Initiated by Dr. Xin Wei, University of Michigan
Ongoing development by the community

TerraMosaic Daily Digest: May 1, 2026

Daily Summary

May 1 is anchored by studies that make water, vegetation, and deformation explicitly observable before they become disaster losses. The strongest landslide contribution reconstructs the Huangya Gully flash flood-debris flow chain as a coupled sequence: antecedent rainfall and vegetation-mediated preferential seepage condition shallow landslide clusters, rapid fluidization initiates the surge, and erosion, scouring, and dam-break effects amplify the final volume. The companion landslide prediction paper moves the same problem into forecasting, showing that hourly antecedent rainfall and satellite leaf-area dynamics can be learned as time-varying controls rather than reduced to static thresholds.

The day broadens from hillslope failure into measurable risk at built and environmental interfaces. Forest protection is formalized as a hydro-mechanical design factor; PS-InSAR is converted from displacement points into building-level deformation signatures; flood susceptibility is mapped nationally for Afghanistan; and UAV, InSAR, satellite gravimetry, tidal-flat altimetry, and multisensor water-quality tools turn sparse or noisy observations into operational hazard products. The engineering papers are less a separate track than the same argument in harder materials: seismic retaining systems, tunnel linings, soft-rock supports, salt caverns, foundations, pipelines, and offshore piles are treated as monitored systems whose risk depends on coupled loading history, spatial variability, and physically constrained inference.

Key Trends

The selected papers point to a practical shift: hazard analysis is being organized around measurable process states rather than static classifications.

  • Slope stability is being recast as coupled ecohydrology: the landslide, forest-protection, vegetated-soil, and fractional-vegetation-cover papers all treat roots, canopy state, suction, infiltration, and rainfall memory as active variables in instability and protection.
  • Remote sensing is becoming diagnostic rather than descriptive: PS-InSAR building signatures, UAV crack localization, mining-subsidence point-cloud filtering, daily GRACE/FO anomalies, annual tidal-flat DEMs, and all-weather mountain LST products convert observations into interpretable states of deformation, mass change, thermal forcing, or structural condition.
  • Risk models are moving from hazard maps to service and network consequences: Afghanistan flood risk, earthquake healthcare flows, gas valve placement, bridge restoration, typhoon grid hardening, warning-siren coverage, and hazardous-chemical routing all optimize consequences under disrupted networks.
  • Geotechnical AI is most convincing when constrained by physics or instrumentation: TBM rock classification, dam-grouting cement-take prediction, pipeline leak detection, DFOS pile-capacity inference, and shield-tunnel seismic random fields each tie learning to measured processes or mechanistic controls.
  • Underground and subsurface systems dominate the engineering signal: tunnel support, leakage, torsion, CAES cavern cycling, cold-region tunnel stress, contaminant barriers, and porous-media clogging show a common emphasis on coupled hydro-thermal-mechanical response over single-load design checks.

Selected Papers

This digest features 70 selected papers from 1,272 papers analyzed. The sequence opens with a high-elevation shallow-landslide cascade and a vegetation-aware landslide prediction model, then moves through forest slope protection, PS-InSAR building risk, Afghanistan flood susceptibility, mining-subsidence DEM filtering, InSAR denoising, supercritical geothermal rock mechanics, cold-region rockfall impact, ENSO warning definitions, severe-wind nowcasting, warning-siren optimization, canal crack detection, earthquake infrastructure networks, typhoon power grids, tunnel and subsurface engineering, and remote-sensing products for mountain temperature, harmful algal blooms, water quality, Antarctic mass anomalies, and tidal-flat topography.

1. Evolution mechanism of the flash flood-debris flow disaster chain triggered by high-elevation shallow landslides: a case study of the Huangya Gully event in Yuzhong, China, on August 7, 2025

Source: Landslides Type: Event Analysis & Process Modelling Geohazard Type: Flash flood-debris flow disaster chain triggered by high-elevation shallow landslides Relevance: 9/10

Core Problem: High-elevation shallow landslide clusters can transform extreme rainfall into flash flood-debris flow chains, but the transition from incipient slope instability to downstream surge amplification is still weakly quantified.

Key Innovation: Multi-source remote sensing, field investigation, and Massflow simulation reconstruct the Huangya Gully event, linking antecedent rainfall, vegetation-enhanced preferential seepage, shallow landslide clustering, rapid fluidization, erosion, scouring, and dam-break amplification into one staged hazard mechanism.

2. A vegetation-aware deep learning model for landslide prediction with antecedent rainfall

Source: Géotechnique (ICE) Type: AI / Landslide Prediction Geohazard Type: Rainfall- and vegetation-controlled landslide initiation Relevance: 9/10

Core Problem: Landslide prediction models often encode rainfall as static thresholds and ignore spatially variable vegetation controls on soil hydrology.

Key Innovation: An interpretable deep learning model extracts 14-day antecedent rainfall dynamics from hourly records and adds satellite leaf-area index and meteorological vegetation signals, improving historical landslide prediction and reducing false alarms in the June 2008 Hong Kong storm case.

3. Introducing a methodology to quantify the hydro-mechanical protective effects of forests on slope stability: application to a deciduous forest in Western Norway

Source: Catena Type: Protection Forest Quantification Geohazard Type: Shallow landslides and forest slope-stability effects Relevance: 8/10

Core Problem: Protection forests are known to stabilize slopes, but management still lacks a design metric that combines hydrological suction and mechanical root reinforcement.

Key Innovation: An unsaturated infinite-slope framework introduces the Forest Factor as the ratio of post-rainfall safety factor inside and outside forest, showing in Western Norway that hydrological reinforcement can exceed root cohesion and remain protective across seasons.

4. From PS-InSAR observations to building-level risk: A data-driven framework for urban structural resilience

Source: Remote Sensing of Env. Type: InSAR Structural Risk Diagnostics Geohazard Type: Urban subsidence and building deformation risk Relevance: 8/10

Core Problem: PS-InSAR resolves millimetric urban deformation, yet many studies stop at displacement maps rather than building-level structural risk.

Key Innovation: The framework defines an InSAR Building Unit Suitability Index and InSAR Building Deformation Signature, translating persistent-scatterer reliability, magnitude, direction, and deformation heterogeneity into settlement, uplift, tilting, torsion, coupled deformation, and four building risk classes in Kunming.

5. Flood Susceptibility and Potential Flood Risk Assessment in Afghanistan Using Morphometric and Socioeconomic Indicators

Source: Remote Sensing (MDPI) Type: Flood Susceptibility & Risk Mapping Geohazard Type: National-scale flood susceptibility and vulnerability Relevance: 8/10

Core Problem: Afghanistan faces severe flood exposure but lacks systematic national risk mapping because of sparse field data and difficult access.

Key Innovation: Remote sensing, GIS, PCA, and AHP are combined to map flood susceptibility, vulnerability, and risk at subbasin and provincial scales, producing a validated national assessment for data-scarce disaster-risk planning.

6. A Novel Dense Image Matching Point Cloud Filtering Algorithm Integrating Visible Light and Progressive Triangulated Irregular Network Densification for High-Accuracy Mining Subsidence Monitoring

Source: Remote Sensing (MDPI) Type: UAV Deformation Monitoring Geohazard Type: Mining subsidence and vegetation-covered DEM filtering Relevance: 7/10

Core Problem: UAV dense image matching supports mining-subsidence monitoring, but vegetation-covered point clouds still introduce DEM errors that mask surface deformation.

Key Innovation: The H-PTD method uses HSV visible-light information to select ground seeds and then applies progressive TIN densification, improving vegetation-aware filtering for high-accuracy subsidence DEM generation.

7. Fast Spatial Denoising of InSAR Interferograms via Empirical Statistical Modeling

Source: Remote Sensing (MDPI) Type: InSAR Processing Geohazard Type: Surface deformation interferogram phase noise Relevance: 7/10

Core Problem: Noisy unwrapped InSAR phases reduce confidence in topography and deformation products, especially where coherence information is incomplete or unreliable.

Key Innovation: Empirical spatial filters based on truncated wrapped statistical models accelerate phase denoising while preserving structural detail, with tests on volcanic and deformation interferograms.

8. High Temperature Friction and Strength of Felsite and Basalt: Implications for Supercritical EGS at Krafla, Iceland

Source: JGR: Solid Earth Type: Rock Mechanics & Geothermal Hazards Geohazard Type: Supercritical volcanic-geothermal rock stability Relevance: 7/10

Core Problem: Drilling near a shallow magma chamber for supercritical geothermal systems requires rock strength and friction constraints at temperatures approaching magmatic conditions.

Key Innovation: High-temperature experiments on felsite and basalt up to 1100 C show sharp thermal weakening and friction reduction, constraining fracture closure, permeability loss, and mechanical risk around the Krafla magma-adjacent system.

9. Dynamic response and residual bearing capacity of RC columns under the actions of steel corrosion, salt-freeze cycles and impact loading: An experimental study

Source: Cold Regions Sci. & Tech. Type: Infrastructure Impact Testing Geohazard Type: Rockfall impact on cold-region bridge columns Relevance: 7/10

Core Problem: Bridge columns in plateau corridors can be degraded by saline soil, freeze-thaw cycling, corrosion, and rockfall impact before residual capacity is assessed.

Key Innovation: Sequential corrosion, salt-freeze, impact, and residual-bearing tests quantify how compound environmental damage lowers dynamic resistance and post-impact capacity of Qinghai-Tibet reinforced-concrete columns.

10. Assessing El Niño definitions and early warning alerts across Australia, Peru, and the U.S.: implications for disaster risk reduction and climate justice

Source: IJDRR Type: Early Warning Governance Geohazard Type: ENSO-linked hydrometeorological hazards Relevance: 7/10

Core Problem: El Nino alerts shape flood, drought, and climate-risk preparedness, but inconsistent definitions can weaken warning timing and public interpretation.

Key Innovation: A cross-country comparison of Australia, Peru, and the United States links ENSO definitions and alert practices to warning usability, disaster-risk reduction, and climate-justice implications.

11. MSF-PhyDRNN: A Physics-Driven Multi-Source Fusion Recurrent Neural Network for Short-Term Thunderstorm Gale Nowcasting

Source: Remote Sensing (MDPI) Type: Severe Weather Nowcasting Geohazard Type: Thunderstorm gale hazards Relevance: 7/10

Core Problem: Short-term thunderstorm gale prediction degrades at high wind thresholds, exactly where warning value is greatest.

Key Innovation: MSF-PhyDRNN fuses radar reflectivity and surface wind fields through physics-driven recurrent dynamics, improving short-term capture of severe convective wind hazards.

12. Optimizing warning siren placement for audibility coverage using acoustic modelling and genetic algorithms

Source: Geomatics, Nat. Haz. & Risk Type: Warning System Optimization Geohazard Type: Multi-hazard outdoor warning audibility Relevance: 7/10

Core Problem: Warning sirens remain important for rapid public alerting, but placement often ignores terrain, propagation, and audibility coverage trade-offs.

Key Innovation: Acoustic modelling and genetic algorithms are coupled to optimize siren locations for spatial audibility coverage across disaster, technological, and security warning contexts.

13. Intelligent detection and 3D visualization for cracks in concrete lining panels of long-distance water conveyance channels: A case study on crack detection in deep-cut sections of the middle route of China ’s South-to-North water diversion project

Source: International Journal of Applied Earth Observation and Geoinformation Type: Infrastructure Crack Monitoring Geohazard Type: Concrete lining cracks in long-distance water conveyance channels Relevance: 7/10

Core Problem: Large water-transfer channels contain extensive concrete linings where crack detection, spatial localization, and 3D visualization remain labor intensive.

Key Innovation: UAV imitation-ground photogrammetry, a dedicated crack dataset, detection models, POS-assisted single-image localization, and textured 3D mesh reconstruction create an integrated inspection workflow for deep-cut canal sections.

14. A damage-cumulative dynamic interaction model for pile-soil-anchor cable systems: Development and verification via shaking table tests

Source: Soil Dyn. & Earthquake Eng. Type: Seismic Reinforcement Modelling Geohazard Type: Bedrock-overburden slope stabilization systems Relevance: 6/10

Core Problem: Pseudo-static retaining-structure design cannot resolve time-history loading or cumulative soil damage in earthquake-exposed reinforced slopes.

Key Innovation: Large-scale shaking-table tests and a damage-cumulative pile-soil-anchor cable interaction model identify elastic, damage-accumulation, and accelerated-development stages in composite slope reinforcement.

15. Prediction of cement take in Bakhtiari and Karun 4 dam foundation sites using ensemble machine learning algorithms

Source: Transportation Geotechnics Type: AI / Dam Foundation Engineering Geohazard Type: Dam foundation grouting and seepage control Relevance: 6/10

Core Problem: Cement take in dam foundations depends on coupled geological, geotechnical, and operational variables that are difficult to estimate before grouting.

Key Innovation: Gradient boosting, AdaBoost, CatBoost, and XGBoost are trained on borehole grouting records from Bakhtiari and Karun 4 dam sites, with SHAP and ICE analysis exposing the controls on cement consumption.

16. A 3D micromechanical multiscale fracture model for fluid flow in coupled fracture-matrix system

Source: Engineering Geology Type: Hydro-Mechanical Fracture Modelling Geohazard Type: Fluid flow in fractured geological media Relevance: 6/10

Core Problem: Deep fractured media contain low-permeability matrices and multiscale high-permeability fractures, making 3D coupled flow representation difficult.

Key Innovation: A micromechanical multiscale fracture model homogenizes matrix microcracking and small fractures while representing larger fracture networks, improving 3D hydro-mechanical simulation for geothermal, reservoir, and waste-disposal settings.

17. A power-law relationship between bulk modulus and permeability in fractured sandstone: Insights from tidal analysis

Source: Engineering Geology Type: Rock Mass Hydrogeomechanics Geohazard Type: Permeability and deformability of fractured sandstone Relevance: 6/10

Core Problem: Permeability and bulk modulus are often measured separately, leaving weak constraints on how fractured rock deformation and flow co-vary in protected sites.

Key Innovation: Tidal-response analysis of borehole water levels in the Yungang Grottoes estimates permeability, storage, and bulk modulus noninvasively, revealing a power-law relation between fracture-enhanced permeability and stiffness loss.

18. Entropy-MIMR-LSTM framework for rain gauge network optimization in mountainous small watersheds: A case study of Fuhuxi Watershed, China

Source: J. Mountain Science Type: Hydrometeorological Network Design Geohazard Type: Extreme rainfall and flood monitoring in mountainous watersheds Relevance: 6/10

Core Problem: Mountainous small watersheds need dense rainfall information for flood warning, but gauge networks are sparse and costly to optimize.

Key Innovation: An Entropy-MIMR-LSTM framework combines information-content analysis, redundancy control, and sequence prediction to optimize rain-gauge placement in the Fuhuxi Watershed.

19. Modeling service flows in earthquake scenarios using a physical–social multilayer network

Source: RESS Type: Disaster Service Flow Modelling Geohazard Type: Earthquake disruption of healthcare access Relevance: 6/10

Core Problem: Earthquake impact assessments often treat casualties, roads, and healthcare capacity separately, obscuring service-flow bottlenecks.

Key Innovation: A physical-social multilayer network couples residential demand, road disruption, and healthcare supply to model service flows under multiple earthquake scenarios.

20. A user-centered, earthquake risk-informed approach to valve placement optimization in urban natural gas distribution networks

Source: RESS Type: Seismic Infrastructure Optimization Geohazard Type: Earthquake risk in urban natural gas networks Relevance: 6/10

Core Problem: Isolation valves reduce leakage after pipeline damage, but they can also disconnect undamaged components and amplify cascading functionality loss.

Key Innovation: A user-centered multi-objective optimization model incorporates earthquake-induced damage, valve cost, leakage risk, and service reliability to guide natural-gas valve placement.

21. Agent-based simulation of socio-technical interactions during community disaster recovery

Source: IJDRR Type: Socio-Technical Recovery Modelling Geohazard Type: Community disaster recovery Relevance: 6/10

Core Problem: Community recovery depends on coupled household decisions and technical-system restoration, not infrastructure repair alone.

Key Innovation: An agent-based model built on the CI-STR framework simulates long-term household recovery by linking human capabilities, disaster impacts, and technical-system recovery trajectories.

22. Graph-based deep reinforcement learning for adaptive and scalable post-earthquake bridge network restoration

Source: RESS Type: Post-Earthquake Network Restoration Geohazard Type: Bridge network repair scheduling Relevance: 6/10

Core Problem: Post-earthquake bridge restoration requires sequential repair choices across damaged networks under limited resources.

Key Innovation: A GNN-DRL framework with top-k action filtering learns scalable bridge repair policies from network topology and dynamic bridge states.

23. Robust calibration with stochastic emulators: hydrological model parameter estimation under uncertain rainfall conditions

Source: RESS Type: Hydrologic Model Calibration Geohazard Type: Flood and pollution peaks under uncertain rainfall Relevance: 6/10

Core Problem: Hydrologic calibration can absorb rainfall uncertainty into parameters, reducing reliability for operational flood or water-quality decisions.

Key Innovation: A stochastic-emulator robust-calibration strategy accounts for aleatory rainfall forcing uncertainty while limiting the computational cost of physically based environmental models.

24. Scenario dynamics-informed enhancement of power grid resilience to typhoons

Source: RESS Type: Infrastructure Resilience Geohazard Type: Typhoon impacts on power grids Relevance: 6/10

Core Problem: Power-grid hardening against typhoons must account for storm dynamics and the evolving response of the grid, not just static exposure.

Key Innovation: Scenario dynamics-informed strategies combine typhoon scenario classification, grid dynamic simulation, and resilience metrics to guide targeted power-grid hardening.

25. Dynamic Causal Graph Network for reliable pipeline leak detection

Source: RESS Type: Pipeline Structural Health Monitoring Geohazard Type: Leak detection in pipeline systems Relevance: 6/10

Core Problem: Pipeline leak diagnosis can be misled by environmental vibration, turbulence, and other confounders that create spurious sensor correlations.

Key Innovation: A Dynamic Causal Graph Network separates causal structure from correlation in multi-sensor data, improving reliability of pipeline structural-health monitoring.

26. An integrated reconstruction–downscaling framework for generating daily, all-weather 30 m land surface temperature in cloudy mountainous regions

Source: International Journal of Applied Earth Observation and Geoinformation Type: Mountain Thermal Remote Sensing Geohazard Type: All-weather land surface temperature in cloudy mountains Relevance: 6/10

Core Problem: Mountain LST products are limited by cloud cover, coarse thermal resolution, and terrain-specific reconstruction errors.

Key Innovation: An integrated reconstruction-downscaling framework generates daily, all-weather 30 m LST in cloudy mountainous regions, improving thermal inputs for hydrology, ecology, and hazard models.

27. Abrupt intensification and spatiotemporal dynamics of large-scale harmful algal blooms in China’s marginal seas mapped using multi-sensor satellite imagery and deep learning

Source: ISPRS J. Photogrammetry Type: Coastal Hazard Monitoring Geohazard Type: Harmful algal bloom dynamics Relevance: 6/10

Core Problem: Short-lived harmful algal blooms are poorly captured by single-sensor satellite records, limiting hazard trend analysis in marginal seas.

Key Innovation: MODIS, VIIRS, ocean-color sensors, and deep learning are integrated to build a 2018-2024 bloom record and quantify abrupt intensification and seasonal dynamics in China's marginal seas.

28. GAIR: Location-aware self-supervised contrastive pre-training with geo-aligned implicit representations

Source: ISPRS J. Photogrammetry Type: GeoAI Representation Learning Geohazard Type: Multimodal geospatial alignment Relevance: 6/10

Core Problem: Geospatial tasks need localized representations across overhead imagery, ground-level views, and vector data, but standard vision transformers provide coarse whole-image embeddings.

Key Innovation: GAIR adds geo-aligned implicit representations to location-aware contrastive pretraining, enabling arbitrary-position features for multimodal geospatial modelling.

29. Disentangling representation and cross-domain adaptation for unsupervised change detection in multimodal remote sensing images

Source: ISPRS J. Photogrammetry Type: Remote Sensing Change Detection Geohazard Type: Multimodal damage and land-cover change assessment Relevance: 6/10

Core Problem: SAR and optical images encode different physics, making unsupervised multimodal change detection unreliable under abrupt damage or land-cover change.

Key Innovation: A disentangled representation and unpaired cross-domain adaptation framework separates modality translation from change representation, reducing dependence on labels and one-to-one image pairs.

30. Seismic response of soil-shield tunnel system considering spatial variability of multivariate soil properties via random field theory

Source: TUST Type: Seismic Tunnel Reliability Geohazard Type: Soil-shield tunnel response under spatially variable ground Relevance: 6/10

Core Problem: Tunnel seismic assessment often assumes homogeneous soil, which can understate response uncertainty in spatially variable ground.

Key Innovation: Random-field modelling of correlated soil properties is embedded in soil-shield tunnel seismic simulations to quantify how ground variability affects structural response.

31. Thermo-hydro-mechanical coupled response and stability analysis of salt cavern CAES with interlayers under high-frequency injection-withdrawal cycling

Source: TUST Type: Subsurface Energy Storage Stability Geohazard Type: Salt cavern CAES under cyclic thermo-hydro-mechanical loading Relevance: 6/10

Core Problem: High-frequency injection-withdrawal cycles in salt cavern compressed-air storage create coupled thermal, hydraulic, and mechanical stability risks, especially with interlayers.

Key Innovation: A fully coupled THM numerical model evaluates temperature-dependent air and rock behaviour under multiple cycling rates, constraining stability of interlayered salt caverns.

32. Stability analysis of primary support and effect of restraining anchor cables in soft rock tunnels

Source: TUST Type: Tunnel Support Mechanics Geohazard Type: Soft-rock tunnel deformation disasters Relevance: 6/10

Core Problem: Soft-rock tunnels can undergo large deformation when primary support buckles or loses confinement, but restraint mechanisms are not well formalized.

Key Innovation: A Muzhailing Tunnel case study analyzes primary-support buckling and proposes restraining anchor-cable composite support to improve stability under deformation pressure.

33. Interpretable prediction of rock mass classification for TBM using bi-modal data

Source: Intl. J. Rock Mech. & Mining Type: AI / Tunnel Geology Geohazard Type: Real-time TBM rock-mass classification Relevance: 6/10

Core Problem: Tunnel boring decisions require surrounding-rock classification during excavation, while single-source TBM data or seismic data alone can be ambiguous.

Key Innovation: DMIANet fuses temporal TBM operation data with horizontal seismic profiling through matrix-profile subsequence extraction and interpretable feature-target attention for rock-mass classification.

34. Coplanar analysis and rapid window mapping based on the automatic identification of rock mass discontinuities from 3D point clouds in underground excavations

Source: Engineering Geology Type: Rock Mass Mapping Geohazard Type: Underground excavation discontinuities Relevance: 5/10

Core Problem: Coplanar discontinuities govern underground excavation stability, but conventional window mapping is slow and subjective.

Key Innovation: Terrestrial laser scanning point clouds, improved watershed clustering, coplanar analysis, and rapid digital window mapping automate discontinuity identification in underground rock outcrops.

35. A novel method for estimating rock joint roughness considering morphological spatial heterogeneity and shear contact characteristics

Source: Engineering Geology Type: Rock Joint Characterization Geohazard Type: Joint roughness and shear contact behaviour Relevance: 5/10

Core Problem: Profile-based roughness metrics underrepresent 3D spatial heterogeneity and evolving asperity contacts during shearing.

Key Innovation: A roughness-estimation method combines simulated 3D morphology, spatial heterogeneity, and shear-contact characteristics to better represent natural rock-joint mechanics.

36. Automated lithological classification and boundary mapping from 3D point clouds

Source: Engineering Geology Type: Lithological Boundary Mapping Geohazard Type: Rock-mass structure and engineering geological boundaries Relevance: 5/10

Core Problem: Lithological boundaries and weathering contrasts are essential for rock-mass assessment, but manual classification from point clouds remains inefficient.

Key Innovation: A TLS point-cloud framework extracts geometric and intensity features to automate lithological classification and boundary mapping for engineering safety assessment.

37. Study on the role of thin spray-on layer (TSL) thickness on roadway surrounding rock under high geostress environment

Source: TUST Type: Underground Support Design Geohazard Type: High-geostress roadway deformation Relevance: 5/10

Core Problem: Deep roadways in high in-situ stress require support systems that absorb failure energy while controlling deformation.

Key Innovation: Field surveys, in-situ stress classification, DFN modelling, and numerical analysis evaluate how thin spray-on liner thickness affects surrounding-rock control in Gaofeng Mine.

38. Torsional behaviors of segmental tunnels under compression-bending-torsion load combinations. Part II: Physics-informed stiffness evaluation and effects of soil-tunnel interaction

Source: TUST Type: Segmental Tunnel Mechanics Geohazard Type: Tunnel torsion and soil-tunnel interaction Relevance: 5/10

Core Problem: Segmental tunnel torsion under compression-bending-torsion loads can damage linings but is rarely included in routine stiffness evaluation.

Key Innovation: Physics-informed stiffness models and soil-tunnel interaction formulations extend model-test findings to evaluate torsional response under arbitrary loading scenarios.

39. Deformation characteristics and excavation strategy optimization of two adjacent deep shafts: Insights from field monitoring and numerical simulation

Source: TUST Type: Urban Excavation Monitoring Geohazard Type: Adjacent deep-shaft excavation in dense cities Relevance: 5/10

Core Problem: Twin or adjacent shafts can interact through ground deformation, yet case histories with detailed monitoring remain limited.

Key Innovation: Field monitoring and numerical simulation of two adjacent deep shafts quantify wall movement, settlement, building response, and optimized time-staggered excavation strategy.

40. Experimental investigation of member-scale composite action, deformation coordination, and waterproofing performance of membrane-bonded composite shell tunnel lining specimens

Source: TUST Type: Tunnel Lining Performance Geohazard Type: Composite shell tunnel lining waterproofing Relevance: 5/10

Core Problem: Membrane-bonded composite tunnel linings need quantitative evaluation of waterproofing, deformation compatibility, and load-bearing capacity.

Key Innovation: Four-point bending tests assess membrane thickness, interface configuration, composite action, and deformation coordination in bonded spray-applied waterproofing systems.

41. Performance evolution and failure mechanism of segmental linings strengthened with prefabricated UHPC panels: Insights from full-ring tests and multi-index evaluation

Source: TUST Type: Tunnel Rehabilitation Geohazard Type: UHPC-strengthened segmental lining failure Relevance: 5/10

Core Problem: Metro shield tunnel strengthening needs efficient prefabricated systems with predictable full-ring structural performance.

Key Innovation: Scaled full-ring tests and multi-index evaluation characterize the load-bearing evolution and failure mechanism of segmental linings strengthened with prefabricated UHPC panels.

42. Tunnelling performance evaluation in complex stratum conditions: Insights from a case study of earth pressure balance shield

Source: TUST Type: Shield Tunnelling Performance Geohazard Type: EPB tunnelling in complex strata Relevance: 5/10

Core Problem: Shield performance metrics are often confounded by changing geology, reducing their use for cost and efficiency control.

Key Innovation: Specific energy is evaluated ring-by-ring after controlling geological variables, revealing EPB shield performance patterns in Beijing Subway Line 8 complex strata.

43. A predictive model for the bearing capacity reduction of shallow foundations due to mechanized tunneling in soft ground

Source: TUST Type: Tunnel-Foundation Interaction Geohazard Type: Bearing-capacity loss above mechanized tunnelling Relevance: 5/10

Core Problem: Surface foundations above advancing soft-ground tunnels can lose bearing capacity through 3D tunnel-passage effects not captured by plane-strain models.

Key Innovation: A predictive model explicitly represents the staged three-dimensional passage of mechanized tunnelling beneath shallow foundations in soft ground.

44. Leakage mechanism of double-gasketed segmental joint in shield tunnel under interacting internal and external water pressures

Source: TUST Type: Tunnel Leakage Modelling Geohazard Type: Double-gasketed segmental joints under water pressure Relevance: 5/10

Core Problem: Deep drainage shield tunnels experience joint rotation and internal-force redistribution under interacting internal and external water pressures.

Key Innovation: A refined 3D joint mechanical model is coupled with a 2D gasket leakage model and dynamic hydraulic loading to simulate progressive contact opening and leakage.

45. Dynamic response and vibration control of soil–structure systems subject to subway and road traffic loads: Mathematical analysis and physical model tests

Source: TUST Type: Traffic-Induced Vibration Control Geohazard Type: Soil-structure vibration from subway and road loads Relevance: 5/10

Core Problem: Subway and road traffic vibrations can couple through soil-structure systems, but joint propagation and isolation effects are difficult to predict.

Key Innovation: Green-function analysis and physical model tests quantify coupled vibration fields and evaluate spring-based isolation strategies.

46. Micromechanical analysis of local scour around monopiles in steady flow using SPH-DEM: effects of cross-sectional shape and orientation

Source: Computers and Geotechnics Type: Coastal & Offshore Foundation Modelling Geohazard Type: Local scour around monopiles Relevance: 5/10

Core Problem: Pile cross-section shape and orientation change scour processes around offshore foundations, but pore-scale mechanisms remain poorly resolved.

Key Innovation: A coupled SPH-DEM model resolves water flow and granular seabed response to explain how pile geometry controls local scour and foundation stability.

47. Coupled water-heat-salt transport in the vadose zone of loess: A numerical framework considering phase changes and ionic fractionation

Source: Computers and Geotechnics Type: Loess Vadose-Zone Modelling Geohazard Type: Water-heat-salt transport in arid loess Relevance: 5/10

Core Problem: Moisture redistribution, salinization, and phase changes interact in deep loess vadose zones, but coupled numerical frameworks remain difficult.

Key Innovation: A REV-scale model couples moisture flow, convective-diffusive solute transport, heat conduction, phase change, and ionic fractionation for unsaturated loess.

48. Pressure gradient hinders particle migration: An LBM-DEM study of clogging in irregular-shaped porous media

Source: Computers and Geotechnics Type: Pore-Scale Transport Modelling Geohazard Type: Clogging in irregular porous media Relevance: 5/10

Core Problem: Particle migration in porous media controls clogging and seepage evolution, but irregular pore geometry complicates mechanism identification.

Key Innovation: An LBM-DEM model shows how pressure gradients influence slurry-particle migration and clogging dynamics in irregular-shaped porous media.

49. A bounding surface non-isothermal hydro-mechanical model for unsaturated fine-grained soils under isotropic stress states

Source: Computers and Geotechnics Type: Thermo-Hydro-Mechanical Soil Modelling Geohazard Type: Unsaturated fine-grained soils under temperature change Relevance: 5/10

Core Problem: Temperature alters retention and volumetric behaviour of unsaturated soils, yet many constitutive models treat hydro-mechanical and thermal effects separately.

Key Innovation: A bounding-surface formulation couples non-isothermal hysteretic retention with isotropic mechanical behaviour under saturated and unsaturated stress states.

50. A novel elastoplastic damage constitutive model for granite residual soil considering structure degradation mechanism and its extended suitability

Source: Computers and Geotechnics Type: Constitutive Modelling Geohazard Type: Granite residual soil structure degradation Relevance: 5/10

Core Problem: Granite residual soil inherits metastable parent-rock fabric and cementation, which conventional Modified Cam Clay formulations do not capture.

Key Innovation: A damage elastoplastic constitutive model adds structure degradation and nonlinear failure behaviour to simulate the mechanical response of structured granite residual soil.

51. Implicit partitioned-integration material point method for saturated porous media with an incompressible stabilization approach

Source: Computers and Geotechnics Type: Numerical Method Geohazard Type: Large-deformation saturated porous media Relevance: 5/10

Core Problem: Explicit two-phase MPM is inefficient for long-term saturated porous-media problems and can suffer stress and pore-pressure oscillations.

Key Innovation: An implicit partitioned-integration MPM with incompressible stabilization improves robustness for coupled solid-fluid deformation at engineering time scales.

52. Modeling the full process evolution of a contaminant plume during containment by a vertical barrier system based on a 1.5-D solution framework

Source: Computers and Geotechnics Type: Contaminant Transport Modelling Geohazard Type: Cut-off wall containment of contaminated sites Relevance: 5/10

Core Problem: Semi-analytical barrier models rarely preserve the full spatial distribution and transient evolution of initial contaminant plumes.

Key Innovation: A 2D advection-dispersion-degradation model with a 1.5D exact solution framework simulates plume evolution during containment by vertical barriers.

53. Analytical stress and displacement solutions of non-circular cold region tunnels under fully bonded and fully smooth contact conditions between lining and surrounding rock

Source: Computers and Geotechnics Type: Cold-Region Tunnel Analysis Geohazard Type: Non-circular tunnel stress and displacement Relevance: 5/10

Core Problem: Cold-region tunnel solutions often simplify non-circular sections as circular and neglect lining-rock contact conditions.

Key Innovation: Conformal mapping and complex-variable solutions derive stress and displacement for non-circular tunnels under fully bonded and smooth contact conditions.

54. Hypoplastic modelling of vegetated sand under monotonic and cyclic loading

Source: Computers and Geotechnics Type: Root-Soil Constitutive Modelling Geohazard Type: Vegetated sand under monotonic and cyclic loading Relevance: 5/10

Core Problem: Existing vegetated-soil models rarely capture root effects on cyclic strain accumulation, stiffness loss, and semifluidized states.

Key Innovation: A hypoplastic model represents root-induced changes in dilatancy, strength, and cyclic stiffness, reproducing monotonic and undrained cyclic behaviour of vegetated sand.

55. Water–salt dual-control discrete element analysis of coupling mechanism between salt swelling and frost heave in sulfate saline soil

Source: Computers and Geotechnics Type: Cold-Region Saline Soil Modelling Geohazard Type: Salt swelling and frost heave Relevance: 5/10

Core Problem: Sulfate saline soils undergo coupled water migration, salt swelling, and frost heave that are difficult to resolve at particle scale.

Key Innovation: A water-salt dual-control DEM framework links particle contact state, local salinity, temperature traversal, and ice-water phase transition to mesoscale freezing response.

56. Unified numerical framework and energy-based analysis for identifying multiple failure modes in microwave-irradiated granite

Source: Computers and Geotechnics Type: Thermo-Mechanical Rock Failure Geohazard Type: Microwave-induced granite fracturing Relevance: 5/10

Core Problem: Microwave rock-breaking studies often focus on cracking while overlooking melting, spalling, and phase-transformation failure modes.

Key Innovation: A unified 3D heterogeneous thermo-mechanical framework classifies multiple granite failure modes under 9 GHz microwave irradiation using energy-based indicators.

57. Inference of pile capacity from distributed strain sensing via PDE-constrained optimization

Source: Computers and Geotechnics Type: Sensor-Based Geotechnical Inference Geohazard Type: Pile capacity from distributed strain sensing Relevance: 5/10

Core Problem: Distributed fiber-optic pile strain data contain physical signatures that are often treated as noise during capacity interpretation.

Key Innovation: A PDE-constrained optimization framework matches DFOS strain while enforcing mechanical equilibrium, inferring pile quality and soil response profiles from field measurements.

58. A finite strain solution incorporating 3D stresses and steady-state seepage for circular tunnels in strain-softening rock

Source: Computers and Geotechnics Type: Tunnel Seepage & Finite-Strain Analysis Geohazard Type: Circular tunnels in strain-softening rock Relevance: 5/10

Core Problem: Three-dimensional excavation effects and seepage coupling are difficult to include in analytical tunnel solutions for strain-softening rock.

Key Innovation: A finite-strain Drucker-Prager solution with incremental stress correction incorporates out-of-plane stress and steady-state seepage for circular tunnels.

59. Bedding effect on sandstone stability using three-point loading: Implication of Longyou Grottoes caves

Source: JRMGE Type: Rock Heritage Stability Geohazard Type: Bedded sandstone cave stability Relevance: 5/10

Core Problem: Weak bedding planes reduce sandstone stability in underground caves, but their strength, orientation, and spacing effects are hard to isolate.

Key Innovation: Weibull-based FDEM simulations of three-point loading quantify bedding-controlled bending strength, acoustic emission, cracking, and energy evolution for Longyou Grottoes sandstone.

60. Size effect on soil desiccation cracking

Source: JRMGE Type: Soil Cracking Mechanics Geohazard Type: Desiccation cracking and size effects Relevance: 5/10

Core Problem: Soil specimen size changes evaporation, shrinkage stress, and crack geometry, but size effects are often described only qualitatively.

Key Innovation: Controlled strip and slab desiccation tests link height, width, and aspect ratio to evaporation rate, shrinkage deformation, and cracking pattern transitions.

61. Modeling of saturated hydraulic conductivity for vegetated soils considering root-soil interactions

Source: JRMGE Type: Ecohydrological Soil Modelling Geohazard Type: Root-soil control of saturated hydraulic conductivity Relevance: 5/10

Core Problem: Plant roots can compact rhizosphere aggregates and open preferential channels, but these competing effects are rarely captured in conductivity models.

Key Innovation: 3D-printed root analogs and new void-ratio functions quantify how root density and orientation modify saturated hydraulic conductivity in vegetated soils.

62. Deformation of servo steel supported zoned excavation pits in Hangzhou soft soil

Source: JRMGE Type: Urban Excavation Monitoring Geohazard Type: Servo-supported zoned excavation in soft soil Relevance: 5/10

Core Problem: Zoned excavation and servo-active supports are widely used to limit deformation near metro infrastructure, but their combined field response remains underdocumented.

Key Innovation: Long-term monitoring of a Hangzhou soft-soil pit near a subway station and tunnels quantifies wall movement, settlement profiles, and deformation controls during zoned excavation.

63. SC-DRGAN: A Cross-Site Transfer Learning Architecture for Tropical Forest Structural Parameters Estimation

Source: IEEE JSTARS Type: AI / Vegetation Structure Remote Sensing Geohazard Type: Tropical forest structure relevant to slope hydrology and biomass Relevance: 5/10

Core Problem: Canopy height and aboveground biomass retrieval from PolInSAR often generalizes poorly across tropical sites with different forest structure.

Key Innovation: SC-DRGAN combines multibaseline PolInSAR, GEDI measurements, similarity calculation, deep residual features, and adversarial domain adaptation for cross-site forest-structure estimation.

64. A Multisensor Framework for Satellite Data Simulation: Generating Representative Datasets for Future ESA Missions—CHIME and LSTM

Source: Remote Sensing (MDPI) Type: Earth Observation Mission Simulation Geohazard Type: Pre-launch proxy data for environmental monitoring Relevance: 5/10

Core Problem: Future hyperspectral and thermal missions need realistic proxy datasets before launch, but current simulations are fragmented by platform and sensor.

Key Innovation: A unified multisensor framework harmonizes PRISMA, EnMAP, HySpex, Sentinel-2, ECOSTRESS, Landsat, and airborne data to simulate CHIME and LSTM products.

65. PFMGAN: A Generative Adversarial Network with Physics Fusion for DTM Generation from Martian Monocular Images

Source: Remote Sensing (MDPI) Type: Topographic Reconstruction Geohazard Type: Physics-aware monocular DTM generation Relevance: 5/10

Core Problem: Monocular planetary DTM generation can mistake illumination and albedo effects for topography, weakening terrain interpretability.

Key Innovation: PFMGAN injects physical priors into a Mix Transformer GAN to reconstruct high-precision DTMs from Martian monocular images across diverse landforms.

66. Interpretable machine learning and uncertainty quantification for high-precision fractional vegetation cover inversion across scales in alpine grasslands

Source: International Journal of Applied Earth Observation and Geoinformation Type: Vegetation Remote Sensing & Uncertainty Geohazard Type: Fractional vegetation cover across alpine grasslands Relevance: 5/10

Core Problem: High-resolution fractional vegetation cover inversion is limited by UAV-satellite spatial mismatch, uncertainty, and dependence on simultaneous field campaigns.

Key Innovation: A multi-source framework combines PSO-SHAP feature selection, uncertainty quantification, and cross-scale modelling to improve alpine grassland FVC inversion.

67. SDGSAT-1 MSI for monitoring water quality parameters in rivers: A comprehensive evaluation and demonstration

Source: International Journal of Applied Earth Observation and Geoinformation Type: Water Quality Remote Sensing Geohazard Type: River turbidity and chlorophyll monitoring Relevance: 5/10

Core Problem: Narrow and optically diverse rivers remain under-monitored despite strong ecological and societal importance.

Key Innovation: SDGSAT-1 MSI is evaluated with synchronized in-situ data and turbidity-adaptive retrievals for total suspended matter and chlorophyll-a in Chinese rivers.

68. Short-term Antarctic ice sheet mass anomalies revealed by daily GRACE/FO satellite gravimetry

Source: International Journal of Applied Earth Observation and Geoinformation Type: Cryosphere Mass Monitoring Geohazard Type: Short-term Antarctic ice-sheet mass anomalies Relevance: 5/10

Core Problem: Monthly GRACE/FO gravity fields miss transient Antarctic mass-change events that affect sea-level and polar hydrology interpretation.

Key Innovation: Monthly, 10-day, and daily GRACE/FO solutions are combined to detect and attribute short-term Antarctic ice-sheet mass anomalies from 2002 to 2024.

69. Integrating ICESat-2 altimetry and multi-source satellite imagery to track annual changes of Jiangsu central tidal flat topography over 2009–2024

Source: International Journal of Applied Earth Observation and Geoinformation Type: Coastal Topographic Monitoring Geohazard Type: Annual tidal-flat elevation change Relevance: 5/10

Core Problem: Tidal-flat DEMs are hard to update frequently because of complex hydrodynamics and limited elevation observations.

Key Innovation: ICESat-2 altimetry and more than 1,000 optical scenes are fused to generate annual 30 m DEMs for Jiangsu central tidal flats from 2009 to 2024.

70. Equity-aware routing optimization for hazardous chemicals considering time-varying conditions and link disruption risks

Source: RESS Type: Risk Optimization Geohazard Type: Hazardous chemical transport under disruption Relevance: 5/10

Core Problem: Hazardous-chemical routing decisions must handle time-varying accident risk, link disruption, and equity in risk distribution simultaneously.

Key Innovation: A comprehensive risk assessment and equity-aware routing optimization framework models time periods, disruption scenarios, and spatial redistribution of transport risk.