Recognition of suitable small watersheds for check dam construction on the Loess Plateau

Check dams are used for soil and water conservation worldwide, especially on the Chinese Loess Plateau, which has suffered from severe soil and water loss. To promote ecological conservation and high-quality development in the Yellow River basin, considerable effort should be devoted to building check dams in the small watersheds where the conditions permit on the Loess Plateau. Therefore, suitable small watersheds must be identified for constructing new check dams. In this study, the weighted overlay method (WOM) of the multi-criteria decision-making analysis method, the random forest (RF), and the support vector machine (SVM) of machine learning methods were first applied to simulate the suitability of small watersheds for check dam construction in the Ganguyi upstream area of the Yanhe River basin on the Loess Plateau. Accuracy assessment showed that the accuracy of the three models was almost perfect and RF had the highest accuracy of 0.95, followed by SVM (0.87), and WOM (0.83). Meanwhile, the impacts of topographic factors and soil erodibility were found to be generally higher than those of the other factors, and 56% of small watersheds had high and very high suitability for check dam construction, 67% of these small watersheds have already been built as check dams. Finally, a total of 108 small watersheds were found to be suitable in the study area when considering their environmental characteristics. The methodology developed in the study can be used as a preliminary planning step for check dam construction and as a foundation for detailed field investigations in the future, which help to realize the sustainable use of soil and water resources.