This Is AuburnElectronic Theses and Dissertations

Evaluation of Soil Effects on Soil Erosion on Off-Road Vehicle Trails using WEPP




Melton, Jonathan

Type of Degree



Civil Engineering


Erosion and sediment loss from the use of off-road vehicles (ORV) is an escalating problem on national forest lands. The increasing number of ORV riders coupled with decreasing riding areas has caused the United States Forest Service and other agencies to begin development of best management practices (BMPs) that will reduce the amount of erosion and sediment loss occurring on trails. In order to quantify the sediment loss from ORV trails, two trail sections of the Kentuck ORV trail system in the Talladega National Forest were equipped with sediment sampling devices and rainfall data collectors. These trail sections were located on a mild slope and a steep slope on the existing trail. Storm data were collected and analyzed for five storms during a nine month period. The storm producing the highest amount of soil erosion produced approximately 0.78 tonnes per hectare and 1.46 tonnes per hectare of sediment from mild and steep slopes, respectfully. These data were used to evaluate an erosion model that would allow land managers to predict sediment loss on general trail sections throughout the trail system. The Water Erosion Prediction Project (WEPP) model was used because of its ability to model forest roads. The instrumented trail sections were modeled using WEPP and compared to sediment and runoff data from field experiments to calibrate the model. Upon calibration, a Nash-Sutcliffe efficiency (NSE) of 0.45 was achieved for runoff while a R2 value of 0.43 and NSE of -0.36 was established for soil erosion. In general, the calibrated model predicted lower runoff, but much higher sediment production, than observed. A second model was calibrated in order to calibrate for sediment production only, which produced a NSE of 0.24. The mild and steep slope data was separated and modeled for each individual slopes. An attempt to validate these models was performed by using the opposite slope for validation. The results from the validation showed that the model was not as good a predictor of expected sediment production for either slope. In addition to the slope comparison, the data was divided into trafficked and non-trafficked data which produced a NSE of -0.11 and 0.76, respectfully. The moderate NSE for the non-trafficked data was expected due to the lack of the additional soil disturbance from the ATV traffic. The calibrated model for sediment only was used to simulate erosion on varying slopes. From these simulations, recommendations for water diversion structure spacing were developed. Water diversion structures carry runoff from the trail into the surrounding forested areas, reducing the velocity and sediment concentration of water flow over the trail surface. The spacing recommendations developed from this study can be used as BMPs to aid land managers in designing systems to reduce the impacts on ORV trails.