TerraMosaic Daily Digest: October 20, 2026
Daily Summary
Anchoring changes how an anti-dip rock slope responds to shaking, as well as when it fails. Model tests show delayed cracking, reduced displacement and a different failure surface after reinforcement. Complementary landslide studies combine spatially held-out susceptibility evaluation in Tongren, inventory-based analysis of the Luhe rainfall event and three-dimensional imaging of possible detachment structures inside La Soufrière volcano.
Coastal protection emerges as a problem of distinct physical mechanisms. Full-scale salt-marsh experiments find a persistent substrate alongside declining vegetation-mediated attenuation during sustained wave exposure. Realistic Reef-Ball layouts dissipate waves mainly through roughness-induced friction, whereas six years of Sussex cliff monitoring reveal episodic mass wasting superimposed on gradual erosion. Together, these studies clarify why protective performance and retreat rates cannot be inferred from a single measure of wave forcing.
Geotechnical risk assessment is also becoming more explicit about material variability and local loading. A laboratory-calibrated liquefaction framework propagates uncertainty in resistance, demand and soil state into triggering probabilities. Cross-fault tunnel models emphasize slip near the crossing, while public videos reveal the long-period response of Bangkok high-rises to a distant earthquake. Mining-fracture observations extend this attention to loading history and evolving underground connectivity.
Key Trends
The strongest studies connect subsurface structure, loading and observable response across laboratory, field and regional scales.
- Slope reinforcement changes the failure mechanism: Shaking-table tests show that anchors alter dynamic response and the geometry of failure in anti-dip bedded slopes. Volcanic ambient-noise tomography adds a complementary route to identifying possible detachments and hydrothermal pathways before they can be represented in stability models.
- Spatial evaluation strengthens susceptibility mapping: Tongren combines nested spatial validation with repeated sampling, while the Luhe inventory documents the geometry and environmental setting of a rainfall-induced cluster. Predictive importance and covariate-adjusted associations answer different questions and remain distinct from causal attribution.
- Coastal protection needs mechanism-specific models: The marsh study separates substrate persistence from vegetation attenuation; Reef-Ball experiments identify frictional roughness rather than frequent wave breaking. Cliff observations add an episodic mass-wasting component that weakens simple forcing–retreat relationships.
- Material testing informs probabilistic assessment: The liquefaction framework carries laboratory-derived variability into triggering probabilities and allows site-representative calibration. Wave–soil flume experiments and cyclic-rock tests complement this approach by resolving how pressure, prior damage and loading sequence change the response.
- Existing observations reveal additional hazard processes: Permanent seismic stations record flood-related river signatures, wide GNSS networks capture earthquake- and tsunami-related ionospheric disturbances, and public videos recover building motion. These studies expand process observations and event reconstruction; prospective warning skill remains a separate task.
Selected Papers
The selected studies connect landslide occurrence and slope structure with failure processes, coastal protection and infrastructure response. Landslide and geotechnical case studies lead the list, followed by routine hazard applications and methods for improving observations or physical models. Papers are ordered by relevance score.
1. Dynamic Response and Failure Behavior of Anchored Anti-dip Bedding Rock Slopes: Shaking Table Tests
Core Problem: Anchors alter both the dynamic response and the failure path of anti-dip bedded rock slopes.
Key Innovation: Shaking-table comparisons show delayed cracking and a changed failure mechanism in anchored models. Cracking began at 0.4g rather than 0.3g, and model failure occurred at 1.3g rather than 0.9g. These are experimental model thresholds, not allowable accelerations for field slopes.
2. Interpretable AutoML and Double Machine Learning for Landslide Susceptibility Assessment in Tongren, Qinghai–Tibet Plateau
Core Problem: Spatially dependent samples and correlated terrain factors complicate both susceptibility evaluation and interpretation.
Key Innovation: Nested spatial cross-validation across 20 sampling realizations gives a mean AUC of 0.953 ± 0.009; the highest-ranked 20% of the area contains 93.27% ± 1.51% of held-out landslides. Double machine learning estimates covariate-adjusted associations; neither SHAP importance nor the random-covariate perturbation test establishes environmental causation.
3. Characteristics and formation controls of a representative rainfall-induced landslide cluster in Southeastern China
Core Problem: Regional landslide clusters require inventory-based characterization of terrain, weathering and rainfall conditions.
Key Innovation: The August 2018 Luhe inventory links a concentrated landslide cluster to weathered granite terrain and rainfall patterns. Estimated landslide area and volume are about 3 km² and 0.0063 km³. Geographical-detector associations identify spatial controls to investigate, rather than independently proving trigger causation.
4. Revealing the Hydrothermal System of La Soufrière Volcano (Guadeloupe) Through Nodal Ambient Noise Tomography
Core Problem: Hydrothermal pathways and potential detachment structures are difficult to resolve inside active volcanic slopes.
Key Innovation: A month of ambient-noise recordings from 39 nodal geophones images La Soufrière's shallow three-dimensional shear-wave structure, including a central low-velocity conduit and possible flank detachments. The geological interpretation informs slope assessment; the images do not establish imminent failure or warning lead time.
5. Analysis of Sea Cliff Erosion Measured Using UAV Photogrammetry
Core Problem: Gradual marine erosion and episodic cliff collapse can produce different relationships between wave forcing and retreat.
Key Innovation: Six years of UAV photogrammetry distinguish lower-cliff erosion from episodic mass wasting at two Sussex chalk-cliff sites. Wave-energy relationships are substantially weaker where mass wasting dominates; storm-period increases in erosion are site- and observation-period-specific, rather than universal retreat multipliers.
6. Probabilistic State Parameter–Based Liquefaction Triggering Model
Core Problem: Case-history-based triggering models can obscure material variability and uncertainty in soil state.
Key Innovation: A hierarchical Bayesian calibration across 20 sands propagates resistance, loading and state-parameter uncertainty into CPTu-based liquefaction probability. Independent case-history discrimination is comparable to, and sometimes better than, conventional procedures. The reported 80% uncertainty reduction comes from a single material-specific calibration example.
7. The contribution of salt marshes to wave attenuation during extreme storms
Core Problem: Coastal protection assessments need to distinguish marsh-platform persistence from vegetation-mediated wave attenuation.
Key Innovation: A full-scale transplanted marsh experiment, supported by global coastal assessment, finds little substrate erosion but declining vegetation attenuation under sustained wave exposure. The vegetation contribution is strongest at lower-energy marsh coasts and remains limited under waves exceeding about 1 m; a stable platform does not imply unchanged vegetation protection.
8. Large-Scale Flume Investigation of Pore Water Pressure Evolution under Wave–Soil–Structure Interaction
Core Problem: Wave loading and seepage can affect foundation stability even when conventional liquefaction thresholds are not exceeded.
Key Innovation: Large-scale solitary-wave flume tests resolve depth- and position-dependent pore-pressure responses with and without a foundation. Neither excess pore pressure nor seepage pressure exceeded the tested critical thresholds. Embedment changes upward seepage beneath the foundation, providing a benchmark for coupled models rather than evidence of observed liquefaction failure.
9. Impact of Fault Slip Distribution on Seismic Damage Assessment for Cross-Fault Tunnels
Core Problem: Uniform fault-slip assumptions can misrepresent the local deformation imposed on a tunnel.
Key Innovation: A Daliang Tunnel case based on the 2022 Menyuan earthquake compares uniform and asperity-based slip descriptions. A shallow-asperity representation better approximates damage from the inverted-slip reference in this case, emphasizing the importance of slip near the crossing rather than a universally optimal fault model.
10. Surface Subsidence Prediction for Inclined Blind Orebodies: An Iterative Slip Model Integrating the Arching Effect of Caved Zones
Core Problem: Inclined blind orebodies can generate asymmetric surface influence zones that simple empirical rules miss.
Key Innovation: An iterative critical-slip search combines a modified Bishop formulation with a Janssen-derived arching reduction. Physical tests and numerical simulations for a Shandong gold-mine case support preliminary influence-zone assessment; the reported 10.8% experimental error is case-specific.
11. Study on Evolution Law of Fracture Field During Coupled Fracture of Multi-layer Key Strata and Prediction of “Two-Zone” Height
Core Problem: Coupled failure of several key strata controls the height and connectivity of mining-induced fracture zones.
Key Innovation: Similarity models and image analysis track fracture development and goaf compaction, supporting a two-zone height prediction method. Field comparisons report errors within 6% in the investigated setting; the result concerns fracture-zone geometry, not a calibrated probability of mine inundation.
12. Far‐Field Response of High‐Rise Buildings in Bangkok Based on Public Videos During the 2025 Mw 7.7 Myanmar Earthquake
Core Problem: Distant sedimentary basins can expose high-rise buildings to prolonged, resonant shaking.
Key Innovation: Public videos of 28 Bangkok buildings during the 2025 Myanmar earthquake yield response periods of 2.8–6.9 s, overlapping the basin's long-period range. Video-derived deformation documents substantial demand, but resonance alone does not establish the cause of a particular building's collapse.
13. Characteristics of Coseismic and tsunami-related ionospheric disturbances from the 2025 Mw 8.7 Kamchatka earthquake
Core Problem: Separating earthquake-driven and tsunami-driven ionospheric signals requires broad spatial observations and independent wave records.
Key Innovation: GNSS records from 2,769 stations identify successive disturbance regimes after the 2025 Kamchatka event, with later tsunami-related signals compared against six DART records. This is a post-event propagation study; the approximately ten-minute onset of the first detected disturbances is not earthquake prediction.
14. A comprehensive laboratory research on wave transmission at Reef-Balls
Core Problem: Wave-transmission formulas for Reef Balls may assume breaking mechanisms that realistic deployments seldom produce.
Key Innovation: A campaign of 1,416 two-dimensional random-wave experiments places the modules directly on the seabed. Observations support macro-roughness friction as the main dissipation mechanism, with only sporadic breaking, motivating a physically interpretable transmission formula for the tested layouts.
15. Applicability and quality evaluation for vibro-compaction reinforcement of coral sand foundations
Core Problem: Weak interlayers and drainage conditions complicate reinforcement of reclaimed coral-sand foundations.
Key Innovation: Cyclic triaxial tests are combined with four field vibro-compaction schemes and in-situ evaluation. A triangular 3.5 m layout with centroid densification and coarse backfill performs best among the tested schemes. Its construction parameters and penetration correlations require site-specific checks before transfer.
16. Regional-scale geotechnical parameter inference with landslide-informed priors: a dual-constraint weakly supervised deep learning framework
Core Problem: Regional slope models need geotechnical parameters where direct measurements are sparse.
Key Innovation: Title-level scope: a dual-constraint weakly supervised framework uses landslide-informed priors for regional geotechnical inference. Parameter identifiability, independent validation and transfer beyond the study region remain unassessed without a matching abstract.
17. Challenges in modelling rock slope failures using a distinct element method: A critical analysis based on the 2023 Brienz/Brinzauls ’Insel’ landslide
Core Problem: Distinct-element simulations can reproduce a failure while retaining uncertainty in structure and parameter choices.
Key Innovation: Title-level scope: a critical modelling analysis of the 2023 Brienz/Brinzauls Insel landslide. The available record does not establish which modelling assumptions were validated or quantify predictive skill.
18. Dynamic landslide hazard assessment through rainfall-driven log-odds updating and MT-InSAR based hazard refinement: The June 2024 event in Pingyuan, Guangdong, China
Core Problem: Rainfall updates and deformation observations need a coherent role in changing landslide hazard estimates.
Key Innovation: Title-level scope: rainfall-driven log-odds updating and MT-InSAR refinement for the June 2024 Pingyuan event. Forecast lead time, prospective skill and the independence of validation data remain unassessed.
19. A terrain-following framework for gravity-driven flow modeling on erodible beds
Core Problem: Runout models must represent both terrain geometry and exchange with an erodible bed.
Key Innovation: Title-level scope: a terrain-following framework for gravity-driven flow modelling on erodible beds. Conservation properties, entrainment performance and field runout accuracy cannot be determined from the available title-level record.
20. Hierarchical controls on failure modes and cliff retreat in flysch sea cliffs: The French Basque coast
Core Problem: Cliff retreat depends on interacting geological controls operating at different scales.
Key Innovation: Title-level scope: hierarchical controls on failure modes and retreat of flysch cliffs along the French Basque coast. The relative importance of each control and the strength of the observational evidence remain unassessed.
21. Climatology and trends of observed daily and hourly extreme precipitation in the French Alps
Core Problem: Daily and hourly rainfall extremes can change differently across the same mountain region.
Key Innovation: Observed French-Alps records and extreme-value analysis show contrasting north–south daily trends and more spatially variable hourly trends. The results support duration- and region-specific hazard inputs, rather than a uniform increase in Alpine rainfall extremes.
22. Seismic analysis of bedload discharge at Tagliamento River during flood events
Core Problem: Existing seismic networks may contain river-process information even when stations were not placed beside channels.
Key Innovation: Three permanent stations up to about 3 km from the Tagliamento River show frequency-dependent flood signatures and stage–amplitude hysteresis. These observations motivate flood-monitoring development; they do not yet provide independently calibrated bedload fluxes or a demonstrated operational warning lead time.
23. An integrated wildfire event database for improved fire science research for the continental United States, 1992–2021
Core Problem: Inconsistent fire sizes, complexes and duplicate event identities affect historical wildfire statistics.
Key Innovation: The continental-US 1992–2021 database reconciles three national sources with regional records and additional reports, explicitly resolving merged, mis-sized and duplicate events. It provides a basis for reassessing extreme-fire activity; completeness and event-level accuracy still depend on the source records. The discussion preprint is distinguished from a final peer-reviewed article.
24. Integrated seismic and electrical imaging with clustering for the hydro-mechanical screening of a high-altitude earthfill dam
Core Problem: Remote alpine dams need non-invasive screening of potential hydraulic and mechanical anomalies.
Key Innovation: A newly available matching abstract describes ERT, seismic refraction and surface-wave measurements at Fourcare dam, integrated through clustering. The anomalies are potentially associated with infiltration; they do not by themselves diagnose active leakage or establish structural safety. This updates an earlier title-level entry.
25. Geophysical site characterization and scenario-based seismic hazard mapping with slope-based topographic amplification factors: case study of the hilltop Trakošćan castle, Croatia
Core Problem: Hilltop heritage sites require treatment of both subsurface contrasts and ridge geometry.
Key Innovation: The matching abstract adds 66 HVSR measurements, eight MASW profiles and four earthquake scenarios at Trakošćan Castle. Combined empirical factors give amplification near 2.0 at steep flanks versus about 1.3 under the code comparison; decoupled factors still require instrumental or 2D/3D validation. This updates an earlier title-level entry.
26. Dynamic Mechanical Response and Energy Dissipation‑Induced Deterioration Characteristics of Sandstone Under Cyclic Loading
Core Problem: Repeated loading can either temporarily compact rock or accumulate damage before a later impact.
Key Innovation: Cyclic preconditioning followed by SHPB impact on mine sandstone reveals non-monotonic strength and stiffness changes. Higher preconditioning eventually degrades capacity, while faster impacts produce more abrupt fragmentation. The laboratory sequence informs pillar-instability mechanisms without defining a field rockburst threshold.
27. Century-scale spatiotemporal variability and trends of rainfall erosivity and erosivity density in an ecologically sensitive tropical monsoon region of India (1901–2020)
Core Problem: Century-scale erosion assessment needs to separate regional rainfall variability from short intense storm processes.
Key Innovation: Monthly gridded rainfall for Kerala supports four erosivity domains and mostly declining long-term trends, with localized highland increases. Arnoldus-based estimates support regional planning; they are not direct measurements of storm energy or sub-hourly erosive intensity.
28. Multi-temporal gully expansion and volumetric soil loss estimation in Kakia and Esamburmbur sub-catchments, Enkare-Narok Basin, Kenya
Core Problem: Gully expansion and volumetric loss need repeated measurements to inform catchment management.
Key Innovation: Title-level scope: multi-temporal gully expansion and soil-loss volumes in two Kenyan sub-catchments. Survey uncertainty, change-detection thresholds and the reported erosion rates remain unassessed without a matching abstract.
29. Nonlinear and seasonal response mechanisms of hydrological drought in coupled human and natural systems
Core Problem: Human activity and natural seasonality can produce nonlinear drought responses.
Key Innovation: Title-level scope: seasonal and nonlinear hydrological-drought mechanisms in coupled human and natural systems. The study's attribution design, geographic coverage and forecasting evidence cannot be assessed from the available record.
30. Runoff fingerprinting reveals transitions in urban underpass flooding from local surface-runoff convergence to sewer-network feedbacks
Core Problem: An underpass can flood through local surface convergence or feedback from the connected sewer network.
Key Innovation: Title-level scope: runoff fingerprinting of transitions between these flooding mechanisms. Data provenance, transition thresholds and independent hydraulic validation remain unassessed without a matching abstract.
31. Mechanisms of lining damage accumulation and bearing capacity degradation under cyclic freeze-thaw action of water-filled crown cavities behind tunnel linings in cold regions
Core Problem: Water trapped in crown cavities may accumulate freeze–thaw damage and reduce lining capacity.
Key Innovation: Title-level scope: lining damage and bearing-capacity degradation under cyclic freezing and thawing of water-filled cavities. The available record establishes the studied mechanism, but not measured capacity losses or safe design limits.
32. Mechanical response of tunnels under sequentially combined seismic loading and fault dislocation: insights from FDM-DEM coupled approach
Core Problem: A tunnel may experience sequential shaking and permanent fault displacement rather than either action alone.
Key Innovation: Title-level scope: coupled FDM–DEM investigation of tunnels under sequential seismic loading and fault dislocation. Interaction effects, validation and damage magnitudes remain unassessed without a matching abstract.
33. Torsional behavior of buildings induced by pounding of base-isolated systems under long-period and pulse-like ground motions
Core Problem: Mass eccentricity and building height can change the consequences of impacts in base-isolated buildings.
Key Innovation: Three-dimensional frame models distinguish severe pulse-driven corner impacts from repeated lower-amplitude long-period pounding. The studied five- and fifteen-storey configurations show different localized and cumulative damage demands; those model results are not universal impact limits for isolated buildings.
34. Seismic resilience assessment and RL-guided restoration optimization of substations considering supply-capacity degradation
Core Problem: Post-earthquake restoration must account for degraded supply capacity rather than only component availability.
Key Innovation: Title-level scope: seismic resilience assessment and reinforcement-learning-guided restoration of substations. Recovery improvements, training assumptions and operational feasibility remain unassessed without a matching abstract.
35. HypoNet Nankai PS: rapid hypocenter determination for the Nankai Trough using physics-informed neural networks with three-dimensional P- and S-wave velocity models
Core Problem: Fast three-dimensional travel-time inference needs to accommodate steeper S-wave gradients as well as P-wave arrivals.
Key Innovation: HypoNet Nankai PS revises PINN training and adds an empirically converted S-wave model, with fast-marching benchmarks and 223 DONET earthquakes. Catalog agreement beneath the network and offshore relocation differences support evaluation, while velocity-model uncertainty and true hypocenter accuracy remain distinct questions.
36. Experimental study on a tracked underwater remediation technique for scour around offshore wind turbine monopile
Core Problem: Scour remediation must operate around submerged monopiles while addressing local sediment removal.
Key Innovation: Title-level scope: a tracked underwater remediation technique tested around offshore wind-turbine monopiles. Remediation efficiency, environmental effects and full-scale performance cannot be inferred from the available record.
37. AHP-based risk assessment of local scour at offshore wind turbine monopiles
Core Problem: Scour assessments must combine several site and structural risk factors transparently.
Key Innovation: Title-level scope: analytic-hierarchy-process assessment of local scour at offshore monopiles. Weight robustness, calibration and observed predictive accuracy remain unassessed without a matching abstract.
38. Experimental study of wave transmission comparison for three pontoon floating breakwater types
Core Problem: Comparing breakwater shapes requires consistent forcing, draft and mooring conditions.
Key Innovation: Eleven pontoon models tested under regular waves show lower transmission for a stepped model than a trapezoidal counterpart of equal dimensions, and performance comparable to a longer rectangular model. Draft has a stronger influence than the tested mooring changes; transfer to irregular extreme seas needs further evaluation.
39. A preliminary study on a synergistic assimilation scheme for multi-band satellite soil moisture data
Core Problem: Combining microwave products requires attention to frequency-dependent uncertainty and vegetation effects.
Key Innovation: A CoLM assimilation experiment combines SMAP, ASCAT and MWRI retrievals, improving mean correlation by about 0.25 and reducing error from 0.068 to 0.058 m³/m³ against Chinese in-situ observations. The result improves a potential hazard-model input, without validating landslide forecasts.
40. A Two-Step Dual-Polarimetric Sparse Bayesian Learning D-TomoSAR Method for Urban Building Deformation Monitoring
Core Problem: Urban radar layover makes adjacent and weak scatterers difficult to separate.
Key Innovation: A coarse-to-fine dual-polarization sparse Bayesian inversion jointly estimates elevation and linear deformation. A 22-image PAZ test reports 2.72 m building-height RMSE for five reference buildings; this height statistic is not a validated deformation-velocity error.
41. RPC Bias Compensation of KOMPSAT-3 Optical Imagery Based on KOMPSAT-5 SAR
Core Problem: Optical-image geolocation can be biased where conventional surveyed image control is difficult to obtain.
Key Innovation: KOMPSAT-5 SAR registration supports RPC correction of one KOMPSAT-3 scene, evaluated against independent GNSS checkpoints. Flat-terrain controls perform better than steep or forested ones, and the single image-pair demonstration does not establish general performance in unstable mountain terrain.
42. Comparison of Assimilating Clear and All-Sky GOES-16 ABI Radiance in a Convective-Scale Model to Improve the Forecast of an MCS Event
Core Problem: Cloud-affected radiances can constrain evolving convection differently from clear-sky observations.
Key Innovation: Five-minute ensemble assimilation of GOES-16 radiances improves cloud and precipitation forecasts for one mesoscale convective system, with larger gains from all-sky data. The case supports meteorological input development rather than a demonstrated flood or landslide warning system.
43. Structure-constrained UMamba3D for seismic fault segmentation and application in the Sichuan Basin
Core Problem: Fault segmentation in seismic volumes needs structural continuity and transfer from synthetic training examples.
Key Innovation: A structure-constrained UMamba3D model is evaluated on synthetic data and a Sichuan Basin seismic volume. Improved interpretation can support geological models; segmented discontinuities do not directly establish active fault slip or earthquake probabilities.
44. ICIBS: Increasing the Context for Ice Boundary Segmentation
Core Problem: Temperate-ice scattering makes radar ice-bottom boundaries weak and difficult to label.
Key Innovation: Spatial and geographical context improves ice-boundary segmentation on IceAnatomy and a new Patagonia campaign, with reported improvement over the baseline on unseen data. Ice-thickness mapping is a supporting input; this study does not directly evaluate glacier failure or outburst-flood hazard.
45. Observation Assimilation Transformer: Probabilistic Weather Analysis from Sparse Station Observations
Core Problem: Sparse station observations need to refine numerical weather fields with calibrated uncertainty.
Key Innovation: The Observation Assimilation Transformer conditions arbitrary query locations on sparse stations and an NWP prior, using probabilistic training and relative geometry. Regional and continental tests improve station-based scores; these results do not establish storm-hazard prediction skill at unobserved disaster sites.
46. FoundPDE: A generative foundation model for data-driven symbolic discovery of partial differential equations
Core Problem: Data-driven equation discovery must generalize beyond predefined symbolic libraries and clean trajectories.
Key Innovation: FoundPDE learns trajectory-to-equation mappings across 20 PDE families and supports few-shot extension. Recognition falls from 97.5% on clean in-domain data to 52.09% at 10% multiplicative noise. No direct geohazard validation is reported, so the foundation-model label does not justify a higher hazard score.
47. Phase-resolving wave surface prediction under typhoon conditions using hierarchical spatiotemporal deep learning model
Core Problem: Phase-resolving wave-surface estimates need to represent rapid spatial and temporal variation under typhoons.
Key Innovation: Title-level scope: hierarchical spatiotemporal deep learning for typhoon wave-surface prediction. Forecast horizon, extreme-wave skill and out-of-event generalization remain unassessed without a matching abstract.
48. Specimen Size Effects on the Dynamic Behavior of Soil Rock Mixtures in Large Scale Triaxial Tests
Core Problem: Laboratory specimen size can bias dynamic properties assigned to heterogeneous soil–rock mixtures.
Key Innovation: Large cyclic triaxial tests, up to 300 × 600 mm, show size-dependent strength, accumulated strain, stiffness and damping. The results motivate scale-aware parameter calibration; they do not directly validate a slope or embankment earthquake prediction.
49. Coupling Acoustic Emission with a Data-Driven Model for Fracture Evolution Investigation of Red Sandstone
Core Problem: Fracture waveforms contain mixed signals that are difficult to associate with evolving damage modes.
Key Innovation: Time–frequency images, learned features and clustering distinguish three acoustic-emission groups during red-sandstone fracture. Their differing energy, frequency and spatial patterns aid specimen-scale interpretation; no field warning performance follows from the clustering alone.
50. A Micromechanical Anisotropic Fatigue Damage Model for Brittle Rocks with Preferential Crack Orientations Under Cyclic Loading
Core Problem: Preferential discontinuities and distributed microcracks interact during repeated loading.
Key Innovation: A thermodynamic micromechanical model links fatigue accumulation to anisotropic damage and reproduces tests on three rock types. Agreement with laboratory deformation and lifetime patterns supports model evaluation, rather than an independently demonstrated field-failure forecast.
51. Pore Structure Evolution and Seepage Behavior of Burnt Rock Mass Under Triaxial Compression
Core Problem: Porosity alone may not describe hydraulic connectivity in thermally altered rock.
Key Innovation: Triaxial CT reconstruction and seepage modelling across four burnt-rock types show that fractures govern connected flow paths, while pores and matrix affect local discharge. Specimen-scale differences provide a basis for field investigation, not a quantified mine-water-inrush prediction.
52. A unified thermodynamics-based constitutive model for granular soils with and without particle breakage
Core Problem: Particle breakage changes soil stress response and the reference critical state.
Key Innovation: A thermodynamically consistent formulation couples plastic deformation and breakage through state variables and dissipation constraints. Drained and undrained triaxial comparisons reproduce stress, pore-pressure and breakage trends; hazard-scale validation is outside the supplied evidence.
53. A unified experimental-numerical assessment of Cable, Pile, and Zone solid rockbolt models in FLAC3D: bond-slip, load transfer, and model selection
Core Problem: Different numerical rockbolt representations can yield different responses despite similar interface calibration.
Key Innovation: Pull-out calibration and independent short-embedment validation compare Cable, Pile and solid Zone models before extension to tunnel support. Nonlinear interfaces capture progressive slip, with a trade-off between global efficiency and local contact detail; model agreement is not a demonstration of tunnel safety.
54. Boundary-aware hard-sample learning for TBM surrounding rock classification via attention-enhanced LightGBM
Core Problem: Transient TBM operations and adjacent-class overlap complicate geological classification from machine records.
Key Innovation: Stable operational segments from the Beishan project support a three-class ensemble-classification task with 94.468% reported accuracy. Boundary errors and limited weak-rock samples remain important; operational classification is not a calibrated collapse or water-inrush forecast.
55. Moisture–dependent evolution of NMR–derived pore structure in saline and non–saline soils during freeze–thaw cycles
Core Problem: Moisture and salinity can redistribute soil pore sizes through repeated freezing and thawing.
Key Innovation: NMR measurements across moisture conditions and 1–15 cycles show strong moisture-dependent pore-size changes. The measured structure informs water-transport interpretation, while slope stability and hydraulic failure thresholds were not established by this experiment.
56. Impact of freeze–thaw systems on cohesive soil pore structure: Numerical Image (NIA) and Factor Analysis (FA)
Core Problem: Open and closed freezing systems can produce different pore-geometry changes in cohesive soils.
Key Innovation: Numerical image analysis and multivariate statistics compare soils after one freeze–thaw cycle, finding larger heterogeneous changes with water migration. The result distinguishes experimental boundary conditions rather than predicting long-term ground degradation from a single cycle.
57. On the mechanical behaviour of a bonding-dominated soft rock
Core Problem: A common reconstituted reference state cannot be assumed for carbonate rocks with different bonding structures.
Key Innovation: Drained triaxial and high-pressure compression tests show convergence of intact and reconstituted Palermo Calcarenite, unlike the weakly bonded comparison material in one-dimensional compression. Cementation-dependent reference behavior matters for constitutive calibration, without directly establishing slope-failure conditions.