downstream impacts are limited. Murphy et al. (2019) developed a modeling framework to predict where debris flow sediment is generated post-wildfire, how much of that sediment is delivered to a stream channel, and how that pulse of sediment propagates downstream through the river network. We are applying that modeling framework to predict post-wildfire risks to water supply reservoirs and aquatic habitat. While most inputs to this modeling framework are available through open source datasets, a significant gap in understanding still exists regarding the grain size distributions (GSD) of post-fire debris flows and the factors influencing these GSDs. This presents a major obstacle in the development of watershed-scale wildfire risk assessment models, as grain size exerts a first-order control on the rates and modes of sediment transport through a river network. Therefore, we have compiled GSDs from previous wildfire studies and conducted new fieldwork measuring GSDs in post-wildfire debris flow deposits to fill this critical knowledge gap. Adding to the 25 GSDs from previous studies spanning the Intermountain West (Idaho, Montana, Utah, Arizona), we measured GSDs from an additional 27 debris flows that occurred in 9 different wildfires across Utah. Altogether this represents the largest and most spatially extensive dataset of post-fire debris flow GSD of which we are aware. Catchments that produced these debris flows vary in upstream burn severity, area, slope, forest type, soils, climate, and geology. These metrics were all extracted as potential predictor variables for our statistical analysis and sourced from NED, UtahWRAP, NOAA, StreamCat, STATSGO, and MTBS databases. We will analyze these data using Random Forest statistical modeling and investigate which landscape metrics exert the most control on post-fire debris flow GSD. We aim to generalize the results of our GSD model and the complete dataset will ultimately be made available in an open-access database.