We use small baseline subset interferometric synthetic aperture radar InSAR (SBAS-InSAR) to estimate the annual soil erosion rate in the catchment area of Shihmen Reservoir in 2015 and then estimate the soil erosion volume of this catchment area. The study data are 20 Sentinel-1 satellite images. The annual soil erosion rates derived from the SBAS-InSAR are validated using those derived from erosion pins. The soil erosion volumes estimated using SBAS-InSAR are compared with those estimated using the universal soil loss equation (USLE). The results reveal that more accurate estimations of the annual soil erosion rate are achieved when at least 12 satellite images of the ascending
orbits are employed. When filtering is not considered, the tension coefficient is 0.25 and block averaging is not used
during data processing; this achieves better SBAS-InSAR results. Compared with erosion pin monitoring results, the
correlation coefficient is 0.56 for natural slopes and 0.51 for remedial slopes. When considering filtering processing,
an approximately 8-km Gaussian filter radius achieves the optimal SBAS-InSAR results; compared with erosion pin
monitoring results, the correlation coefficient is 0.67 for natural slopes and 0.57 for remedial slopes. Finally, soil erosion volumes estimated using SBAS-InSAR and USLE display similar trends.
Key Words: SBAS-InSAR, Shihmen Reservoir, soil erosion