Modeling of the interactions between forest vegetation, disturbances, and sediment yields

The controls of forest vegetation, wildfires, and harvest vegetation disturbances on the frequency and magnitude of sediment delivery from a small watershed (similar to3.9 km(2)) in the Idaho batholith are investigated through numerical modeling. The model simulates soil development based on continuous bedrock weathering and the divergence of diffusive sediment transport on hillslopes. Soil removal is due to episodic gully erosion, shallow landsliding, and debris flow generation. In the model, forest vegetation provides root cohesion and surface resistance to channel initiation. Forest fires and harvests reduce the vegetation. Vegetation loss leaves the land susceptible to erosion and landsliding until the vegetation cover reestablishes in time. Simulation results compare well with field observations of event sediment yields and long-term averages over similar to10,000 years. When vegetation is not disturbed by wildfires over thousands of years, sediment delivery is modeled to be less frequent but with larger event magnitudes. Increased values of root cohesion (representing denser forests) lead to higher event magnitudes. Wildfires appear to control the timing of sediment delivery. Compared to undisturbed forests, erosion is concentrated during the periods with low erosion thresholds, often called accelerated erosion periods, following wildfires. Our modeling suggests that drainage density is inversely proportional to root cohesion and that reduced forest cover due to wildfires increases the drainage density. We compare the sediment yields under anthropogenic (harvest) and natural (wildfire) disturbances. Disturbances due to forest harvesting appear to increase the frequency of sediment delivery; however, the sediment delivery following wildfires seems to be more severe. These modeling-based findings have implications for engineering design and environmental management, where sediment inputs to streams and the fluctuations and episodicity of these inputs are of concern.