TerraMosaic Daily Digest: July 3, 2026
Daily Summary
The July 3 literature pivots from landslide inventories toward the physical controls that determine when slopes, faults, and engineered ground systems become unstable. The strongest landslide-specific contribution is an analytical rainfall-slope framework that combines suction loss, groundwater rise, wetting-induced weight gain, water-filled tension cracks, and raindrop impact in one limit-equilibrium formulation. A complementary Natural Hazards study uses interpretable machine learning and multitemporal remote sensing to explain aboveground biomass recovery after landslide disturbance, linking geomorphic disturbance to carbon and ecosystem resilience rather than treating landslide scars as a static end state.
Earthquake and volcanic hazards provide the highest-impact source-process signal. Science Advances reports a simple but consequential relation between subduction fault dip and giant-earthquake occurrence, emphasizing ultralow-angle geometries and evolving stress alignment. A JGR: Solid Earth study reconstructs clustered paleoearthquakes on the Liuyuan fault and attributes the clustering to fluid-driven fault-valve behavior. InSAR observations at Piton de la Fournaise resolve 151 small active fault structures during rift-zone intrusions, connecting centimeter-scale deformation, hydrothermal alteration, and flank stability.
The methods layer is broad but coherent. Interpretable Kolmogorov-Arnold networks expose rainfall controls on urban flood depth; large-sample hydrological benchmarking shows how a compact model set can preserve predictive skill under uncertainty; neural-network reparameterized tomography reduces subjective regularization in seismic imaging. Rock, tunnel, storage, and structural studies focus on fracture-flow coupling, thermal damage, karst seepage, weak interlayers, rockburst, carbon-storage geomechanics, stress heterogeneity, and earthquake strengthening. Remote-sensing papers on object reasoning, dataset validation, SAR lake dynamics, InSAR atmospheric correction, and construction-year mapping extend the issue toward auditable geospatial evidence.
Key Trends
Five movements define this issue: geometry-aware earthquake science, mechanism-rich rainfall-slope analysis, interpretable hydro-geomorphic modeling, coupled underground-rock processes, and remote-sensing AI focused on evidence reliability.
- Fault geometry is becoming a first-order hazard variable: The Science Advances giant-earthquake study, the Liuyuan paleoearthquake reconstruction, North China waveform tomography, and Piton de la Fournaise InSAR analysis all connect hazard potential to geometry, stress state, and hidden structural weakness.
- Rainfall-slope analysis is becoming more mechanism-complete: The rainfall-induced slope framework explicitly couples suction, groundwater, cracks, surface loading, and wetting effects, while biomass recovery and soil-erosion papers track the longer geomorphic and ecological aftermath of slope disturbance.
- Hydro-geomorphic models are being judged by interpretability and uncertainty: XKAN flood modeling, hydrological model benchmarking, karst-groundwater tracing, Himalayan drought severity analysis, and SAR flood-season lake monitoring emphasize drivers, model adequacy, and regime-specific response rather than black-box prediction alone.
- Underground hazard studies are converging on coupled rock processes: Coal seepage damage, thermal sandstone fracture, fracture stiffness-flow relations, karst tunnel seepage, weak-interlayer roadway failure, rockburst, carbon-storage poroelasticity, and heterogeneous in-situ stress studies all frame instability as coupled damage, water, heat, and stress evolution.
- Remote-sensing AI is shifting toward reasoning, validation, and correction: Referring remote-sensing segmentation, dual satellite-dataset validation, InSAR delay correction, and construction-year mapping target the reliability of geospatial evidence, not only pixel-level accuracy.
Selected Papers
The selected papers cover ultralow-angle megathrust faults, fluid-driven earthquake clustering, rainfall-induced slope instability, landslide-disturbance biomass recovery, volcanic InSAR fault detection, full-waveform and neural-network seismic tomography, Himalayan erosion, active-margin sediment strength, interpretable urban-flood modeling, hydrological benchmark selection, karst groundwater, underground rock damage, tunnel seepage and support, rockburst, carbon-storage geomechanics, earthquake building strengthening, soil erosion, permafrost thaw, offshore scour, remote-sensing reasoning, and satellite-dataset validation. This issue contains 39 selected papers from 1963 papers analyzed.
1. Ultralow-angle faults produce giant earthquakes
Core Problem: Forecasting the largest subduction earthquakes remains limited by incomplete understanding of why some faults grow into giant ruptures.
Key Innovation: Shows that earthquake size depends strongly on fault dip, with extreme events favored on ultralow-angle faults aligned with the regional stress field.
2. Unified Analytical Framework for Rainfall-Induced Slope Instability Considering Crack and Raindrop Effects
Core Problem: Common analytical slope models underrepresent the combined hydraulic, geometric, and surface-loading pathways that control rainfall-triggered failure.
Key Innovation: Unifies suction loss, groundwater rise, wetting-induced weight gain, water-filled tension cracks, and raindrop impact in a limit-equilibrium framework for unsaturated slopes.
3. Uncovering drivers of aboveground biomass recovery after landslide disturbances in tropical montane forest area using explainable machine learning
Core Problem: Post-landslide forest recovery controls carbon storage and ecosystem resilience, but the drivers of biomass recovery across mapped scar populations remain unclear.
Key Innovation: Combines multitemporal remote sensing, XGBoost, and SHAP analysis across 608 landslide scars to identify landslide age, elevation, and residual vegetation as dominant recovery controls.
4. Fluid-Driven Fault-Valve Behavior Controls Clustered Paleoearthquakes in a Stable Continental Region: Insights From the Liuyuan Fault, Beishan
Core Problem: Stable continental interiors can host rare but damaging earthquakes, yet their recurrence is often treated as time independent despite sparse evidence.
Key Innovation: Uses paleoseismic trenching, IRSL and cosmogenic dating, rockfall ages, and fault-zone microstructures to link clustered ruptures to fluid-driven fault-valve behavior.
5. InSAR Evidence of Small-Scale Faults Activated During Rift Zone Intrusions at Piton de la Fournaise Volcano
Core Problem: Small faults and fractures on volcanic edifices influence intrusion pathways and flank stability but are difficult to resolve from broad deformation fields.
Key Innovation: Filters InSAR observations from 30 intrusions to identify 151 centimeter-scale active faults and relate them to rift dilation, hydrothermal circulation, and edifice weakening.
6. High-Resolution Tomography for the Upper-Mantle Structure in North China by Full-Waveform Inversion of Teleseismic P Wave and Its Codas
Core Problem: The deep structure beneath the North China Craton remains hard to resolve with conventional teleseismic imaging alone.
Key Innovation: Uses full-waveform inversion of teleseismic P waves and codas to map crustal-thickness variation and low-velocity upwelling beneath the Datong volcanic group.
7. Interpretable Kolmogorov-Arnold Networks for revealing key drivers of urban flood inundation: a case study of the Longgang River Basin in Shenzhen, China
Core Problem: Urban flood depth depends on interacting rainfall descriptors, but conventional data-driven models often hide which drivers dominate under different intensity regimes.
Key Innovation: Develops an interpretable XKAN surrogate with derivative-based sensitivity measures to expose rainfall controls on flood depth and support intensity-specific warning analysis.
8. Benchmarking and Selecting Optimal Hydrological Models for Large-Sample Applications Considering Complexity and Uncertainty
Core Problem: Large-sample rainfall-runoff prediction requires model selection that accounts for structural uncertainty, computational cost, and regional diversity.
Key Innovation: Benchmarks 47 conceptual models across 159 watersheds and shows that a compact set of storage-aware models can preserve reliable streamflow and flood prediction.
9. Global Cooling Driving the Increase in Erosion Rates Across the Himalaya Since the Quaternary
Core Problem: The influence of Quaternary cooling on Himalayan erosion rates remains debated because marine and long-distance sediment records can obscure source signals.
Key Innovation: Integrates luminescence, cosmogenic dating, geochemistry, apatite fission-track data, and thermo-kinematic modeling to show accelerated erosion after about 2 Ma.
10. Field and Experimental Observations Show Elevated Shear Strength of Active Margin Sediments Requires Processes Beyond Normal Consolidation
Core Problem: Near-seafloor sediments on active margins show elevated shear strength, but laboratory consolidation alone cannot explain the field observations.
Key Innovation: Compares active- and passive-margin sediments and argues that repeated earthquakes and non-uniaxial stress paths strengthen active-margin sediments beyond normal consolidation.
11. Seismic Traveltime Tomography With Deep Neural Network Reparameterization: The Method and Its Application to Central Chile
Core Problem: Traveltime tomography is sensitive to subjective grid and regularization choices, especially where earthquake-source and station coverage are uneven.
Key Innovation: Uses an untrained deep neural network to reparameterize velocity models, providing implicit spatial regularization and clearer slab imaging in central Chile.
12. Fracture evolution and mechanical response of coal with damage induced by oil-gas seepage
Core Problem: Long-term oil-gas seepage can alter coal integrity and fracture behavior, affecting extraction safety in co-storage zones.
Key Innovation: Combines CT, uniaxial compression, acoustic emission, and fractal analysis to quantify seepage-induced damage and identify precursors to coal instability.
13. A fractional-order thermo-mechanical micromechanical model for quasi-brittle rocks under elevated temperature and confining pressure
Core Problem: Quasi-brittle rock behavior under high temperature and confinement requires models that capture nonlinear deformation, damage, and non-coaxiality.
Key Innovation: Builds a fractional-order micromechanical model with temperature-dependent parameters and validates it against high-temperature triaxial tests.
14. Thermo-mechanically coupled damage constitutive model and cross-scale fracturing mechanisms in argillaceous sandstone
Core Problem: Deep underground engineering needs constitutive models that link thermo-mechanical loading to fracture evolution across scales.
Key Innovation: Combines triaxial tests, Weibull damage mechanics, SEM, and acoustic emission to model brittle-ductile transition and cross-scale sandstone fracturing.
15. Insights into the Characteristics and Mechanism of Heterogeneous In Situ Stress: A Case Study from the YD Copper Mine, China
Core Problem: Heterogeneous in-situ stress fields can drive deep mining failures, but stress magnitude and orientation are often poorly constrained between boreholes.
Key Innovation: Integrates hydraulic fracturing, borehole imaging, rock tests, focal mechanisms, and finite-element simulations to map stress heterogeneity and its lithologic controls.
16. Analytical Relationships Between Normal Stress and Fluid Flow for Single Fractures Based on the Two-Part Hooke's Model
Core Problem: Predicting fluid flow through stressed fractures requires a mechanistic link between normal stiffness and aperture closure.
Key Innovation: Derives a closed-form stiffness-flow relationship that separates soft- and hard-aperture contributions and explains multiple observed flow regimes.
17. Bearing Capacity and Settlement of Footings above Rounded Corner Rectangular Tunnels
Core Problem: Sharp tunnel corners can concentrate stress and reduce bearing capacity for overlying footing systems.
Key Innovation: Uses finite-element analysis to show how rounded tunnel corners redistribute stress, raise bearing capacity under shallow cover, and smooth failure surfaces.
18. Ring-shaped structure of tunnel surrounding rock: Characteristics and load-bearing mechanisms
Core Problem: Tunnel design requires a quantitative description of how surrounding rock transfers load as excavation-induced failure evolves.
Key Innovation: Develops a ring-shaped-structure theory linking stress migration, progressive deformation, support stiffness, and load-bearing zones around tunnels.
19. Numerical simulation and field monitoring on seepage evolution of surrounding rock in tunnels adjacent to high-pressure karst caves
Core Problem: Water-rich karst environments can impose hydraulic loads on tunnel linings, but the controlling seepage parameters are difficult to isolate in the field.
Key Innovation: Uses controlled numerical simulation and field monitoring to relate seepage coefficient, water pressure, and lining response around tunnels near high-pressure caves.
20. Mesoscopic instability mechanism and grouting reinforcement of deep roadway surrounding rock with weak interlayers
Core Problem: Deep roadways with weak interlayers can lose internal support through progressive cracking, roof displacement, floor heave, and rib shrinkage.
Key Innovation: Uses particle-flow simulation and field evidence to identify crack-network breakdown and evaluate grouting as a way to rebuild surrounding-rock support.
21. Topographically corrected land surface temperatures reveal a 2005 regime shift and prolonged thawing in the Yellow River Source Region (1981-2020)
Core Problem: Rugged terrain and sparse stations make it difficult to resolve thawing-freezing dynamics in the Yellow River source region.
Key Innovation: Builds a 1-km, 40-year topographically corrected temperature dataset and identifies a 2005 thawing regime shift linked to longer unfrozen periods.
22. Spatiotemporal evolution and influencing factors of soil erosion sensitivity in the red soil region of Southern China
Core Problem: Regional erosion planning needs sensitivity maps that separate stable natural controls from changing human and land-use influences.
Key Innovation: Uses spatial principal component analysis and geographical detectors to map soil-erosion sensitivity and identify evolving topographic, climatic, vegetation, and human controls.
23. Automatic prediction of soil erosion in arid regions using multi-sensor remote sensing integrated with object-based image analysis and deep neural networks
Core Problem: Sand-dune movement and active desert processes threaten arid ecosystems, but erosion prediction depends on scale-sensitive landform extraction.
Key Innovation: Combines OBIA, CNNs, FAHP, SAGAN-enhanced imagery, and spectral-topographic analysis to predict erosion-linked dune dynamics in arid regions.
24. A model-driven method for predicting the global response of base-isolated structures based on local monitoring data from the isolation layer
Core Problem: Rapid post-earthquake assessment of base-isolated buildings is limited when dense floor-by-floor sensor networks are unavailable.
Key Innovation: Reconstructs global structural response from isolation-layer monitoring and ground motion using an MDOF model, bilinear constitutive laws, and PSO calibration.
25. Quantifying the Role of Karst Groundwater on Mountain River Discharge
Core Problem: Snow-fed karst basins depend on groundwater-stream exchanges, yet flow paths and residence times remain uncertain under seasonal recharge shifts.
Key Innovation: Combines discharge, solute, isotope mass balances, and spring tracer data to quantify distributed groundwater exchange and baseflow buffering capacity.
26. Tidal current-induced local scour around offshore tripile foundations: A CFD study of flow-structure-sediment interaction
Core Problem: Tripile foundations for offshore infrastructure face local scour under tidal currents, but complex pile-group flow and sediment interactions remain underconstrained.
Key Innovation: Develops a three-dimensional CFD model to evaluate scour evolution around tripile foundations under varying flow intensity, water depth, spacing, and attack angle.
27. Cyclic and post-cyclic behaviour of compacted and intact completely decomposed soils subjected to blasting vibrations
Core Problem: Blasting vibrations in underground construction can raise pore pressure in decomposed soils, but thresholds for slope-stability concern remain uncertain.
Key Innovation: Uses cyclic and post-cyclic triaxial tests linked to peak particle velocity to quantify pore-pressure buildup and shear-strength response.
28. Experimental study on the fracture process and failure mechanism of impact rockburst
Core Problem: Impact rockburst hazard depends on burial-depth-controlled energy accumulation and fracture evolution, but the failure sequence remains difficult to observe directly.
Key Innovation: Uses laboratory impact experiments and ANN-assisted crack classification to identify four fracture stages and explain stronger fragment ejection at greater burial depths.
29. Multi-output spatio-temporal surrogate modelling for two geotechnical benchmark problems
Core Problem: Repeated geotechnical uncertainty analyses are limited by the cost of multi-physics simulations that produce time-dependent scalar, vector, and spatial-field outputs.
Key Innovation: Builds a multi-output neural surrogate with shared temporal representation and output-specific reconstruction for stress, pore pressure, and other geotechnical responses.
30. Transversely Isotropic Poroelasticity Modeling for Geomechanical Risk Analysis of Geological Carbon Storage
Core Problem: Carbon-storage risk analyses can underestimate failure potential when reservoir and caprock transverse isotropy is simplified as isotropic behavior.
Key Innovation: Uses 3D transversely isotropic poroelastic modeling to show how mechanical and permeability anisotropy change effective stress, failure timing, and seal risk near injection wells.
31. Constitutive Parameter Calibration for Enhanced Cohesive Zone Models: Simulating Anisotropic Deformation-Failure in Layered Soft Rock Tunnels
Core Problem: Layered soft-rock tunnels fail through anisotropic deformation and shear-dominated fracture that standard cohesive-zone models may not capture.
Key Innovation: Enhances FEM-CZM modeling with Mohr-Coulomb shear failure, calibrated constitutive parameters, and nonlinear compaction to simulate dip-angle and lateral-pressure effects on tunnel stability.
32. Strengthening strategies for mitigating damage induced by high ground stories: field data and analytical findings from RC buildings in the Kahramanmaraş earthquakes
Core Problem: High ground stories create stiffness and strength irregularities that contributed to RC building damage during the 2023 Kahramanmaraş earthquakes.
Key Innovation: Combines field evidence and pushover analyses to compare FRP and RC jacketing strategies, showing that multi-story strengthening better prevents brittle damage transfer.
33. Referring Remote Sensing Image Segmentation With Positive-Incentive Visual Adaptation and Logical Reasoning
Core Problem: Vision-language models struggle with remote-sensing segmentation because spatial expressions, clutter, and domain shift weaken cross-modal alignment.
Key Innovation: Combines positive-incentive visual adaptation with explicit referring-logic reasoning to improve prompt-guided segmentation of complex aerial scenes.
34. A Dual Validation Framework for Curating Machine Learning-Ready Satellite Datasets: A Scalable Pipeline and Stratified Analysis
Core Problem: Earth-observation foundation models need large datasets whose quality and difficulty are validated rather than assumed from aggregate benchmark scores.
Key Innovation: Introduces intrinsic and extrinsic validation, a dataset difficulty index, and scalable Zarr-based curation to expose difficulty-stratified model failures.
35. Integrating Gross Error Identification with Deep Learning for InSAR Topography-Dependent Delay Correction: A Case Study of the Baihetan Hydropower Station Area
Core Problem: Topography-dependent atmospheric delays can obscure subtle deformation signals in steep hydropower-station terrain.
Key Innovation: Combines gross-error identification with deep learning to improve InSAR delay correction for deformation monitoring around the Baihetan hydropower station area.
36. Spatiotemporal Dynamics of Dongting Lake During the Flood Season Using Long Time Series SAR Imagery on Google Earth Engine
Core Problem: Large floodplain lakes require cloud-robust, long-time-series monitoring to capture seasonal inundation dynamics during flood periods.
Key Innovation: Uses long time series SAR imagery on Google Earth Engine to characterize flood-season spatiotemporal water dynamics in Dongting Lake.
37. Mapping Building Construction Year from Landsat in Data-Scarce, Cloud-Prone Regions: A Parsimonious Spatial Triage Tool for Physical Vulnerability Screening
Core Problem: Disaster vulnerability screening in data-scarce regions is constrained by missing building-age information and persistent cloud contamination.
Key Innovation: Uses Landsat-based spatial triage to infer building construction year and support first-pass physical vulnerability screening where cadastral data are limited.
38. Multi-scale meteorological and hydrological drought severity in a high-altitude Himalayan basin: An intensity-duration-frequency approach using SPI and SSI (1970-2016)
Core Problem: Snow- and glacier-fed Himalayan basins can show different meteorological and hydrological drought responses across time scales.
Key Innovation: Applies SPI, SSI, and intensity-duration-frequency analysis to quantify drought severity, duration, and return-period deficits in the Astore Basin.
39. Constructing a risk probability model for resilience construction in mountainous communities in Southwest China based on field theory
Core Problem: Mountain communities need risk-probability tools that identify which resilience dimensions most strongly constrain preparedness and recovery.
Key Innovation: Builds a Bayesian-network resilience risk model for ten mountainous communities and identifies information access, training, and economic factors as key leverage points.