Influences of vegetation disturbance on hydrogeomorphic response following wildfire

Quantifying the linkages between vegetation disturbance by fire and the changes in hydrologic processes leading to post-fire erosional response remains a challenge. We measured the influence of fire severity, defined as vegetation disturbance (using a satellite-derived vegetation disturbance index, VDI), landscape features that organize hydrologic flow pathways (relief and elongation ratios), and pre-fire vegetation type on the probability of the occurrence of post-fire gully rejuvenation (GR). We combined field surveys across 270 burned low-order catchments (112 occurrences of GR) and geospatial analysis to generate a probabilistic model through logistic regression. VDI alone discriminated well between catchments where GR did and did not occur (area under the curve=0.78, model accuracy=0.72). The strong effect of vegetation disturbance on GR suggests that vegetation exerts a primary influence on the occurrence of infiltration excess run-off and post-fire erosion and that major gully erosion will not occur until fire consumes aboveground biomass. Other topographic and local factors also influenced GR response, including catchment elongation, per cent pre-fire shrub, mid-slope riparian vegetation, armoured headwaters, firehose effects, and concentration of severe burn in source areas. These factors highlight the need to consider vegetation effects in concert with local topography and site conditions to understand the propensity for flow accumulation leading to GR. We present a process-based conceptual hydrologic model where vegetation loss from fire decreases rainfall attenuation and surface roughness, leading to accelerated flow accumulation and erosion; these effects are also influenced by interactions between fire severity and landscape structure. The VDI metric provides a continuous measure of vegetation disturbance and, when placed in a hydrologic context, may improve quantitative analysis of burned-area susceptibility to erosive rainfall, hazard prediction, ecological effects of fire, landform evolution, and sensitivity to climate change.