Bulletin of Engineering Geology and the Environment最新文献

Earthquake-induced landslides pose significant threats to human life and property. Assessing their hazard reveals the probability of regional landslide occurrence under seismic action, providing a basis for pre-earthquake planning and post-earthquake rescue operations. This study integrates landslide data from 26 historical earthquakes through efficient data preprocessing. A Fault Direction Effect (FDE) factor is introduced, quantitatively characterizing the angle between natural slope aspect and fault rupture direction, establishing a seismic geological effect evaluation index with defined physical significance. We constructed a multi-factor database comprising 18 disaster-inducing factors and conducted statistical analyses of the spatial distribution of earthquake-induced landslides. Based on the Newmark model, we developed eighteen coupled Newmark and machine learning models suitable for shallow translational failures, hereafter referred to as the Newmark_i-X dual-drive models. Using metrics such as AUC, precision, recall, F1 score, and Kappa coefficient, supplemented by leave-one-out cross-validation, and taking the 2008 Wenchuan, 2015 Nepal, and 2010 Haiti earthquake data, the best performing model (improved Newmark coupled with Backpropagation Neural Network(BPNN) incorporating peak ground acceleration (PGA)) showed significant performance improvements: pre-earthquake prediction AUC reached 0.88 and post-earthquake inversion AUC 0.94, outperforming traditional single-method models in both accuracy and stability. By employing big data analysis, machine learning, and model integration techniques, this study develops a multi-technique earthquake-induced landslide hazard evaluation system, offering quantitative support for regional disaster risk management.

Loess collapse is highly concealed and abrupt. Consequently, traditional displacement monitoring is ineffective in capturing early instability signals and disaster warning timeliness is poor, a key challenge in loess disaster prevention. Hence, this paper focused on nonforce field monitoring and uses acoustic emission (AE) technology to conduct uniaxial compression tests on Malan loess. Tests explored AE responses during loess deformation/failure, analyzed AE parameter anomalies near failure, and established precursor criteria. Results show distinct AE signals throughout the process are strongly correlated with failure stages. Near peak strength, the AE ringing count rises rapidly and peaks before specimen failure, which confirms AE anomalies precede macroscopic strain changes. Before peak strength, the AE b-value drops abruptly, high-frequency signals disappear, and low-frequency signals surge. This paper identified three reliable AE-based failure precursors: sudden increase in AE ringing count, abrupt b-value drop, and “high-frequency disappearance–low-frequency surge.” This paper confirmed AE monitoring’s feasibility for loess collapse early warning and provides a technical basis to enhance disaster prevention efficiency and protect infrastructure/ecology in loess regions.

The three-bench method is extensively applied in deeply buried soft rock tunnel construction. Its multiple construction steps cause significant spatial and temporal effects on surrounding rock stress evolution, which brings challenges for tunnel excavation and support design. This study monitors stress evolution of Muzhailing highway tunnel via the Rheological Stress Recovery method, conducts creep-based numerical simulation for entire excavation process, and compares results using the Stress Disturbance Index (SDI) analysis method. The results reveal: (1) Stress disturbance starts when the upper bench is 8 m from the monitoring cross-section and stabilizes post invert arch lining. (2) The range of the stress disturbance zone is 2.5 times the tunnel radius R₀ (R₀ = 6.48 m), with stress states varying by depth—σ_z > σ_y > σ_x at depths < 0.3 R₀, σ_y > σ_z > σ_x at depths of 0.3–1 R₀, and σ_y > σ_x > σ_z at depths > 1 R₀. (3) Overall, σ_x exhibits a decreasing trend with the excavation progression, while σ_z exhibits an increasing trend. The peak values of σ_x, σ_y, σ_z are 16.74 MPa, 17.02 MPa, and 16.55 MPa, respectively. (4) Middle bench and left lower bench excavations induce significant stress variation amplitudes and stress average variation rates. (5) Through the SDI analysis method, the monitoring and simulation results exhibit good consistency in evolutionary patterns.

Impulse wave generated by the failure of large deposits in reservoirs directly affect dam safety. This study takes a large potential reservoir deposit in southwest China’s Lancang River as an example, determines its failure mode using centrifuge tests, investigates the propagation attenuation characteristics of impulse waves generated by RS deposit failure by constructing 1:200 large-scale physical simulation tests and numerical simulations based on fluid‒solid coupling, and explores the mechanism and influencing factors of impulse wave propagation attenuation in mountainous reservoir channels. The failure of the deposits produced maximum wave amplitude in excess of 30 m in the river channel. In this study, we obtained the ratio of wave amplitude attenuation along the propagation of impulse waves generated by RS deposit failure and analyzed the wave attenuation characteristics: waves are generated, climb toward the opposite shore and then fall back and show periodic attenuation, the wave form propagating downstream gradually breaks irregularly, the waves are in sharp attenuation within 1.5 km from the deposits, after which they slowly attenuate but with small fluctuations in some areas, and the attenuation ratio is the largest in the nearly straight river channel. When the impulse wave propagates in the river channel of mountain reservoirs, its wave amplitude as a whole decreases with increasing propagation distance, but the degree of attenuation is different in different areas. Its propagation is mainly affected by the obstruction of the bend, frictional resistance, bank bounce interference, and the width change of the river channel.

River blockage serves as a critical link in the cascading hazards of post-earthquake debris flows. The early identification of potential debris flows that may block the river is essential for post-earthquake disaster prevention and risk reduction. Based on multi-source remote sensing images, field surveys and data collection, we compiled a dataset of river blockage events caused by debris flows following the 2008 Wenchuan earthquake. Simultaneously, we identified the primary controlling factors of the river blockages by post-earthquake debris flows and developed the prediction models based on data-driven and artificial intelligence. The results show that the root mean square error of Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM) were 0.141, 0.134 and 0.086, respectively. Among the three models, LightGBM performed with higher prediction accuracy and lower variability. Based on Shapley Additive Explanation (SHAP), we found that the co-seismic landslide area connectivity (LSN) and the slope of the deposition area (FANS) are the first-order dominant controlling factors of the river blockage hazards caused by post-earthquake debris flows, followed by the topographic roughness index (TRI), the stream power index (SPI), the average slope of the catchment (CAS), the hypsometric integral (HI), the flow accumulation ratio (FR), and the stream steepness index (KSN). Subsequently, we further examined the relationship between the debris flow river-blockage index (PBR) and its influencing factors, and determined factor thresholds that correspond to the high hazard of river blockage. This research can provide insights for the prediction and risk mitigation of river blockage hazards caused by post-earthquake debris flows.

In recent years, it has been pointed out that there is a relative aquifuge in the top portion of the Ordovician limestone aquifer. To further explore the water-resisting property of Ordovician limestone, systematic tests were conducted on the O-7 borehole samples, including well logging curves analysis, physical and mechanical property tests, X-ray diffraction (XRD), nuclear magnetic resonance (NMR), and scanning electron microscopy (SEM). The results confirmed the existence of the relatively aquifuge. Compared to conventional Ordovician limestone samples, the presence of minerals such as kaolinite, a large proportion of micropores, and the filling of fissures under the microscopic structure jointly revealed the mechanism of enhanced water-resisting property: macropores and fissures were filled by weathering substances and decomposed into mesopores or micropores, thereby forming an impermeable texture in the rock, and thus creating an impermeable structure on a large scale. Based on the quantified thickness of the aquifuge, the potential risk of the lower coal seam mining in Yangcun Coal Mine was assessed using the water inrush coefficient method. The existence of the aquifuge would expand the field coverage of the mining safety zone. This indicates that the existence of the aquifuge and further quantification has great practical significance for reducing the management cost of the Ordovician limestone aquifer and raising the mining limit.

Alpine grasslands in the seasonally frozen source region of the Yellow River experience severe sandification, intensified by freeze‒thaw (FT) cycles. In this study, the efficacy of two organic stabilizers, polyacrylamide (PAM) and modified polyvinyl alcohol (SH), in reinforcing local sandy soil was investigated. Through laboratory experiments, including direct shear tests, unconfined compressive strength (UCS) tests, X-ray diffraction (XRD), scanning electron microscopy (SEM), and nuclear magnetic resonance (NMR), the mechanical properties and microstructural evolution of the stabilized soil in terms of different additive dosages and numbers of FT cycles were evaluated. Both stabilizers significantly increased the cohesion and UCS of the soil before FT cycling. After ten cycles, compared with PAM, SH performed markedly better, exhibiting lower cohesion loss and a unique, substantial increase in UCS. Microstructural analysis indicated that although both stabilizers initially reinforced the soil through encapsulation, cementation, and entanglement, they subsequently experienced macroscopic degradation due to adverse pore structure evolution, particularly an increase in total and macroporosity that occurred primarily during the first two cycles. In contrast, SH developed a denser and more robust network under FT cycling, which presumably accounted for its distinct increase in UCS. Considering both mechanical performance and cost, a 3% dosage of SH was recommended for sand fixation projects in the Yellow River source region.

The subgrade clay experiences cyclic dynamic loads from passing vehicles and intermittent stages when no vehicles are present. In this study, the deformation characteristics of freeze–thaw silty clay in the Xining subgrade under intermittent stages/cyclic loading were analyzed by dynamic triaxial test considering different loading patterns. The results show that the deformation of silty clay increases as the cyclic dynamic stress ratio (CSR) increases. Additionally, the deformation further increases as the freezing temperature decreases. Furthermore, the intermittent stage facilitates partial recovery of axial strain. Pore pressure, which show hysteresis compared to axial strain, are governed by the degree of freeze–thaw deterioration and stress state. By combining the time hardening method with multiple regression analysis, a deformation prediction model that incorporates freeze–thaw cycles and intermittent stages is proposed, and model predictions closely match experimental results. An evaluation method to quantify the intermittent behavior is established, which focuses on analyzing deformation evolution during the intermittent stage and reveals how pore pressure regulates axial strain. This study can provide a theoretical reference for settlement prediction under traffic loads and long-term performance assessment of subgrade engineering in cold regions.

Slope failure poses a significant threat to infrastructure, particularly in regions with heterogeneous soil profiles and steep embankments. This study explores the use of vetiver grass (Chrysopogon zizanioides) as a sustainable bioengineering solution to enhance slope stability across varying geotechnical conditions. The stabilizing effects of vetiver root systems were evaluated on slopes composed of laterite soil and sand-laterite mixtures using GeoSlope’s SLOPE/W module. Slope configurations with angles of 10°, 20°, 30°, 45°, 60°, and 75° were analyzed to assess performance under diverse geometries. Soil properties—cohesion, internal friction angle, and unit weight—were derived from laboratory testing and supplemented with literature values. Vetiver reinforcement was modeled through root-induced cohesion enhancement and adjusted shear strength parameters. The results demonstrate that vetiver grass significantly improves the Factor of Safety (FoS) across all tested conditions, particularly in cohesive soils at moderate slope angles. Reinforced slopes showed delayed failure onset, reduced shear strain, and greater resistance to sliding. The 20% sand-laterite mix with vetiver yielded the most favorable stability outcomes. These findings confirm the potential of Vetiver grass as an environmentally sustainable bioengineering approach for stabilizing vulnerable embankments, especially in areas prone to rainfall-induced failures and erosion. The study contributes to advancing sustainable and practical alternatives to conventional slope reinforcement techniques.

Currently, image processing methods based on segmentation do not achieve sufficient accuracy to reliably capture the actual dimensions of drill core samples, limiting their ability to provide precise quantitative information on the extracted material’s true length. Although core trays are standardized in 50 cm segments, the core samples themselves often vary in size. The measurement of the Length of Intact Rock Core Pieces (LIRCP) is still carried out using empirical tools such as measuring tapes or rulers, and in some cases, unreliable algorithms, leading to operational delays and increased costs despite LIRCP being a key parameter for calculating the Rock Quality Designation (RQD). Furthermore, the core cutting process and LIRCP measurement are typically performed in separate locations. This project aimed to design and implement an integrated system capable of accurately calculating both the actual length and volume of core samples. The methodology was based on the YOLOv11 segmentation model, chosen for its high detection accuracy and speed, from which length and volume were estimated through 2D pixel analysis. This process was supported by experimental validation and statistical error analysis. The system achieved 98% accuracy in estimating both parameters. This study successfully developed and implemented an integrated system capable of accurately calculating the actual length and volume of drill core samples, overcoming the limitations of traditional image processing and manual methods. This innovation significantly improves operational efficiency by reducing time and costs, while providing precise real time data to support informed geological decision making.