Evaluation and Modeling of Runoff and Sediment Retention in Vegetative Filter Strips

Abstract

Soil erosion can carry sediment-bound pollutants (phosphorus, pesticides, etc.) to nearby water bodies. Vegetative Filter Strips (VFS) have been proven to be an effective Best Management Practice (BMP) in agricultural fields to reduce off-site sediment losses. However, there are only a few studies on VFS application on a steep slope, which is more common at construction sites, especially on highway construction projects. Due to the inefficiency of currently used perimeter barrier BMPs, such as silt fences, it is imperative to explore the feasibility of VFS as a perimeter barrier at construction sites. Many factors impact the effectiveness of VFS, such as rainfall intensity, topography, buffer dimensions, vegetation species and density, runoff flow rates, antecedent soil moisture, etc. Therefore, it is necessary to evaluate the performance of VFS with various combinations of the above parameters to improve their effectiveness. However, it may not be feasible to assess the role of all the factors in a field setting. Modeling can be a valuable tool to predict the performance of VFS and is beneficial to the design of VFS across a range of conditions. In this study, a series of field tests were conducted at the Erosion Control Research and Training Center of the University of Illinois at Urbana-Champaign. Two test plots were established on a berm, both having a slope of 3:1. Vegetative buffers with different lengths (5 ft., 10 ft., and 15 ft.) were tested. Runoff delivery ratio (RDR) and sediment delivery ratio (SDR) were quantified. Results showed that the efficiency of VFS in trapping sediment increased as the buffer length increased. However, the increase in trapping efficiency was smaller when the buffer length was over 10 ft. under the test conditions in this study. The results indicated that VFS had at most nearly 40% RDR and 30% SDR. The 10 ft. long VFS were able to trap nearly 85% of the runoff and more than 90% of the sediment. Data from the field tests were also used to calibrate the Vegetative Filter Strip Modeling Systems (VFSMOD) model. Some key input parameters (vertical saturated hydraulic conductivity, soil water content, stem spacing, Manning‘s roughness, and slope) were measured or computed during the field tests. The predicted values from VFSMOD and the corresponding observed values of RDR and SDR were compared using the percent bias (PBIAS), the mean absolute error (MAE), and the normalized root mean square error (nRMSE). The results indicated that VFSMOD performed well in predicting RDR for all the tests, and SDR for most tests with a slight underestimation in the shortest buffer (5 ft.). A total of 156 scenarios were simulated in VFSMOD after the model parameterization. To find the appropriate buffer length (L), under a certain slope gradient (S) and stem spacing (SS), a relationship between RDR or SDR and L, S and SS was established using regression analysis. The equations: <fig><graphic xlink:href=23091_files/23091-00.jpg id=ID_99619a03-24f0-44cc-b5c0-8cb7fcda0365></graphic></fig> with generic coefficients were proposed as prediction models. The coefficients can be reparametrized based on topographic conditions and soil type. This quantitative model can help designers to determine how long of a vegetative buffer is needed to achieve a desired RDR and/or SDR under a certain situation.