Gully morphology, hillslope erosion, and precipitation characteristics in the Appalachian Valley and Ridge province, southeastern USA

• Precipitation-related gully erosion in Ultisols is examined in three morphological settings.
• Channel and sidewall erosion is related to antecedent precipitation, accumulation, and duration.
• Erosion in interfluves is not related to precipitation parameters.
• Precipitation intensity is not a useful predictor for soil erosion in this setting.
• Synoptic scale is appropriate for evaluating precipitation related erosion.

This study investigates gully erosion on an east Tennessee hillslope in a humid subtropical climate. The study area is deeply gullied in Ultisols (Acrisol, according to the World Reference Base for Soil), with thirty years of undisturbed erosional history with no efforts to correct or halt the erosion. The objectives are (1) to examine how different gully morphologies (channel, sidewall, and interfluve) behave in response to precipitation-driven erosion, and (2) to identify an appropriate temporal scale at which precipitation-driven erosion can be measured to improve soil loss prediction. Precipitation parameters (total accumulation, duration, average intensity, maximum intensity) extracted from data collected at an on-site weather station were statistically correlated with erosion data. Erosion data were collected from erosion pins installed in four gully systems at 78 locations spanning three different morphological settings: interfluves, channels, and sidewalls. Kruskal–Wallis non-parametric tests and Mann–Whitney U-tests indicated that different morphological settings within the gully system responded differently to precipitation (p < 0.00). For channels and sidewalls, regression models relating erosion and precipitation parameters retained antecedent precipitation and precipitation accumulation or duration (R2 = 0.50, p < 0.00 for channels, R2 = 0.28, p < 0.00 for sidewalls) but precipitation intensity variables were not retained in the models. For interfluves, less than 20% of variability in erosion data could be explained by precipitation parameters. Precipitation duration and accumulation (including antecedent precipitation accumulation) were more important than precipitation intensity in initiating and propagating erosion in this geomorphic and climatic setting, but other factors including mass wasting and eolian erosion are likely contributors to erosion. High correlation coefficients between aggregate precipitation parameters and erosion indicate that a suitable temporal scale to relate precipitation to soil erosion is the synoptic time-scale. This scale captures natural precipitation cycles and corresponding measurable soil erosion.