Estimation of sediment yield and effectiveness of level stone bunds to reduce sediment loss in the Gumara-Maksegnit watershed, Nile Basin, Ethiopia

Purpose Soil erosion is one of the most significant environmental problems in the Ethiopian highlands. The study was conducted in the main Gumara-Maksegnit watershed and at the sub-catchment scale with and without soil and water conservation (SWC) measures to estimate event-based sediment yield and evaluate the effectiveness of level stone bunds on sediment reduction. Materials and methods The runoff volume and peak discharge were recorded manually with the help of a time series of pictures using the DOERR digital camera. The sediment concentration was determined from manual runoff samples and grid-based soil samples were collected at 162 points. The slope map and satellite image were obtained from the USGS Earth Explorer database with 30-m and 15-m spatial resolutions, respectively. Finally, all six-model factors were combined using the raster calculator in map algebra through the MUSLE framework. The paired t test parametric procedures were applied to test whether the means of event-based predicted and observed sediment yield values are different. Result and discussion The average event-based observed and estimated sediment yields were 0.5581 and 0.4031 Mg ha −1 for the Gumara-Maksegnit watershed, 0.5125 and 0.4194 Mg ha −1 for the treated (with SWC) sub-catchment, and 1.0694 and 1.0150 Mg ha −1 for the untreated (without SWC) sub-catchment, respectively. The observed and estimated sediment losses within the main watershed and sub-catchments were not significantly different. However, comparing the treated and untreated sub-catchments, the sediment losses were highly significantly different and the stone bunds have a capacity for reducing sediment loss by 58.8%. In general, the MUSLE model performed well to estimate sediment yield in the study area with R 2 values of 0.62, 0.72, and 0.7 and NSE values of 0.53, 0.71, and 0.34 for treated, untreated, and the main Gumara-Maksegnit watershed, respectively. Conclusion The results showed that stone bunds have a capacity for reducing sediment loss by 58.8% as compared with untreated sub-catchment. Therefore, successful implementation of stone bunds in the study area, as well as similar agro-ecologies, has a great benefit to enhance land productivity. Meanwhile, the MUSLE model was well suited for reliable applications of sediment yield estimation in the study area.