Comparison of Sediment Yield Models on Watersheds in Mississippi

ABSTRACT Knowledge of existing erosion models is important for researchers, action agencies and consultants to understand how a model might be applied for their use. This study compared the simulated results from the models CREAMS, SWRRB, EPIC, ANSWERS and AGNPS with measured data of runoff and sediment yield on an annual and storm rainfall event basis. Results were obtained from a parallel, tile outlet terrace, a flatland watershed, and a small watershed with 3 to 8% slopes—all located in Mississippi. The comparisons showed that no one model worked well in every situation of runoff and sediment yield on the watersheds. Overall, CREAMS and SWRRB produced results that were close to the measured values more often then the other models, even though SWRRB required simpler inputs. AGNPS, a single-event model that has no capability for updating parameters, except for a modification that was made to update the USLE's C-factor, also produced results that were close to the measured values in many situations. EPIC produced results that were similar to the measured runoff in most situations, but not as good as the other models in predicting sediment yield. This might be from the fact that EPIC computes its own C-factor, which for Mississippi may need to be adjusted. Results obtained by ANSWERS were not close to the measured values in most situations. This probably occurred because the model was designed to simulate single storm rainfall events and was modified to simulate the events continuously, without updates of soil moisture between events.