TerraMosaic Daily Digest: October 5, 2026
This catch-up issue is part 2 of a three-issue series allocated from the October 8, 2026 collection. Each issue reviews 429 input records. October 4–6 are editorial issue dates rather than exact paper-publication windows.
Daily Summary
The October 5 catch-up issue follows damage from concealed rock-mass changes to cascading movement. A Minjiang River case reconstructs staggered initiation of three secondary slide zones through progressive loading and stress transfer. Along the upper Yellow River, Scoops3D-derived stability information and reservoir water-level dynamics improve a CNN assessment over the component models. Rock-bridge experiments detect internal shear damage well before peak resistance, while multi-block rockfall tests show that collision-driven dissipation across a group can coexist with longer extreme trajectories for individual blocks.
Different observation systems expose different parts of the evolving state. Campi Flegrei gas records and deformation are linked through a thermodynamic interpretation of hydrothermal pressurization and condensation heating. A slope-radar framework separates nonstationary deformation regimes without treating a fitted sublinear curve as proof of imminent failure. Ground-constrained oilfield inversion reduces ambiguity in subsidence estimates, and building-scale InSAR decomposes settlement, tilt and curvature. A LiDAR–InSAR mining study improves the original LiDAR result but remains less precise overall than InSAR alone, making spatial coverage the main complementarity.
Underground risk assessment increasingly combines geology with mechanically explicit observations. Drilling records reconstruct structural planes, and TBM noise provides a passive source for surface–borehole forward prospecting. Physics-constrained microseismic localization addresses speed and velocity-model uncertainty together. Tailings-dam sensitivity analysis identifies where uncertain resistance parameters matter most, while a coupled pressure–temperature diagram translates fault-reactivation conditions into operational screening constraints.
Supporting methods address layered-rock creep, laboratory-scale freeze–thaw deformation and accumulated creep–fatigue damage. A masonry drift-capacity formulation replaces abrupt failure-mode categories with a continuous relation calibrated against cyclic wall tests, connecting model structure to seismic vulnerability assessment. Verified title-level records cover probabilistic landslide stratigraphy, weak-interlayer MPM, coal-fire cavities, roadway rockburst, tunnel gush hazards and compound bridge or dam demand. Their verified scope is reported without attributing unassessed methods or results.
Key Trends
The strongest papers resolve how internal damage, joint geometry, hydrology and uncertain measurements propagate into observable deformation and hazard extent.
- Failure is being treated as a sequence: Cascade initiation, rock-bridge deterioration and colliding blocks reveal how localized changes develop into larger motion or extended impact corridors.
- Subsurface information is becoming usable during construction: Measurement-while-drilling, passive TBM seismic sources and physics-constrained localization connect hidden geology with operational monitoring.
- Sensor fusion needs task-specific validation: Combining LiDAR, InSAR and ground observations can reduce ambiguity or improve coverage without outperforming every input sensor on the same accuracy metric.
- Uncertainty is entering engineering decisions explicitly: Reservoir-slope physics, tailings resistance, calibration uncertainty and coupled fault-stability envelopes make assumptions inspectable before design choices.
Selected Papers
Progressive landslide initiation, reservoir-corridor stability, rock-bridge damage and multi-block rockfall anchor this selection. The monitoring papers connect volcanic degassing, radar displacement, subsidence and hidden structural planes to evolving mechanical states. Companion studies address mine microseismicity, tunnel prospecting and seismic response, tailings reliability and fault reactivation. Subsurface modelling and other title-level records are included for their verified scope, with unassessed results clearly identified.
1. Initiation mechanism of a large-scale cascade landslide: a case study of progressive failure
Core Problem: Secondary slide zones may initiate sequentially as loading and stress redistribute.
Key Innovation: Field observations and simulations reconstruct staggered movement of three zones into the Minjiang River, linking progressive loading and saturation to mobility rather than assuming a simultaneous rigid-body release.
2. Regional landslide hazard assessment in the upper Yellow River using an integrated Scoops3D-CNN model
Core Problem: Regional susceptibility needs stability physics and reservoir-operation effects.
Key Innovation: Scoops3D information and statistical controls are blended in a CNN, including maximum water-level change rate; test AUC 0.89 exceeds the separate physical and CNN models in the studied upper Yellow River corridor.
3. Ultrasonic Detection of Shear-Induced Damage in Rock Bridges: Experimental Analysis and DEM Verification
Core Problem: Nonpersistent joints can retain apparently intact bridges while internal damage develops.
Key Innovation: Shear tests, ultrasonic measurements and DEM detect initial damage at roughly 50–70% of peak strength and explain geometry-dependent failure, supporting earlier recognition of rock-bridge degradation.
4. Quantifying Hazard Amplification in Multi-Block Rockfalls Based on Group Suppression and Individual Enhancement
Core Problem: Energy dissipation across a group can coexist with amplified extreme individual trajectories.
Key Innovation: Controlled tests show inter-block collisions redistribute energy and yield enhancement coefficients of 1.14–1.73, explaining why single-block models can underestimate hazard reach within the experimental configurations.
5. Escalating soil CO₂ degassing correlates with ground uplift and seismicity during the ongoing unrest of Campi Flegrei (2005–2025)
Core Problem: Accelerating caldera uplift needs mechanisms consistent with gas and seismic observations.
Key Innovation: Two decades of Campi Flegrei soil degassing correlate with uplift; thermodynamic modelling links fluid pressurization and condensation heating to deformation, supporting concern about escalating unrest without forecasting an eruption date.
6. Spectral Element Extension for Direct Simulation of Rotational Ground Motions
Core Problem: Translation maps alone can miss the locations of greatest rotational ground motion.
Key Innovation: A SPECFEM3D extension computes rotations from displacement gradients and passes analytical validation; basin simulations show low-velocity sediments amplify and redistribute peak rotation rates differently from peak velocity.
7. An efficient computational framework for seismic reliability analysis of submarine slopes
Core Problem: Spatially variable strength creates high-dimensional seismic reliability problems.
Key Innovation: PLS reduces 203 inputs to nine latent variables before SVR prediction, achieving R² 0.9751 in pseudo-dynamic Bishop calculations; the surrogate inherits the assumptions of that reference slope model.
8. Integrating Multisource Observations for Spatiotemporal Subsidence Monitoring and Constrained High-precision Reservoir Parameter Inversion in the Shuguang Oilfield
Core Problem: Single-view InSAR leaves geometric ambiguities in reservoir and deformation estimates.
Key Innovation: Ground-constrained Bayesian inversion supplements 2017–2025 InSAR and UAV damage mapping, reducing model residual variance and parameter uncertainty while documenting cumulative deformation up to 120 cm.
9. Integrated Multipattern Building Risk Assessment Based on MT-InSAR Measurements
Core Problem: One displacement indicator cannot distinguish uniform settlement, tilt and curvature.
Key Innovation: Multi-temporal InSAR is decomposed into engineering-relevant building deformation components and evaluated against component-specific criteria in Hong Kong reclaimed areas, identifying potential construction- and consolidation-related risk.
10. Flood Mapping Using Sentinel-1 Dual-Polarimetric SAR Data in Agricultural Landscapes
Core Problem: Sentinel-1 flood monitoring needs descriptors adapted to agricultural scattering.
Key Innovation: A surface-scattering flood descriptor is implemented in Earth Engine and evaluated for floods in Bihar and Odisha, linking high values to flood-inducing rainfall across two regions.
11. Invited perspective: Uncertainties in natural systems may be uncomfortable, but ignoring them would be absurd
Core Problem: Recognized sources of uncertainty are often hidden inside one forecast number.
Key Innovation: The perspective proposes a complete forecast that keeps uncertainty sources distinct and connects their representation to model evaluation, communication and decisions; it is a conceptual framework rather than a validated predictive algorithm.
12. A 256-year daily rainfall record reveals multiscale rainy-season variability and emerging hydroclimatic hazard over northern East Asia
Core Problem: Short instrumental records cannot separate long modulation from recent rainy-season changes.
Key Innovation: A Seoul daily record spanning 1770–2025 resolves multidecadal rainy-season variability and its conjunction with stronger extremes; near-term duration outlooks are exploratory rather than established forecasts.
13. Dynamic evolution characteristics and laws of ground settlement due to shield tunneling in water-rich sandy pebble strata
Core Problem: Water-rich sandy-pebble strata can evolve from local cavity formation to surface settlement.
Key Innovation: A three-zone model, collapse criterion and similarity tests resolve cracking, collapse-arch formation and restabilization, with laboratory-scale height and settlement estimates consistent with measurements.
14. A Fusion Framework Integrating UAV LiDAR and SBAS-InSAR for Enhanced Mining-Induced Surface Subsidence Modeling
Core Problem: InSAR precision and LiDAR coverage differ across deformation gradients.
Key Innovation: Bias correction and residual kriging reduce LiDAR RMSE from 181 to 120 mm, while remaining above InSAR's 57 mm; the gain lies in spatial coverage of steep gradients, not uniformly superior accuracy.
15. Mechanics-consistent interpretation of non-stationary rock mass deformation using SSR monitoring and regime-based analysis
Core Problem: Nonstationary radar displacement can make fixed velocity thresholds misleading.
Key Innovation: Mechanically interpreted features, clustering and change-point analysis separate deformation regimes; three clusters are a motivated modelling choice, and the fitted sublinear exponent is not evidence of imminent failure.
16. Contrasting heavy rainfall events and associated mechanisms over the north western Himalayas of India
Core Problem: Short intense storms and persistent multi-day rainfall can arise from different synoptic pathways.
Key Innovation: Two events contrast low-pressure-system interaction with a slow upper-level trough and sustained monsoon moisture transport, clarifying distinct trigger configurations in complex Himalayan terrain.
17. Study of key driving factors in failure mechanisms of land subsidence and ground fissures in the southwest Tehran plain-Iran using SAR interferometry
Core Problem: Mapped deformation must be related to observed surface damage and groundwater use.
Key Innovation: A seven-year Tehran deformation field is compared with 369 field locations and 49 documented features, identifying strong spatial associations with extraction and land use alongside geological controls.
18. TriPhy-GF—A Triple-Physics-Constrained Graph Neural Network for Accurate and Efficient Microseismic Event Localization in Coal Mines
Core Problem: Fast event localization must remain consistent with uncertain wave speeds.
Key Innovation: TriPhy-GF embeds travel-time physics in graph topology, supervision and residual loss, approaching grid-search accuracy with millisecond inference on numerical and Dongtan mine data.
19. Data-Driven 3D Reconstruction of Rock Mass Structural Planes from Measurement-While-Drilling Data
Core Problem: Hidden discontinuities need rapid three-dimensional characterization during excavation.
Key Innovation: Laboratory drilling and spatially grouped change points reconstruct structural planes from MWD; Xiwuling field predictions agree with borehole images, supporting geology-aware stability interpretation.
20. Reliability-Based Sensitivity Analysis of Tailings Dam Resistance Parameters
Core Problem: Uncertain strength parameters can dominate dam reliability differently under local and global sensitivity measures.
Key Innovation: FORM and surrogate-assisted Sobol analysis identify residual-soil friction as a leading control, with embankment friction also important globally; conclusions apply to the tested Brazilian dam.
21. A novel approach for the estimation of seismic drift capacity in unreinforced masonry (URM) shear walls based on governing structural parameters
Core Problem: Failure-mode categories can introduce artificial jumps in estimated masonry drift capacity.
Key Innovation: A numerical matrix and independent cyclic-wall database support a continuous parameter-based formulation linking axial load, slenderness and moment distribution to seismic deformation capacity, with the stated standards application attributed to the authors.
22. Probabilistic landslide stratigraphic modelling by uncertainty fusion of surface wave inversion and borehole data
Core Problem: The verified study combines surface-wave inversion and borehole information to represent uncertain landslide stratigraphy.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
23. Borehole-derived 3D geological model for B-spline MPM analysis of landslides with weak interlayers
Core Problem: The verified study examines how a borehole-derived three-dimensional geological model supports landslide simulation with weak layers.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
24. A unified pressure-temperature stability diagram for assessing injection-induced fault reactivation potential
Core Problem: Concurrent cooling and pressurization can change the stress regime and susceptible fault orientation.
Key Innovation: A poro-thermo-elastic diagram maps three-dimensional Mohr–Coulomb constraints into operational pressure–temperature envelopes for conceptual storage and geothermal settings, supporting early screening rather than site certification.
25. Targeting Flood Resilience Investments Through a Readiness Gaps Index: Evidence from Mississippi
Core Problem: High exposure and weak institutional readiness can coincide at different administrative scales.
Key Innovation: A Mississippi index combines socioeconomic and stakeholder-informed capacity indicators, flags about 37% of units for potential priority intervention and tests weighting sensitivity without claiming demonstrated investment outcomes.
26. Mapping subsurface combustion cavities in underground coal fires using nonlocal DS-InSAR and physics-based inversion
Core Problem: The verified study maps subsurface combustion cavities using nonlocal distributed-scatterer InSAR and physical inversion.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
27. Mitigating urban tunneling risks: Detecting adverse geology with a tunnel-surface-borehole observation system using TBM drilling noise source
Core Problem: Adverse geology ahead of urban tunnel faces needs broader seismic illumination.
Key Innovation: Surface and borehole arrays use TBM drilling noise to constrain P- and S-wave structure; synthetic tests and a power-tunnel application support anomaly imaging before excavation.
28. Static-dynamic triggering mechanisms of rockburst in straight-walled arch roadways under different principal stress orientations and impact disturbances
Core Problem: The verified study examines stress orientation and impact disturbance as joint controls on roadway rockburst triggering.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
29. Nonlinear sea level rise and storm surge interactions amplify future coastal flood hazards in the world’s largest urban delta
Core Problem: The verified study assesses nonlinear interaction between sea-level rise and storm surge in an urban delta.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
30. From end-of-pipe detention to source–process regulation: multi-objective optimization of gray–green infrastructure for urban flood mitigation
Core Problem: The verified study compares source–process regulation and detention through multi-objective infrastructure optimization.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
31. An Interpretable XGBoost-SHAP framework for rapid regional assessment of liquefaction-induced lateral spreading in mildly sloping ground
Core Problem: Rapid regional displacement estimates need physics-based training and independent experimental checks.
Key Innovation: XGBoost trained on 50,000 calibrated FEM cases is additionally checked against five unseen centrifuge tests; a hypothetical regional application demonstrates speed, not completed regional field validation.
32. Bayesian framework for uncertainty-informed seismic capacity modeling of embankment dams for fragility and risk assessment
Core Problem: The verified study concerns uncertainty-informed embankment-dam capacity, fragility and risk; the mismatched raw abstract is excluded.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
33. Seismic fragility assessment of scoured bridges under coupled horizontal-vertical ground motions
Core Problem: The verified study evaluates scoured bridges under combined horizontal and vertical ground motion.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
34. Mechanism of gush disasters caused by structural failure of shield tunnel through water-rich sand stratum: A 3D MPM Study Combining Point Cloud Data Processing Method
Core Problem: The verified study examines structural tunnel failure and gush mechanisms in water-rich sand through three-dimensional MPM.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
35. Discrepancies in Extreme Daily and Monthly Precipitation Trends Between Observations and Earth System Models
Core Problem: Regional precipitation trends can disagree even when global extreme-rainfall trends appear consistent.
Key Innovation: Earth-system ensembles underestimate the magnitudes of observed daily-extreme trends and overstate the spatial prevalence of increasing monthly extremes, exposing scale-dependent limits for regional hazard projections.
36. Projected 5G Interference Can Degrade Passive Microwave Radiance Assimilation and Alter High‐Impact Weather Forecasts
Core Problem: Telecommunication leakage near a passive-microwave band may corrupt moisture observations.
Key Innovation: A projected 5G scenario perturbs assimilated radiances and Hurricane Ida-remnant rainfall forecasts through both retained biases and observation rejection; results concern the tested scenario, not observed present-day forecast degradation.
37. Resolution Dependence in a Global Atmospheric Simulation From km to 220 m Grid Spacing
Core Problem: Kilometre-scale convection and turbulence choices can produce unrealistic local rainfall concentrations.
Key Innovation: A 220 m global simulation weakens spurious localized precipitation and shows partial convergence across turbulence schemes, while cloud responses remain scheme-dependent; downstream hazard-model gains are not directly tested.
38. SoftSEEPS improves ML-based precipitation forecasting
Core Problem: Training rainfall predictors with squared error may not optimize categorical precipitation skill.
Key Innovation: SoftSEEPS makes the SEEPS score differentiable and trains an IMERG precipitation decoder jointly with RMSE, exposing trade-offs between the two objectives rather than claiming a new hazard forecast system.
39. Artificial intelligence pathways from weather to climate
Core Problem: Weather skill does not establish credible responses to altered climate forcing.
Key Innovation: The review identifies explicit independently variable forcing and stress tests on extremes and counterfactuals as minimum requirements for AI climate models, distinguishing forecast emulation from intervention-ready climate projection.
40. Attention OffsetNet: Translation-Free Transfer Learning for Optical–SAR Registration With Full-Scene Subpixel Refinement
Core Problem: Radiometric differences complicate reliable cross-sensor geometric alignment.
Key Innovation: Transfer learning improves local offset estimation, while RANSAC aggregation produces scene-level subpixel registration; local network RMSE remains several pixels and must not be presented as subpixel inference.
41. Pysammos 1.0.0: a discrete-to-continuum transformation Python tool to analyse the rheology of granular materials
Core Problem: Particle simulations need consistent continuum fields for constitutive interpretation.
Key Innovation: Pysammos coarse-grains polydisperse DEM–CFD results into reusable analysis formats and supports separate phases and nonspherical particles, offering open processing rather than a new validated landslide warning model.
42. InSAR-Derived Surface-Deformation-Constrained Machine Learning for Permafrost Occurrence Probability Mapping in the Eastern Kunlun Mountains
Core Problem: Uneven training samples limit permafrost distribution mapping.
Key Innovation: InSAR annual rates and seasonal amplitudes constrain GBM training samples and agree with 59 field-positive sites; that agreement does not by itself validate non-permafrost specificity or every region.
43. Large-Scale Testing for Characterisation of Cable Bolt System Behaviour: A Novel Laboratory Approach to Simulate Relative Joint Shear Sliding
Core Problem: Joint sliding challenges support components differently from simple pullout loading.
Key Innovation: A multiaxis test system separates pretension, grouting, anchorage and lateral restraint effects; grouting governs load transfer while pretension benefits diminish after larger-displacement cracking and yielding.
44. Laboratory Study on the Anisotropic Creep Behavior and Failure Mechanism of Layered Rocks Under Confining Pressure
Core Problem: Bedding and confinement jointly shape time-dependent failure.
Key Innovation: Sandstone creep tests identify nonmonotonic confinement effects on anisotropy and a tensile-to-shear transition, motivating bedding-aware support assessment within the investigated stress range.
45. Experimental Investigation of Self-Expanding Polyurethane Grout Propagation in Two-Dimensional Fracture Networks
Core Problem: Connected fractures redistribute self-expanding grout unlike isolated fractures.
Key Innovation: Two-dimensional tests vary aperture, intersection angle and temperature and reveal heterogeneous cured density; geometry explains filling differences while proposed local-heat effects remain unmeasured.
46. Mesoscale numerical modelling of freeze–thaw deformation in closed- and open-system soils using a Voronoi–Delaunay lattice
Core Problem: Fast laboratory-scale deformation models need clear limits on coupled freezing physics.
Key Innovation: A Voronoi–Delaunay framework passes conduction and mechanical benchmarks and fits heave/settlement endpoints; it remains a small-strain reduced formulation rather than complete Darcy, effective-stress and ice-lens mechanics.
47. Time-series anomaly detection and interval-based risk identification for station roof slab displacement based on multi-model consistency fusion
Core Problem: The verified study examines station-roof displacement risk using consistency across multiple time-series models.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
48. StreamFormer: A multi-resolution transformer for scalable global river streamflow prediction
Core Problem: The verified StreamFormer study develops multiresolution global streamflow prediction; flood-warning performance is not yet established from the retrieved evidence.
Key Innovation: Not assessed: a reliable abstract matching the verified record was unavailable. Findings and validation performance cannot be established from bibliographic evidence alone.
49. A viscoplastic-fatigue peridynamic model and the GPU-accelerated simulation of creep-fatigue failure in geomaterials
Core Problem: Repeated loading and viscoplastic creep evolve together in three-dimensional rock damage.
Key Innovation: A bond-level model reproduces experimental salt damage and reaches about 21-fold speedup in its finest reported discretization, with cavern simulations indicating creep-dominated wall damage.
50. Dynamic strength degradation and microstructural evolution of cement-stabilized soil embankment subgrades under wet–dry cycling
Core Problem: Moisture cycling and traffic loading can jointly degrade stabilized soil bonds.
Key Innovation: Cyclic triaxial tests and NMR link pore-network deterioration with reduced dynamic strength and identify compaction-dependent resistance; reported parameter changes remain tied to the tested specimens and cycles.