TerraMosaic Daily Digest: June 1, 2026
Daily Summary
The June 1 papers concentrate on the measurement chain that turns slope motion into hazard knowledge. The strongest landslide studies do not merely classify terrain: they quantify uncertainty in susceptibility inputs, track slow-moving failures from optical imagery, and update reservoir-slope reliability from observed responses. A parallel group of studies treats landslide motion as a coupled process, linking rainfall lag, glacier retreat, flood-generated mass flow, debris-flow channel geometry, and boulder impact mechanics to the evolution of failure and consequence.
The wider geohazard literature extends this process view into floods, earthquakes, rock instability, mining deformation, tailings dams, tunnels, and unsaturated or frozen soils. Methodological progress is clearest where remote sensing and machine learning are constrained by mechanics, sensor geometry, or hydrological forcing: optimized InSAR deformation retrieval, terrain-adaptive LiDAR-SLAM in underground settings, hierarchical Bayesian CPTU interpretation, and large vision-language grounding for Earth observation all improve the observability of terrain state. The issue is therefore less a catalogue of isolated hazards than a set of tools for estimating when, where, and why geomorphic systems become unstable.
Key Trends
Five movements define the issue: state-estimating landslide workflows, hydrology as dynamic forcing, measurable debris-flow consequence, coupled geotechnical reliability, and remote-sensing AI tied to physical observability.
- Landslide mapping is moving from detection to state estimation: TerraTrack, error-aware susceptibility mapping, reservoir-slope Bayesian updating, and LLM-assisted post-disaster reporting all shift landslide workflows toward explicit displacement, uncertainty, reliability, and interpretable event descriptions.
- Hydrology is treated as a dynamic forcing field: Rainfall-lag displacement prediction, paraglacial water availability, reservoir-slope reliability updating, flood-susceptibility modelling, and global precipitation benchmarking all place water input and routing at the center of hazard inference.
- Runout, impact, and channel morphology are becoming measurable targets: The debris-flow papers emphasize motion and consequence: three-dimensional landslide-flow simulation, UAV-SLAM channel mapping, and boulder-impact mitigation each connect source-area instability to downstream exposure.
- Geotechnical hazard models are becoming more probabilistic and coupled: Anchored slopes, nonlinear strength criteria, tailings dams, expansive-soil cracking, unsaturated footing problems, and CPTU parameter inference all show a move away from single deterministic safety factors toward evolving material state, boundary condition, and uncertainty.
- Remote-sensing AI is useful when it improves physical observability: The transferable AI papers matter here because they support terrain-state measurement: optimized InSAR corridor deformation retrieval, satellite stereo, airborne LiDAR adaptation, and multi-image visual grounding can all feed hazard models with denser or more interpretable observations.
Selected Papers
This issue contains 40 selected papers from 1,212 papers analyzed. The selected papers are ordered by relevance score and emphasize landslide deformation monitoring, susceptibility uncertainty, rainfall and glacier controls on slope motion, debris-flow impact and mapping, flood and earthquake risk, and remote-sensing methods that improve terrain-state observation.
1. A workflow to identify and monitor slow-moving landslides through spaceborne optical feature tracking
Core Problem: Slow-moving landslides require repeatable displacement monitoring that remains accessible beyond expert local computing environments.
Key Innovation: Introduces TerraTrack, an open-source cloud workflow for Sentinel-2 optical feature tracking that outputs landslide masks, velocity maps, and displacement time series.
2. A novel error inversion method for landslide susceptibility assessment based on topography-derived and spatially projected factors
Core Problem: Susceptibility maps inherit errors from DEM-derived continuous factors and rasterized discrete factors, yet these errors are rarely quantified systematically.
Key Innovation: Builds an error inversion framework to identify suitable spatial precision for both continuous terrain factors and spatially projected categorical factors.
3. Constrained Bayesian inference of geotechnical parameters for response prediction and reliability updating of reservoir slopes
Core Problem: Near-dam slope failure can trigger landslide-surge-dam-breach-flood chains, but parameter uncertainty is difficult to update from monitoring responses.
Key Innovation: Formulates constrained Bayesian inference to invert geotechnical parameters, predict slope response, and update reservoir-slope reliability.
4. From Landslide Detection to Multi-Source LLM-Based Reporting: A Complete Framework for Rapid Assessment of Post-Disaster Scenarios
Core Problem: Post-event landslide response remains slowed by manual interpretation, field risk, and fragmented geospatial evidence.
Key Innovation: Combines enhanced SegFormer landslide localization, geo-attribute extraction, and an instruction-tuned LLM for structured rapid assessment reports.
5. Displacement prediction model for rainfall-induced landslides based on PDL-VMD-SVR: A case study of the southern slope landslide at Fushun West Open-pit Mine, China
Core Problem: Rainfall-triggered displacement shows lagged and abrupt behavior that static rainfall-displacement models fail to capture.
Key Innovation: Combines polynomial distributed lag modelling, variational mode decomposition, and support vector regression for rainfall-driven landslide displacement forecasting.
6. Daily to annual controls on paraglacial slope stability at Portage Glacier, Alaska
Core Problem: Retreating glaciers destabilize adjacent slopes through interacting annual thinning, lake expansion, and daily to seasonal water availability.
Key Innovation: Uses camera, seismic, geodetic, DEM, and environmental records to isolate controls on paraglacial instability motion at multiple time scales.
7. Three-dimensional numerical modeling of the movement of the flood mixture in conditions of complex river topography by taking into account the moving river bed
Core Problem: Landslide-water mixtures moving through complex river relief are difficult to simulate because bed motion and non-Newtonian material behavior interact.
Key Innovation: Uses a three-dimensional VOF-based model with moving-bed representation to simulate landslide-flow waves and debris-flow arrival characteristics.
8. AscDAMs 2.0: advanced SLAM-and-UAV-based channel detection and mapping system
Core Problem: Debris-flow gullies require high-resolution channel morphology, but satellite imagery and backpack mapping can fail in steep, semi-enclosed terrain.
Key Innovation: Extends AscDAMs into an autonomous UAV-SLAM mapping system for debris-flow channel detection and morphology measurement.
9. Mitigation of boulder-laden debris flow impacts on frame structures using an EVA cushion layer: A numerical study
Core Problem: Frame structures exposed to boulder-rich debris flows can fail through intense localized impact and energy transfer.
Key Innovation: Numerically evaluates an EVA cushion layer that reduces peak impact force and structural damage under high-velocity boulder-flow loading.
10. Effect of embankment construction quality on stability of Aketao tailings dam under rainfall-seismic coupling
Core Problem: Raised tailings dams can fail catastrophically when construction quality, rainfall infiltration, and seismic loading interact.
Key Innovation: Combines field survey, laboratory testing, and numerical simulation to test drainage and density controls on rainfall-seismic tailings-dam stability.
11. Slope Stability Analysis Using Continuously Evolving Equivalent Strength for Nonlinear Failure Criteria
Core Problem: Nonlinear strength effects are often added after conventional slope-stability calculations, limiting their role in defining critical mechanisms.
Key Innovation: Embeds evolving equivalent strength directly into a discrete upper-bound framework to solve slip surface, strength distribution, and safety factor jointly.
12. Seismic stability of anchored slopes: a comprehensive framework for coupled viscoelastic-rheological effects in anchoring cables and geomaterials
Core Problem: Anchored slopes lose prestress over time while earthquakes can mobilize or overload anchor systems.
Key Innovation: Develops a coupled framework for seismic anchored-slope stability that includes viscoelastic soil behavior, rheological anchor response, prestress loss, and permanent deformation.
13. Effects of ground-based camera deployment parameters on SfM-Based 3D reconstruction of deformation features on stepped slopes
Core Problem: Ground-based SfM on stepped slopes is sensitive to occlusion, camera height, distance, and overlap, limiting deformation-feature reconstruction.
Key Innovation: Uses scaled physical modelling and field validation to define camera-deployment controls for reconstructing cracks and deformation features on stepped slopes.
14. The Combined Influence of Weathering, Anisotropy, and Cyclic Loading on Barton's Peak Friction Angle: Scaling Intact Rock Behaviour to Rock Joint
Core Problem: Mountain rock instability depends on joint shear strength, but weathering, anisotropy, and cyclic loading are not easily transferred from intact-rock tests to joint-scale parameters.
Key Innovation: Evaluates how those controls alter Barton's peak friction angle, improving scaling from intact rock behavior to joint stability assessment.
15. Tracking 3D desiccation crack evolution and permeability dynamics in expansive soil
Core Problem: Desiccation cracking changes infiltration pathways in expansive soils, but its internal 3D evolution and permeability impact are poorly resolved.
Key Innovation: Combines X-ray CT, infiltration testing, and 3D seepage simulation to track crack-network growth and permeability dynamics.
16. Bearing Capacity of Strip Footings near Vertical Slopes in Cross Anisotropic Unsaturated Soil under Transient Flow
Core Problem: Transient infiltration changes suction stress and unit weight near vertical cuts, altering footing capacity and slope-adjacent infrastructure safety.
Key Innovation: Integrates transient-flow suction stress and cross anisotropy into finite-element lower-bound limit analysis for strip footings near vertical slopes.
17. Metaheuristic-optimized ANN for flood susceptibility mapping: a case study in the Khiavchai Watershed, Iran
Core Problem: Watershed-scale flood susceptibility depends on terrain, hydrology, land cover, and SAR-derived inventory information that interact nonlinearly.
Key Innovation: Compares metaheuristic-optimized ANN models using Sentinel-1/2, Landsat, DEM, and hydrologic factors for flood susceptibility mapping.
18. Integrating GIS-based and AHP-enhanced approach for identifying earthquake zones in Northwest Vietnam
Core Problem: Seismic-source identification in mountainous regions requires integrating geological, geomorphological, geophysical, and hazard-conditioning factors.
Key Innovation: Combines AHP, Garson's algorithm, and neural-network weighting to identify earthquake zones in Northwest Vietnam.
19. Numerical modelling of the stress-strain responses of clean and laponite-treated Hostun sands
Core Problem: Loose sands require treatments that improve liquefaction resistance, but constitutive models must capture treated and untreated cyclic response.
Key Innovation: Models clean and laponite-treated sands using OpenSees and Plaxis constitutive frameworks calibrated against cyclic triaxial tests.
20. Independence verification of peak-strength strain energy storage index from rock specimen shape effects
Core Problem: The strain energy storage index is widely used for rockburst proneness, but specimen-shape effects can bias its interpretation.
Key Innovation: Uses theoretical derivation and compression tests on granite and red sandstone to test whether the index is independent of specimen geometry.
21. Acoustic Emission-Microcurrent Response and Energy Dissipation Characteristics of Coal Under Impact Loading
Core Problem: Rockburst early warning needs coupled physical signals that reveal energy dissipation during impact-induced coal failure.
Key Innovation: Synchronously monitors acoustic emission and microcurrent under impact loading to characterize acousto-electric response and energy release.
22. Rapid prediction of roof weighting behavior in karst underground coal mines using simulation-based deep learning
Core Problem: Karst peak-cluster mining creates heterogeneous overburden stress redistribution and unexpected roof weighting.
Key Innovation: Uses simulation-based deep learning for rapid roof-weighting prediction in karst underground coal mines.
23. Lagged backward-compatible physics-informed neural networks for unsaturated soil consolidation analysis
Core Problem: Unsaturated consolidation under long-term loading is computationally demanding and important for near-surface ground deformation.
Key Innovation: Develops a lagged backward-compatible PINN for forward simulation and inversion of unsaturated soil consolidation.
24. Curved Megathrust Geometry and Locking Heterogeneity Contributed to the Rupture of the 2025 Mw 8.8 Kamchatka Earthquake, as Inferred from Geodesy and Seismic Data
Core Problem: Large megathrust rupture interpretation requires resolving fault geometry and locking heterogeneity from geodetic and seismic observations.
Key Innovation: Combines InSAR, GPS, and seismic evidence to connect curved megathrust geometry and locking patterns to the 2025 Kamchatka rupture.
25. Refining InSAR Deformation Retrieval for the South-to-North Water Diversion via Buffer Optimization
Core Problem: Ultra-long water-diversion infrastructure is vulnerable to deformation, but InSAR buffers can mix corridor and background signals.
Key Innovation: Optimizes buffer design for SBAS-InSAR deformation retrieval along the South-to-North Water Diversion corridor.
26. Hydro-mechanical coupling between pore pressure and crustal deformation following earthquakes: insights from groundwater tidal response and InSAR analysis
Core Problem: Pore-pressure changes after earthquakes can couple groundwater response and crustal deformation, but direct field evidence is limited.
Key Innovation: Combines groundwater tidal response and InSAR analysis to infer hydro-mechanical coupling in the shallow crust after earthquakes.
27. The impact of spatial resolution on hourly flood modeling in large watersheds
Core Problem: Flood modelling in large watersheds depends strongly on spatial resolution, especially for hourly simulations at internal stations.
Key Innovation: Evaluates multiple resolutions in the Jialing River Basin and uses XGBoost to identify flood characteristics sensitive to resolution.
28. Comprehensive Global Assessment of 24 Gridded Precipitation Datasets Across 18 428 Catchments Using Hydrological Modeling
Core Problem: Hazard models depend on precipitation products, but users lack global evidence on dataset performance across diverse catchments.
Key Innovation: Benchmarks 24 gridded precipitation datasets through hydrological modelling across 18,428 catchments.
29. Enhancing Empirical Predictions from CPTU Data Using a Hierarchical Bayesian Framework with Adaptive Model Weighting
Core Problem: CPTU-based empirical correlations are practical but often deterministic and weakly calibrated to site-specific uncertainty.
Key Innovation: Uses hierarchical Bayesian modelling and adaptive model weighting to improve undrained-strength prediction and uncertainty estimates.
30. Assessing hyperspectral LiDAR capability for understory soil spectral retrieval and property estimation
Core Problem: Understory and soil-surface properties are hard to retrieve where canopy blocks passive optical sensing, yet these properties affect hydrology and shallow-slope behavior.
Key Innovation: Tests hyperspectral LiDAR for soil spectral retrieval and property estimation in vegetation-soil scenes.
31. Terrain-Adaptive LiDAR-IMU SLAM with Multi-Feature fusion for complex underground environments
Core Problem: Underground safety monitoring requires robust mapping where geometry, intensity, and ground constraints vary strongly.
Key Innovation: Develops a terrain-adaptive LiDAR-IMU SLAM framework that fuses intensity, ground structure, and inertial information for complex underground environments.
32. Beyond single and earthbound: Exploring multi-image grounding in remote sensing with Large Vision-Language Models
Core Problem: Vision-language models struggle with fine-grained, spatially grounded interpretation of remote-sensing scenes across multiple images.
Key Innovation: Develops multi-image grounding for remote sensing with large vision-language models, supporting spatial reasoning that can transfer to hazard-scene interpretation.
33. APCoTTA: Continual Test-Time Adaptation for semantic segmentation of airborne LiDAR point clouds
Core Problem: Airborne LiDAR segmentation models degrade under changing sensors, domains, and terrain scenes.
Key Innovation: Introduces continual test-time adaptation for airborne LiDAR point-cloud segmentation, reducing domain-shift risk in operational terrain mapping.
34. HUGStereo: A Hierarchical Fusion and Geometry-Aware Refinement Network for High-Resolution Satellite Stereo Matching
Core Problem: Satellite stereo matching must handle occlusion, repetitive texture, and scale variation that affect terrain reconstruction.
Key Innovation: Uses hierarchical feature and cost-volume fusion with geometry-aware refinement for high-resolution satellite disparity estimation.
35. SCENet: Semantic Consistency-Enhanced Decoding With Cross-Scale Tokens for High-Resolution Remote Sensing Image Segmentation
Core Problem: High-resolution segmentation suffers semantic drift and degraded boundaries during multiscale feature reconstruction.
Key Innovation: Constrains decoding with semantic consistency and cross-scale tokens, a transferable improvement for high-resolution hazard-scene segmentation.
36. Task-driven remote sensing image captioning for earth understanding: From multi-task dataset to task-conditioned framework
Core Problem: Remote-sensing captions are often visually plausible but weakly tied to downstream tasks such as response, monitoring, and environmental interpretation.
Key Innovation: Introduces a multi-task dataset and task-conditioned captioning framework that can support more actionable Earth-observation descriptions.
37. Impact of climate change on hydroclimatic extremes across major river basins of Peninsular India
Core Problem: Future hazard planning requires basin-scale evidence on changing drought and intense-rainfall extremes.
Key Innovation: Assesses climate-change impacts on hydroclimatic extremes across major Peninsular Indian river basins.
38. Revealing Coastal Storm-Wave Transformations From SWOT HR Observation: The English Channel Case Study
Core Problem: Nearshore storm-wave refraction and diffraction are poorly resolved by conventional satellite altimetry.
Key Innovation: Uses SWOT high-rate observations to measure sub-kilometric coastal wave transformations during Storm Mathis.
39. Local-level flood disaster governance in Nepal: Insights from the Tinau River Basin
Core Problem: Local flood-risk governance in developing regions is shaped by policy, institutional capacity, and implementation constraints.
Key Innovation: Examines local flood-governance strengths and weaknesses in the Tinau River Basin, Nepal.
40. Assessment of the Relationship Between Seismic Vulnerability and Seismic Risk Perception: A Case Study of Peshawar, Pakistan
Core Problem: Urban seismic risk management depends on both built-environment vulnerability and public risk perception.
Key Innovation: Assesses how seismic vulnerability and perceived risk relate in Peshawar, linking hazard exposure with preparedness behavior.