Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6409196 | Geoderma | 2013 | 9 Pages |
â¢WEPP-interrill erosion prediction values were lower than the measured ones.â¢The model tended to overestimate lower erosion rates.â¢WEPP underestimated higher erosion rates.â¢Interrill erodibility was predicted on the basis of stream power.
Modeling soil erosion contributes to the understanding of the erosion processes and needs to predict accurately the erosion rates under different environmental conditions. Few studies have investigated the WEPP's applicability for arid and semi-arid conditions that differ from those where the model was developed. This research was carried out to evaluate and improve the WEPP model for arid and semiarid regions for interrill erosion using a rainfall simulator at plot scale. The results showed that measured interrill erosion rates ranged from 9.3 Ã 10â 6 to 89.6 Ã 10â 6 kg mâ 2 sâ 1. In comparison, the WEPP-interrill erosion prediction values were on average 14.5 times lower than the measured ones, which can be ascribed to a systematic error. To calibrate the model, two regression equations were presented to estimate baseline interrill erodibility (Kib) and adjusted interrill erodibility (Kiadj). Accordingly, the baseline interrill erodibility was influenced by surface rock cover much more than any other soil property. Also, Kiadj was functioned as sediment concentration and sand percent. Model validation showed that after calibrating Kib and Kiadj, the coefficient of efficiency enhanced to 0.31 and 0.90, respectively. However, the model tended to overestimate the lower erosion rates and to underestimate the higher erosion rates. To modify interrill erodibility and to achieve a process-based parameter, a simple equation was developed based on stream power. The coefficient of efficiency of 0.61 was obtained for the proposed approach. In general, the model predictions enhanced after calibration, but the model tendency to underestimate high erosion rates was observed. One of the reasons for low accuracy of the WEPP was its dependence on soil empirical parameters used in predicting the sub-model components. The model prediction was found to be improved through process-based calibration of soil parameters.