Effects of gradient,distance, curvature and aspect on steep burned and unburned hillslope soil erosion and deposition

Wildfire denudes vegetation and impacts chemical and physical soil properties, which can alter hillslope erosion rates. Post‐wildfire erosion can also contribute disproportionately to long‐term erosion rates and landscape evolution. Post‐fire hillslope erosion rates remain difficult to predict and document at the hillslope scale. Here we use ²¹⁰Pb_aex (lead‐210 mineral‐adsorbed excess) inventories to describe net sediment erosion on steep, convex hillslopes in three basins (unburned, moderately and severely burned) in mountainous central Idaho. We analyzed nearly 300 soil samples for ²¹⁰Pb_aex content with alpha spectrometry and related net sediment erosion to burn severity, aspect, gradient, curvature and distance from ridgetop. We also tested our data against models for advective, linear and non‐linear diffusive erosion. Statistically lower net soil losses on north‐ versus south‐facing unburned hillslopes suggest that greater vegetative cover and soil cohesion on north‐facing slopes decrease erosion. On burned hillslopes, erosion differences between aspects were less apparent and net erosion was more variable, indicating that vegetation influences erosion magnitude and fire drives erosion variability. We estimated net soil losses throughout the length of unburned hillslopes, including through a footslope transition to concave form. In contrast, on burned hillslopes, the subtle shift from convex to concave form was associated with deposition of a post‐fire erosion pulse. Such overall patterns of erosion and deposition are consistent with predictions from a non‐linear diffusion equation. This finding also suggests that concave sections of overall convex hillslopes affect post‐disturbance soil erosion and deposition. Despite these patterns, no strong relationships were evident between local net soil losses and gradient, curvature, distance from ridgetop, or erosion predicted with advection or diffusion equations. The observed relationship between gradient and erosion is therefore likely more complex or stochastic than often described theoretically, especially over relatively short timescales (60–100 years). Copyright © 2016 John Wiley & Sons, Ltd.