TerraMosaic Daily Digest: May 10, 2026
Daily Summary
The May 10 literature is led by a clear movement from static hazard mapping toward dynamic, history-conditioned process models. The Science study on the 2025 Mandalay earthquake links satellite-observed rupture with multi-millennial simulations of Sagaing fault segmentation, turning a single destructive event into evidence for persistent rupture boundaries and recurrence structure. For landslide and mass-movement research, the strongest contributions are similarly process-centered: long L-band InSAR deformation histories resolve slow landslides together with subsidence, volcanic unrest, and tectonic strain; mountain sediment-connectivity metrics locate flood-mobilized source areas; debris-flow modelling separates the timing and hydraulic role of upstream and downstream sub-basins; and strain-softening material-point simulations translate seismic demand into slope-collapse probability and displacement.
Flood, coastal, cryospheric, wildfire, and infrastructure studies broaden the same theme into decision-relevant modelling. Compound-flood analysis combines rainfall, storm surge, and shallow groundwater; HAND inundation experiments quantify how DEM bias cascades into flood extent; dune simulations show when erodible morphology reduces inundation; and counterfactual rainfall-runoff modelling strengthens flood-frequency estimates in data-poor headwaters. The engineering papers emphasize state-dependent risk rather than single design loads: tunnel leakage erosion, excavation suction loss, liquefaction, levee uncertainty, railway subgrade fragility, and converter-station reinforcement are all treated as coupled deformation, hydraulic, or network-recovery problems. The day also includes a compact group of remote-sensing and AI papers whose value lies in measurement infrastructure: optical-SAR change detection, lithology segmentation, river-ice classification, cryosphere spectral indexing, and UAV embodied intelligence extend the observational base for future hazard workflows.
Key Trends
The methodological signal is a shift toward coupled process constraints: fault history, sediment connectivity, hydrological compounding, and soil-structure interaction are increasingly treated as variables to be inferred, simulated, and optimized together.
- Dynamic rupture and slope mechanics are being made explicit. The top-ranked studies replace static hazard descriptors with models of process memory: fault segmentation over millennia and strain-softening seismic slope displacement.
- Flood hazard is treated as compound and routed. Rainfall, surge, groundwater, sediment supply, sub-basin runoff timing, DEM uncertainty, and dune morphodynamics are each shown to change the spatial footprint of flood or debris-flow impact.
- Observation quality is now central to hazard inference. L-band InSAR, ICESat-2, harmonized Landsat-Sentinel data, optical-SAR fusion, lithology segmentation, and cryosphere spectral indexing all sharpen the measurements that downstream hazard models depend on.
- Engineering risk is shifting toward coupled service and deformation states. Seismic infrastructure fragility, tunnel leakage, liquefaction, levee uncertainty, bridge scour, and pile response are framed as evolving system states rather than isolated factor-of-safety calculations.
- Decision support is moving closer to users and institutions. Personalized flood alerts, tsunami shelter optimization, wildfire fuel-treatment economics, and networked infrastructure recovery connect physical risk estimates to decisions that must be made under time, cost, and comprehension constraints.
Selected Papers
This issue contains 54 selected papers from 1,933 papers analyzed. The leading papers connect dynamic Sagaing fault segmentation, L-band InSAR geohazard monitoring, sediment-connectivity mapping, debris-flow propagation, slow bentonitic-clay landslide reactivation, and seismic slope-displacement modelling. The broader set adds compound flooding, counterfactual flood-frequency analysis, dune-based coastal mitigation, tsunami shelter optimization, wildfire fuel-treatment economics, tunnel leakage erosion, excavation stability under infiltration, liquefaction assessment, levee reliability, cryospheric runoff modelling, extreme-precipitation nowcasting, and remote-sensing AI for disturbance, lithology, river ice, cryosphere indexing, and UAV inspection.
1. Dynamic segmentation of the Sagaing fault
Core Problem: Large strike-slip earthquakes expose how persistent fault segmentation, historical slip-rate structure, and rupture dynamics jointly shape recurrence patterns.
Key Innovation: Satellite observations and 5,000-year earthquake-cycle simulations link the 2025 Mandalay rupture to long-term segmentation on the Sagaing fault, with direct analogues for other major plate-boundary faults.
2. Nine‐Year L‐Band InSAR Time Series of Tectonic and Non‐Tectonic Surface Deformation in Northern California
Core Problem: NISAR-era deformation monitoring needs evidence that long L-band time series can resolve multiple hazards despite ionospheric and decorrelation noise.
Key Innovation: A nine-year ALOS-2 workflow combines split-spectrum ionospheric correction, adaptive networking, filtering, multilooking, and SBAS inversion to recover tectonic strain, volcanic deformation, aquifer subsidence, and slow landslides across northern California.
3. Optimising the identification of river sediment supply areas and their connectivity in the mountainous river basins of central Italy
Core Problem: Post-event sediment supply areas in steep basins are difficult to distinguish from channel transport corridors and hillslope instability zones.
Key Innovation: SPI, index of connectivity, and stream-length gradient metrics are combined with NDVI change, UAV surveys, and field validation to locate sediment sources after the 2022 Burano River flash flood.
4. Contributions of sub-basins runoff to debris flow propagation at a basin scale: A numerical study in the Meilong Basin, China
Core Problem: Debris-flow forecasts often treat basin runoff as spatially uniform, obscuring how tributary timing and location control propagation.
Key Innovation: A cascade model for the Meilong debris flow separates upstream energy supply from downstream water contribution, showing how sub-basin runoff reorganizes entrainment and propagation velocity.
5. Investigation of a slow-moving landslide in Cretaceous bentonitic clay
Core Problem: Slow reactivation of old clay-rich landslides can damage settlements while remaining hard to diagnose from surface deformation alone.
Key Innovation: Field mapping, laboratory mineralogy, FEM back-analysis, and displacement monitoring identify a paleolandslide in smectite-rich bentonitic clay with low residual strength and rainfall-access fissures.
6. Earthquake-induced displacement model for earth slopes incorporating soil softening effect based on material point method
Core Problem: Permanent slope displacement models need to distinguish recoverable deformation from collapse under strength degradation.
Key Innovation: Material point simulations with strain-softening generate predictive functions for collapse probability and displacement using PGV, yield acceleration, and degradation parameters.
7. Flood Hazard and Risk Assessment in the Mpanga River Catchment Using Integrated Hydrological Modeling and Decision Support Tools
Core Problem: Operational flood-risk planning in tropical basins requires joint hydrodynamic hazard and social-physical vulnerability information.
Key Innovation: RRI flood simulation calibrated against observations is coupled with AHP-based exposure and vulnerability layers to map flood risk in Uganda's Mpanga River basin.
8. Hazard potential of compound flooding from rainfall, storm surge, and groundwater in coastal New York and Connecticut
Core Problem: Urban coastal flood maps often understate risk by treating rainfall, surge, and groundwater as separate hazards.
Key Innovation: Historical co-occurrence analysis, joint return periods, return-period adjustment factors, and a pseudo-trivariate hazard score quantify compound flooding in New York and Connecticut.
9. Morphodynamic controls on the performance of dune-based coastal flood mitigation under sea-level rise
Core Problem: Static flood mapping cannot determine when erodible dunes reduce or amplify inundation during extreme waves.
Key Innovation: XBeach simulations and event-frequency analysis show that morphodynamic feedbacks can reduce flood extent substantially and that static-dune assumptions are not consistently conservative.
10. Comparative assessment of analytical approaches for optimizing tsunami vertical-evacuation shelter locations
Core Problem: Tsunami shelter planning needs methods that balance computational cost, urban realism, and life-safety outcomes.
Key Innovation: GIS, agent-based, casualty-driven agent-based, and urban-feature regression approaches are compared across 15 Japanese coastal municipalities to quantify performance and transferability.
11. Evaluation of amplitude–duration parameters for rapid magnitude estimation in earthquake early warning systems
Core Problem: Early-warning systems require magnitude estimates before rupture duration and network coverage are fully resolved.
Key Innovation: Peak displacement, cumulative absolute velocity, and integral-squared velocity parameters are tested on more than 21,000 Japan waveforms to improve rapid magnitude estimation.
12. Considering rainfall events from a neighborhood improves local flood frequency analysis
Core Problem: Small catchments often lack enough extreme floods for robust high-return-period estimates.
Key Innovation: Radar precipitation and hydrological modelling generate local counterfactual floods from nearby storm events across more than 13,000 German headwater catchments, improving high-quantile estimates.
13. Evaluating the Vertical Accuracy of Global DEMs Using ICESat-2 and Its Cascading Impact on HAND-Based Flood Modeling in a Low-Gradient Coastal Plain
Core Problem: Inundation estimates from HAND models can be dominated by vertical DEM errors at shallow thresholds.
Key Innovation: ICESat-2 ATL08 elevations are used to benchmark eight DEMs and trace how elevation uncertainty propagates into flood extent in a coastal Chinese city.
14. Hybrid-cryo: a modular data-physics hybrid framework for cryospheric runoff simulation
Core Problem: High-mountain runoff models struggle to represent cryosphere processes while remaining learnable from sparse observations.
Key Innovation: Hybrid-Cryo couples physical skeletons with data-driven components to simulate glacier- and permafrost-influenced runoff in cryospheric basins.
15. Wildfire damages and the cost-effective role of forest fuel treatments
Core Problem: Wildfire mitigation requires evidence on where fuel treatments reduce damages at costs that justify implementation.
Key Innovation: Large-scale treatment and damage analyses are synthesized to identify cost-effective restoration strategies and policy reforms for wildfire-risk reduction.
16. Investigation of dynamic effect on soil erosion and ground settlement due to tunnel leakage
Core Problem: Tunnel leakage can erode surrounding soil and produce progressive ground settlement that is difficult to capture in design.
Key Innovation: Coupled experiments and simulations characterize erosion, void formation, and settlement evolution around tunnel leakage pathways.
17. Subsurface imaging in urban environments: Insights from multi-method geophysical surveys in the historical center of Messina (southern Italy)
Core Problem: Urban seismic-risk assessment is limited by sparse subsurface constraints beneath built environments.
Key Innovation: Integrated geophysical imaging in Messina improves the stratigraphic and dynamic ground model needed for microzonation.
18. Threshold responses of runoff-sediment dynamics to hydrological connectivity
Core Problem: Watershed sediment response can switch abruptly when rainfall, runoff, and connectivity thresholds are crossed.
Key Innovation: A 2015-2024 event archive identifies hydrological connectivity thresholds that separate runoff-sediment response regimes in a southeast China watershed.
19. Unified Probabilistic Evaluation of Gravelly Soil Liquefaction Triggering by Dynamic Cone Penetration Test
Core Problem: Gravelly soils remain difficult to classify for liquefaction because standard penetration-based methods are poorly suited to coarse materials.
Key Innovation: Dynamic penetration testing is calibrated for liquefaction triggering assessment in gravel-rich deposits.
20. Cementing Marine Sands Wetted with Chemically Distinct Seawaters via Electrodeposition
Core Problem: Coastal erosion countermeasures need low-impact methods for strengthening marine sands in place.
Key Innovation: Electrodeposition-induced cementation is evaluated as a technique for improving marine-sand resistance to erosion.
21. Establishing an evidence base to inform the development of personalised flood alerts
Core Problem: Flood warnings often fail when message wording does not match recipient context, comprehension, or perceived urgency.
Key Innovation: A controlled participant study shows that personalized flood-alert messages improve comprehension, perceived risk, and urgency compared with generic warnings.
22. Deforestation-induced drying lowers Amazon climate threshold
Core Problem: Amazon climate thresholds cannot be assessed from warming alone because deforestation alters moisture recycling and rainfall stability.
Key Innovation: A dynamical-systems model shows that deforestation can lower the system-wide drying threshold to near current warming trajectories by disrupting atmospheric moisture transport.
23. Mid-Holocene retreat of the Greenland Ice Sheet indicated by subglacial methane release
Core Problem: Past ice-sheet retreat provides constraints on future cryosphere response and subglacial carbon release.
Key Innovation: Methane released from 26 Greenland meltwater streams is used to infer mid-Holocene ice retreat and short-lived subglacial methane export.
24. LAMES: A Large-Scale and Artisanal Mining Environmental Segmentation Dataset
Core Problem: Mining-induced terrain disturbance needs benchmark data for consistent remote-sensing extraction.
Key Innovation: LAMES provides a labelled remote-sensing dataset and segmentation baseline for mapping mining-related environmental change.
25. Geological information in shield tunnelling: exploration, estimation, prediction, and perspectives
Core Problem: Shield tunnelling requires rapid updating of ground conditions ahead of the cutterhead to reduce collapse, clogging, and settlement risk.
Key Innovation: A review organizes geological exploration, machine inference, and prediction methods into a workflow for data-driven shield-tunnelling decision support.
26. Failure Mechanism of Soil Excavations Under Unsaturated Steady-State Flow Conditions Based on Slip-Line Method
Core Problem: Rainfall infiltration reduces matric suction and can shift excavation-support loads toward failure.
Key Innovation: A slip-line solution for unsaturated steady-state flow quantifies how infiltration changes active earth pressure and excavation stability.
27. Seismic-risk-based differentiated reinforcement analysis for ultra-high-voltage converter stations
Core Problem: Critical power infrastructure needs seismic strengthening strategies that balance fragility reduction, functionality, and cost.
Key Innovation: Fragility analysis, functional loss, life-cycle cost, and particle-swarm optimization are coupled to select reinforcement strategies for converter-station equipment.
28. Seismic fragility assessment of high-speed railway subgrade: an integrated framework of safety and serviceability
Core Problem: Railway embankments and subgrades require fragility models that connect seismic demand to serviceability loss.
Key Innovation: Numerical fragility analysis quantifies seismic vulnerability for high-speed railway subgrade systems.
29. Fluid-solid coupling analysis of levee soils considering parameter uncertainty and spatial variability
Core Problem: Levee performance depends on uncertain coupled seepage and deformation properties that are rarely propagated through reliability analysis.
Key Innovation: A fluid-solid coupled framework quantifies parameter uncertainty and reliability for levee soils.
30. G2Lcast: A global-to-local adversarial network for high-fidelity precipitation nowcasting
Core Problem: Nowcasting systems often smooth extremes, degrading flood-warning value.
Key Innovation: G2Lcast combines global and local adversarial learning with wavelet compression and an energy bypass to sharpen high-intensity precipitation forecasts.
31. Near-real-time mapping for causal agents of forest disturbances in China using harmonized Landsat and Sentinel-2 dataset
Core Problem: Rapid attribution of forest disturbance agents is needed for hazard response and ecosystem-risk monitoring.
Key Innovation: Harmonized Landsat-Sentinel time series provide near-real-time maps of wildfire, logging, and stress disturbance agents across China.
32. Impact of inter-story pounding on the seismic demands of adjacent reinforced concrete buildings
Core Problem: Closely spaced buildings can transfer damaging seismic impacts across floor levels.
Key Innovation: Numerical analysis evaluates inter-story pounding mechanisms and structural response in adjacent RC buildings.
33. A high-fidelity and scalable computational platform for large-scale SSI analysis of seismically isolated NPPs
Core Problem: Nuclear facilities require platform-level seismic models that include realistic soil-structure interaction.
Key Innovation: Soil-structure interaction analysis is applied to nuclear power plant platform systems under earthquake loading.
34. Resilience analysis of road transportation-electric power system considering interdependencies under earthquake effects
Core Problem: Post-earthquake recovery depends on coupled road and power networks rather than isolated infrastructure components.
Key Innovation: A resilience model links transportation accessibility and power restoration to quantify interdependent recovery dynamics.
35. Mechanical behaviour, deformation mechanisms and support optimization of tunnels in gentle dip layered rock after freeze–thaw cycles
Core Problem: Freeze-thaw cycles alter surrounding-rock behavior and tunnel-lining demand in cold-region layered strata.
Key Innovation: Model tests and simulations quantify lining response in gently dipping layered rock after freeze-thaw cycling.
36. Macro and mesoscopic mechanical responses analysis of EPB shield under bias-load effects in soil-rock composite strata using FDM-DEM coupled method
Core Problem: Earth-pressure-balance shields experience asymmetric loads in soil-rock composite formations.
Key Innovation: Coupled FDM-DEM simulation resolves bias-load effects around EPB shield tunnelling.
37. Mechanical behaviour of Australian superfine silica sand: from monotonic to high-cyclic loading
Core Problem: High-cyclic loading of sand controls liquefaction and accumulated deformation in offshore and seismic settings.
Key Innovation: Australian superfine silica sand is tested under monotonic and high-cyclic loading to characterize pore-pressure and deformation response.
38. Deep Learning Surrogate for Undrained Cyclic Response of Sands: Stability and Generalization
Core Problem: Constitutive models for cyclic sand response can be costly when embedded in probabilistic geotechnical analysis.
Key Innovation: An LSTM surrogate predicts pore-pressure buildup and post-liquefaction response in undrained cyclic sand loading.
39. Modeling nonlinear propagation behavior of multi-type secondary cracks in rocks with contact retrieval by quasi-state-based peridynamics
Core Problem: Rock-mass damage models must capture secondary crack development under complex loading.
Key Innovation: A quasi-static bond-based peridynamic method simulates secondary crack propagation in rock-like materials.
40. Iterative indicator kriging using probability and pseudo-information entropy drop for spatial distribution estimation of geological discrete variables
Core Problem: Subsurface engineering models require uncertainty-aware interpolation of discrete geological variables.
Key Innovation: Iterative indicator kriging is developed for categorical geological-variable simulation.
41. A novel hydro-mechanical viscoelastic-plastic model for mechanical deconstruction and morphology evolution in clay compaction
Core Problem: Clay core walls undergo coupled compaction, creep, and hydraulic effects that influence dam safety.
Key Innovation: A hydro-mechanical viscoelastic-plastic model represents deformation and compaction behavior in clay core-wall materials.
42. Transformation of dredged silt into fluidized solidified soil for cross-sea bridge scour protection
Core Problem: Bridge foundations need scour-protection materials that are durable and compatible with dredged-sediment reuse.
Key Innovation: Dredged silt is transformed into fluidized solidified soil and evaluated for scour-protection performance.
43. Deep Learning Surrogate Model for Mud-Protected Bored Piles in Layered Silty Deposits with Application to Preliminary Design
Core Problem: Pile design in layered ground needs rapid surrogates for nonlinear soil-structure response.
Key Innovation: A deep-learning surrogate predicts bored-pile behavior in layered silty deposits.
44. Reconstructing a Century of Urban Growth Through Deep Learning-Based Colorization and Segmentation of Historical Aerial and Satellite Imagery: Les Sables-d’Olonne, France (1920–2024)
Core Problem: Long-term coastal exposure assessments need historical urban-growth trajectories that predate modern satellite records.
Key Innovation: Historical imagery is used to reconstruct urban growth around Les Sables-d'Olonne and relate expansion to coastal climate risk.
45. Spatio-temporal shoreline assessment of a dynamic deltaic coast: integrating DSAS and ecological indices for Konark Coastal Belt, India (2005–2025)
Core Problem: Coastal management requires long-term shoreline-change evidence at erosion-prone tourist and heritage sites.
Key Innovation: DSAS shoreline metrics and ecological indices quantify 2005-2025 erosion-accretion patterns along the Konark coast.
46. Scour hole remediation around a monopile by installing a net-like siltation-promoting mat
Core Problem: Scour around monopiles threatens offshore foundation stability and requires practical remediation.
Key Innovation: A net-like siltation mat is tested for scour-hole infilling and erosion reduction around a monopile.
47. Snow phenology-guided regionalization for parameter optimization: enhancing runoff simulation in an alpine basin
Core Problem: Hydrological parameter transfer in alpine basins is limited when snow timing differs across terrain.
Key Innovation: Remote-sensing snow phenology and clustering define surrogate cells for optimizing VIC runoff parameters.
48. Enhancing large-scale streamflow prediction by coupling structural inductive bias with meta-learning
Core Problem: Streamflow prediction needs to generalize across basins with sparse local observations.
Key Innovation: Meta-learning and entity-aware LSTMs improve discharge prediction across CAMELS basins.
49. A bias-corrected deep learning-based data assimilation method for riverine bathymetry inversion
Core Problem: Flood and navigation models need bathymetry where direct surveys are incomplete.
Key Innovation: A bias-corrected deep-learning ensemble smoother improves riverine bathymetry inversion.
50. A Dual-Modal Mixture-of-Experts Attention U-Net (DMoE-AttU-Net) for Change Detection Using Heterogeneous Optical and SAR Remote Sensing Images
Core Problem: Hazard and exposure mapping benefit from change detectors that exploit complementary optical and SAR information.
Key Innovation: A mixture-of-experts attention U-Net fuses optical and SAR inputs for improved semantic change detection.
51. AFPN-ResUNet: A Residual Attention Mechanism-Guided Asymptotic Feature Pyramid Network for Complex Outcrop Lithology Segmentation
Core Problem: Slope and ground models depend on lithological boundaries that are difficult to digitize consistently.
Key Innovation: AFPN-ResUNet improves lithology segmentation in outcrop images.
52. Optical River Ice Spectral Subclassification on the Tibetan Plateau: A Landsat 5–9 and Sentinel-2 Benchmark with Interpretable Machine Learning
Core Problem: Cold-region flood and infrastructure hazards require reliable discrimination of river-ice states.
Key Innovation: A Landsat-Sentinel benchmark and spectral analysis distinguish river-ice subclasses across the Tibetan Plateau.
53. A Continuous Cryosphere Index for Snow and Ice Reflectance
Core Problem: Binary snow masks miss the continuous spectral variability needed for cryosphere monitoring.
Key Innovation: More than 140 million EMIT spectra are used to construct a continuous cryosphere index for snow and ice discrimination.
54. Embodied AI in the Sky: A Comparative Review of UAV Embodied AI, from Autonomous Remote Sensing to Task Execution
Core Problem: Disaster inspection needs UAV systems that can perceive, navigate, and adapt beyond static mapping missions.
Key Innovation: A review synthesizes embodied-AI architectures for UAV remote sensing, including perception, planning, and deployment constraints relevant to hazard response.