Modeling Uncertainty of Runoff and Sediment Yield Using a Distributed Hydrologic Model

The event- and physically-based runoff and erosion model KINEROS2 is applied to assess the impact of uncertainty in model parameters on simulated hydrographs and sediment discharge in a small experimental watershed. The net capillary drive parameter, which affects soil infiltration, is shown to be approximately lognormally distributed, and related statistics of this parameter for all soil texture class are computed and tabulated. The model output response to uncertainty in soil hydraulic and channel roughness parameters is evaluated by performing Monte Carlo (MC) simulations based on the parameters ’ statistics obtained from the literature. The results show the extent to which uncertainty in the saturated hydraulic conductivity, net capillary drive, and initial relative saturation influences the simulated cumulative distributions of peak sediment discharge and sediment yield. The distribution of the simulated time to peak sediment discharge is dominated by uncertainty in the channel and plane Manning’s roughness coefficients. Comparison of the simulated median and uncertainty ( ± 25 % and 45 % quartiles) with observed values of runoff and sediment discharge, for two, large and small rainfall events, indicate that the model performs properly and can be calibrated based on the range of soil parameters reported in the literature