Introduction: Soil erosion is the separation, transport, and successive destruction of soil on the surface of the earth by an erosive agent such as water, wind, or glacier. Therefore, reducing soil erosion and managing it in watersheds to reduce its losses is a very important task in the field of watershed management. Accordingly, many models have been developed to simulate soil erosion. In the present study, the effectiveness of each of the USLE, RUSLE, and MUSLE models in preparing a soil erosion zoning map of the Kardeh Dam catchment area in Khorasan Razavi province and identifying high-risk erosion areas has been investigated.

Methodology: In the present study, the amount of soil erosion in the Kardeh dam catchment is calculated using the Universal Soil Erosion Equation (USLE) model and its modified (MUSLE) and revised (RUSLE) versions. In the USLE and RUSLE, the amount of soil erosion is the product of six factors, rain erosion (R), soil erodibility (K), slope length (L), slope degree (S), vegetation management (C) and the protective operation (P). However, in the MUSLE, the rain erosion factor is replaced by the momentary flood peak discharge (qp) and the total volume of flood discharge (Q) to estimate the erosion magnitude. The above models were developed in ArcGIS software under the SATEEC system, and the sediment load of the basin and additional assessments were made.

Results and Discussion: The erosion map derived from rainfall data collected at eight stations indicates that the R coefficient in the Kardeh Dam catchment area ranges from 2.8 to 3.8 (MJ mm ha⁻¹ h⁻¹ yr⁻¹). Approximately 65% of the Kardeh basin area, including the northern region as well as parts of the eastern, western, and central areas, exhibits a high erosion coefficient ranging from 3.41 to 3.78. Based on the K factor map, the soil erodibility coefficient in the Kardeh Dam catchment area varied from 0.21 to 0.42 (MJ mm ha⁻¹ h⁻¹ yr⁻¹) and about 40% of the northern and southern parts of the basin have a high erodibility coefficient. The LS coefficient in the three models varies between 0.76-12.42, 0.00-9.50 and 0.07-21.90, respectively. The observed NDVI index and C factor values in the Kardeh Dam basin showed that about 92 percent of the basin area has a low NDVI coefficient and high C, only about one percent has a high NDVI and low C coefficient, and about 92.2 percent of the Kardeh Dam basin has no vegetation cover. Integrating the above factors in ArcGIS, the range of soil erosion changes for the three models was obtained as 0.14-1.18, 0.00-2.95, and 0.01-5.19, respectively. In the USLE model, 33 percent of the basin, in RUSLE about 15 percent, and in MUSLE 30 percent of the basin showed high and very high erosion. In addition, the R2 coefficient value between the RUSLE and MUSLE models was 0.87, which indicates the correlation of these two models in overestimating the soil erosion rate of the Kardeh Dam basin. Based on the observed relative error, the USLE model based on the Boyce method had a better estimate of the sediment load compared to other applied models.

Conclusion: Comparison of the three studied models in estimating erosion and sediment load showed that the USLE model with the lowest relative error, the best performance and the erosion rate estimated with this model was also accepted. Performance evaluation of the soil erosion models RUSLE and MUSLE showed a large difference between the results of both models compared to the USLE model. The research findings will be useful for regional authorities to better analyze the soil protection measures in the Kardeh Dam catchment area.