TerraMosaic Daily Digest: May 8, 2026
Daily Summary
May 8's papers are led by a compact set of slope and mass-movement studies in which deformation monitoring is tied directly to mechanics. The Mosaic Creek work links field structure, InSAR evidence, and finite-element modelling for a potential deep-seated slope deformation in a deglaciated volcanic valley; the Taiwan mountain-road system converts landslide hazard into life-safety, closure, and maintenance economics; and the leakage-induced slope-failure and debris-flow rheology papers resolve how suction loss, clay-sand composition, and shear-rate transitions govern the shift from incipient instability to post-failure motion. Rockfall laser scanning adds an operational measurement threshold to this mechanics-centred group.
The broader signal is a move toward actionable hazard variables rather than descriptive maps alone. Sentinel-1 workflows are used to derive near-real-time subsidence, flood extents, and warning thresholds; mining subsidence is constrained with moment tensors and stress modelling; flood, windstorm, and extreme-precipitation studies propagate uncertainty into building damage, economic loss, and event footprints; and erosion papers connect hydrological connectivity, fine-sediment entrainment, tree mortality, and ice deformation to measurable geomorphic response. The lower-scored but useful methods extend this direction through DSM generation, UAV mosaicking, building-height extraction, all-weather LST fusion, physics-informed tunnel and underground-infrastructure models, and more diagnostic land-surface model evaluation.
Key Trends
The methodological centre of gravity is shifting from hazard presence to measurable deformation, threshold, loss, and recovery variables.
- Slope failure is being treated as a coupled sensing-mechanics problem: deep-seated deformation mapping, leakage-induced SPH simulation, debris-flow rheology, and rockfall laser scanning all link observed geometry or point-cloud change to process controls on failure and mobility.
- SAR workflows are becoming operational decision tools: CISA subsidence mapping, Sentinel-1 flood-threshold estimation, and SAR-boosted flood susceptibility move beyond event delineation toward near-real-time deformation products and warning thresholds.
- Risk models increasingly carry uncertainty and consequences through the full chain: the Cagayan Bayesian flood-impact study, Dutch windstorm loss model, U.S. building-level flood scoring, Taiwan landslide-road system, and global extreme-precipitation attribution connect forcing uncertainty to damage, access, exposure, or economic loss.
- Geomorphic change is being resolved through transport and connectivity: Loess Plateau erosion modelling, muddy-sediment entrainment, old-growth forest erosion, Baikal ice deformation, aeolian dune dynamics, and terrace uplift dating all quantify pathways by which material is stored, released, or rearranged.
- Infrastructure geohazard modelling is absorbing physics-informed learning: underground path-loss prediction, tunnel-deformation PINNs, tunnel-vibration acceleration, adjacent-excavation resilience, cement-treated soil transformers, and all-weather terrain LST products show a practical convergence between physical constraints and data-driven prediction.
Selected Papers
This issue contains 42 selected papers from 1,132 papers analyzed. The leading papers connect deep-seated slope deformation, cloud-based mountain-road landslide risk, leakage-induced slope failure, debris-flow rheology, rockfall laser scanning, near-real-time InSAR subsidence, SAR-derived flood thresholds, flood susceptibility, mining-induced fault reactivation, Bayesian flood-impact uncertainty, and the expanding footprint of extreme precipitation. The broader set adds windstorm and building-level flood risk, post-disaster recovery, hillslope erosion, hydrological-connectivity sediment modelling, fine-sediment entrainment, ice-deformation signals, DSM generation, UAV mosaicking, exposure mapping, underground monitoring communications, tunnel-deformation prediction, all-weather land-surface temperature fusion, and climate or geomorphic context for hazard analysis.
1. Insights from finite element modelling and remote sensing on a potential deep-seated slope deformation in a deglaciated valley in the southern Coast Mountains, British Columbia
Core Problem: Potential gravitational deformation zones in high mountain volcanic terrain can evolve into large-volume failures, but field indicators, InSAR motion, and mechanical controls are difficult to reconcile.
Key Innovation: Field mapping, remote sensing, InSAR, and finite-element modelling are integrated for Mosaic Creek Valley, linking antislope lineaments, deformation kinematics, deglaciated-valley geometry, and material controls.
2. Cloud-based risk and economic assessment system for landslide damage of mountain roads
Core Problem: Mountain-road agencies need to prioritize slope maintenance under typhoon and geologic landslide risk, but conventional inventories rarely translate hazard into life-safety, access, and economic consequences.
Key Innovation: A cloud-based assessment system combines environmental conditions, stabilization measures, historical landslide cases, road-closure losses, and maintenance decision support for Taiwan's mountain roads.
3. Modelling of unsaturated slope failures induced by water pipe leakage using smoothed particle hydrodynamics
Core Problem: Pipe leakage can rapidly destroy matric suction in unsaturated slopes and trigger failure without visible warning, while many models stop before the post-failure stage.
Key Innovation: A fully coupled hydro-mechanical SPH framework with an unsaturated critical-state constitutive model simulates both leakage-driven strength loss and post-failure slope motion.
4. Effects of clay and sand contents on phase transition and rheological shift behaviors in debris flows
Core Problem: Debris-flow initiation and runout depend on how granular composition shifts a mixture between solid-like, transitional, and liquid states.
Key Innovation: Rheological experiments quantify critical shear-rate thresholds, shear-thinning to shear-thickening transitions, and the opposing roles of clay and sand content in structural strength and flow behaviour.
5. Rockfall monitoring with three terrestrial and two mobile laser scanners: a comparison of detection levels
Core Problem: Rockfall monitoring requires detection thresholds that remain credible across terrestrial and mobile laser-scanning platforms.
Key Innovation: Three terrestrial and two mobile laser scanners are compared using 3DMASC filtering, ICP registration, M3C2 change detection, controlled displacements, and rockfall-volume validation.
6. Sentinel-1 Consecutive Interferogram Stacking Approach (CISA) for High-Resolution and Near-Real-Time Ground Subsidence Mapping
Core Problem: Operational subsidence monitoring often trades resolution for noise suppression, limiting small-scale urban and tectonic deformation detection.
Key Innovation: The CISA workflow stacks consecutive full-resolution Sentinel-1 interferograms to preserve native spatial detail while improving near-real-time deformation mapping in Pakistan's capital region.
7. Estimation of Flood Thresholds for Hydrological Warning Purposes Using Sentinel-1 SAR Imagery-Based Modeling in the Tumbes River Basin (PERU)
Core Problem: Cloud cover and sparse gauges limit flood warning in dry tropical basins during extreme events.
Key Innovation: A Google Earth Engine Sentinel-1 framework calibrates Otsu-based flood extraction against PlanetScope imagery and derives hydrological warning thresholds for the Tumbes River Basin.
8. SAR-based flood detection and different ensemble boosting techniques for multi-factor flood susceptibility modelling: a case study in Uttar Dinajpur, West Bengal, India
Core Problem: Flood susceptibility remains difficult to model because conditioning factors interact nonlinearly and flood extents must be observed reliably.
Key Innovation: Sentinel-1 flood detection is coupled with RFE-selected factors and five boosting models, including CatBoost, LightGBM, and XGBoost, for Uttar Dinajpur district.
9. Geophysical observations and numerical modeling constrain mining-induced subsidence and fault reactivation: A case study
Core Problem: Mining in faulted rock masses can reactivate pre-existing structures, producing coupled subsidence and seismic hazard.
Key Innovation: Microseismic moment-tensor inversion and 3-D numerical modelling jointly resolve shear-dominated fault rupture, tensile overburden fracturing, and evolving differential-stress controls at Xinjulong coal mine.
10. Quantifying uncertainty in climate change impacts: A Bayesian approach for the Cagayan River Basin
Core Problem: Rare-event rainfall uncertainty propagates nonlinearly through inundation and damage estimates, making deterministic flood-risk projections fragile.
Key Innovation: Bayesian extreme-value inference, CMIP6 model averaging, rainfall-runoff modelling, and damage-chain uncertainty propagation are linked for the Cagayan River Basin.
11. Observed Increase in the Spatial Extent of Extreme Precipitation Events Attributed to Anthropogenic Climate Change
Core Problem: The spatial footprint of extreme precipitation determines disaster losses, but global trends and external forcing attribution remain underconstrained.
Key Innovation: Global observations and climate simulations show increasing event extent across multiple continents and attribute the expansion primarily to anthropogenic climate change.
12. An empirically derived high-resolution spatial windstorm risk model: an application for the Netherlands
Core Problem: Windstorm risk models need to connect high-resolution gust climatology, exposure, vulnerability, and insurance loss in a transferable framework.
Key Innovation: Downscaled reanalysis, extreme-value return levels, detailed building attributes, vulnerability curves, and claim-ratio functions are integrated for the Netherlands.
13. National-level Resilience-Informed Risk Scoring Approach: Generalized Socio-Physical Flood Risk Factor for Buildings
Core Problem: Community flood-risk screening often omits building vulnerability, social vulnerability, and resilience capacity in the same scoring framework.
Key Innovation: A scalable geospatial approach combines FEMA flood layers, National Structure Inventory attributes, fragility functions, social vulnerability, and community resilience factors.
14. Can pre-disaster community resilience enhance post-disaster recovery among older adults? A natural experiment in Iwanuma City after the 2011 Great East Japan Earthquake and Tsunami
Core Problem: Whether pre-disaster community resilience measurably improves long-term recovery remains difficult to isolate from post-event conditions.
Key Innovation: A natural experiment in Iwanuma City uses pre-2011 baseline data, post-tsunami follow-up, and multilevel modelling across 98 districts and 3,523 older residents.
15. Accelerated erosion and tree mortality in a primeval temperate forest dominated by large trees, Washington, USA
Core Problem: Large trees can regulate soil production and hillslope transport, but mortality-driven erosion changes are rarely quantified with independent timescale constraints.
Key Innovation: Repeated tree censuses are combined with 239+240Pu and 10Be soil-redistribution estimates to compare recent and long-term erosion in Wind River old-growth forest.
16. The spatial and temporal heterogeneity and driving mechanism of soil erosion in the loess plateau based on hydrological connectivity
Core Problem: Soil erosion in the Loess Plateau reflects climate, land-use change, and conservation projects, but their pathways through hydrological connectivity are not easily separated.
Key Innovation: A MUSLE-IC-SDR framework is coupled with geographical detectors, geographically weighted regression, and path analysis to attribute spatial-temporal erosion change in the Malian River Basin.
17. Entrainment of Muddy Sediment From Immobile Gravel River Beds
Core Problem: Fine sediment stored within gravel beds may be re-entrained even when the surface gravel layer remains immobile, complicating sediment-transport assumptions.
Key Innovation: Field and process analysis show how pore-fill elevation relative to gravel tops controls mud entrainment thresholds without full gravel mobilization.
18. Deterministic structures in functionals of Baikal ice deformations
Core Problem: Ice shocks are difficult to anticipate because deformation statistics can contain deterministic structure hidden within high-resolution measurement series.
Key Innovation: Functionals derived from Lake Baikal ice-cover deformation measurements are analysed to identify deterministic patterns preceding ice-shock events.
19. Accelerated uplift of the Rhenish Massif (central Europe) since 700–800 ka revealed by isochron-burial dating of strath terraces
Core Problem: Terrace correlations in the Rhenish Massif have uncertain age control, limiting estimates of Quaternary uplift and valley incision.
Key Innovation: Isochron-burial dating of Older and Younger Main Terrace deposits provides new terrace ages and constrains accelerated uplift since roughly 700-800 ka.
20. SAT-MAK: Digital Surface Model Generation from Satellite Imagery Using Multi-Type Aggregated Keypoints and Weighted Clustering
Core Problem: Large-format satellite DSM generation is limited by uneven keypoint distribution and unstable matching across diverse imagery.
Key Innovation: SAT-MAK combines multi-type aggregated keypoints, weighted clustering, and refined matching to improve high-resolution satellite-derived DSM reconstruction.
21. LargeStitch: Efficient Seamless Stitching of Large-Size Aerial Images via Deep Matching and Seam-Band Fusion
Core Problem: Large UAV image stitching is slowed by dense-feature extraction, misalignment, and visible seam artefacts.
Key Innovation: LargeStitch uses deep matching, batch alignment, and seam-band fusion to produce efficient seamless panoramas from large aerial images.
22. Robust Recursive Fusion of Multi-Resolution Multispectral Images with Location-Aware Neural Network
Core Problem: Multi-resolution image fusion can be corrupted by cloud, shadow, and sensor outliers when recursively updating high-resolution products.
Key Innovation: A location-aware neural network models image dynamics while explicitly representing pixel contamination probability for robust recursive fusion.
23. Urban Building Footprints Extraction and Heights Estimation from High-Resolution Spaceborne Remote Sensing Imagery Using a CNN-Transformer Network
Core Problem: Urban exposure models require footprints and heights from high-resolution imagery despite shadows, occlusions, and heterogeneous structures.
Key Innovation: SECT-Net extracts footprints and shadows from Jilin-1 imagery and estimates heights for more than 750,000 Shanghai buildings using a shadow-based workflow.
24. A Benchmark Evaluation of Intelligent Identification Models for Marine Outfalls
Core Problem: Large-scale marine outfall inspection remains manual and weakly standardized despite direct relevance to coastal environmental risk.
Key Innovation: A multi-source remote-sensing fusion dataset and model benchmark compare YOLO and RT-DETR variants for outfall detection along Zhanjiang's southern coast.
25. Evaluating Field Sampling Design and LiDAR-Based Approaches for Woody Vegetation Assessment in Reclaimed Wellsite Certification
Core Problem: Reclaimed wellsite certification can be biased by judgment-based field plots and limited sampling intensity.
Key Innovation: Conventional plots, random plots, UAV-LiDAR, and airborne laser scanning are compared for woody vegetation density and height assessment on reclaimed Alberta sites.
26. Optimization of Sampling Density in Regional-Scale Soil Organic Matter Mapping and Its Impact on Prediction Accuracy
Core Problem: Digital soil mapping must balance field-sampling cost against prediction reliability.
Key Innovation: Sentinel-2 variables, environmental covariates, XGBoost, and stratified random sampling identify a lower-cost sampling density that preserves SOM prediction patterns.
27. A physics-informed machine-learning model for wireless signal path loss prediction in underground infrastructure monitoring
Core Problem: Buried infrastructure monitoring depends on stable data transmission, but heterogeneous subsurface attenuation is poorly captured by classical path-loss formulas.
Key Innovation: A physics-informed machine-learning model combines recalibrated Friis and Fresnel baselines with residual learning from in-situ underground path-loss measurements.
28. Research on MPICP system in tunnel excavation process based on multi-physical field coupling
Core Problem: Conventional ventilation may not sufficiently suppress coupled dust and gas exposure during tunnel excavation.
Key Innovation: Multi-physical-field simulation and field measurements are used to optimize a same-side pressurized-extraction ventilation mode and an intelligent collaborative purification system.
29. Rapid prediction of metro tunnel vibration via linear expression method and component mode synthesis
Core Problem: High-fidelity tunnel vibration models are computationally expensive, limiting rapid evaluation of underground rail impacts.
Key Innovation: A linear expression method and component mode synthesis accelerate coupled train-track-tunnel-soil vibration prediction while matching finite-element and field data.
30. The resilience assessment model for existing tunnels under the disturbance of adjacent tunnel excavation
Core Problem: Adjacent tunnel excavation can disturb existing tunnels, but resilience assessment needs both disturbance intensity and recovery duration.
Key Innovation: A 3-D resilience model defines ground disturbance degree, disturbance intensity, functional recovery, resilience index, and level thresholds for existing tunnel systems.
31. Prediction of tunnelling-induced longitudinal deformation based on knowledge-guided neural network
Core Problem: Data-driven tunnel-deformation models struggle when monitoring and geological data are sparse.
Key Innovation: A physics-informed neural network embeds governing equations and boundary conditions through automatic differentiation to predict longitudinal deformation with sparse observations.
32. Surface transient waves induced by an arbitrary-speed moving force on a transversely isotropic half-space
Core Problem: Moving loads can induce transient surface waves whose behaviour in anisotropic half-spaces is analytically difficult to resolve.
Key Innovation: A Galilean-transformed potential-function solution with Fourier-Laplace transforms derives wave response under arbitrary-speed moving vertical force.
33. Indirect tensile strength of cement-treated soil under diverse curing conditions: From experimental characterization to prior-fitted transformer-based prediction
Core Problem: Cement-treated soil tensile strength varies with cement content, curing temperature, and curing duration, limiting rapid design estimation.
Key Innovation: Brazilian splitting tests, prior-fitted transformer prediction, and explainable AI quantify indirect tensile strength under diverse curing conditions.
34. Coupling land surface modeling with multi-scale adaptive fusion: Generating hourly 1-km all-weather LST in cloud-prone complex terrain
Core Problem: High-frequency land-surface temperature products in complex terrain suffer from cloud gaps and topographic texture blurring.
Key Innovation: Noah-MP physical anchors are coupled with terrain-driven error decoupling and multi-scale adaptive weighting to generate hourly 1-km all-weather LST.
35. Introducing the Model Fidelity Metric (MFM) for robust and diagnostic land surface model evaluation
Core Problem: Common efficiency metrics can obscure magnitude, phase, and distributional errors in skewed land-surface simulations.
Key Innovation: The Model Fidelity Metric introduces a diagnostic, robust evaluation score designed for non-normal land-surface modelling data.
36. Global Diurnal Precipitation Cycle in the AI Model GraphCast and a 5‐km Unified Model: Challenges and Opportunities
Core Problem: AI weather models must reproduce diurnal precipitation timing before their rainfall fields can be trusted for hazard forcing.
Key Innovation: GraphCast, ERA5, IMERG, and a 5-km Unified Model are compared globally, revealing inherited and amplified timing biases in boreal-summer precipitation.
37. Volcanic Forcing Overwhelmed AMV's Influence on Eurasian Winter Temperatures During the Little Ice Age
Core Problem: Long-term temperature teleconnections can be masked by volcanic forcing, complicating historical climate-hazard interpretation.
Key Innovation: CESM Last Millennium Ensemble analysis shows volcanic forcing overwhelmed AMV-related Eurasian winter temperature patterns during the Little Ice Age.
38. Morphological response of asymmetric barchan dunes to bimodal winds and the driving mechanism of surface airflow
Core Problem: Asymmetric barchan dunes have airflow and sediment-routing patterns that are difficult to infer from static morphology.
Key Innovation: Field measurements in the Taklamakan Desert link reverse airflow, limb geometry, sediment properties, and bimodal wind forcing to asymmetric dune evolution.
39. Morphological dynamics of complex longitudinal dunes in the Kumtagh Desert: Synergistic effects of tri-directional winds and sediment grain size
Core Problem: Longitudinal-dune formation models often underrepresent sediment grain-size control and tri-directional wind forcing.
Key Innovation: Satellite morphology, sedimentology, and aeolian dynamics are integrated to explain complex longitudinal dunes in the Kumtagh Desert.
40. Mining impacts on heavy metals in Qingni River Basin (China): Integrating patterns, mineral carriers, and diatom responses
Core Problem: Mining-derived heavy metals require integrated evidence linking spatial patterns, mineral carriers, and biological response.
Key Innovation: Spatial analysis, TIMA-EDS mineralogy, and diatom indicators connect Cd, Pb, and Zn enrichment to historical lead-zinc mining in the Qingni River Basin.
41. Effects of salinity on speciation and mobility of potentially toxic elements in soil porewater of estuarine wetlands
Core Problem: Sea-level rise and salinity fluctuations can remobilize toxic elements from wetland porewater, but speciation changes are underresolved.
Key Innovation: High-resolution dialysis and stage-filter fractionation quantify how 0-35 psu salinity shifts colloidal and dissolved As, Se, V, and other element fractions.
42. Siderite Concretions in Svalbard Lake Sediments Capture 7,000 Years of Extreme Arctic Cold Season Climate Change
Core Problem: Most Arctic proxies emphasize growing-season conditions despite strongest future warming being expected in winter.
Key Innovation: Microscale core scanning of siderite concretions in Svalbard lake sediments reconstructs about 6.7-7 kyr of extreme cold-season climate variability.