PREDICTING GULLY HEAD LOCATION FOLLOWING POST-FIRE GULLY REJUVENATION

The spatial arrangement and hydrologic connectivity of wildfire burn mosaics are expected to influence post-fire erosion and gully rejuvenation (GR). However, analysis of empirical data to support this hypothesis is lacking. Intact vegetation inhibits erosive overland flow and triggering of gully rejuvenation by attenuating rainfall impact, temporarily storing water, and resisting runoff. Wildfire creates patches of varying degrees of vegetation disturbance, described as fire severity, where the magnitude of loss is expected to control the potential for erosive flow and the threshold for channel incision and the formation of a fresh gully head. Terrain steepness controls the potential energy that drives downslope flow while terrain curvature controls flow convergence. Steepness and curvature combined control flow accumulation. This paper presents an analytical method that uses field measurements and geospatial modelling to predict the location of gully heads generated by post-fire erosion relative to terrain steepness and curvature and the structural connectivity of vegetation disturbed by fire. Locations of 38 gully heads were mapped in the Sapphire Mountains of Western Montana, USA following fire in 2000 and GR in 2001. Steepness and curvature metrics were derived from 1m resolution airborne LiDAR. Fire severity was derived from full-scale burned area reflectance classification (BARC256) imagery. The analysis was conducted using a materials accumulation algorithm in SAGA GIS treating fire severity as the material of interest. In demonstration trials, fire severity accumulated non-linearly along flowpaths with a surge of accumulation in the primary catchment channel proximate to the mapped gully heads. This suggests that the surge reflects the threshold of erosive forces needed to incise a fresh gully and that the surge resulted from hydrologically connected patches of disturbed vegetation where flow was unrestrained. These results have implications for debris hazard prediction as material scoured from the channel starting from the gully head constitutes the majority of debris volume. These results may aid development of an integrated systems understanding of post-fire erosion that incorporates vegetation and geomorphic controls on debris flow generation and landscape evolution.