Wildfire burned 350,000 acres of forested mountain lands in Western Montana during 2000. The following summer, short duration - high intensity convection storms triggered extreme gully erosion in a large number of steep ephemeral drainages that had been severely burned. Through field study and air photo interpretation, this study investigated 200 first and second order drainages within two 5000-acre basins that were both burned and hit by the intense rains. I introduce the Burn Severity Distribution Index (BSDI), a weighted average of burn severity, to compare fire impact between drainages. BSDI were generated in a GIS from each of two sources, as mapped by Burn Area Emergency Rehabilitation (BAER) teams and as derived by another researcher who applied Composite Burn Index (CBI) and Normalized Burn Ratio (NBR) techniques to remote sensed imagery. All drainages were compared by a series of geomorphometrics and landscape characteristics to determine the dominant features associated with severe gully erosion following wildfire.

Gully erosion with deep incision occurred almost exclusively in the drainages with high BSDI values, drainages that were severely burned over the vast majority of the upper extents of the drainage area. Channel slope and spatial extent of severe burn are the two most important factors controlling gully rejuvenation following wildfire. More erodible soils lead to channel incision at lower slope angles. Slope and contributing area above the upslope terminus of the eroded gully are inversely related. Roads close drainage divides attenuated flows, forcing gully erosion to terminate at the road and re-initiate below the road crossing. Where roads crossed eroding gullies lower on the slope, the road prism was fully removed, adding material to flow volumes. Large woody material functioned as geomorphic agents by focusing flow and initiating rills, creating nick-points within gully channels, and forming jams that attenuated or altogether arrested debris-laden flows. The burn severity map derived from remote sensed imagery was generally more precise than the BAER map but displayed inaccurate classifications as verified by field observations. Nonetheless, refinement of CBI/NBR techniques will provide land managers with an important tool to assess the risk of extreme erosion following wildfire.