Landslides along the Three Gorges Reservoir in China pose a threat to coastal residents and waterway safety. To reduce false positive misjudgments caused by a sudden local change in the landslide deformation curve, in this paper, we propose an effective method for predicting the deformation trend of reservoir bank landslides. We take reservoir bank landslides in the Wanzhou District of the Three Gorges Reservoir area as the research object. The Time-Series Interferometric Synthetic Aperture Radar (InSAR) method and 62 Sentinel-1A images from 2018 to 2022 were selected for landslide deformation monitoring, and the Hurst index was calculated to characterize the deformation trend. Furthermore, we propose a method for predicting the deformation trend based on the statistical distribution of deformation rates and the physical significance of the Hurst index. After the field survey and Global Positioning System (GPS) verification, the Time-Series InSAR results are shown to be reliable. We take the Sifangbei landslide as a representative case to analyze the validation results. It is found that the determined Sifangbei landslide deformation trend is consistent with the conclusions for the region. In addition, the deformation trend of a reservoir bank slope has obvious spatial and temporal differences. Changes in the reservoir water level and concentrated rainfall play roles similar to those of catalysts. The proposed method, involving the combination of Time-Series InSAR and the Hurst index, can effectively monitor deformation and predict the stability trend of reservoir bank landslides. The presented research results provide new ideas and solutions for landslide prevention and risk mitigation in reservoir areas.

Rainfall threshold is an effective way for landslide early warning (LEW). Many threshold calculation models based on statistical principles have been proposed, or have been applied in national or regional early warning of geological hazard. However, due to global warming, frequent extreme rainfall and other factors, the triggering conditions of geological hazards have changed. And a fixed single rainfall threshold may no longer be applicable. In addition, the traditional threshold model requires a large number of high-quality landslide records in the region, and it is easy to ignore the heterogeneity of geological environment. Therefore, taking the slope unit as the object, we explore the dynamic updating model of rainfall threshold based on machine learning. Based on 170 high-risk slope units in Lin 'an District of Zhejiang Province and 65625 warning data from 2021 to 2023, we collected landslide records, rainfall station information and hourly rainfall data simultaneously. According to E-D threshold model and effective antecedent rainfall model, the general law of rainfall-induced landslide in Lin 'an District is derived, which is used as prior knowledge for model training. Then, through the warning data recorded in each slope unit, Decision Tree (DT), Bayesian Ridge (BR), Support Vector Machine (SVM), and Multi-Layer Perceptron (MLP) are selected as base learners, and ensemble strategies such as Bagging, Boosting, and Stacking are considered to dynamic updating of rainfall thresholds. Taking R2, Mean Square Error (MSE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) as evaluation metrics, the Stacking model shows the best prediction performance. In addition, two new landslides occurred in 170 slope units in 2024 were used to verify, and it was found that the updated threshold reduced the redundant workload and gave an accurate early warning of landslides. However, the volume of warning data and its distribution on different rainfall indicators are important factors affecting the threshold update. The accuracy of updating threshold needs to be tested with more experience and practice. The warning data reflects the response of slope to rainfall under different rainfall conditions, which is of great significance to the threshold update of slope unit. The dynamic updating of rainfall thresholds using machine learning meets the application requirements of current climate change and provides new ideas for disaster prevention and mitigation in the new era.

Slow-moving landslides pose prolonged and severe damage to buildings. The assessment of building vulnerability is a critical step in quantifying the risk of slow-moving landslides. Accurately retrieving the real displacement of landslides is essential for developing a more reliable vulnerability model of buildings. However, remote sensing observations acquired from a single orbit are insufficient to capture the three-dimensional (3D) displacement of landslides, which consequently limits the advancement of building vulnerability modeling.To this aim, this study integrates multi-source displacement monitoring, including synthetic aperture radar (SAR) imagery, optical imagery, and global navigation satellite system (GNSS), to assess the vulnerability of masonry buildings affected by slow-moving landslides. A total of 32 ENVISAT and 193 Sentinel-1A images were collected and processed using the multi-temporal differential interferometric SAR (MT-InSAR) technique to derive the vertical displacement of the landslide. Meanwhile, horizontal displacement was estimated from high-resolution optical imagery through pixel offset tracking. Then the real displacement of landslide can be retrieved by combining GNSS observations. We take 6 slow-moving landslides and masonry buildings on them in the Three Gorges Reservoir Area (TGRA) of China as the research objects. About 50 damaged masonry buildings were found out among near 500 residential buildings on landslides in field survey. These buildings are classified considering crack width on walls. Based on this, we then developed vulnerability curves using Polynomial, Exponential, Logistic, Weibull, and Sigmoid nonlinear regression functions.The results indicate that building damage responds more acutely to surface deformation in the vertical direction. In terms of fitting functions, the performance of different functions is affected by the choice of intensity parameter. Vulnerability curve derived from cumulative displacement is more suitable for the slow-moving landslides with low temporal heterogeneity in movement. Furthermore, 3D displacement measurements contribute to a comprehensive understanding of landslide movement characteristics and facilitates the development of reliable building vulnerability models. This provides essential guidance for the quantitative assessment of slow-moving landslide risk to buildings.