Predicting gully rejuvenation after wildfire using remotely sensed burn severity data

The loss of surface vegetation and reduced infiltration caused by wildfires can trigger gully rejuvenation, resulting in damage to downstream aquatic resources and risk to human life and property. We developed a spatially explicit metric of burn severity - the Burn Severity Distribution Index (BSDI) - and tested its ability to predict post-fire gully rejuvenation in 1st and 2nd order basins burned in the 2000 Valley Complex fires in the Sapphire Mountains of western Montana. The BSDI was derived from burn severity data interpreted from Landsat 7 satellite imagery using the Normalized Burn Ratio (NBR) method, and ranged from 0.0 for completely unburned basins to 4.0 for basins burned entirely at high severity. In July 2001 rainstorms with peak 30-minute intensities of up to 17 mm h^{- 1} triggered gully rejuvenation in 66 of the 171 basins examined. The frequency of gully rejuvenation was higher in basins with higher BSDI values, increasing from zero for basins with a BSDI less than 1.3 to 67% for basins with a BSDI greater than 3.0. Binary logistic regression indicated that BSDI was a more significant predictor of gully rejuvenation than basin morphometric variables. The absence of gully rejuvenation in several basins with a high BSDI was attributed to low gradient, dense riparian vegetation, or concentration of high burn severity at lower elevations in the basin. The presence of gully rejuvenation in several basins with a low BSDI was associated with false negative NBR classification errors in northwest aspects, and concentration of severe burn impacts in the drainage headslopes. BSDI is a useful metric for predicting gully rejuvenation after wildfire. The use of the BSDI in Burned Area Emergency Response team assessments could improve the planning, implementation, and monitoring of burned area recovery treatments.