Pub Date : 2026-01-31DOI: 10.1016/j.enggeo.2026.108598
Qi Liu , Ze Zhang , Xu Chunguang , Qingkai Yan , Zhiyuan Wang , Yaqi Zhang , Anderi Zhang , Torgovkin Nikolai
Freeze–thaw cycles (FTCs) profoundly influence the mechanical stability of sandy soils in cold regions, yet the micro-mechanisms governing their grain size and morphology evolution remain insufficiently understood. This study investigates the fragmentation and morphological evolution of quartz sand subjected to 1000 FTCs. The results reveal that freeze–thaw action drives a progressive ‘coarsening-to-fining’ shift in grain size distribution. This transition reaches stage-specific dynamic equilibria through successive particle breakage and abrasion. Morphologically, angular particles undergo selective edge abrasion, exhibiting a progressive transition from angular to structurally regular geometries. We identify a critical particle size for freeze-thaw weathering of quartz that lies within the coarse silt range (0.01–0.05 mm). The fining process reflects a gradual reduction of lattice defects until particles reach a stable size with enhanced resistance to breakage. These findings systematically elucidate the particle fragmentation supply behavior of quartz sand under freeze-thaw weathering and provide a microstructural basis for improving predictive models of freeze-thaw related geohazards in cold regions.
{"title":"Progressive modification of quartz sand under freeze-thaw weathering: Identification of critical particle size","authors":"Qi Liu , Ze Zhang , Xu Chunguang , Qingkai Yan , Zhiyuan Wang , Yaqi Zhang , Anderi Zhang , Torgovkin Nikolai","doi":"10.1016/j.enggeo.2026.108598","DOIUrl":"10.1016/j.enggeo.2026.108598","url":null,"abstract":"<div><div>Freeze–thaw cycles (FTCs) profoundly influence the mechanical stability of sandy soils in cold regions, yet the micro-mechanisms governing their grain size and morphology evolution remain insufficiently understood. This study investigates the fragmentation and morphological evolution of quartz sand subjected to 1000 FTCs. The results reveal that freeze–thaw action drives a progressive ‘coarsening-to-fining’ shift in grain size distribution. This transition reaches stage-specific dynamic equilibria through successive particle breakage and abrasion. Morphologically, angular particles undergo selective edge abrasion, exhibiting a progressive transition from angular to structurally regular geometries. We identify a critical particle size for freeze-thaw weathering of quartz that lies within the coarse silt range (0.01–0.05 mm). The fining process reflects a gradual reduction of lattice defects until particles reach a stable size with enhanced resistance to breakage. These findings systematically elucidate the particle fragmentation supply behavior of quartz sand under freeze-thaw weathering and provide a microstructural basis for improving predictive models of freeze-thaw related geohazards in cold regions.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108598"},"PeriodicalIF":8.4,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146095632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-31DOI: 10.1016/j.enggeo.2026.108599
A.L. Achu , C.D. Aju , Jobin Thomas , Girish Gopinath
Landslide activity is driven by complex interactions among geo-environmental factors, yet most machine-learning-based landslide susceptibility models primarily rely on topographic variables derived from digital elevation models (DEMs), often neglecting the role of depth-dependent soil profile characteristics. This study addresses this gap and advances landslide predictive capabilities by integrating soil geotechnical and hydrological properties at multiple soil profile depths using a Random Forest (RF) model framework coupled with the Shapely additive explanations (SHAP)-based explainable artificial intelligence (XAI) for model interpretability. Demonstrated in the Muthirapuzha River Basin (MRB) of the southern Western Ghats (SWG), India, the study compares both grid unit-based (GUB) and slope unit-based (SUB) mapping approaches. Results suggest that integrating soil properties at multiple depths (10 cm, 110 cm, and 210 cm) significantly improves model accuracy and minimises overestimation compared to a model relying solely on topographic variables. Key predictors included field capacity (FCY), chemical index of alteration (CIA), liquid limit (LLT), and unsaturated hydraulic conductivity (Kunsat), alongside topographic factors, such as slope angle and topographic wetness index (TWI). SUB approach outperforms GUB in terms of area under the receiver operating characteristic curve (AUROC) and provides a better understanding of landslide depth and volume. SHAP values and waterfall plots are critical in interpreting model predictions and elucidating feature contributions, enhancing their potential for site-specific landslide risk assessments. The consistency of variable importance rankings across mapping units further reinforces the robustness of the selected predictors. This study highlights the critical role of soil profile characteristics in landslide susceptibility modelling and advocates integrating XAI techniques to enable transparent, physically meaningful predictions in mountainous regions.
{"title":"Catena matters: Enhancing landslide prediction with soil profile characteristics and explainable AI","authors":"A.L. Achu , C.D. Aju , Jobin Thomas , Girish Gopinath","doi":"10.1016/j.enggeo.2026.108599","DOIUrl":"10.1016/j.enggeo.2026.108599","url":null,"abstract":"<div><div>Landslide activity is driven by complex interactions among geo-environmental factors, yet most machine-learning-based landslide susceptibility models primarily rely on topographic variables derived from digital elevation models (DEMs), often neglecting the role of depth-dependent soil profile characteristics. This study addresses this gap and advances landslide predictive capabilities by integrating soil geotechnical and hydrological properties at multiple soil profile depths using a Random Forest (RF) model framework coupled with the Shapely additive explanations (SHAP)-based explainable artificial intelligence (XAI) for model interpretability. Demonstrated in the Muthirapuzha River Basin (MRB) of the southern Western Ghats (SWG), India, the study compares both grid unit-based (GUB) and slope unit-based (SUB) mapping approaches. Results suggest that integrating soil properties at multiple depths (10 cm, 110 cm, and 210 cm) significantly improves model accuracy and minimises overestimation compared to a model relying solely on topographic variables. Key predictors included field capacity (FCY), chemical index of alteration (CIA), liquid limit (LLT), and unsaturated hydraulic conductivity (K<sub>unsat</sub>), alongside topographic factors, such as slope angle and topographic wetness index (TWI). SUB approach outperforms GUB in terms of area under the receiver operating characteristic curve (AUROC) and provides a better understanding of landslide depth and volume. SHAP values and waterfall plots are critical in interpreting model predictions and elucidating feature contributions, enhancing their potential for site-specific landslide risk assessments. The consistency of variable importance rankings across mapping units further reinforces the robustness of the selected predictors. This study highlights the critical role of soil profile characteristics in landslide susceptibility modelling and advocates integrating XAI techniques to enable transparent, physically meaningful predictions in mountainous regions.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108599"},"PeriodicalIF":8.4,"publicationDate":"2026-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146095631","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-30DOI: 10.1016/j.enggeo.2026.108602
Duo Wang , Qin Zhang , Guanwen Huang , Yuan Du
The highly nonlinear and spatiotemporal nature of landslide deformation poses significant challenges to the accurate estimation of landslide creep parameters. This study proposes a novel GNSS-based method to invert landslide creep parameters, integrating both spatial and temporal characteristics. First, the Burgers creep constitutive model is employed to describe the time-dependent deformation behavior of the landslide. Next, an orthogonal experimental design is used to conduct numerical creep simulations and generate synthetic displacement time series for model training. Based on these data, a Spatiotemporal Graph Convolutional Network (STGCN) is constructed to capture both spatial correlations and temporal dynamics. Finally, the inverted parameters are validated through forward numerical simulations. The case study results indicate that the Burgers creep constitutive model effectively reproduces nonlinear creep behavior and captures the spatial evolution of deformation. The simulated results show close agreement with the monitored displacements, yielding an average Mean Absolute Error (MAE) of 0.010 m. Compared with the traditional back-propagation neural network (BPNN), the STGCN reduces the MAE by 54.5%, thereby confirming the reliability of the proposed method. The results demonstrate that this approach provides a powerful tool for simulating the spatiotemporal evolution of landslides.
{"title":"STGCN-based inversion of landslide creep parameters using GNSS displacement time series","authors":"Duo Wang , Qin Zhang , Guanwen Huang , Yuan Du","doi":"10.1016/j.enggeo.2026.108602","DOIUrl":"10.1016/j.enggeo.2026.108602","url":null,"abstract":"<div><div>The highly nonlinear and spatiotemporal nature of landslide deformation poses significant challenges to the accurate estimation of landslide creep parameters. This study proposes a novel GNSS-based method to invert landslide creep parameters, integrating both spatial and temporal characteristics. First, the Burgers creep constitutive model is employed to describe the time-dependent deformation behavior of the landslide. Next, an orthogonal experimental design is used to conduct numerical creep simulations and generate synthetic displacement time series for model training. Based on these data, a Spatiotemporal Graph Convolutional Network (STGCN) is constructed to capture both spatial correlations and temporal dynamics. Finally, the inverted parameters are validated through forward numerical simulations. The case study results indicate that the Burgers creep constitutive model effectively reproduces nonlinear creep behavior and captures the spatial evolution of deformation. The simulated results show close agreement with the monitored displacements, yielding an average Mean Absolute Error (MAE) of 0.010 m. Compared with the traditional back-propagation neural network (BPNN), the STGCN reduces the MAE by 54.5%, thereby confirming the reliability of the proposed method. The results demonstrate that this approach provides a powerful tool for simulating the spatiotemporal evolution of landslides.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108602"},"PeriodicalIF":8.4,"publicationDate":"2026-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146095633","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.enggeo.2026.108597
Weiwei Zhu , Shengwen Qi , Xupeng He , Bowen Zheng , Songfeng Guo , Yu Zou , Wenhai Lei , Wang Zhang , Hussein Hoteit , Moran Wang , Manchao He , Wenjiao Xiao
Fracture network connectivity fundamentally controls subsurface fluid flow and rock mass behavior across spatial scales, yet determining the representative elementary volume (REV) remains a core challenge in geological system characterization. This study investigates scale-dependent connectivity through systematic analysis of natural outcrop data and artificial discrete fracture networks (DFNs). We implement a novel connectivity metric, , integrating both intra-cluster connectivity and inter-cluster interactions, and propose the Standard Deviation Stability Criterion (SDSC) for objective REV determination using second-order statistical measures. Analysis of 63 natural outcrop maps and various artificial DFN configurations reveals several key findings. First, fracture network connectivity exhibits pronounced scale-dependence with REV values approaching the same order of magnitude as the investigated systems, with mean REV values of 0.586 for natural outcrops and exceeding 0.2 for artificial networks. Second, preferential orientations increase REV requirements, particularly under stress conditions where only critically stressed fractures remain permeable, with fracture clustering further amplifying this effect. Third, in-situ stress conditions substantially increase REV requirements, with values nearly doubling when only critically stressed fractures remain active. Complete sealing creates the most challenging REV determination due to orientation selectivity, while partial sealing provides intermediate behavior by preserving orientation diversity. These findings demonstrate that obtaining representative volumes through conventional sampling presents fundamental limitations and provide critical insights for enhancing predictive models in subsurface engineering and environmental applications.
{"title":"Scale-dependent connectivity behavior in multi-clustered fracture systems","authors":"Weiwei Zhu , Shengwen Qi , Xupeng He , Bowen Zheng , Songfeng Guo , Yu Zou , Wenhai Lei , Wang Zhang , Hussein Hoteit , Moran Wang , Manchao He , Wenjiao Xiao","doi":"10.1016/j.enggeo.2026.108597","DOIUrl":"10.1016/j.enggeo.2026.108597","url":null,"abstract":"<div><div>Fracture network connectivity fundamentally controls subsurface fluid flow and rock mass behavior across spatial scales, yet determining the representative elementary volume (REV) remains a core challenge in geological system characterization. This study investigates scale-dependent connectivity through systematic analysis of natural outcrop data and artificial discrete fracture networks (DFNs). We implement a novel connectivity metric, <span><math><msub><mrow><mi>C</mi></mrow><mrow><mi>t</mi></mrow></msub></math></span>, integrating both intra-cluster connectivity and inter-cluster interactions, and propose the Standard Deviation Stability Criterion (SDSC) for objective REV determination using second-order statistical measures. Analysis of 63 natural outcrop maps and various artificial DFN configurations reveals several key findings. First, fracture network connectivity exhibits pronounced scale-dependence with REV values approaching the same order of magnitude as the investigated systems, with mean REV values of 0.586 for natural outcrops and exceeding 0.2 for artificial networks. Second, preferential orientations increase REV requirements, particularly under stress conditions where only critically stressed fractures remain permeable, with fracture clustering further amplifying this effect. Third, in-situ stress conditions substantially increase REV requirements, with values nearly doubling when only critically stressed fractures remain active. Complete sealing creates the most challenging REV determination due to orientation selectivity, while partial sealing provides intermediate behavior by preserving orientation diversity. These findings demonstrate that obtaining representative volumes through conventional sampling presents fundamental limitations and provide critical insights for enhancing predictive models in subsurface engineering and environmental applications.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108597"},"PeriodicalIF":8.4,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146071827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.enggeo.2026.108596
Bowen Tai , Zurun Yue , Pengcheng Wang , Jingpeng Liu
The exacerbation of frost damage in subgrade structures of high-speed railways (HSR) in cold regions, often triggered by extreme climatic events such as severe cold spells, heavy snowfall, and intense rainfall infiltration. To ensure the operational integrity of HSR in seasonally frozen soil regions, it is imperative to investigate the impacts of extreme climate conditions on the stability of typical anti-frost subgrades. This study employs an integrated methodology combining field monitoring, model development, numerical simulations, and theoretical analysis. First, the differential influences of various climatic scenarios on the hydrothermal behavior of seasonally frozen soil are examined. Subsequently, the coupled water-heat-deformation characteristics of a standard anti-frost subgrade structure are analyzed, leading to the development of a novel fully coupled water-heat-strain model. Finally, the model is utilized to predict and assess the structural stability under extreme climate events. Key findings include: (1) marked differential responses in the hydrothermal regime of seasonally frozen soil under varying climate conditions; (2) a time-lag in variations of temperature and moisture with increasing depth; (3) synergistic effects of compound extreme weather events significantly aggravate subgrade damage; and (4) the necessity of holistic consideration of extreme climate, engineering geological conditions and slope effect in the optimal design of anti-frost layers. These insights not only advance the mechanistic understanding of frost deformation processes under extreme climate, but also provide valuable guidelines for the optimized design of anti-frost infrastructures in cold regions.
{"title":"Investigating the catastrophe mechanism and evolution of anti-frost subgrade in high-speed railways under extreme climatic events","authors":"Bowen Tai , Zurun Yue , Pengcheng Wang , Jingpeng Liu","doi":"10.1016/j.enggeo.2026.108596","DOIUrl":"10.1016/j.enggeo.2026.108596","url":null,"abstract":"<div><div>The exacerbation of frost damage in subgrade structures of high-speed railways (HSR) in cold regions, often triggered by extreme climatic events such as severe cold spells, heavy snowfall, and intense rainfall infiltration. To ensure the operational integrity of HSR in seasonally frozen soil regions, it is imperative to investigate the impacts of extreme climate conditions on the stability of typical anti-frost subgrades. This study employs an integrated methodology combining field monitoring, model development, numerical simulations, and theoretical analysis. First, the differential influences of various climatic scenarios on the hydrothermal behavior of seasonally frozen soil are examined. Subsequently, the coupled water-heat-deformation characteristics of a standard anti-frost subgrade structure are analyzed, leading to the development of a novel fully coupled water-heat-strain model. Finally, the model is utilized to predict and assess the structural stability under extreme climate events. Key findings include: (1) marked differential responses in the hydrothermal regime of seasonally frozen soil under varying climate conditions; (2) a time-lag in variations of temperature and moisture with increasing depth; (3) synergistic effects of compound extreme weather events significantly aggravate subgrade damage; and (4) the necessity of holistic consideration of extreme climate, engineering geological conditions and slope effect in the optimal design of anti-frost layers. These insights not only advance the mechanistic understanding of frost deformation processes under extreme climate, but also provide valuable guidelines for the optimized design of anti-frost infrastructures in cold regions.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108596"},"PeriodicalIF":8.4,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072691","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.enggeo.2026.108591
Hongyun Fan , Yuguang Fu , Wei Shen , Xiangyu Chang
Soil-structure interaction (SSI) is commonly encountered in various geohazards such as landslides and debris flows. To understand and mitigate these hazards, it is essential to simulate the interaction between soil and structures with accuracy. However, existing coupled numerical methods often represent structural motion using particle-based models, which limits their ability to precisely capture the dynamic interaction mechanisms between soil and structures. To address this limitation, this study proposes a novel coupled simulation framework that integrates the three-dimensional Discontinuous Deformation Analysis (3D DDA) with the Material Point Method (3D MPM), leveraging the strengths of 3D DDA in modeling structural motion and the capability of MPM in capturing large deformation of geomaterials. First, a contact detection and force computation scheme between MPM particles and DDA blocks is established by incorporating bounding box techniques and a penalty spring model, enabling accurate simulation of soil–structure interaction processes. Subsequently, the proposed coupling method is applied to simulate a series of benchmark scenarios, including soil collapse, soil collapse with embedded blocks, block impact on soil, and soil impact on blocks. The simulation results are validated against experimental data, demonstrating the accuracy and robustness of the proposed approach. Finally, the coupling method is employed to investigate the collapse behavior of buildings subjected to landslide impact, with a particular focus on the influence of landslide height on structural collapse mechanisms. By clarifying the underlying mechanisms, the findings contribute theoretical knowledge that supports efforts to prevent and mitigate landslide-induced hazards.
{"title":"A coupled 3D DDA-MPM framework for soil-structure interaction modeling and its application in geotechnical hazards modeling","authors":"Hongyun Fan , Yuguang Fu , Wei Shen , Xiangyu Chang","doi":"10.1016/j.enggeo.2026.108591","DOIUrl":"10.1016/j.enggeo.2026.108591","url":null,"abstract":"<div><div>Soil-structure interaction (SSI) is commonly encountered in various geohazards such as landslides and debris flows. To understand and mitigate these hazards, it is essential to simulate the interaction between soil and structures with accuracy. However, existing coupled numerical methods often represent structural motion using particle-based models, which limits their ability to precisely capture the dynamic interaction mechanisms between soil and structures. To address this limitation, this study proposes a novel coupled simulation framework that integrates the three-dimensional Discontinuous Deformation Analysis (3D DDA) with the Material Point Method (3D MPM), leveraging the strengths of 3D DDA in modeling structural motion and the capability of MPM in capturing large deformation of geomaterials. First, a contact detection and force computation scheme between MPM particles and DDA blocks is established by incorporating bounding box techniques and a penalty spring model, enabling accurate simulation of soil–structure interaction processes. Subsequently, the proposed coupling method is applied to simulate a series of benchmark scenarios, including soil collapse, soil collapse with embedded blocks, block impact on soil, and soil impact on blocks. The simulation results are validated against experimental data, demonstrating the accuracy and robustness of the proposed approach. Finally, the coupling method is employed to investigate the collapse behavior of buildings subjected to landslide impact, with a particular focus on the influence of landslide height on structural collapse mechanisms. By clarifying the underlying mechanisms, the findings contribute theoretical knowledge that supports efforts to prevent and mitigate landslide-induced hazards.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108591"},"PeriodicalIF":8.4,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072685","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-29DOI: 10.1016/j.enggeo.2026.108601
Sofie Axéen, Johanna Merisalu, Ezra Haaf, Lars Rosén
Leakage of groundwater and subsequent pore pressure reduction can cause consolidation in subsidence sensitive soils and subsequently pose damage risks to the built environment. This study presents the first systematic, quantitative evaluation of how geological conceptualization – specifically the inclusion or exclusion of permeable sand lenses within glaciomarine clay deposits - affects simulated pore pressure reduction due to groundwater leakage into deep excavations. By employing Multiple Point Statistics (MPS) to generate alternative geological models and integrating these with MODFLOW-NWT transient groundwater simulations, we reveal that the presence and hydraulic connectivity of sand lenses significantly influence the rate and magnitude of pore pressure reduction in clay, which has significant consequences for settlement magnitudes. These findings underscore the importance of explicitly accounting for geological heterogeneity and uncertainty in risk assessment for urban excavations, a factor often neglected in conventional engineering geology practice when assessing settlement hazards and their consequences for the surrounding areas.
{"title":"Impact of geological conceptualization in predicting pore pressure reduction from urban excavations","authors":"Sofie Axéen, Johanna Merisalu, Ezra Haaf, Lars Rosén","doi":"10.1016/j.enggeo.2026.108601","DOIUrl":"10.1016/j.enggeo.2026.108601","url":null,"abstract":"<div><div>Leakage of groundwater and subsequent pore pressure reduction can cause consolidation in subsidence sensitive soils and subsequently pose damage risks to the built environment. This study presents the first systematic, quantitative evaluation of how geological conceptualization – specifically the inclusion or exclusion of permeable sand lenses within glaciomarine clay deposits - affects simulated pore pressure reduction due to groundwater leakage into deep excavations. By employing Multiple Point Statistics (MPS) to generate alternative geological models and integrating these with MODFLOW-NWT transient groundwater simulations, we reveal that the presence and hydraulic connectivity of sand lenses significantly influence the rate and magnitude of pore pressure reduction in clay, which has significant consequences for settlement magnitudes. These findings underscore the importance of explicitly accounting for geological heterogeneity and uncertainty in risk assessment for urban excavations, a factor often neglected in conventional engineering geology practice when assessing settlement hazards and their consequences for the surrounding areas.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108601"},"PeriodicalIF":8.4,"publicationDate":"2026-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072686","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-28DOI: 10.1016/j.enggeo.2026.108595
Jiewei Zhan , Changle Pu , Zhaoyue Yu , Yongqiang Liu , Jianbing Peng
The discrete fracture network (DFN) modeling technique is a critical method for revealing the three-dimensional structural characteristics of rock masses and predicting the connectivity and stability of fractured rock masses. Constrained by the dominant bedding planes, discontinuities in layered rock masses often intersect with bedding planes to form characteristic T-type topological structures. Considering that existing DFN modeling techniques are unable to accurately reproduce this structural characteristic, this paper proposes an innovative hierarchical method for spatial structural modeling of layered rock masses. First, a three-dimensional fusion model of outcrop is constructed using optical images and point cloud data collected by UAV photogrammetry, on which the geometric parameters of discontinuities are extracted. On the basis of the interpreted discontinuity data, a characterization study is subsequently conducted on the orientation, major axis rotation angle, size, and spatial point distribution of the discontinuities. By introducing a hierarchical modeling method based on the sequence of bedding planes, strata-bound discontinuities and non-strata-bound discontinuities, the limitations of traditional methods in simulating the unique intersection relationships of discontinuities in layered rock masses is effectively addressed. In addition, the Latin hypercube sampling is employed to determine the position of non-strata-bound discontinuities, which effectively reduces the edge effects in the DFN modeling process. Finally, a layered rock mass discrete fracture network model is constructed using an outcrop from a highway slope in Chongqing as a case study, and the effectiveness of the proposed method is validated through both geometric characterization and topological structure analysis. This work provides a universal methodology for spatial structural modeling of layered rock masses and has good application prospects.
{"title":"Modeling the spatial structural network of layered rock masses using an innovative hierarchical method","authors":"Jiewei Zhan , Changle Pu , Zhaoyue Yu , Yongqiang Liu , Jianbing Peng","doi":"10.1016/j.enggeo.2026.108595","DOIUrl":"10.1016/j.enggeo.2026.108595","url":null,"abstract":"<div><div>The discrete fracture network (DFN) modeling technique is a critical method for revealing the three-dimensional structural characteristics of rock masses and predicting the connectivity and stability of fractured rock masses. Constrained by the dominant bedding planes, discontinuities in layered rock masses often intersect with bedding planes to form characteristic T-type topological structures. Considering that existing DFN modeling techniques are unable to accurately reproduce this structural characteristic, this paper proposes an innovative hierarchical method for spatial structural modeling of layered rock masses. First, a three-dimensional fusion model of outcrop is constructed using optical images and point cloud data collected by UAV photogrammetry, on which the geometric parameters of discontinuities are extracted. On the basis of the interpreted discontinuity data, a characterization study is subsequently conducted on the orientation, major axis rotation angle, size, and spatial point distribution of the discontinuities. By introducing a hierarchical modeling method based on the sequence of bedding planes, strata-bound discontinuities and non-strata-bound discontinuities, the limitations of traditional methods in simulating the unique intersection relationships of discontinuities in layered rock masses is effectively addressed. In addition, the Latin hypercube sampling is employed to determine the position of non-strata-bound discontinuities, which effectively reduces the edge effects in the DFN modeling process. Finally, a layered rock mass discrete fracture network model is constructed using an outcrop from a highway slope in Chongqing as a case study, and the effectiveness of the proposed method is validated through both geometric characterization and topological structure analysis. This work provides a universal methodology for spatial structural modeling of layered rock masses and has good application prospects.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108595"},"PeriodicalIF":8.4,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-28DOI: 10.1016/j.enggeo.2026.108589
Nirandoal Cheng , Mohd Ashraf Mohamad Ismail , Fatin Nadhirah Ahmad Pauzi , Yasuhiro Yokota , Hayato Tobe
Accurate mapping of rock mass discontinuities is critical for geotechnical assessments but remains challenging in steep or complex terrains using conventional or oblique photogrammetry. This study introduces a multiscale videogrammetry approach integrating UAV-mounted and handheld 4 K video capture to produce high-resolution 3D models. A coded target-based semi-georeferencing tool enables spatial alignment in a local coordinate system without GNSS. Point clouds were analyzed using semi-automated plane detection, supported by manual trace mapping and stereonet-based clustering. The multiscale model achieved a ground sampling distance of 0.27 cm/pixel and point cloud density of 47,000 pts./m2 over 20 times higher than the oblique dataset. Orientation accuracy showed RMSE values of 2.16° for dip and 6.52° for dip direction. Compared to conventional methods, the multiscale approach captured more complete joint distributions and higher structural detail, particularly in recessed or overhanging zones. Kinematic analysis revealed a broader range of failure modes, including planar, wedge, and toppling failures. This study demonstrates that multiscale videogrammetry, combined with semi-georeferencing and trace-based analysis, provides a scalable, accurate, and flexible workflow for discontinuity detection and structural interpretation in complex geological environments.
{"title":"Exploring multiscale videogrammetry techniques for analyzing rock mass discontinuities in geological formations","authors":"Nirandoal Cheng , Mohd Ashraf Mohamad Ismail , Fatin Nadhirah Ahmad Pauzi , Yasuhiro Yokota , Hayato Tobe","doi":"10.1016/j.enggeo.2026.108589","DOIUrl":"10.1016/j.enggeo.2026.108589","url":null,"abstract":"<div><div>Accurate mapping of rock mass discontinuities is critical for geotechnical assessments but remains challenging in steep or complex terrains using conventional or oblique photogrammetry. This study introduces a multiscale videogrammetry approach integrating UAV-mounted and handheld 4 K video capture to produce high-resolution 3D models. A coded target-based semi-georeferencing tool enables spatial alignment in a local coordinate system without GNSS. Point clouds were analyzed using semi-automated plane detection, supported by manual trace mapping and stereonet-based clustering. The multiscale model achieved a ground sampling distance of 0.27 cm/pixel and point cloud density of 47,000 pts./m<sup>2</sup> over 20 times higher than the oblique dataset. Orientation accuracy showed RMSE values of 2.16° for dip and 6.52° for dip direction. Compared to conventional methods, the multiscale approach captured more complete joint distributions and higher structural detail, particularly in recessed or overhanging zones. Kinematic analysis revealed a broader range of failure modes, including planar, wedge, and toppling failures. This study demonstrates that multiscale videogrammetry, combined with semi-georeferencing and trace-based analysis, provides a scalable, accurate, and flexible workflow for discontinuity detection and structural interpretation in complex geological environments.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"363 ","pages":"Article 108589"},"PeriodicalIF":8.4,"publicationDate":"2026-01-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072692","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2026-01-27DOI: 10.1016/j.enggeo.2026.108594
Shijiao Yang , Qing Du , Jianchang Yan , Wenhua Liu , Jiancheng Huang , Danli Li
Strain localization, spanning from microscopic mineral fabrics to crustal-scale fault zones, fundamentally controls failure modes in natural geological systems and rock engineering. While individual measurement and modeling techniques have advanced significantly, an integrated framework bridging these approaches remains lacking. This review systematically synthesizes multi-scale measurement technologies, numerical simulation methods, and AI-driven prediction approaches for rock strain localization. Contact-based techniques including strain gauges, LVDT, distributed optical fiber sensing, and acoustic emission are examined alongside non-contact optical methods such as digital image correlation and X-ray computed tomography. Continuum and discontinuum numerical frameworks are compared, and AI methodologies from conventional machine learning to physics-informed neural networks are evaluated, with adaptability analysis for different monitoring data types. Three critical insights emerge: (1) multi-source data fusion is essential under geological heterogeneity; (2) physics-based constraints ensure data-driven model reliability; and (3) a gap persists between post-failure analysis and predictive capability. These findings inform rockburst warning, tunnel support design, slope stability assessment, and reservoir management. This review provides a framework for advancing from phenomenological description to mechanistic prediction and from laboratory understanding to engineering geological application.
{"title":"Strain localization in rock: From multi-scale measurement to AI-driven prediction","authors":"Shijiao Yang , Qing Du , Jianchang Yan , Wenhua Liu , Jiancheng Huang , Danli Li","doi":"10.1016/j.enggeo.2026.108594","DOIUrl":"10.1016/j.enggeo.2026.108594","url":null,"abstract":"<div><div>Strain localization, spanning from microscopic mineral fabrics to crustal-scale fault zones, fundamentally controls failure modes in natural geological systems and rock engineering. While individual measurement and modeling techniques have advanced significantly, an integrated framework bridging these approaches remains lacking. This review systematically synthesizes multi-scale measurement technologies, numerical simulation methods, and AI-driven prediction approaches for rock strain localization. Contact-based techniques including strain gauges, LVDT, distributed optical fiber sensing, and acoustic emission are examined alongside non-contact optical methods such as digital image correlation and X-ray computed tomography. Continuum and discontinuum numerical frameworks are compared, and AI methodologies from conventional machine learning to physics-informed neural networks are evaluated, with adaptability analysis for different monitoring data types. Three critical insights emerge: (1) multi-source data fusion is essential under geological heterogeneity; (2) physics-based constraints ensure data-driven model reliability; and (3) a gap persists between post-failure analysis and predictive capability. These findings inform rockburst warning, tunnel support design, slope stability assessment, and reservoir management. This review provides a framework for advancing from phenomenological description to mechanistic prediction and from laboratory understanding to engineering geological application.</div></div>","PeriodicalId":11567,"journal":{"name":"Engineering Geology","volume":"364 ","pages":"Article 108594"},"PeriodicalIF":8.4,"publicationDate":"2026-01-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146072693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号