An improved monitoring strategy for shallow rainfall-induced landslides under critical site-specific recognition

Traditional methods for the monitoring of rainfall-induced shallow landslides have the drawbacks of high cost due to the uniform distribution of monitoring points and the lag of early warning index response. To address these issues, we conducted an optimization study on monitoring strategy based on the slope hydrological responses in Xide County, Sichuan Province, China. In this study, the finite element method was employed to simulate the rainfall infiltration process in a series of orthogonal tests. Then, the susceptibility analysis of different hydrological responses was carried out to identify different controlling factors. On this basis, the slope instability mechanisms under different conditions were investigated through laboratory tests, and the corresponding monitoring strategies were successfully proposed. The results of sensitivity analysis show that the slope hydrological response is most sensitive to slope height, and the control effect on hydrological response is enhanced with the increasing of slope height. In the changing process of slope height, the landslide hydrological response can be divided into three different models. Laboratory tests are used to reveal the landslide hydrological response characteristics under different models. The results show that as the increasing of slope height, the main deformation zone gradually migrates from slope top to slope toe, and the rainwater gathers in the main deformation zone though preferential infiltration channels. The fine particles in the soil move down to the slope toe continuously under the action of seepage force, resulting in the expansion of tensile cracks in the slope. In addition, rainwater accumulation produces higher seepage force, which speeds up the migration of fine particles. The migration of fine particles also promotes the expansion of slope surface cracks, which is conducive to the accumulation of rainwater. Under this cycle interaction, the effect of slope self-weight and hydrodynamic driving force in different part of the slope exerts significant influence on slope failure mode, leading to the change from thrust-load-caused landslides to retrogressive landslides. During the critical sliding stage, the displacement and water content of the main slope deformation area increase dramatically in advance. Therefore, we consider the main slope deformation zone as the key monitoring area, which can be identified through coupling sensitivity analysis and laboratory tests. This monitoring method can timely capture the change of monitoring indexes reflecting the slope state and tackle the lag problem of early warning index of sudden landslides, which not only reduce the high cost but also increase the accuracy and reliability of landslide early warning. This study provides new insights for the monitoring strategy of rainfall-induced landslides.