TerraMosaic Daily Digest: May 23, 2026
Daily Summary
May 23 emphasizes papers that quantify how climate and subsurface state variables convert into damage. The strongest contribution is the Nature Geoscience analysis of the 2025 Scottish megafire, which shows that severe burning of peatlands, not only vegetation combustion, dominated event-scale carbon release. A second cluster links climate forcing to water and slope processes: Central Texas rainfall was shaped by competing soil-moisture and Gulf SST controls; large floods exchanged more than 31% of basin water with floodplains in some networks; and glacier retreat in the Italian Alps produced proglacial lake growth, dead-ice subsidence, debris-flow activity, and divergent sediment-release potential.
The landslide-adjacent and geotechnical papers are unusually mechanistic. Gully-erosion AI is assessed at field scale through 158 studies, exposing the gap between high model skill and weak geomorphic grounding. Slope reliability is made computationally tractable by reducing both random-field and slip-surface dimensions; tunnel face collapse is identified in real time from TBM muck and machine signals; frozen soil-rock mixtures, ice-wedge loess, and shield-tunnel freezing are tested under cryogenic loading; and tunnel, bridge, pipeline, railway, and underground-lining studies translate seismic or construction disturbances into measurable failure modes. Remote sensing contributes through physical observables rather than generic classification alone: L-band InSAR feasibility for snow-water retrieval, LiDAR ground filtering for bare-earth terrain, SAR stereo topography, DSM-to-DTM reconstruction, wetland-change mapping, SAR coherence-change uncertainty, sea-ice concentration fusion, and daily reservoir-water-area inference.
Key Trends
The day's papers converge on five methodological moves: resolving pre-event fuel and moisture states, measuring cryosphere-driven weakening, isolating failure surfaces and early signals, retrieving physical state variables from remote sensing, and requiring AI outputs to remain transferable and physically interpretable.
- Climate extremes are treated as state-dependent physical events: the Scottish megafire, Central Texas storm, river-floodplain exchange, Amazon fire-deforestation evapotranspiration loss, pyrocumulonimbus smoke forcing, rip-current prediction, and hybrid coastal run-up experiments show that hazard magnitude depends on antecedent moisture, fuel condition, morphology, and feedback with the atmosphere or shoreline.
- Cold-region papers move from descriptive thaw response to testable hydro-mechanical parameters: frozen-soil isotope modelling, L-band InSAR SWE mapping, proglacial DEM differencing, frozen soil-rock triaxial tests, ice-wedge loess experiments, frozen-rock strength prediction, and artificial-ground-freezing models identify controls on runoff partitioning, weakening, and infrastructure deformation.
- Slope and underground risk studies emphasize failure surfaces, early signals, and coupled infrastructure response: representative slip-surface reliability, TBM face-collapse recognition, tunnel seismic analysis, bridge source-path-terrain response, twin-tunnel earthquake-train loading, loess tableland scattering, railway GPR defect detection, underground-lining hyperspectral inspection, and drilling-based in situ stress estimation convert subsurface behavior into actionable observables.
- Remote sensing is strongest where it retrieves a physical quantity: L-band InSAR snow water, LiDAR ground filtering, SAR stereo radargrammetry, DSM-to-DTM residual learning, SAR coherence deviation modelling, reservoir water-area inference, waterbody connectivity preservation, wetland time-series mapping, sea-ice concentration fusion, and salinization mapping produce variables that can enter hydrologic, geomorphic, or hazard models.
- AI methods are being judged by transferability and physical consistency: the gully-erosion review, RSCLIP, VG-DETR, AI fault prediction, multimodal water-quality forecasting, rock shear-strength prediction, rock CT mineral segmentation, DPM stratigraphic classification, physics-regularized compaction prediction, and acoustic-emission crushing analysis point toward interpretable controls, domain adaptation, and physically meaningful outputs.
Selected Papers
This issue contains 65 selected papers from 1,257 papers analyzed. The leading papers focus on quantified hazard mechanics: peatland carbon loss from the 2025 Scottish megafire, AI-based gully-erosion science, paraglacial proglacial adjustment with debris-flow and dead-ice signatures, isotope-constrained frozen-soil hydrology, representative-slip-surface slope reliability, real-time TBM face-collapse identification, frozen soil-rock mixture mechanics, Central Texas extreme rainfall, river-floodplain exchange during major floods, and rapid earthquake impact assessment. The broader set follows fire, water, ice, sediment, and infrastructure hazards through remote sensing and mechanics, including L-band InSAR snow-water feasibility, SAR and LiDAR terrain retrieval, wetland and sea-ice mapping, reservoir water-area inference, soil salinization, tunnel and bridge seismic response, liquefaction mitigation, railway subgrade defects, underground lining inspection, drilling-based stress estimation, groundwater contamination, biomineralization, and rock or soil damage under cyclic, thermal, chemical, and freeze-thaw forcing.
1. Widespread peat carbon losses driven by the 2025 Scottish megafire
Core Problem: Carbon-dense peatlands are increasingly exposed to severe fire under warmer and drier climates, but event-scale peat combustion and recovery times remain difficult to quantify.
Key Innovation: A Nature Geoscience study combines remote sensing, field measurements, and modelling to show that the 2025 Scottish megafire emitted 38,600 MgC, with peat combustion contributing nearly 85% of total emissions.
2. Gully erosion studies using artificial intelligence approaches: A systematic review
Core Problem: AI-based gully-erosion studies have expanded rapidly, but their physical interpretability, spatial transferability, and validation standards remain uneven.
Key Innovation: A systematic review of 158 studies quantifies the dominance of susceptibility mapping, highlights strong ensemble and semantic-segmentation performance, and identifies the need to couple AI with process-based geomorphology.
3. Assessing the paraglacial evolution magnitude affecting proglacial areas through multi-temporal digital models and historical archives
Core Problem: Proglacial landscapes are reorganizing rapidly as glaciers retreat, but long-term magnitudes of paraglacial adjustment are hard to measure consistently.
Key Innovation: Historical archives, field surveys, SfM, digital photogrammetry, DEM differencing, and point-cloud comparison quantify 150 years of Aurona and Leone glacier-margin evolution, including debris-flow tracks, lake formation, buried ice, and sediment-load contrasts.
4. Isotope-aided frozen soil hydrological modeling reveals freeze–thaw controls on runoff partitioning in a mountainous catchment of the upper Heihe River, China
Core Problem: Mountain runoff partitioning in frozen-soil catchments remains uncertain when isotope constraints and explicit freeze-thaw processes are treated separately.
Key Innovation: FLEXTopo-iso is extended into FLEXCryo-iso, showing that stable isotopes improve source separation while explicit frozen-soil physics is needed to reproduce runoff generation in the upper Heihe River.
5. A novel representative slip surface-based dual-dimensionality reduction strategy for slope reliability analysis with spatially variable soils
Core Problem: Slope reliability analysis with random fields is constrained by the simultaneous high dimensionality of soil variables and candidate slip surfaces.
Key Innovation: The RSS-SBR-MGPR framework identifies representative slip surfaces, reduces random-field inputs, augments training samples, and estimates slope failure probability with far fewer stability evaluations.
6. A real-time identification method for tunnel face collapse based on the TBM muck analysis system
Core Problem: TBM tunnel face collapse requires rapid recognition before visual inspection or delayed geological interpretation can protect construction crews and machinery.
Key Innovation: A TBM muck-analysis system integrates operating parameters, muck size and shape, muck mass flow, and rock-mass information in a two-layer cascade model for real-time face-collapse identification.
7. Mechanical behavior of unsaturated frozen soil-rock mixtures: Insights from cryogenic triaxial tests
Core Problem: Frozen soil-rock mixtures control the stability of cold-region structures and slopes, but their strength evolution depends jointly on temperature, rock content, confinement, and deformation stage.
Key Innovation: Drained cryogenic triaxial tests quantify stress-strain response, volumetric behavior, shear-strength envelopes, particle breakage, and critical-state tendencies across freezing temperatures and rock-block contents.
8. Influence of Surface Conditions on the 04 July 2025 Extreme Storms in Central Texas
Core Problem: The July 4, 2025 Central Texas flood-producing storm requires attribution of how antecedent soil moisture and Gulf sea-surface anomalies shaped rainfall intensity.
Key Innovation: Convection-permitting simulations show that wet antecedent soils enhanced storm rainfall, whereas warm Gulf SST anomalies suppressed rainfall through circulation changes, isolating competing surface controls on an extreme event.
9. River‐Floodplain Exchange Influences Basin‐Scale Mixing During Large Fluvial Flood Events
Core Problem: Floodplain inundation is locally understood but poorly constrained at network scale, limiting estimates of residence time, exchange flux, and biogeochemical processing during large floods.
Key Innovation: Machine learning predicts exchange duration and magnitude for more than 1.8 million U.S. river reaches and shows that over 31% of basin water can exchange with floodplains during large events.
10. DEVELOPING A RAPID EARTHQUAKE IMPACT ASSESSMENT PROCEDURE AT EUROPEAN SCALE
Core Problem: Europe lacks a harmonized rapid earthquake-impact workflow that can translate shaking into comparable residential damage and loss estimates across national systems.
Key Innovation: A continental procedure links ShakeMapEU intensity estimates with a spatially disaggregated ESRM20 exposure database and EMS-98 vulnerability classes for rapid post-event impact assessment.
11. Evapotranspiration declines prolonged by deforestation and fire in South American biomes
Core Problem: Land-atmosphere coupling after deforestation, fire, and drought is difficult to partition across South American biomes.
Key Innovation: A high-resolution data-constrained hydrological model shows that deforestation-induced evapotranspiration declines persist 21-22% longer than fire or drought alone, while compound disturbance intensifies and prolongs losses.
12. Feasibility Mapping of L‐Band InSAR for SWE Retrievals Across the Western United States
Core Problem: Future L-band InSAR snow retrievals need spatially explicit feasibility estimates because forest cover, wet snow, and incidence angle constrain sensitivity.
Key Innovation: A western U.S. feasibility map quantifies where L-band InSAR can retrieve SWE change through the snow season and provides mission guidance for NISAR-era snow hydrology.
13. Does Geocentric Sea‐Level Rise in the Maritime Continent Reveal a Tectonic Fingerprint?
Core Problem: Maritime Continent sea-level trends are commonly separated into oceanographic and vertical-land-motion terms, but tectonic deformation may also affect geocentric sea level.
Key Innovation: Satellite sea level, GRACE geoid change, and process decomposition identify an unexplained trend aligned with the Sumatra-Andaman rupture zone, suggesting a measurable tectonic signal.
14. Large Particle Size and Thick Coating Influence Pyrocumulonimbus Smoke Radiative Forcing and Stratospheric Warming: Insights From the 2019–2020 Australian Megafires
Core Problem: The climatic influence of pyrocumulonimbus smoke depends on particle size and coating, which are often poorly represented in radiative calculations.
Key Innovation: Core-shell Mie calculations coupled to transport and radiative-transfer models show that large coated smoke particles from the 2019-2020 Australian megafires better explain observed forcing and stratospheric warming.
15. Shifting Controls on Erosion‐Induced Soil Carbon Loss in the Yellow River Basin
Core Problem: Soil erosion can export large soil-organic-carbon stocks, but long-term controls shift when check dams and vegetation recovery reconfigure sediment pathways.
Key Innovation: A Yellow River Basin analysis identifies changing controls on eroded SOC yield and clarifies how conservation infrastructure and revegetation alter carbon-loss pathways.
16. GRACE TWSA Data Assimilation in Land Surface Models‐The Role of State Vector and Model Uncertainty Representation
Core Problem: GRACE/GRACE-FO data assimilation frameworks differ in state vectors and uncertainty representation, obscuring why storage increments differ among land-surface models.
Key Innovation: A controlled evaluation in the Community Land Model tests multiple state-vector configurations and ensemble perturbation methods, clarifying how TWS observations are translated into vertical storage updates.
17. Hardness Toxicity and Multi‐Route, Multi‐Demographic Health Risks of Nitrate‐Contaminated Groundwater: A Call for Sustainable Environmental and Public Health Protection
Core Problem: Groundwater nitrate and hardness hazards require multi-route and age-specific exposure analysis rather than single-threshold water-quality screening.
Key Innovation: Physicochemical sampling, USEPA health-risk models, and ArcGIS mapping identify Ghanaian agricultural hotspots with nitrate up to 720 mg/L and high oral hazard quotients for infants.
18. Machine learning and pre-simulation models for rip current prediction at Duck, NC: comparison and integration
Core Problem: Rip-current warnings need models that balance local morphology, wave forcing, transferability, and sparse lifeguard observations.
Key Innovation: A Duck, North Carolina comparison evaluates NOAA logistic regression, pre-simulation hydrodynamic predictors, and integrated machine-learning approaches for operational rip-current likelihood prediction.
19. Experimental assessment of hybrid solutions for coastal protection and existing wave run-up prediction formulations
Core Problem: Wave run-up formulas developed for hard structures are often applied to hybrid saltmarsh-grey systems without adequate validation.
Key Innovation: Full-scale experiments across 12 hybrid protection typologies and 340 irregular-wave tests quantify how artificial saltmarsh elements alter run-up exceedance behaviour.
20. Biogeodynamics-Ice sheet-Geneva-MITgcm (BIG-MITgcm, v1.0): a simulation tool for exploring climate states with a representation of global ice sheets
Core Problem: Multi-millennial climate simulations require ice sheets, vegetation, deep ocean, and atmosphere to evolve together without the cost of full Earth-system models.
Key Innovation: BIG-MITgcm v1.0 couples biogeodynamics, ice sheets, and MITgcm to explore long-timescale climate attractors and nonlinear cryosphere feedbacks.
21. Derivation and validation of estimation model of rainfall kinetic energy under the canopy
Core Problem: Soil-erosion models need under-canopy raindrop kinetic energy, but drip and splash components are rarely parameterized from canopy structure.
Key Innovation: A canopy-stratified model validated against nine LiDAR and raindrop-spectrum datasets estimates dripping and splashing kinetic energy under different vegetation structures.
22. Predicting shear strength of quartzite using portable index tests: a machine learning approach
Core Problem: Rock slope and underground design require cohesion and friction angle, yet direct shear-strength testing is costly and slow.
Key Innovation: PSO-SVM, DBO-Random Forest, and ADAM-FNN models predict quartzite friction angle and cohesion from portable point-load, Schmidt hammer, and S-wave velocity indices.
23. AI fault prediction for ultra-deepwater OBN seismic data: advancing structural interpretation in carbonate reservoirs
Core Problem: Subtle fault networks in deepwater carbonate reservoirs are difficult to image with conventional seismic interpretation under noise and complex topography.
Key Innovation: An enhanced HR-Net workflow using OBN seismic data, structure-oriented filtering, skeletonization, and ant tracking automates high-resolution 3D fault prediction.
24. A Gradient-Based Index for Multiscale Mapping of Land Degradation in Brazil
Core Problem: Trend-based land-degradation categories can miss cumulative severity and misclassify degraded areas as stable.
Key Innovation: A 500 m Land Degradation Index integrates deforestation age, gross primary productivity, and soil organic carbon to map Brazil's degradation gradient from 2001 to 2021.
25. Stereo Radargrammetry Using Deep Learning-Based Image Matching with Fine-Tuned Model on Synthetic Aperture Radar Images
Core Problem: Stereo SAR radargrammetry struggles in mountainous and vegetated terrain because template matching fails under SAR-specific geometric modulation.
Key Innovation: A transformer-based RoMa matching workflow is fine-tuned on automatically generated SAR patches and reference DSM projections to improve dense 3D topographic reconstruction.
26. DSM-to-DTM Reconstruction Using Only DSM-Derived Inputs with Residual Learning and CSF Priors
Core Problem: Global DSM products retain canopy and above-ground signals that bias hydrologic and geomorphic terrain analyses.
Key Innovation: A residual-learning framework predicts DSM-DTM differences using DSM-derived terrain features and CSF priors, reconstructing bare-earth DTM without external inputs at inference time.
27. A Terrain-Feature-Aware Multiscale PTD for Airborne LiDAR Ground Filtering
Core Problem: Conventional progressive TIN densification can smooth fine terrain features and degrade DTM quality in complex terrain.
Key Innovation: A multiscale PTD method uses terrain-feature-aware seed selection, adaptive thresholds, and geometry-guided filtering to preserve fine-scale ground morphology for downstream hazard analysis.
28. RSCLIP for Training-Free Open-Vocabulary Remote Sensing Image Semantic Segmentation
Core Problem: Open-vocabulary segmentation is promising for remote sensing but struggles with local detail and object-scale variation in aerial and satellite images.
Key Innovation: RSCLIP adapts CLIP for training-free remote-sensing segmentation through neighbour-aware patches, semantic-correlation enhancement, and multihead multiscale attention.
29. VFM-Guided Semi-Supervised Detection Transformer Under Source-Free Constraints for Remote Sensing Object Detection
Core Problem: Source-free remote-sensing object detection often collapses under noisy pseudo-labels when original source data cannot be shared.
Key Innovation: VG-DETR uses vision-foundation-model guidance inside a semi-supervised detection transformer to stabilize cross-domain adaptation with limited target-domain labels.
30. Spectrally Derived Soil Salinization Information Extraction and Analysis of Driving Factors: A Case Study of Zhanhua District, Yellow River Delta
Core Problem: Long-term salinization mapping must connect spectral indicators with interpretable drivers and future climate scenarios.
Key Innovation: Remote-sensing salinity indices, XGBoost-SHAP attribution, and CMIP6 scenarios map 1993-2023 salinization dynamics in the Yellow River Delta and project future trajectories.
31. Deviation-guided dual-model machine learning SAR coherence change detection framework for extracting changes in low-coherence areas
Core Problem: SAR coherence change detection often confuses real surface change with decorrelation from geometry, time, noise, and co-registration errors.
Key Innovation: A dual-model machine-learning framework estimates coherence, predicts estimation deviation, interprets uncertainty with SHAP, and improves change extraction in low-coherence areas.
32. Annual wetland changes across China (2016–2024) from multisource temporal data and knowledge-driven machine learning approach
Core Problem: High-resolution annual wetland change data remain scarce at national scale, limiting conservation planning and hydrologic accounting.
Key Innovation: A knowledge-driven machine-learning framework uses more than 2.59 million Sentinel images to produce 10 m annual wetland maps across China from 2016 to 2024.
33. Sea ice concentration estimation via physical information-guided multi-source data fusion and spatial continuity preservation
Core Problem: Sea-ice concentration retrievals often concatenate active and passive microwave features without preserving physical dependencies and spatial continuity.
Key Innovation: PIMS-Net decouples physically distinct sensor branches and enforces spatial continuity to improve multi-source sea-ice concentration estimation.
34. Daily inference of reservoir surface water area at large scales via a systematic deep learning framework synergistically integrating multisource remote sensing imagery
Core Problem: Single-sensor satellites cannot resolve daily reservoir water-area dynamics at both high spatial and temporal resolution.
Key Innovation: A deep-learning framework fuses Sentinel-1, MODIS, and Landsat-derived information to reconstruct daily cloud-free Landsat-like imagery and extract reservoir surface water area at large scale.
35. Changes in sediment transport capacity in the lower Yellow River: Analysis using an improved formula based on energy theory
Core Problem: Traditional sediment-transport formulas are poorly suited to the lower Yellow River after Xiaolangdi Reservoir altered flow-sediment conditions and channel morphology.
Key Innovation: An energy-based sediment transport capacity formula is recalibrated with post-2000 observations and validated against 2011-2022 data under low-sediment-flow regimes.
36. Precise waterbody mapping at the city scale using a novel connectivity preservation and boundary refinement network based on medium-resolution remote sensing imagery
Core Problem: Medium-resolution waterbody extraction at city scale suffers from broken connectivity and rough boundaries in complex scenes.
Key Innovation: CB-Net introduces dense dynamic snake convolution and boundary refinement to preserve waterbody geometry and connectivity from medium-resolution remote-sensing imagery.
37. Multimodal deep-learning–driven water-quality forecasting and scenario assessment: shifting from reactive monitoring to proactive basin management
Core Problem: River-water-quality forecasting must integrate static watershed attributes, dynamic hydroclimate drivers, land use, population, and intervention scenarios under sparse monitoring.
Key Innovation: A multimodal encoder-decoder framework trained on more than 4 million samples from 1801 Chinese stations forecasts key parameters and supports modular local fine-tuning with only 1% local data.
38. Efficient MLM-DRM framework for seismic performance analysis of 3D tunnels in complex geologic media
Core Problem: Long tunnels crossing stiffness contrasts experience spatially variable earthquake motions that are expensive to model at full scale.
Key Innovation: An efficient MLM-DRM framework represents asynchronous rupture and wave-passage effects for seismic performance analysis of 3D tunnels in complex geologic media.
39. Macro-micro responses of ice-wedge loess subjected to freeze-thaw cycles
Core Problem: Ice-wedge geometry may accelerate freeze-thaw weakening of fissured loess, but its macro- and micro-scale controls are poorly constrained.
Key Innovation: Triaxial tests and SEM show that freeze-thaw cycles reduce strength and that 60-degree ice wedges produce the strongest degradation in loess samples.
40. New strategy for determining the dynamic compressive strength of sandstone in cold regions: A practical interpretable ensemble model
Core Problem: Dynamic compressive strength of freeze-thawed rock is critical for cold-region rock stability but difficult to measure rapidly.
Key Innovation: An interpretable ensemble model combining ANN/RF with energy valley and spider wasp optimization predicts frozen-thawed sandstone dynamic compressive strength and identifies controlling variables.
41. Cryogenic freezing aging strength model coupled with stochastic medium theory for shield tunnelling undercrossing existing metro line
Core Problem: Ultra-close shield tunnelling beneath existing metro lines requires time-dependent frozen-soil strength and frost-heave prediction that conventional AGF design lacks.
Key Innovation: Low-temperature direct shear tests, a cryogenic aging strength model, and stochastic medium theory predict full-cycle deformation for a Suzhou metro undercrossing case.
42. Full-process seismic response analysis of long-span cable-stayed-suspension bridges considering source-path-terrain-structure effects
Core Problem: Long-span cable-stayed-suspension bridges require seismic assessment that includes rupture process, propagation path, and local topography simultaneously.
Key Innovation: A full-process simulation links kinematic hybrid source modelling, spectral-element wavefield propagation, station validation, and structural response analysis for railway bridge design.
43. Performance evaluation of natural rubber latex treatment across various sand types for seismic energy dissipation
Core Problem: Liquefaction mitigation must be evaluated across sand types using an energy framework rather than a single material response.
Key Innovation: Cyclic tests show how natural rubber latex treatment changes energy dissipation and cyclic resistance across river and clean sands, informing sand-specific liquefaction mitigation.
44. Dynamic response of twin-tunnel metro tunnel under the coupling effect of multi-directional earthquake and train loads
Core Problem: Metro tunnels may experience earthquakes during train operation, yet combined train and multidirectional seismic loading remains insufficiently characterized.
Key Innovation: Field vibration measurements and 3D numerical simulations quantify acceleration, stress, and displacement response of a Qingdao twin-tunnel system under train-only and coupled earthquake-train loading.
45. Scattering of near-source SH-waves by non-isosceles trapezoidal loess tableland: semi-analytical method and parametric study
Core Problem: Steep loess tablelands can amplify or shield near-source seismic waves, but irregular trapezoidal geometries are analytically difficult.
Key Innovation: A semi-analytical wave-function expansion with non-integer Bessel terms quantifies SH-wave scattering, amplification, and shielding for non-isosceles loess tablelands.
46. Automated railway subgrade defect detection using GPR and an enhanced YOLOv11 framework
Core Problem: Manual GPR interpretation limits operational detection of settlement, subsidence, mud pumping, and water accumulation in railway subgrades.
Key Innovation: YOLOv11-DSConv-CCA uses dynamic snake convolution and criss-cross attention to detect multiple GPR defect classes at 90 FPS on 25 km of in-service railway data.
47. Automated nondestructive evaluation of compressive strength of underground lining structure using hyperspectral imaging and deep neural networks
Core Problem: Underground concrete lining strength is difficult to map without destructive tests in confined tunnel and utility-corridor environments.
Key Innovation: Hyperspectral imaging and deep neural regression generate two-dimensional compressive-strength heatmaps and identify cracked, spalling, and leaking lining regions in field validation.
48. Effect of confining pressures on the drilling parameters and specific energy: Insights into determining in situ stress using drilling monitoring techniques
Core Problem: Conventional in situ stress tests occur after drilling and ignore mechanical information generated during the drilling process itself.
Key Innovation: Monitored thrust, rotation speed, torque, drilling speed, and specific energy from hydraulic-rotary drilling tests are linked to confining pressure and rock strength for stress estimation.
49. Sustainable engineering decisions toward nature-based solutions for flood protection infrastructure: An integrated system dynamics framework
Core Problem: Flood-protection upgrades are shaped by behavioural and organizational feedbacks that conventional technical decision models often ignore.
Key Innovation: A system-dynamics framework integrates value-belief-norm theory, innovation diffusion, interviews, and surveys to model adoption of nature-based flood-protection solutions.
50. Estimating whole-of-event consequences in interdependent infrastructure systems
Core Problem: Hazard assessments underestimate total consequences when they measure disruption only near the initial damage rather than over the full recovery trajectory.
Key Innovation: A time-dependent framework estimates service disruption and restoration across interdependent electricity, water, and wastewater networks using property-days without service and related metrics.
51. Post-Earthquake Functionality Assessment and Resilience Enhancement of Interdependent Electric Power and Mobile Communication Networks
Core Problem: Post-earthquake functionality depends on bidirectional infrastructure interdependence and emergency restoration priorities that are often omitted from single-network fragility models.
Key Innovation: A coupled power-communication model simulates Mw 6.0-7.0 scenarios, identifies power-driven communication failure propagation, and optimizes pre-disaster resilience investment.
52. Sulfate-reducing bacteria mediated mobilization and remediation of arsenic in groundwater: A critical review
Core Problem: Arsenic contamination in groundwater depends on microbial sulfur cycling, but mobilization and remediation mechanisms remain fragmented across hydrogeochemical contexts.
Key Innovation: An Earth-Science Reviews synthesis evaluates sulfate-reducing bacteria effects on arsenic release, immobilization, sulfide-mineral oxidation, and remediation pathways.
53. A review of biomineralization mechanisms and theories in MICP and EICP
Core Problem: Biomineralization-based ground improvement needs mechanisms that connect pore-scale precipitation, bacterial transport, crystal growth, and field-scale mechanical behaviour.
Key Innovation: A review organizes MICP/EICP mechanisms around microfluidic evidence, nucleation theory, particle-scale biocementation, and multiscale bio-chemo-hydro-mechanical upscaling.
54. Quantile and extreme threshold methods for geotechnical reliability-based design with inseparable parameters
Core Problem: Reliability-based geotechnical design struggles when design parameters are inseparable from performance functions.
Key Innovation: Vector-updated quantile and extreme threshold methods estimate design parameter thresholds in a synthesized variable space, delivering accurate RBD outcomes in seconds.
55. Spatial Variability in Trace Organic Compound Reactivity During Urban River Infiltration Into an Alluvial Aquifer
Core Problem: Trace organic compounds enter groundwater along losing urban rivers, but reactivity varies spatially along river-to-aquifer flowpaths.
Key Innovation: High-resolution chemical observations and convolution-based Monte Carlo travel-time modelling estimate spatial distributions of apparent removal rates for 33 trace organic compounds.
56. A dual-time-scale softening constitutive modeling and parameter inversion for frozen coarse-grained soil
Core Problem: Frozen coarse-grained soil exhibits nonlinear strain softening controlled by load transfer among sand, gravel, ice, and localized damage.
Key Innovation: A dual-time-scale constitutive model combines homogenization before peak strength with post-peak skeleton and fracture-friction domains for parameter inversion.
57. Mathematical modeling for interactions and transport of particulate bioavailable phosphorus with sediment dynamics
Core Problem: Particulate bioavailable phosphorus transport is difficult to represent because it couples hydrodynamics, sediment mineral properties, and watershed inputs.
Key Innovation: A hydrodynamic-sediment-phosphorus model linked with SWAT simulates PBAP transport in the Yarlung Tsangpo River and resolves spatiotemporal nutrient-sediment interactions.
58. A Curvature-Aware Enhancement and Geometrically Constrained Growing Method for Narrow River Extraction From TG2-InIRA Imagery
Core Problem: Narrow rivers are hard to extract from near-nadir InIRA imagery because speckle, weak contrast, and sinuous geometry obscure channel continuity.
Key Innovation: A curvature-aware framework combines structure-adaptive despeckling, multiscale feature enhancement, and geometrically constrained region growing to delineate fine river networks.
59. Enhancing Mineral Segmentation in Rock CT Images via Hessian-Guided Morphological Feature Learning with Deep Neural Networks
Core Problem: Rock CT mineral segmentation suffers from low-contrast boundaries and limited morphological awareness in deep learning models.
Key Innovation: A Hessian-guided segmentation framework adds eigenvalue-ratio regularization and Grayscale-Hessian inputs to improve boundary coherence and 3D mineral reconstruction.
60. Advancing drilling parameter reliability: A data-driven comparison of MWD and DPM for stratigraphic identification
Core Problem: MWD penetration-rate fluctuations can obscure stratigraphic boundaries and weaken mechanical interpretation during drilling.
Key Innovation: A comparison of measurement-while-drilling and drilling-process-monitoring data shows that SVM analysis of DPM parameters improves interface detection resolution by 20.57% and 38.01% in two drillholes.
61. Characterizing particle crushing behavior of dry granular soils by acoustic emission
Core Problem: Particle crushing during shearing is difficult to quantify continuously from grain-size data alone.
Key Innovation: Staged triaxial tests show that high-frequency acoustic emission hits correlate linearly with relative breakage and identify early crushing during yielding and peak-stress phases.
62. Anisotropic mechanical behaviour of Callovo-Oxfordian claystone under triaxial lateral unloading
Core Problem: Excavation-induced damage in Callovo-Oxfordian claystone is anisotropic even under near-isotropic drift stress states.
Key Innovation: Triaxial lateral-unloading tests at different bedding orientations quantify strength anisotropy, progressive damage, and time-dependent behaviour relevant to repository drift excavation.
63. Loading-unloading responses of non-persistent flawed granite subjected to cyclic disturbances
Core Problem: Flawed rock masses in underground construction accumulate damage under cyclic loading and unloading, but failure mode depends on stress path and flaw geometry.
Key Innovation: Triaxial tests and a fracture-mechanics model show that confining-pressure unloading strongly weakens flawed granite and shifts failure toward shear-tensile composite modes.
64. Impact of pharmaceutical (diclofenac) contamination on the compressibility and strength of residual soil: Towards a resilient urban infrastructure
Core Problem: Emerging pharmaceutical contaminants can alter geotechnical properties, yet their impact on residual soil compressibility and strength is poorly constrained.
Key Innovation: Consolidation, infusion, and direct-shear tests show how diclofenac concentration, cation type, clay mineralogy, and diffuse-double-layer effects change residual-soil volume and strength behaviour.
65. Physics-regularized neural networks for predicting soil compaction parameters
Core Problem: Machine-learning estimates of optimum moisture content and maximum dry unit weight can violate basic geotechnical constraints such as the zero-air-voids line.
Key Innovation: Physics-regularized neural networks couple OMC and MDUW predictors with geotechnical consistency penalties, reducing physically inadmissible soil-compaction predictions across external validation splits.