TerraMosaic Daily Digest: June 29, 2026
Daily Summary
The June 29 literature is anchored by landslide systems in which material state changes during the hazard itself. The strongest papers reconstruct the Ashigong ancient landslide-dammed lake, quantify melting-driven settlement and overtopping in ice-soil landslide dams, track rock-glacier lakes that can precede outburst floods and debris flows, localize rockfall sources with distributed acoustic sensing, and update post-seismic debris-flow rainfall thresholds with environmental covariates. The common contribution is not another static hazard map; it is a shift toward observable transitions: dam formation, thaw settlement, lake growth and drainage, impact localization, and changing post-earthquake threshold behavior.
A second group treats hydrometeorological and seismic hazards as coupled infrastructure problems. Alpine sub-daily precipitation extremes, multivariate design storms, semi-arid storm-flood scaling, the September 2024 Danube flood, Hurricane Milton wave changes, bridge collapse under scour-earthquake interaction, soil-structure interaction on soft alluvium, high-arch-dam reliability, seepage control, and suffusion in gap-graded soils all connect forcing to a specific failure pathway. These papers are most useful where they keep the chain intact from load or flow to material response, structural demand, or sediment transport.
The methods papers are valuable when they make missing states measurable rather than merely adding model complexity. Borehole-ERT fusion constrains landslide stratigraphy; UAV photogrammetry resolves unstable rock-mass geometry; machine learning is used to interpret rate-dependent rock strength, LiDAR segmentation under changing observation quality, SAR pretraining, satellite change captioning, geospatial image synthesis, and soil particle-size estimation. Together the issue emphasizes reviewable hazard evidence: time-dependent forcing, evolving material state, explicit uncertainty, and observations that can be tied back to a physical mechanism.
Key Trends
Five movements define this issue: time-dependent landslide transitions, sub-daily hydrometeorological forcing, coupled infrastructure risk, fused subsurface evidence, and AI methods constrained by sensing geometry or physics.
- Landslide hazards are being described through state transitions: Dammed-lake formation, ice-soil dam thaw settlement, rock-glacier lake growth, rockfall acoustic localization, and post-seismic threshold decay all require timing and material change, not only susceptibility classes.
- Hydrometeorological extremes are becoming more local and sub-daily: Alpine precipitation bias correction, design-storm construction, storm-flood scaling, Danube flood comparison, and Hurricane Milton wave analysis all target the temporal scales that control warning, routing, and impact.
- Infrastructure risk is being treated as a coupled-hazard problem: Bridge scour-earthquake interaction, soil-structure interaction on soft deposits, high-arch-dam reliability, seepage barriers, suffusion, offshore monopiles, and salt-cavern fracturing show that single-load design assumptions are increasingly inadequate.
- Subsurface models are leaning on fused evidence: Borehole-ERT fusion, UAV rock-mass photogrammetry, hydrogeological flow matching, particle-size inference, and tunnel seepage monitoring all use complementary observations to reduce the ambiguity of hidden ground conditions.
- AI is strongest when constrained by sensing geometry or physics: LiDAR test-time adaptation, SAR pretraining, satellite change-captioning benchmarks, geospatial primitive control, and soil-image particle-size estimation are relevant because they improve what can be measured, transferred, or checked.
Selected Papers
The selected papers cover ancient landslide-dammed lakes, thawing ice-soil landslide dams, rock-glacier lake hazards, DAS rockfall localization, post-seismic debris-flow thresholds, UAV rock-mass photogrammetry, borehole-ERT landslide stratigraphy, alpine precipitation extremes, design storms, storm-flood scaling, major river flooding, hurricane waves, bridge scour-earthquake interaction, soil-structure interaction, high-arch-dam reliability, seepage barriers, suffusion, freeze-thaw rock degradation, jointed-rock creep, salt-cavern integrity, hillslope erosion, coastal erosion, dry-wet extremes, hydrogeological inversion, LiDAR adaptation, SAR pretraining, satellite change understanding, geospatial image synthesis, and low-cost soil characterization. This issue contains 38 selected papers from 2371 papers analyzed.
1. Formation process of the Ashigong ancient landslide-dammed lake in Guide Basin: impact on geomorphological evolution
Core Problem: The Guide Basin preserves a large ancient landslide and dammed-lake system, but its timing, activity phases, and geomorphic consequences remain uncertain.
Key Innovation: Combines field investigation, OSL dating, UAV three-dimensional modeling, and GIS analysis to reconstruct the Ashigong landslide-dammed lake and its role in basin evolution.
2. Melting-induced settlement and overtopping evolution of ice-soil landslide dams
Core Problem: Ice-rich landslide dams can lose elevation and strength during thaw, but the link between internal ice melting, settlement, overtopping, and breach evolution is poorly quantified.
Key Innovation: Couples thaw-settlement experiments with numerical simulation to show how melting-driven deformation controls overtopping development in ice-soil landslide dams.
3. Satellite-based hazard assessment of rock glacier lakes: examples from the European Alps
Core Problem: Permafrost degradation is forming lakes on active rock glaciers, but their short-lived growth, drainage, and downstream hazard potential are difficult to monitor from the ground.
Key Innovation: Uses Sentinel-2 and PlanetScope imagery to detect and track rock-glacier lakes, including cases linked to outburst flooding and debris-flow initiation.
4. High-precision rockfall source localization in acoustically complex environments using distributed acoustic sensing
Core Problem: Distributed acoustic sensing can monitor long slopes, but ambient wavefields obscure rockfall impact signatures and degrade source localization.
Key Innovation: Integrates adaptive signal isolation with time-difference-of-arrival localization to improve rockfall source estimates in acoustically complex field settings.
5. Dynamic changes in rainfall thresholds for post-seismic debris flows using logistic regression and resampling
Core Problem: Rainfall thresholds after earthquakes evolve as vegetation and loose material recover, yet many threshold models ignore environmental covariates and imbalanced trigger/non-trigger samples.
Key Innovation: Builds a logistic-regression and resampling framework that incorporates rainfall, vegetation, and material distribution to estimate dynamic post-seismic debris-flow thresholds.
6. Fracture characterization and kinematic simulation of unstable rock mass using UAV nap-of-the-object photogrammetry
Core Problem: High-steep slopes require detailed fracture geometry for rockfall assessment, but conventional surveys can be incomplete or unsafe.
Key Innovation: Uses close-range UAV nap-of-the-object photogrammetry to map joints, reconstruct unstable blocks, and simulate likely rockfall movement.
7. A borehole-geophysical data fusion method for stratigraphic modeling and landslide stability assessment: a case study
Core Problem: Sparse boreholes alone leave large uncertainty in stratigraphic models used for landslide stability analysis.
Key Innovation: Fuses borehole lithology with ERT resistivity constraints in a probabilistic stratigraphic framework for landslide stability modeling.
8. Scale-dependent biases in Alpine sub-daily areal precipitation extremes: added value of convection-permitting models
Core Problem: Short-duration rainfall extremes control flash floods and slope failures, but model biases vary strongly with accumulation time and areal scale in alpine terrain.
Key Innovation: Evaluates where convection-permitting models improve sub-daily areal extreme precipitation estimates relevant to rapid hydrologic hazards.
9. Analysis of bridge failure direction and collapse probability under multi-hazard scour and earthquake effects
Core Problem: Bridge collapse risk in hurricane- and earthquake-prone regions depends on joint scour and seismic loading, not either hazard in isolation.
Key Innovation: Develops a probabilistic multi-hazard framework for bridge failure direction and collapse likelihood under river scour and subduction-zone earthquakes.
10. Monitoring structural-plane slip using bolt axial force: insights from accelerated shear experiments
Core Problem: Slip along structural planes governs many rock-slope and underground instabilities, but direct warning indicators remain difficult to instrument.
Key Innovation: Uses accelerated shear experiments to link bolt axial-force evolution to structural-plane slip, offering a measurable proxy for incipient instability.
11. Multivariate design storms using vine copulas and Gaussian process regression
Core Problem: Flood and landslide-triggering rainfall cannot be represented by a single rainfall depth when duration, intensity, and temporal pattern co-vary.
Key Innovation: Combines vine copulas and Gaussian-process regression to construct multivariate design storms that preserve dependence among rainfall characteristics.
12. Scenario-conditioned flow matching for probabilistic generation of three-component ground-motion waveforms
Core Problem: Performance-based seismic risk analysis requires realistic three-component waveforms conditioned on source, path, and site, not only scalar intensity measures.
Key Innovation: Introduces a two-stage probabilistic ground-motion generator that combines physics-informed PGA modeling with flow matching in a wavelet-packet space.
13. Impact of soil-structure interaction on seismic vulnerability assessment of irregular masonry-infilled RC buildings on soft soils
Core Problem: Irregular reinforced-concrete buildings on soft alluvial deposits can have different seismic vulnerability when soil-structure interaction is represented explicitly.
Key Innovation: Uses nonlinear three-dimensional finite-element modeling and time-history analysis to quantify how soft-soil interaction changes vulnerability estimates.
14. Effects of acid-base environments and freeze-thaw cycles on dynamic properties and microstructures of granite
Core Problem: Cold-region rock structures experience coupled chemical corrosion, freeze-thaw weathering, and dynamic disturbance, but their combined damage mechanism is difficult to isolate.
Key Innovation: Combines SHPB testing, microstructural analysis, and particle-flow modeling to quantify granite degradation under acid-base and freeze-thaw conditions.
15. Rate-dependent compressive and tensile strength of rocks: from prediction to explanation using machine learning
Core Problem: Rock strength under rapid loading is rate dependent, but empirical predictions often lack interpretability across lithologies and test conditions.
Key Innovation: Uses machine-learning prediction and explanation to identify controls on compressive and tensile strength under different loading rates.
16. Grain-scale failure mechanism of heterogeneous rock under tensile loading: insights from image-based DEM modeling
Core Problem: Macroscopic tensile failure in heterogeneous rocks emerges from grain-scale fabric and mineral contrasts that are hard to resolve experimentally.
Key Innovation: Uses image-based DEM modeling to connect mineral-scale heterogeneity with fracture initiation and propagation under tensile loading.
17. Investigation of suspended cutoff walls on seepage and pollutant containment in saturated sites
Core Problem: Suspended cutoff walls can redirect groundwater flow, but their insertion depth controls both seepage bypass and contaminant breakthrough.
Key Innovation: Combines field-scale numerical coupling with centrifuge verification to assess cutoff-wall performance against pollutant migration.
18. Shear creep testing and constitutive modeling of jointed sandstone with varying joint angles under submerged conditions
Core Problem: Submerged jointed sandstone can creep along structural planes, affecting slopes, tunnels, and reservoirs, but joint angle effects are rarely quantified.
Key Innovation: Uses shear-creep testing and constitutive modeling to show how joint orientation controls time-dependent deformation under submerged conditions.
19. Reliability analysis of high arch dams using multi-objective parameter inversion and Bayesian model averaging
Core Problem: High arch dam safety assessment depends on uncertain model parameters and competing calibration objectives.
Key Innovation: Combines multi-objective inversion with Bayesian model averaging to quantify reliability while accounting for parameter and model uncertainty.
20. Continuum modeling of suffusion and filtration in underfilled gap-graded soils with an immobile skeleton
Core Problem: Underfilled gap-graded soils can lose fine particles while the coarse skeleton remains nearly immobile, complicating continuum representation of suffusion.
Key Innovation: Develops a continuum model for coupled suffusion and filtration that separates mobile fines from the load-bearing skeleton.
21. Modelling thermomechanical fracturing in underground hydrogen storage in salt caverns
Core Problem: Rapid gas cycling in salt caverns can create thermal stresses that threaten storage integrity through fracture initiation.
Key Innovation: Models thermomechanical fracture development during underground hydrogen storage to evaluate cavern stability under operational pressure-temperature changes.
22. Determining soil erosion rates on a grazed Australian hillslope: comparison of two landform evolution models with field-based methods
Core Problem: Landform evolution models are widely used in rehabilitation and hazard contexts, but few studies benchmark them against field-based erosion rates.
Key Innovation: Compares SIBERIA, SSSPAM, RUSLE, sediment-trap data, and 137Cs estimates to evaluate modeled erosion rates on a LiDAR-derived hillslope.
23. Environmental modulation of storm-flood response scale in a semi-arid urbanizing basin
Core Problem: Urbanizing semi-arid basins can shift storm-flood response scales as land surface and rainfall controls change.
Key Innovation: Uses stability-validated power-law modeling to identify environmental controls on the scale of storm-flood response.
24. The September 2024 Danube flood compared to the 1899, 2002, and 2013 events
Core Problem: The September 2024 Danube flood requires comparison with historical flood events to separate meteorological extremity, catchment memory, and climate context.
Key Innovation: Places the 2024 event against the 1899, 2002, and 2013 floods through hydrometeorological analysis under changing climate conditions.
25. Rapid changes in ocean surface waves across the eye of Hurricane Milton (2024)
Core Problem: Wave fields can change sharply across hurricane eyes, affecting coastal and offshore hazard loads at short time scales.
Key Innovation: Analyzes rapid spatial changes in ocean surface waves across Hurricane Milton's eye to constrain extreme wave-forcing structure.
26. Coarse sediment transport in a near-terminus subglacial channel
Core Problem: Sediment movement near glacier termini affects channel morphology, meltwater routing, and downstream sediment pulses, but direct constraints remain rare.
Key Innovation: Quantifies coarse sediment transport in a near-terminus subglacial channel to improve understanding of cryosphere-linked sediment dynamics.
27. Rapid changes in global river particulate organic carbon flux
Core Problem: River particulate organic carbon flux responds to changing hydrology, erosion, and sediment transport, with implications for Earth-surface feedbacks.
Key Innovation: Provides a global analysis of rapid changes in river particulate organic carbon flux, linking sediment movement to biogeochemical cycling.
28. Monitoring coastal erosion-driven LULC and shoreline changes along the Lagos Coastal Corridor with Sentinel-2 imagery and DSAS
Core Problem: Urban coastal corridors require consistent evidence of shoreline retreat and land-cover change driven by erosion.
Key Innovation: Combines Sentinel-2 imagery with DSAS shoreline analysis to quantify erosion-driven LULC and shoreline change along the Lagos coast.
29. Projected shifts in dry-wet extremes and possible associated drivers over the Hindu Kush Himalaya
Core Problem: The Hindu Kush Himalaya faces compound drought, flood, and slope-hazard implications as dry-wet extremes shift.
Key Innovation: Projects future dry-wet extreme changes and diagnoses associated drivers across the Hindu Kush Himalayan region.
30. Soil moisture impact on convective initiation
Core Problem: Soil moisture can modulate convective storm initiation, affecting the rainfall extremes that drive floods and slope failures.
Key Innovation: Evaluates how soil-moisture patterns influence convective initiation and thus short-duration precipitation hazard potential.
31. A flow matching-based deep generative framework for end-to-end hydrogeological parameter inversion in subsurface multiphase flow
Core Problem: Subsurface multiphase-flow assessment is limited by uncertain hydrogeological parameters and expensive inverse modeling.
Key Innovation: Uses flow matching to build an end-to-end generative inversion framework for hydrogeological parameters in multiphase flow systems.
32. Scale-specific controls of annual suspended sediment yield in the Yellow River Basin
Core Problem: Sediment-yield controls vary across spatial scales, complicating basin-wide erosion and sediment management.
Key Innovation: Identifies scale-specific controls on annual suspended sediment yield across the Yellow River Basin.
33. No Adaptation Without Observation: observability-constrained test-time prompt tuning for LiDAR semantic segmentation
Core Problem: LiDAR segmentation degrades under changing field conditions, while pseudo-label adaptation can amplify unreliable observations.
Key Innovation: Weights test-time prompt tuning by depth-consistent observability to stabilize adaptation under range-dependent sparsity and occlusion.
34. JL1-CC&QA: extending the JL1-CD benchmark with change captioning and question answering
Core Problem: Remote-sensing change detection often stops at pixel masks, while disaster monitoring needs language-level descriptions and question answering.
Key Innovation: Extends a satellite change-detection benchmark with captioning and QA tasks to support more interpretable bi-temporal scene analysis.
35. SAMBA: a scatter-guided masked bidirectional Mamba foundation model for SAR target recognition
Core Problem: SAR interpretation remains data-limited, and generic masking schemes ignore radar scattering physics.
Key Innovation: Introduces scatter-guided self-supervised pretraining with a bidirectional Mamba encoder for efficient SAR target representation learning.
36. TerraDiT-Omega: unified spatial control for satellite image synthesis with any geospatial primitive
Core Problem: Synthetic satellite imagery is difficult to condition on heterogeneous geospatial primitives such as points, lines, masks, or polygons.
Key Innovation: Proposes a unified spatial-control framework for satellite-image synthesis that can accept multiple geospatial primitive types.
37. AeroVerse-SatAgent: UAV-satellite collaborative spatial reasoning
Core Problem: Geohazard assessment often needs coordinated local UAV views and broader satellite context, but spatial reasoning across scales remains difficult.
Key Innovation: Develops an agentic UAV-satellite reasoning framework inspired by dual visual pathways for collaborative aerial and orbital interpretation.
38. GRAI3: generalizable soil particle-size distribution from smartphone images
Core Problem: Particle-size distribution is fundamental to erosion, infiltration, and geotechnical behavior, but laboratory measurement is slow and spatially sparse.
Key Innovation: Uses smartphone imagery to infer soil particle-size distributions, aiming for low-cost and field-deployable soil characterization.