Kapo in Wensu County, Xinjiang of China, is a high and steep slope formed by fluvial deposition. Highly susceptible to heavy rainfall, it faces prominent erosion hazards and soil loss. Combined with the region’s arid climate and the slope soil’s significant water sensitivity and collapsibility, local slope prevention and control is extremely challenging. To conduct efficient, targeted slope management, this study uses physical rainfall simulation tests to explore surface erosion and deformation characteristics of the area’s interbedded silt-fine sand slopes under rainfall. Results show that rainfall volume and soil properties are the primary drivers of slope erosion. As rainfall duration increases, the slope’s rainwater infiltration rate first rises then declines. Gradually saturated by infiltration, the slope’s erosion resistance weakens, and surface damage intensifies. The sand layer has better permeability than the soil layer; coarse sand also mitigates raindrop-induced splash erosion, which has stronger erosion resistance, erosion initiates in the soil layer beneath the sand layer. Thus, influenced by soil properties and rainfall infiltration, slopes in this region exhibit a retrogressive failure mode: surface erosion, local damage, mid-slope failure, top-slope failure.

Numerous debris flow disasters with difficult drainage exist in northern Xinjiang of China. Civil engineering facilities such as provincial highways and gas stations built on alluvial fans at the mountain outlets block the normal discharge of debris flows into the main river, making it impossible to implement drainage measures downstream. Therefore, studying mitigation methods and engineering effects for debris flows in areas with inadequate or difficult drainage conditions is of great significance. This study focuses on the Coxon ditch in Burqin County, Altay Prefecture, Xinjiang, China. At present, a water-intercepting seepage drainage zone has been established downstream of the Coxon ditch. Through two sets of physical model tests under different rainfall intensities, and based on process imaging and monitoring data, we investigate the variation characteristics of soil pressure, pore water pressure, and moisture content during the seepage drainage process of debris flows in drainage-challenged areas. The seepage drainage patterns are summarized, and the underlying mechanisms are revealed. The results show that the drainage process can be divided into five stages: under the condition of 50-year rainfall intensity confluence, which is, channel surface erosion stage, ditch bank erosion stage, seepage discharge stage, seepage deceleration stage, and seepage end stage; under the condition of 100-year rainfall intensity confluence, which is, seepage drainage stage, seepage deceleration stage, water overflow stage, accumulation starting behind the dam stage, and seepage end stage. The soil pressure curve in the upstream channel first rises and then declines, while downstream, soil pressure increases due to sediment deposition. The pore water pressure curve correlates with the moisture content curve, but its rising phase lags significantly behind the soil pressure curve. After water-sediment separation, the seepage zone retains sediment while draining water, which slightly reduces the seepage rate. However, the overall seepage drainage process remains efficient, demonstrating the high feasibility of seepage-based mitigation engineering.

The Ili River Valley is a key development area in Xinjiang with dense population, favorable environmental and economic conditions, and one of the regions in China where loess landslides are widely distributed. Affected by factors such as climate and topography, landslide disasters occur frequently in this area, but the landslide response to rainfall events has not been fully demonstrated by monitoring data. Therefore, based on the monitoring data of the Karaheisu landslide in Xinyuan County during the rainfall period from April to September in 2021, and using the SPSSPro platform, this study employs mathematical statistics and grey correlation analysis methods to comprehensively analyze the correlation and association degree between displacement and rainfall amount, soil water content, and determine the response characteristics of soil water content, vertical displacement, and horizontal displacement of loess-bedrock landslides to rainfall. The results show that landslide deformation has a strong correlation with rainfall amount and water content, with the rainfall infiltration depth basically stabilized at 1.2~2.0 m; the degree of association between landslide displacement and rainfall amount is greater than that with soil water content.