Research on slope-oriented prediction method for shallow loess landslides

Shallow loess landslides, as one of the widely distributed and high-frequency geologic hazards, have brought great economic losses and ecological damage to human society. In this study, Qinzhou District, Tianshui City, Gansu Province, is taken as the study area, and the Scoops3D model is used to predict the occurrence of loess landslides in the area based on the DEM data of the area. Bishop’s simplified method and box search method were used to calculate and analyze the landslide stability in the study area. The landslide prediction results of the Scoops3D model of this paper are compared and analyzed under different DEM data resolutions. Subsequently, local environmental data are collected to study the correlation between environmental impact factors and shallow loess landslides. Finally, the prediction accuracy of the shallow loess landslide prediction model based on Scoops3D in this paper is tested by comparing the difference between the prediction results of the Scoops3D model of this paper and other prediction models with the actual results. The resolution of the DEM data has an important influence on the prediction results of the Scoops3D model, and the accuracy of the high-resolution DEM prediction results is higher than that of the low-resolution prediction results. There is a significant correlation between landslide displacement and humidity and cumulative precipitation, and the difference between the predicted and measured values of the GA-BP and GA-Elman models is within 8 mm, and the difference gradually increases. The difference between the predicted and measured values of the Scoops3D model in this paper is between 0.00 and 2.30 mm, and the prediction effect is optimal.