WILDFIRE AND MONSOON INDUCED SHORT-TERM HILLSLOPE DEGRADATION RATES COMPARED WITH MODEL LONG-TERM RATES

Event-scale erosion data were collected in the San Francisco Volcanic Field, AZ, from 1996 to 2000, to establish natural short-term hillslope degradation rates and compare these with existing computational models of long-term rates. This Neogene scoria cone field consists of late Miocene to Holocene volcanic cones, with a wide range of degradation stages. The semiarid climate makes the area exceptionally suitable for studies of extreme erosion events, because of the annual intense July and August monsoon season. Being the first study to collect an integrated set of natural hillslope degradation data, the results can be used to validate and improve existing computational models. Data on rainfall, rainsplash, overland flow and soil creep were collected with the help of splashboards, erosion pins, erosion plots, raingauges and channel cross-sections on slopes of 0⁰ to 38⁰ . Rainsplash data are consistent with a long-term diffusion model of slope degradation, but measured sediment fluxes are slightly higher than those predicted by such a model. Rainfall events cause formation of rilling and gullying by overland flow while dry periods cause diffusional infilling of the gullies and rills, as observed in successive channel cross-section and erosion pin measurements. Data on degradation by overland flow are not consistent with any type of diffusive-degradation model but have been compared to a channel-forming computational model, producing satisfactory results regarding flow speed, path and channel development, and erosion/deposition patterns. A severe forest fire burned one of two hillslope observation sites in 1996. This has given us the opportunity to observe the impact of fire on erosion due to drizzle (15 mm h^-1) in areas having different severities of fire damage. The degradation has been accelerated in areas of high fire severity by almost one order of magnitude. In particular, the rate of degradation by overland and pipe flow increased. A multiple regression of the field data including slope angle, vegetation, precipitation and fire severity fits data on degradation rate at alpha=1 when fire severity is included in the regression.