Impact of Uncertainty on the Design of Vegetative Filter Strips

Design of vegetative filter strips for trapping sediment and sediment-borne chemicals can be done on a storm event basis using VFSMOD. This approach requires confidence in a number of input parameters such as soil infiltration characteristics, surface topography, vegetative composition and roughness, and incoming sediment load and particle size distribution. We have developed a simple program, UH, to estimate runoff hydrographs and sediment transport from a source area based on the NRCS Unit Hydrograph Method and the Modified Universal Soil Loss Equation (MUSLE). VFSMOD uses this information from the source area to predict the amount of sediment trapped in the filter strip. For a given design case, we demonstrate the use of an integrated design tool to 1) identify and rank the input parameters of UH and VFSMOD relative to their sensitivity on sediment trapping, 2) develop probability density functions for the most sensitive input parameters, 3) use Monte Carlo Simulation to sample the input parameters and develop a probability density function for sediment trapping. We will demonstrate the use of confidence intervals for sediment trapping based on uncertainty in the inputs parameters for the source area and the vegetative filter strip.