Field model experiments and numerical analysis of rainfall-induced shallow loess landslides

Shallow loess landslide induced by rainfall is a widespread form of slope failure on the Loess Plateau of China. The hydrological response, evolutionary mechanism and failure processes of unsaturated loess slopes under precipitation in a field scale are currently lacking. Three groups of full-scale field experiments with different rainfall intensities were conducted in Jingyang County, Northwest China. Instrumentations including soil moisture sensors, tensiometers and earth pressure cells were installed in the slope. This paper presents the results from three types of sensors to reveal the surface infiltration process and deformation characteristics of unsaturated loess slopes. In addition, SEEP/W and SLOPE/W were selected to quantify the effects of soil suction strength on the stability of the homogeneous loess slope. The results show that the maximum infiltration depth of homogeneous loess slope after two days of continuous heavy rainfall is about 0.6 m. The infiltration depth and seepage velocity on the slope crest and slope toe are faster than the middle of the slope; The decrease amplitude and decrease speed of matric suction increase with the increase of rainfall intensity. Moreover, it is found that the rainfall intensity only affect the development of slope deformation and failure in terms of the time scale, but has little influence on the failure mode; Tensile cracks at the slope shoulder develope more rapidly due to the creep deformation of slope, the deformation and failure at the slope shoulder are the key factors controlling the occurrence of loess landslides. In addition, the proportion of suction strength in shear strength is limited, and the finiteness of the suction strength determines that its attenuation will not determine the stability of homogeneous loess slope without preferential infiltration channels. However, the increase of positive pore water pressure at the potential slide surface may be the key factor to determine the slope failure.