TerraMosaic Daily Digest: May 18, 2026
Daily Summary
The May 18 literature is anchored by direct slope-process studies. Its strongest contribution is methodological: open, repeatable monitoring of slow-moving landslides from Sentinel-2 optical feature tracking is paired with reservoir-slope hydro-mechanical modelling, regional tree-ring chronologies, landslide-dam stability assessment, rock-ice avalanche reconstruction, and probabilistic seismic slope analysis. Together these studies treat landslides as evolving systems whose displacement, triggering, hydraulic boundary conditions, and cascading consequences can be measured rather than inferred only from static susceptibility.
The surrounding geohazard literature sharpens the forcing and infrastructure context. Flood papers quantify how mountain urbanization, watershed model resolution, coastal vulnerability, wetland backwater dynamics, and Mediterranean atmospheric rivers alter hazard exposure. Cold-region and underground-infrastructure studies trace heat, water, seismic loading, and joint behavior through embankments, high-geothermal tunnels, subsea tunnels, shield linings, and foundation pits. Remote-sensing and AI papers are most valuable where they improve geologic observables: soil moisture, precipitation covariance, lithology under vegetation, optical-SAR retrieval, and civil-infrastructure crack segmentation.
Key Trends
The common thread is process-state measurement: landslide velocity, reservoir seepage, reactivation chronology, dam stability, rainfall infiltration, seismic slope safety, flood depth, soil moisture, and infrastructure deformation.
- Landslide studies emphasize monitoring and mechanics over static mapping: TerraTrack, reservoir toppling analysis, dendrochronology, landslide-dam assessment, rock-ice avalanche reconstruction, multi-source deep learning, active-learning slope digital twins, and probabilistic seismic stability broaden measurable landslide state variables.
- Hydrologic hazard papers resolve where forcing becomes damaging: mountain flood modelling, haor early warning, hourly model-resolution tests, coastal FVI mapping, atmospheric-river detection, soil-moisture prediction, AMSR2-to-SMAP translation, and anisotropic interpolation refine the water variables controlling hazards.
- Infrastructure hazard research follows coupled degradation paths: high-geothermal tunnel synthesis, subsea tunnel seismic damage, shield joints, undercrossing settlement, rheological linings, retaining-wall PINNs, pile-soil interaction, and outage recurrence modelling connect loading to structural performance.
- Cold-region and subsurface papers treat heat and fluid flow as geotechnical drivers: permafrost embankment drainage, hydrochemical coal-rock weakening, fracture clogging, non-Newtonian fracture flow, hydraulic-fracture bedding interaction, gas-water migration, and energy-pile thermomechanics track transport-driven material change.
- Remote-sensing advances matter where they strengthen geologic or damage observables: PolAlpha-Net, GeoMamba, South China rainfall microphysics, Zelongnong UAV and satellite reconstruction, the civil-infrastructure crack dataset, and lightweight PolSAR classification improve lithology, retrieval, precipitation structure, mass-flow interpretation, damage segmentation, or land-cover evidence.
Selected Papers
This issue contains 50 selected papers from 2,013 papers analyzed. The leading papers are centered on landslide process measurement and coupled geotechnical response: open Sentinel-2 monitoring of slow-moving landslides, reservoir-slope deformation under water-level and rainfall forcing, landslide-dam stability, regional reactivation chronology, multi-source deep-learning landslide mapping, excavation-slope digital twins, probabilistic seismic slope stability, rainfall-driven reinforced-slope behavior, fissured rock-slope fracture propagation, and rock-ice avalanche erosion. The broader set extends those controls into flood exposure, atmospheric-river rainfall, soil-moisture prediction, permafrost embankment drainage, high-geothermal tunnels, subsea and shield-tunnel seismic response, foundation deformation, fracture-network flow, lithology mapping, and infrastructure damage detection.
1. A workflow to identify and monitor slow-moving landslides through spaceborne optical feature tracking
Core Problem: Slow landslides require repeatable regional monitoring, but optical feature tracking workflows remain difficult to deploy at scale.
Key Innovation: TerraTrack provides an open-source, cloud-based Sentinel-2 workflow that produces landslide masks, velocity maps, and displacement time series in Google Earth Engine or Python.
2. Deformation evolution and stability analysis of the huangcaoping toppling slope in a hydropower reservoir
Core Problem: Anti-dip bedded reservoir slopes deform under coupled reservoir-level fluctuation, rainfall infiltration, and groundwater response.
Key Innovation: GNSS, SBAS-InSAR, field mapping, and a hydro-mechanically coupled UDEC model link reservoir stage, seepage evolution, rainfall, and progressive toppling deformation at Huangcaoping.
3. Landslide dam activity in Flysch Zone of the Eastern Carpathian, Romania: distribution, stability, and failure mechanism assessment
Core Problem: Landslide dams in flysch mountain belts can persist for decades but also fail suddenly, yet their stability controls are poorly constrained.
Key Innovation: A 21-dam Eastern Carpathian inventory relates dam distribution, triggering factors, morpho-hydrology, stability indices, and failure mechanisms.
4. Regional dendrochronology of slope movements in the Bohemian Cretaceous Basin, one of Central European landslide-prone areas
Core Problem: Regional landslide hazard assessment needs long reactivation chronologies beyond sparse instrumental records.
Key Innovation: Tree-ring dating across 12 landslides and 366 disturbed spruce trees reconstructs 20 reactivation events over roughly 70 years in the Bohemian Cretaceous Basin.
5. Landslide Detection and Mapping Using Deep Learning Across Multi-Source Satellite Data and Geographic Regions
Core Problem: Automated landslide mapping must generalize across geographic regions and sensor types rather than rely on a single optical image source.
Key Innovation: Sentinel-2 multispectral imagery, ALOS PALSAR-derived slope, and DEM layers are integrated with U-Net, DeepLabV3+, and ResNet-style models for cross-regional landslide detection.
6. Adaptive Excavation Slope Model Updating Framework Accelerated by Pareto-Guided Active Learning
Core Problem: Slope digital twins become unreliable when excavation-induced parameter changes cannot be synchronized quickly with monitoring data.
Key Innovation: Gaussian-process regression, adaptive optimization, and Pareto-guided active learning accelerate reservoir-slope model updating for security assessment during construction.
7. Probabilistic three-dimensional seismic slope stability analysis using Krill Herd Algorithm-optimized Random Forest networks and modified pseudo-dynamic methods
Core Problem: Seismic slope assessment must account for stochastic ground motion, wave propagation, damping, and spatially variable soil properties.
Key Innovation: A modified pseudo-dynamic method, 3D limit analysis, and Krill-Herd-optimized Random Forest metamodel deliver rapid probabilistic slope safety evaluation.
8. Rainfall-induced stability of geobag reinforced expansive soil slopes: DEM investigation and simplified stability analysis
Core Problem: Geobag-reinforced expansive soil slopes are vulnerable to rainfall-driven swelling and strength degradation that are difficult to resolve at particle scale.
Key Innovation: MatDEM simulations and simplified stability analysis connect rainfall infiltration, surcharge loading, swelling, and geobag reinforcement mechanisms.
9. Investigation on the Crack Propagation of Edge-Fissured Slope-Shaped Specimens: Coupling Effects of Bedding Planes and Edge Fissures
Core Problem: Bedding and edge fissures jointly control crack propagation and stability in slope-shaped rock masses.
Key Innovation: 3D sand printing, DIC full-field strain monitoring, acoustic-emission analysis, and PFC modelling reveal crack paths and force-chain evolution across bedding dips.
10. Mountainous Urban Development Unintentionally Reshapes Spatial Patterns of Flood Disaster Risk
Core Problem: Mountainous cities can shift high-value development into floodplains when terrain constrains expansion.
Key Innovation: Field surveys, Telemac-2D hydraulics, and LightGBM-SHAP attribution quantify how urban growth, drainage density, vegetation, and topography reshape flood exposure from 2000 to 2020.
11. HaorFloodAlert: Deseasonalized ML Ensemble for 72-Hour Flood Prediction in Bangladesh Haor Wetlands
Core Problem: Bangladesh haor wetlands experience rapid backwater floods that riverine warning systems miss.
Key Innovation: A deseasonalized RF-XGBoost ensemble combines local hydrometeorology with an upstream Sentinel-1 SAR river proxy to forecast 72-hour flood probability.
12. The impact of spatial resolution on hourly flood modeling in large watersheds
Core Problem: Large-basin flood forecasts lack clear guidance on spatial resolution for both outlet and internal station accuracy.
Key Innovation: Five model resolutions in the 157,041 km2 Jialing River Basin and XGBoost sensitivity analysis show where finer grids improve hourly flood simulation.
13. Degradation-induced disaster mechanisms and prevention–control technologies in high-geothermal tunnels: a state-of-the-art review
Core Problem: High-geothermal tunnels face coupled thermal, hydraulic, mechanical, and support-degradation hazards during construction and operation.
Key Innovation: A state-of-the-art synthesis classifies thermal hazards and links high-geothermal conditions to deformation, support instability, water inrush, monitoring, and early-warning technologies.
14. From flow evolution to topographic erosion: insights of the Zelongnong valley rock-ice avalanche
Core Problem: Glacier retreat and permafrost degradation are increasing catastrophic high-mountain mass-flow cascades.
Key Innovation: Field investigation, UAV surveys, satellite imagery, and deposit analysis reconstruct the Zelongnong rock-ice avalanche transformation into a debris flow and quantify channel-topography-driven erosion.
15. Impact of Temporal and Spatial Resolution in Slope-Plant-Atmosphere Interaction Modelling
Core Problem: Hydro-mechanical slope models are sensitive to atmospheric and vegetation boundary conditions, yet temporal and spatial simplifications remain common.
Key Innovation: Fully coupled analyses of a representative cut slope compare daily versus monthly weather, dynamic versus static vegetation growth, and depth-distributed transpiration.
16. Evolution Characteristics of Overlying Strata Caving and Failure Under Sublevel Caving Mining: A Field Monitoring Study
Core Problem: Sublevel caving requires real-time knowledge of overlying caving and fractured zones to control ground pressure.
Key Innovation: Microseismic monitoring, roadway observations, and borehole data reconstruct caving-zone and fracture-zone evolution at Dahongshan Iron Mine.
17. Escalating wildfire risks in Brazilian biomes under climate change scenarios
Core Problem: Climate change may sharply alter wildfire risk across the Amazon, Cerrado, and other Brazilian biomes.
Key Innovation: Geoprocessing, land-use, vegetation, climate, topographic data, and SSP projections map national-scale future wildfire risk.
18. Detection of atmospheric rivers affecting the western Mediterranean and producing extreme rainfall over northern-central Italy
Core Problem: Western Mediterranean atmospheric rivers can focus moisture transport into northern Italy and the Alpine chain, producing severe rainfall.
Key Innovation: An adapted detection method identifies atmospheric rivers affecting the western Mediterranean and links them to extreme rainfall episodes over northern-central Italy.
19. Interpretable soil moisture prediction with a knowledge-guided deep learning approach
Core Problem: Multi-depth soil moisture prediction is difficult because vertical water transport is nonlinear and boundary conditions vary strongly.
Key Innovation: Non-local neural networks with knowledge-guided operations forecast simultaneous multi-depth soil moisture while preserving vertical physical relationships.
20. ASTRA: AMSR2-to-SMAP Image-to-Image TRAnslation for soil moisture conversion and radiative transfer model-based interpretation
Core Problem: AMSR2 soil-moisture retrievals suffer from vegetation attenuation and fixed global parameters relative to SMAP L-band observations.
Key Innovation: ASTRA uses Swin-Unet image-to-image translation and radiative-transfer interpretation to convert AMSR2 observations into SMAP-consistent soil moisture.
21. Mitigating the adverse impact of heat advection by water flow on embankment built on permafrost: using the Peclet and Nusselt numbers as key design parameters
Core Problem: Transport embankments can redirect surface and subsurface water, advect heat into permafrost, and destabilize the ground.
Key Innovation: Field observations support Peclet- and Nusselt-number drainage criteria for reducing water-flow heat advection around embankments.
22. Joint behavior and damage evolution of subsea shield tunnels with staggered assembly under three-dimensional seismic excitation
Core Problem: Subsea shield tunnels are exposed to marine soils, seawater interaction, and three-dimensional seismic loading.
Key Innovation: Finite-element analyses of the Shantou Bay Tunnel combine nonlinear marine soils, coupled acoustic-structure seawater effects, and staggered segmental assembly.
23. Explainable mechanics-guided machine learning-based modeling of seismic performance of rocking foundations
Core Problem: Rocking-foundation design needs fast models that retain mechanics rather than purely empirical predictions.
Key Innovation: OpenSees simulations of shaking-table and centrifuge tests are embedded as features in explainable ML models for energy dissipation and acceleration amplification.
24. Hazards and Resilience of Subsea Telecommunications Connections for Small Islands
Core Problem: Small islands depend heavily on subsea cables but face nearshore environmental and anthropogenic hazards that global assessments often miss.
Key Innovation: A 40-year database of 5,113 cable faults and spatial hazard exposure analysis quantify resilience constraints for 24 island and island-group settings.
25. Vulnerability to floods as a decision-making tool in coastal urban zones
Core Problem: Coastal cities need flood-vulnerability maps that combine rainfall, tide exposure, elevation, infrastructure, and social sensitivity.
Key Innovation: AHP, GIS, census indicators, and a Delphi urgency-feasibility matrix produce an FVI-based risk-management prioritization for Guayaquil.
26. Cracks in the Foundation: A Civil Infrastructure Dataset to Challenge Vision Foundation Models
Core Problem: Pixel-level defect segmentation is limited by scarce expert annotations and strong shape-dependence in texture-poor infrastructure materials.
Key Innovation: CiF introduces a large expert-annotated civil-infrastructure defect dataset for testing vision foundation models on crack and defect segmentation.
27. Improving precipitation interpolation using anisotropic variograms derived from convection-permitting regional climate model simulations
Core Problem: Daily precipitation covariance is hard to estimate in regions with sparse gauges or limited radar coverage.
Key Innovation: Convection-permitting regional climate simulations are used to derive anisotropic variograms for improved precipitation interpolation.
28. Characterizing Stratiform and Convective Precipitation Based on Multi-Source Observations in South Coastal China During 2022–2023
Core Problem: Typhoon-prone South China needs event-scale precipitation microphysics to interpret extreme rainfall generation.
Key Innovation: Parsivel disdrometer and Micro Rain Radar observations in Zhuhai separate stratiform and convective rainfall structure during 2022-2023.
29. PolAlpha-Net: A Lithology Segmentation Network Guided by Polarimetric Decomposition Knowledge for Dual-Polarization SAR Images in Vegetated Areas
Core Problem: Lithology mapping under vegetation remains difficult for optical remote sensing and conventional SAR classifiers.
Key Innovation: PolAlpha-Net injects polarimetric decomposition knowledge into dual-polarization ALOS-1 SAR deep segmentation for vegetated lithological mapping.
30. GeoMamba: A Geometry-driven MambaVision Framework and Dataset for Fine-grained Optical-SAR Object Retrieval
Core Problem: Optical-SAR retrieval is hindered by modality gaps, speckle noise, and unaligned samples.
Key Innovation: GeoMamba adds geometric feature guidance to a MambaVision retrieval framework and dataset for fine-grained optical-SAR object retrieval.
31. Shaking table test and numerical simulation of dynamic interaction for single pile-soil-structure model
Core Problem: Pile foundations in non-bedrock sites can fail under large-stiffness superstructures during earthquakes.
Key Innovation: Shaking-table tests and numerical simulation reveal inertia interaction, stress response, and seismic damage mechanisms in a pile-soil-structure system.
32. Mechanical degradation and acoustic emission in coal–rock under hydrochemical scouring-dissolution: damage and fracture mechanism
Core Problem: Deep coal-rock roadways are weakened by hydrochemical scouring and dissolution under complex groundwater conditions.
Key Innovation: Compression tests, DIC, NMR, SEM-EDS, XRD, and acoustic emission quantify mechanical degradation and fracture mechanisms under varied chemistry and flow.
33. Effects of Flow Intensity and Boundary Conditions on Transport and Clogging of Coal Fines in Fracture Networks: Insights from CFD–DEM Simulations
Core Problem: Coal fines transported through fracture networks can clog flow paths and alter permeability under different hydraulic boundary conditions.
Key Innovation: A semi-resolved CFD-DEM framework with DLVO adhesion simulates fines transport and clogging in a realistic heterogeneous fracture network.
34. Stress Transfer Around Cross Passages in Shallow Tunnels: Effects of Volume Loss and Cross Passage Shape
Core Problem: Cross-passage excavation changes stress redistribution and deformation in shallow tunnel linings.
Key Innovation: Three-dimensional finite-element analyses quantify how volume loss, cross-passage geometry, and soil density control hoop and longitudinal stress transfer.
35. Reliability Analysis of Shallow Foundations under Undrained Conditions Using Random Finite Element Limit Analysis and LRFD
Core Problem: Foundation design under undrained conditions needs reliability estimates that include spatial variability in shear strength.
Key Innovation: Random finite-element limit analysis and LRFD scenarios estimate failure probability across load ratios, correlation lengths, and strength variability.
36. Improved downhole seismic data processing through seismic attributes and joint analysis with HVSR: enhanced trace interpretability and extended investigation depth
Core Problem: Downhole seismic traces can be misinterpreted when complex propagation and scattering obscure shear-wave arrivals.
Key Innovation: Seismic attributes and joint HVSR analysis improve phase-arrival recognition and extend investigation depth in downhole surveys.
37. A streamlined yet rigorous CQC response spectrum formulation for SSI systems with viscous-spring boundaries under spatially variable seismic excitations
Core Problem: Multi-support response-spectrum methods for SSI systems are often too complex for practical viscous-boundary applications.
Key Innovation: A hybrid-excitation multi-support response-spectrum method derives a streamlined CQC formulation for viscous-spring boundary SSI systems.
38. Mechanical behavior and failure mechanisms of a novel composite DDCI-inclined bolt joint in large-diameter shield tunnels
Core Problem: Segmental joints weaken large-diameter shield tunnels as diameter and segment thickness increase.
Key Innovation: A composite DDCI-inclined bolt joint is evaluated with refined modelling based on the Haitai Yangtze River Tunnel.
39. Large-Eddy Simulation of non-Newtonian fluid flow in fracture networks
Core Problem: Fracture-flow models often assume laminar Newtonian behavior and miss inertia-rheology interactions.
Key Innovation: Large-eddy simulation resolves coupled inertial and non-Newtonian effects in field-based fracture networks.
40. Research on the Interaction Laws between Hydraulic Fractures and Bedding Planes in Heterogeneous Reservoirs under Various Combined Conditions Based on the Phase-Field Method
Core Problem: Hydraulic fractures interact with bedding planes through coupled mechanical, geological, and treatment conditions.
Key Innovation: A hydro-mechanical phase-field model evaluates fracture-bedding interaction under combined heterogeneity and bedding scenarios.
41. Hydro-mechanical modeling of gas–water two-phase migration in fractured shale reservoirs by combining discrete hydraulic fractures with multi-continuum representation
Core Problem: Gas-water migration after fracturing is controlled by multiscale matrix, natural fracture, and hydraulic fracture systems.
Key Innovation: A discrete-continuum framework explicitly models hydraulic fractures while representing natural fractures and matrix pores with multi-continuum mechanics.
42. Theoretical prediction method for settlement of existing tunnel induced by undercrossing
Core Problem: New undercrossing tunnels can disturb existing overlying tunnels and threaten structural integrity.
Key Innovation: A Hamilton-principle beam model and Rayleigh-Ritz solution predict settlement while accounting for axial force and boundary conditions.
43. A viscoelastic-plastic composite lining tunnel mechanical analysis model that considers excavation stress release and stratum rheology
Core Problem: Weak strata can produce support failure through excavation stress release, creep, and plastic rock behavior.
Key Innovation: A viscoelastic-plastic composite-lining analytical model combines virtual support pressure, Maxwell and Kelvin-Voigt creep, and Hoek-Brown yielding.
44. Deflection of foundation pit retaining wall in heterogeneous soil strata: Intelligent prediction methods based on physics-informed machine learning
Core Problem: Foundation-pit wall deflection threatens adjacent built environments and is difficult to predict in three dimensions.
Key Innovation: A derived PDE constrains PINN and data-driven PINN models for retaining-wall deflection in heterogeneous soil strata.
45. Evaluating the effects of grid hardening strategies on weather-driven power outage risk: A case study over connecticut
Core Problem: Power-outage risk from severe weather is hard to extrapolate beyond short historical outage records.
Key Innovation: A framework combines validated data-driven impact models with parametric statistics to estimate recurrence intervals and grid-hardening risk reduction.
46. A Method for Land-Cover Classification of Fully Polarimetric SAR Images by Fusing LiteDSANet and Polarization Feature-Guided DenseCRF
Core Problem: PolSAR classification must balance noise robustness, spatial consistency, and lightweight model design.
Key Innovation: LiteDSANet and polarization-feature-guided DenseCRF improve fully polarimetric SAR land-cover classification.
47. Precipitation Forecasting for Hydrologic Modeling in West-Central Florida using Seasonal Climate Outlooks
Core Problem: NOAA seasonal precipitation outlooks need local-scale evaluation before use in basin hydrologic decisions.
Key Innovation: Localized 0.5-month lead outlooks are evaluated for Florida basins using hindcasts and observed precipitation.
48. Zeolite-based soil stabilization: A review
Core Problem: Soil stabilization needs lower-carbon binders that can also support environmental remediation.
Key Innovation: A review synthesizes zeolite, especially clinoptilolite, for cementitious soil improvement, ion exchange, and remediation functions.
49. Anti-seepage performance of polymer-enhanced three-layer cover system of landfill: Field monitoring and numerical modelling
Core Problem: Landfill cover systems must limit rainwater infiltration and leachate generation over field operating conditions.
Key Innovation: Field monitoring and numerical modelling compare waste-soil, sand-bentonite, and sand-polymer-bentonite cover layers at a leather-sludge dump site.
50. Long-term thermomechanical performance of energy piles in sand: A semi-analytical solution using interface constitutive model
Core Problem: Energy piles combine foundation and heat-exchange roles but cyclic temperature changes can affect long-term structural integrity.
Key Innovation: A semi-analytical load-transfer solution with cyclic interface constitutive laws models pile-side and pile-tip response in sand.