Water level fluctuations in the reservoir deteriorate soils and rocks on the bank landslides by dryingwetting (D-W) cycles, which results in a significant decrease in mechanical properties. A comprehensive understanding of deterioration mechanism of sliding-zone soils is of great significance for interpreting the deformation behavior of landslides. However, quantitative investigation on the deterioration characteristics of soils considering the structural evolution under D-W cycles is still limited. Here, we carry out a series of laboratory tests to characterize the multi-scale deterioration of sliding-zone soils and reveal the mechanism of shear strength decay under D-W cycles. Firstly, we describe the micropores into five grades by scanning electron microscope and observe a critical change in porosity after the first three cycles. We categorize the mesoscale cracks into five classes using digital photography and observe a stepwise increase in crack area ratio. Secondly, we propose a shear strength decay model based on fractal theory which is verified by the results of consolidated undrained triaxial tests. Cohesion and friction angle of sliding-zone soils are found to show different decay patterns resulting from the staged evolution of structure. Then, structural deterioration processes including cementation destruction, pores expansion, aggregations decomposition, and clusters assembly are considered to occur to decay the shear strength differently. Finally, a three-stage deterioration mechanism associated with four structural deterioration processes is revealed, which helps to better interpret the intrinsic mechanism of shear strength decay. These findings provide the theoretical basis for the further accurate evaluation of reservoir landslides stability under water level fluctuations. (c) 2025 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/).

In landslide studies, particle size is a key quantitative indicator, reflecting the formation and development of the sliding zone. It plays a crucial role in understanding the mechanisms and evolutionary processes that lead to landslide occurrences. Precise measurement of particle size is crucial. This study centered on soil samples from the Lanniqing landslide in Southwest China. To begin, seven distinct methods were used to preprocess the soil samples. Next, the particle size frequency distribution was measured using the Mastersizer 2000 laser particle size analyzer. Key parameters, including median particle size, mean particle size, sorting coefficient, skewness, and kurtosis, were then compared and analyzed to determine the most appropriate preprocessing method for evaluating the characteristics of the soil samples. The mechanism of landslide occurrence was subsequently analyzed by examining the particle size characteristics, mechanical properties, and mineral composition of the soil samples. The results suggested that method C provides the most reliable analysis of particle size characteristics in soil samples. The observed coarsening of coarse particles, along with a significant increase in clay content within the sliding zone, indicates that the sliding surface has undergone multiple shear and compression events. The interplay of the upper traffic load and slope cutting at the front edge set the stage for the Lanniqing landslide, prompting the initial development of potential sliding surfaces. Rainfall acts as a catalyst for slope instability. The high clay content, combined with the formation of a low-permeability layer rich in clay minerals on the sliding surface, leads to excessive pore water pressure and mineral lubrication. These factors inherently trigger and accelerate the occurrence of the landslide.