Landslides, prevalent in mountainous areas, are typically triggered by tectonic movements, climatic changes, and human activities. They pose catastrophic risks, especially when occurring near settlements and infrastructure. Therefore, detecting, monitoring, and predicting landslide deformations is essential for geo-risk mitigation. The mainstream of the previous studies have often focused on deterministic models for immediate landslide prediction. However, in most of them, the aspect of prediction uncertainties are not sufficiently addressed. This paper introduces an innovative probabilistic prediction method using a Variational Autoencoder (VAE) combined with Gated Recurrent Unit (GRU) to forecast landslide deformations from a generative standpoint. Our approach consists of two main elements: firstly, training the VAE-GRU model to maximize the variational lower bound on the likelihood of historical precipitation data; secondly, using the learned approximated posterior distribution to predict imminent deformations from a generative angle. To assess the prediction quality, we use four widely-used metrics: Prediction Interval Coverage Probability (PICP), Prediction Interval Normalized Average Width (PINAW), Coverage Width-Based Criterion (CWC), and Prediction Interval Normalized Root Mean Square Width (PINRW). The results demonstrate that our proposed VAE-GRU framework surpasses traditional state-of-the-art (SOTA) probabilistic deformation prediction algorithms in terms of accuracy and reliability.

The Baishi landslide was located in the western part of Beichuan County, Sichuan Province, China. The landslide experienced multiple minor collapses at the front part, accompanying with numerous tensile cracks. To comprehensively grasp the stability conditions and predict the moment of failure of the landslide, deformation monitoring of the landslide has been carried out from the moment that the landslide was reported until it failed. This study analyzed the different phases of landslide deformation and its failure mechanism through the analysis of monitoring data. The result showed that the failure manifests both the retrogressive and advancing features. The landslide was divided into several zones based on the spatial variation of the deformation characteristics. Moreover, the improved tangential angle criterion is applied to categorize the deformation phases of a landslide. Investigating the surface displacement vectors and vector angles of landslides plays a significant role for ascertaining the failure and sliding mechanism. The monitoring results revealed that the front part of the landslide played a key role in the stability of the landslide. Therefore, the monitoring data from this zone were crucial for predicting the moment of complete landslide failure.

This study investigates the impact of long-term water immersion on the mechanical and hydrochemical properties of cemented calcareous soil, emphasizing the critical role of carbonate content in mechanical performance. Utilizing hydrochemical analysis and triaxial testing, the research revealed that prolonged immersion disrupts the acid-base balance of the solution, resulting in an increased concentration of ions and chemicals. Significant dissolution of carbonates and soluble minerals occurs, which reacts with carbon dioxide to generate bicarbonate ions, thereby elevating the alkalinity of the soaking solution. Additionally, the gradual dissolution of clay minerals compromises the cementitious structure, leading to particle reorientation and interlocking. The study quantitatively assesses the changes in soil properties, demonstrating a substantial reduction in soil cohesion by up to 86.1% and an increase in the internal friction angle by 37.5%. Furthermore, the gradual dissolution of clay minerals compromises the cementitious structure, resulting in particle reorientation and interlocking that contribute to the observed mechanical changes. The findings underscore the importance of understanding the effects of extended immersion on the stability and engineering applicability of cemented calcareous soils.

This paper systematically investigates the dynamic mechanical response and fragment characteristics of sandstone with asymmetrical conjugate fissures subjected to preexisting static stress based on the split Hopkinson pressure bar apparatus. The cross-fissured sandstone exhibits significant variations in mechanical behavior when exposed to higher dynamic strain rates under identical static pre-stress. The cross-fissured sandstone with a higher dynamic strain rate is characterized by a greater coupled strength under the same static pre-stress; for a given dynamic load, the highest coupled strength occurs under the static pre-stress of 60% UCS. The failure mode of the cross-fissured sandstone is predominantly governed by the dynamic strain rate, independent of static pre-stress variations considered in this investigation. At lower dynamic strain rates, the specimens typically exhibit a mixed tensile-shear failure mode, characterized by the dominance of larger fragments in the broken specimens. In contrast, under high dynamic impacts, the sandstone tends to fail in a shear-dominated manner, resulting in smaller fragments with a more uniform size distribution. Furthermore, the study explores how varying dynamic strain rates and static pre-stress influence the fragment characteristics of the cross-fissured sandstone. Higher dynamic strain rates and increased static pre-stress generally lead to smaller mean fragment sizes. This phenomenon is quantitatively described by fitting fragment size distributions using the Generalized Extreme Value (GEV) distribution, revealing a decrease in the location parameter (μ) and an increase in fractal dimension. These metrics indicate that higher dynamic strain rates and static pre-stress result in sandstone specimens breaking into smaller fragments with a more homogeneous size distribution. These findings contribute to a deeper understanding of rock dynamics, with potential implications for engineering applications involving similar geological configuration under dynamic loading conditions.

Preexisting fissures within a natural rock mass significantly affect the mechanical responses of underground engineering rocks when subjected to coupled static-dynamic loads. Understanding the dynamic mechanical responses of prestressed fissured rocks is beneficial for optimizing the stability of rock engineering structures. Based on the split Hopkinson pressure bar (SHPB) device, a series of coupled static-dynamic compression tests were systematically performed to investigate the mechanical properties, the failure behaviors, the fragment characteristics, and the energy dissipation of unparallel-fissured rocks. Results show that the coupled dynamic strength increases with the applied strain rate for a given static prestress, while the elastic modulus does not exhibit any clear loading rate dependence. Under the given strain rate, the fissured sandstone has the highest strength under the static prestress of 0.6 uniaxial compressive strength (σ_s), followed by the static prestress of 0.4 σ_s, 0.2 σ_s, and 0.8 σ_s. The failure modes of fissured sandstone highly depend on the dynamic strain rate, and the ultimate failure pattern gradually changes from the mixed tensile-shear failure mode to the shear failure dominant mode with increasing strain rate. With an increasing dynamic strain rate, fissured rocks should absorb more energy in a short time, which results in more crack formation in the failure process. A higher static prestress can induce greater energy dissipation density and energy utilization efficiency when the static prestress is less than 0.8 σ_s. In addition, the relationships between energy dissipation and mechanical response of fissured sandstones under coupled static-dynamic loads are established and discussed.