Abstract
In forested ecosystems, wildfires often effect changes in a watershed’s hydrologic response to precipitation. Elevated runoff causes stream discharge to spike during high intensity, short duration rainfall and snowmelt events. This discharge can damage or destroy road infrastructure such as culverts, and make road impassable, resulting in expensive repairs. Forest managers use hydrologic models to identify areas of highest risk to this damage and prescribe treatments to mitigate the risk. We measured peak flow at 12 road crossings following five large wildfires in 2017 in Western Montana and used the data to assess the peak flow estimates of four commonly used hydrologic models: Runoff Curve Number, USGS Regression Equations, WEPPcloud-PEP (Post-fire Erosion Prediction), WEPPcloud-Disturbed. Our data showed that annual peak flow in this region occurs primarily during the spring runoff, though small, isolated rainfall events are capable of producing large spikes in streamflow volume. The Runoff Curve Number method tended to greatly over-estimate peak flows generated by rainfall events we observed and is unable to model runoff events caused by snowmelt. USGS Regression equations underpredicted the changes to post-fire hydrology. When detailed soil information was used, WEPPcloud-Disturbed was able to reasonably predict runoff for both rainfall and snowmelt events which is critical in snowmelt dominated watersheds.