Quantification of stream sediment inputs from steep forested mountains

Mountain environments are regions of high erosion rates due to steep slopes and extreme weather and climate forcing. In steep, soil-mantled forested mountains, long-term sediment yield is often dominated by rare catastrophic erosion events. Some of these catastrophic erosion events are linked to natural and anthropogenic watershed disturbances. They may cause significant property damage and life loss. Episodic sediment delivery to streams disturbs the ecologic environment and aquatic habitats. The frequency and magnitude of mountain sediment yields is a complex function of topography, climate, and vegetation cover. This dissertation focuses on understanding of the erosion processes on steep slopes and the factors controlling catastrophic erosion events from a theoretical and empirical perspective based on field data analysis.; The dissertation is a collection of three papers. Field information used in the papers is compiled from study watersheds in the Idaho batholith. The first paper presents a probabilistic approach for channel initiation due to overland flow and shows the application of the theory using channel head data from two watersheds. The probability distribution of area-slope thresholds (aSα) derived from the probabilistic theory is compared to and found to match well with the observed area-slope thresholds at channel head locations in the field. The second paper reports field observations of soil loss volumes due to gully formation during a single thunderstorm event. The paper theoretically relates sediment transport capacity to runoff rate, slope and drainage area, then uses the soil loss observations to calibrate the nonlinear relationship between sediment transport and drainage area and slope. The paper also shows that the concavity index of the gully profiles obtained from the area and slope exponents of the calibrated sediment transport function using field data agrees well with the observed profile concavity of the gullies. Finally, the last paper describes a numerical framework for modeling the frequency and magnitude of sediment yields in the Idaho batholith. Simulation results show good correspondence with field observations of event sediment yields and long-term averages of soil loss over time scales up to 10,000 years. The model underscores the influence of forest vegetation and vegetation disturbances on erosion. The findings of the papers contribute to the understanding of how mountain ranges erode and have applicability in modeling the erosion response to land use and climate changes.