TerraMosaic Daily Digest: July 23, 2026
Daily Summary
Several studies resolve slope failure as an evolving sequence rather than a terminal displacement. At Illgraben, seismic-attribute-guided self-supervised learning extracts weak rockfall signals from noise and reconstructs a growth-acceleration-quiescence-burst sequence before collapse without site-specific retraining. At Casale-Colla, multitemporal imagery, UAV-LiDAR, GNSS, MT-InSAR, boreholes and geophysics delineate five interacting landslide sub-units and sliding surfaces to about 40 m depth. A coupled thermo-hydro-mechanical model of a Tibetan rock/ice avalanche links rainfall and rapid warming to destabilization, then reproduces the transition from avalanche to fluidized debris flow.
Volcanic and seismic monitoring increasingly targets low-amplitude signals that conventional thresholds suppress. SSTAR uses two decades of MODIS thermal infrared observations to recover diffuse anomalies of roughly 1 K over several square kilometres before multiple Shishaldin eruptions, while preserving causal checks against post-eruptive contamination. Sentinel-1 at Masaya resolves simultaneous deep-reservoir inflation and shallow-source deflation, and fourteen years of Whakaari sulfur-dioxide observations distinguish conduit sealing from deep supply. In Alaska, retrospective replay of 530 earthquakes exposes a depth-dependent limitation of the EPIC early-warning algorithm: its fixed 8 km source depth strongly underestimates magnitudes for earthquakes deeper than 100 km.
Persistent climate regimes and inherited urban structure reshape exposure well beyond the instant of an event. Fire-weather waves occupy only 4% of days but coincide with 26% of burned area and nearly half of the most energetic forest fires. High Mountain Asia is dominated by continental snow climate, yet its southeastern margins shifted toward warmer, wetter regimes during 1980-2018, altering the basis for avalanche susceptibility. Northeast Pacific marine heatwaves intensify landfalling atmospheric rivers thermodynamically even when synoptic circulation is held fixed. Across 202 US cities, historically redlined neighbourhoods carry a disproportionate flood burden in roughly three-quarters of cities; in China, terrain exposure and social capacity explain 43.7% and 53.0% of fatal-landslide impact inequality, respectively.
Key Trends
Five shifts connect adaptive runout simulation, observability-aware sensing, persistent hazard regimes, mechanics-constrained inference and explicit exposure.
- Weak precursors are reconstructed as sequences: Rockfall seismology, diffuse volcanic thermal anomalies and long-term degassing records identify ordered changes before failure or eruption instead of relying on a single threshold exceedance.
- Observation geometry becomes part of the inference: InSAR line-of-sight limits, DAS coupling, UAV viewing conditions and cloud-contaminated thermal records are modelled or cross-checked explicitly rather than treated as neutral measurements.
- Failure mechanics bridge initiation and runout: Bedding-controlled toppling, thermo-hydro-mechanical rock/ice failure, entrainment-aware material-point models and cross-scale SPH connect destabilization to impact.
- Persistent regimes replace isolated extremes: Fire-weather waves, snow-climate zones, compound drought states and ocean thermal anomalies reveal cumulative forcing that event maxima and annual means obscure.
- Exposure retains social and infrastructural history: Fatal-landslide inequality, redlining-era flood exposure, building flammability and infrastructure fragility show that comparable physical hazards do not produce comparable consequences.
Selected Papers
This issue is led by seismic deep learning that reconstructs the rockfall sequence preceding rockslide collapse, an open satellite framework for subtle volcanic thermal precursors, a coupled simulation of climate-forced rock/ice-avalanche failure and runout, and complementary evidence that fatal-landslide and urban-flood impacts inherit social inequality. Global fire-weather waves, changing snow climates in High Mountain Asia and marine-heatwave amplification of atmospheric rivers extend the same emphasis from short-lived events to persistent hazard regimes.
1. Fire weather waves drive extreme fires globally
Core Problem: Short-lived fire-weather extremes do not explain why a small fraction of days produces a disproportionate share of the world’s most destructive forest fires.
Key Innovation: The study defines persistent fire-weather waves and resolves their global fire response: 4% of days coincide with 26% of burned area and nearly half of the top 1% most energetic forest fires.
2. Unexpected spatial inequality of fatal landslide impacts
Core Problem: Fatality totals and rates reveal where landslides kill but not why their human consequences are distributed unequally among places.
Key Innovation: County-level partitioning of China’s fatal-landslide database attributes 43.7% of spatial inequality to terrain exposure and 53.0% to social capacity, with a sharp nonlinear rise above 1,950 people per square kilometre.
3. Disproportionate exposure to extreme floods in historically redlined U.S. communities
Core Problem: Flood maps do not reveal whether exposure is distributed systematically along the boundaries created by historic housing policy.
Key Innovation: High-resolution flood modelling and dasymetric population mapping across 202 US cities show that D-graded neighbourhoods are disproportionately exposed in roughly three-quarters of cities, even away from major rivers and coasts.
4. Seismic Insights Into the Role of Rockfall in Rockslide Destruction Processes
Core Problem: Weak rockfall signals are difficult to separate from continuous seismic noise, leaving the activity sequence that precedes rockslide collapse poorly resolved.
Key Innovation: Seismic-attribute-guided self-supervised learning treats detection as time-series segmentation and, without retraining, reconstructs an Illgraben growth-acceleration-quiescence-burst sequence absent from the existing catalogue.
5. SSTAR: a user-friendly framework for detecting and monitoring subtle thermal precursors to volcanic eruptions—application to Shishaldin, Alaska
Core Problem: Diffuse volcanic warming can precede eruption by months to years but remains below the hotspot thresholds used by conventional satellite systems.
Key Innovation: SSTAR processes two decades of MODIS thermal infrared data with reference-area correction and complementary causal filters, recovering roughly 1 K spatially coherent anomalies before multiple Shishaldin eruptions and releasing the workflow for near-real-time use.
6. Reinterpreting the Recurrence Time–Moment Relation of Global Repeating Earthquakes: A Scattered Data Cloud Bounded by Crack‐Model Physics
Core Problem: A single global power-law fit to repeating-earthquake recurrence times conflicts with classical crack-model scaling and ignores the catalogue’s large intrinsic scatter.
Key Innovation: Expanded catalogues are treated as a bounded distribution rather than a regression line; after repeater selection and creep-rate correction, a robust one-third-slope lower envelope recovers the steady-creep crack-model limit.
7. Double Trouble: Two Active Magma Reservoirs Identified at Masaya Volcano, Nicaragua, Using Satellite Geodesy
Core Problem: Masaya’s recent deformation cannot be explained by one source because the caldera interior and Santiago crater move in opposite directions.
Key Innovation: A 2018-2024 Sentinel-1 time series and geodetic inversion resolve concurrent inflation of a reservoir roughly 3.2 km deep and deflation of a second source roughly 200 m beneath the active crater.
8. Explaining building damage from wildfires in California
Core Problem: Wildfire building-loss models often mix environmental exposure, weather and construction effects without testing their separate causal contributions or spatial transfer.
Key Innovation: More than 100,000 California inspections, a composite flammability rating and a 200 m spatial dead zone quantify the incremental value of buildings and weather, with the full model reaching 88% accuracy.
9. A GPU-accelerated adaptive active-grid MPM for high-resolution 3D modelling of large-scale landslides: Case study of the 2025 Blatten ice-rock avalanche
Core Problem: High-resolution three-dimensional simulation of large avalanches is limited by fixed-grid memory and repeated particle-grid transfers over mostly inactive terrain.
Key Innovation: An adaptive active-grid GPU MPM activates only blocks around the moving mass and uses shared-memory transfer, reproducing the main runout features of the 9.3-million-cubic-metre 2025 Blatten avalanche.
10. InSAR Techniques for Landslide Study: A Review of Methods, Challenges, and Emerging Trends
Core Problem: Landslide InSAR methods have proliferated, but their operating limits and progression toward intelligent, multisource monitoring remain fragmented across technique-specific literature.
Key Innovation: The review organizes differential, persistent-scatterer, small-baseline, distributed-scatterer and derivative InSAR methods around applicability, geometric distortion, decorrelation, atmospheric delay and three-dimensional recovery.
11. A deep-seated roto-translational landslide in flysch rocks in the Northern Apennines (the Casale-Colla landslide, Ferriere, Italy)
Core Problem: The geometry and internal kinematics of the active Casale-Colla deep-seated landslide are too complex for a single surface-displacement product to resolve.
Key Innovation: Multitemporal imagery, UAV-LiDAR, GNSS, MT-InSAR, boreholes and geophysics delineate five interacting sub-units, sliding surfaces to about 40 m depth and a downslope velocity decrease consistent with stacked overthrusting.
12. A multi-source monitoring framework for mining-induced geohazards integrating time-series InSAR and microseismic observations
Core Problem: Mining deformation, subsurface fracture activity and operational motion are usually monitored separately, obscuring how surface and underground evidence jointly diagnose emerging geohazards.
Key Innovation: Time-series InSAR and microseismic observations are co-interpreted in a multi-source framework that connects deformation evolution with the location and timing of mining-induced seismic activity.
13. A study of rainfall-flood compound risk in Wuchang Historic City based on different climate scenarios
Core Problem: Flood risk to historic urban fabric cannot be projected from rainfall alone because development pathways alter drainage, land cover and the location of irreplaceable heritage.
Key Innovation: CMIP6 scenarios, FLUS land-use simulation and MIKE-SHE hydrology are coupled with heritage locations to resolve scenario-dependent rainstorm-flood risk across Wuchang Historic City.
14. Evaluation of ERA5-Land Against Ground-Based Precipitation Observations in Supporting Landslide Analysis in Pannonian Croatia
Core Problem: Open reanalysis precipitation is attractive for regional landslide analysis, but its ability to reproduce the local rainfall that triggers failures must be tested against gauges.
Key Innovation: ERA5-Land daily precipitation is evaluated against ground observations in Pannonian Croatia with explicit attention to the timing and magnitude of landslide-triggering rainfall.
15. Exploring the Use of the EPIC Earthquake Early Warning Algorithm in Alaska
Core Problem: An early-warning algorithm tuned to shallow West Coast earthquakes may fail systematically in Alaska's sparse network and deep subduction-zone seismicity.
Key Innovation: Simulated real-time replay of 530 earthquakes shows 345 successful detections and isolates a fixed 8 km source-depth assumption as the main cause of strong magnitude underestimation for events deeper than 100 km.
16. Numerical simulation of the stability and dynamics of a glacier under climatic forcing: a case study from Tibet, China
Core Problem: The climatic path from glacier retreat to bedrock-slope failure, and the subsequent transformation of rock and ice into debris flow, is rarely represented in one quantitative model.
Key Innovation: A coupled thermo-hydro-mechanical and runout framework links rainfall and rapid warming to interface failure, then reproduces a roughly 120 s avalanche-to-debris-flow cascade with a 59 m/s peak front speed.
17. Rapid Post-earthquake Damage Detection Using UAV Imagery and YOLO: A Case Study from the Kahramanmaras Earthquake
Core Problem: Post-earthquake field inspection cannot rapidly map distributed building damage over a large urban footprint.
Key Innovation: UAV imagery and a YOLO detector are combined for rapid damage localization in the 2023 Kahramanmaras earthquake case, translating overhead imagery into building-scale screening.
18. The Use of Geophysical Surveys in the Study of a Landslide-Prone Area near the Village of Dolan in the Almaty Region of Kazakhstan
Core Problem: Individual geophysical surveys can give non-unique interpretations of a landslide-prone slope where the body and potential failure zones are poorly constrained.
Key Innovation: Electrical, seismic and magnetic surveys are interpreted together so geoelectric and velocity structure provide a mutually constrained assessment of the Dolan slope.
19. Toppling deformation mechanisms in layered rock slopes controlled by bedding dip angle
Core Problem: Layered rock slopes topple through different crack paths, but the governing role of bedding dip in the transition from local deformation to global instability is unresolved.
Key Innovation: Base-friction tests, numerical simulation and image tracking show toe-upward cracking at 45-60 degrees, crest-downward cracking at 75 degrees and progressively deeper, wider toppling as bedding steepens.
20. Towards Transparent Geohazard Model: XAI for Ground Deformation Susceptibility in Rhenish Coalfields, Germany
Core Problem: Ground-deformation susceptibility models can classify mining-affected terrain accurately while leaving the spatial reasons for each class opaque.
Key Innovation: European Ground Motion Service, geological, hydrological and geospatial predictors feed a four-level LightGBM model whose XAI diagnostics expose the controls on subsidence susceptibility in the Rhenish coalfields.
21. Width-dominant rupture of an Mw 7.4 earthquake off Aomori Prefecture, Japan, on 8 December 2025 and its relationship with fault scaling
Core Problem: The offshore extent of the 2025 Aomori rupture is weakly constrained by land observations, limiting tests of whether large interplate earthquakes scale mainly in length or width.
Key Innovation: Deep-ocean pressure records invert the tsunami source and support a roughly 40 km by 116 km fault model, resolving strong up-dip slip and a width-dominant rupture geometry.
22. Marine heatwaves in the Northeast Pacific intensify landfalling atmospheric rivers on the west coast of North America
Core Problem: Atmospheric-river intensification during marine heatwaves may reflect altered circulation or direct air-sea thermodynamics, with different implications for predictability.
Key Innovation: Coupled ocean-atmosphere ensembles hold synoptic circulation fixed and show that warm Northeast Pacific sea surfaces increase evaporation, vapour transport and coastal precipitation through a direct thermodynamic pathway.
23. Mapping distribution and evolution of snow climate in High Mountain Asia during 1980–2018
Core Problem: Avalanche assessment across High Mountain Asia lacks a spatially continuous account of snow-climate regimes and how their controls change under warming.
Key Innovation: Reanalysis from 1980-2018 maps maritime, transitional and continental snow climates, detects a warmer-wetter shift along southeastern margins and separates temperature, snowfall and North Atlantic Oscillation controls.
24. A morphometry driven FAHP machine learning framework for flood, erosion and landslide susceptibility in Chitral sub‐basins, Pakistan
Core Problem: Regional multi-hazard screening in the data-sparse Chitral sub-basins needs a defensible link between drainage morphometry and susceptibility to flood, erosion and landslides.
Key Innovation: FAHP weights morphometric controls, a composite priority index ranks nine sub-basins and leave-one-out tests with three machine-learning models evaluate the ranking without a separate large inventory.
25. Vs Decrease Between Hydraulic Fracturing Stages Revealed by Ambient Noise Analysis
Core Problem: Hydraulic-fracturing seismicity cannot be interpreted fully without tracking stimulation-induced changes in the surrounding velocity structure.
Key Innovation: Stage-matched ambient-noise tomography with identical ray coverage detects an average phase-velocity decline of about 2%, consistent with 5-10% shear-wave-speed reductions around injection depth.
26. Shear Fault Formation in Granite Under High and Transient Co‐Seismic Strain Rates Unveiled by Ultra‐Fast X‐Ray Imaging
Core Problem: The sequence by which dynamic rupture damages granite, forms gouge and weakens a fault occurs too rapidly for conventional laboratory imaging.
Key Innovation: Confined impact experiments and sub-microsecond synchrotron X-ray imaging show volumetric damage followed within 5 microseconds by shear localization and gouge formation before substantial slip.
27. Insights Into Magmatic Degassing, Hydrothermal Dynamics, and Eruption Precursors From Long Term SO2 Emission Monitoring at Whakaari/White Island, New Zealand
Core Problem: Open-vent volcanoes release gas continuously, obscuring the long-period reservoir and sealing processes that distinguish background degassing from eruption preparation.
Key Innovation: Fourteen years of sulfur-dioxide data identify five phases, persistent 565- and 1,275-day cycles and a multi-month 2019 baseline rise above 800 tonnes per day before eruption.
28. InSAR Observability-Guided Multimodal Fusion Framework for Landslide Hazard Detection in Open-Pit Mines
Core Problem: Open-pit InSAR contains fragmented deformation and operational motion, while radar geometry leaves terrain-dependent blind spots that can mimic or hide slope precursors.
Key Innovation: An observability-guided framework combines deformation time series with optical structure and terrain information to distinguish hazard-related acceleration from low-magnitude mining displacement.
29. HGTN: Hybrid GAT-Transformer Network With Spatial–Semantic Features for Landslide Susceptibility Assessment
Core Problem: Landslide-susceptibility graphs capture local neighbourhoods but often miss long-range relations among conditioning factors and terrain units.
Key Innovation: HGTN couples graph attention, channel selection and a Vision Transformer so local anisotropic interactions and global spatial-semantic dependencies are learned jointly.
30. Synergistic impacts of climate and land use/land cover changes on the spatiotemporal evolution of landslides in Kerala
Core Problem: Future landslide susceptibility is usually projected from climate or land-use change alone, although rainfall, vegetation loading, root cohesion and hydraulic conductivity evolve together.
Key Innovation: A TRIGRS-PLUS-CMIP6 chain dynamically updates hydro-mechanical parameters under joint 2026-2035 climate and land-cover scenarios after achieving a baseline AUC of 0.91.
31. Stakeholder-integrated landslide susceptibility mapping in the Western Ghats, India: integrating community perception with AHP-based GIS modelling
Core Problem: Conventional susceptibility maps omit local knowledge about triggers and impacts, weakening their correspondence with lived hazard conditions.
Key Innovation: Responses from 322 stakeholders are analysed alongside eleven GIS factors in a consistency-checked AHP model, placing about 40% of the district in high or very high susceptibility classes.
32. Landslide-terrace interaction mechanisms of two rapid loess landslide cases: insights from MPM modeling
Core Problem: Rapid long-runout and short-runout loess landslides encounter terraces differently, but the entrainment mechanics governing their contrasting mobility are unresolved.
Key Innovation: Entrainment-aware MPM simulations identify terrace shearing by the Dongfeng flow and terrace thrusting by the Shutangwang slide as distinct interaction modes.
33. Probability assessment of potential earthquake-induced landslides in the Namche Barwa Syntaxis, eastern Himalaya
Core Problem: Regional earthquake-induced-landslide maps require a probability of slope failure rather than a deterministic susceptibility class detached from seismic exceedance.
Key Innovation: Probabilistic seismic hazard is coupled with Newmark displacement to map 50-year instability probabilities and isolate high-risk valleys around the Namche Barwa syntaxis.
34. Debris flow events on 5th August 2025 in Dharali region, Uttarakhand Himalaya, India: insights from remote sensing data
Core Problem: The successive 5 August 2025 Dharali debris flows occurred in steep glaciated terrain where direct observations cannot isolate rainfall, thaw and lake-water contributions.
Key Innovation: Pre- and post-event imagery, drainage geometry and historical evidence reconstruct roughly 20 hectares of deposition and support a combined heavy-rainfall, permafrost and meltwater trigger.
35. An enhanced methodological framework for objective assessment of regional land subsidence risks: a case study in Cangzhou city, China
Core Problem: Subsidence-risk maps lose spatial detail through contour interpolation and inherit subjective class thresholds and weighting choices.
Key Innovation: High-resolution deformation rasters, information quantity, combined AHP-entropy weights, natural breaks and sensitivity tests form an objective hazard-vulnerability-exposure workflow.
36. Modeling Oblique Rockfall Impacts Using a Viscoplastic Hardening-Based Contact Framework
Core Problem: Rockfall impact prediction lacks a unified treatment of nonlinear normal and tangential response when irregular blocks strike deformable soil obliquely.
Key Innovation: A viscoelastoplastic hardening model calibrated by small tests and validated at large scale represents sticking-sliding transitions and the energy-dependent distribution of the hardening coefficient.
37. Analysis of Rockfall Motion Characteristics Based on a 3D DDA Stochastic Model: A Case Study of the Houyu Tunnel in Fujian Province
Core Problem: Protection design must account for stochastic rebound and lateral dispersion produced jointly by block shape and three-dimensional channelized terrain.
Key Innovation: A 3D DDA model represents polyhedral blocks and random lateral impact disturbance, quantifying how six sphericities alter speed, bounce height, runout and valley confinement.
38. Analysis and Intelligent Prediction of Mechanical Properties of Similar Materials Under Hydro-dynamic Coupling and the Application to Complex Rock Mass Engineering Model Tests
Core Problem: Physical models of slopes and tunnels lack tunable analogue materials that reproduce coupled seepage and dynamic loading across hydraulic and mechanical scales.
Key Innovation: Similarity criteria, laboratory characterization and a MIMO-PSO-BP predictor define a hydro-dynamic material whose adjustable permeability and stiffness were demonstrated in slope shaking-table and tunnel-excavation tests.
39. Capturing Typhoon‑Driven Causality: A Dynamic Graph Convolutional Network Approach to Interpretable Storm Surge Forecasting
Core Problem: Storm-surge networks change with typhoon position and wind forcing, so static station graphs cannot reproduce causal propagation or peak timing.
Key Innovation: A dynamic graph-learning module constrained by surge propagation laws reduces 72-hour RMSE by 14.4% and peak error by 37.9% relative to a static graph.
40. Karst drought propagation regulated by surface-subsurface 3D structures: A global synthesis and future paradigms
Core Problem: Karst drought cannot be explained by climate and demand alone because the same deficit propagates differently through conduits, epikarst and matrix storage.
Key Innovation: A global synthesis frames surface-subsurface three-dimensional structure as a nonlinear regulator that can rapidly amplify atmospheric drought or buffer it through porous storage.
41. A cross-scale SPH coupling framework based on temporal dynamic boundaries for assessing debris flow mitigation structures
Core Problem: Debris-flow assessment rarely carries catchment-scale propagation into the local fluid-structure interaction needed to evaluate check-dam damage.
Key Innovation: Depth-integrated SPH supplies a temporal dynamic inflow boundary to a full Navier-Stokes SPH model, preserving kinematics while resolving local impact and structural failure.
42. A Comprehensive Survey of Satellite-Based Wildfire Indicators and Spatiotemporal Modeling Approaches: Past, Present, and Future
Core Problem: Satellite wildfire products and predictive models have evolved rapidly, but their indicators, temporal assumptions and regional limitations remain scattered across separate literatures.
Key Innovation: The survey connects MODIS- and VIIRS-era fire indicators with spatiotemporal modelling, tracing the shift toward multisource data fusion and deep learning while identifying unresolved coverage and transferability gaps.
43. A Hybrid Approach of Frequency Ratio Model and Geospatial Technology for Landslide Susceptibility Mapping: A case study
Core Problem: A terrain with recurrent landslide exposure lacks a spatially explicit estimate of how individual conditioning factors combine into susceptibility.
Key Innovation: Frequency ratios and GIS integrate the conditioning factors into five susceptibility classes, assigning 11.5% of the area to very high and 23.2% to high susceptibility and evaluating the final map by its ROC curve.
44. A novel identification method for megadroughts based on return period and drought loss rate
Core Problem: Megadrought definitions rarely connect duration, severity and affected area to the losses that distinguish exceptional droughts from ordinary long events.
Key Innovation: A three-threshold run analysis and copula model link duration-severity-area return periods to drought loss rate, deriving explicit thresholds and identifying seven union-type and one co-occurring megadrought in Shaanxi during 1961-2022.
45. Climate change hotspots in Africa: spatiotemporal trends of rainfall, temperature, and drought using GIS and statistical modelling
Core Problem: African climate assessments are often regional or short-term, limiting continent-wide comparison of century-scale observations with high-emissions projections.
Key Innovation: CRU TS 4.06 observations for 1901-2021 and MRI-ESM2.0 SSP5-8.5 projections map rainfall, temperature and drought jointly, resolving strong regional contrasts that require region-specific adaptation.
46. Collaborative Multimodal Drone-Based Remote Sensing for Levee Piping Detection
Core Problem: Early levee piping produces weak thermal signatures that are easily confused with heterogeneous background surfaces.
Key Innovation: Collaborative UAV sensing combines infrared saliency detection with visible and multisource verification, retaining candidate cold anomalies only when independent spatial evidence supports seepage.
47. Comparison of Earthquake Wavefield Measurements from DAS and a Broadband Seismic Array in an Urban Setting
Core Problem: Urban dark-fibre DAS records depend strongly on cable orientation and coupling, so their ground-motion amplitudes and wavefield directions cannot be assumed equivalent to broadband seismometers.
Key Innovation: A 95 km DAS array is compared with a collocated 16 km broadband array for an M5.2 earthquake, quantifying heterogeneous amplitude ratios and testing recovery of Rayleigh-wave direction of arrival.
48. DPCD-Net: Data-Physics Co-Driven Network for Tropical Cyclone Trajectory and Intensity Prediction
Core Problem: Data-driven tropical-cyclone forecasts struggle to represent the coupled evolution of environmental fields, storm structure, track and intensity.
Key Innovation: DPCD-Net combines learned atmospheric representations with physical constraints in a generative framework that predicts cyclone trajectory and intensity together rather than as independent targets.
49. Drought resilient groundwater systems: evidence from semi-arid areas of northern Ethiopia
Core Problem: It is unclear whether decades of small, distributed recharge interventions can protect semi-arid groundwater systems during an extreme regional drought.
Key Innovation: Four years of regional evidence and monitoring of 78 wells and 41 springs show that soil-water conservation, harvesting structures, dams and a natural lake buffered the 2015-2016 Tigray drought by sustaining recharge.
50. Drought vulnerability assessment of the Beni Mellal-Khenifra region, Morocco
Core Problem: Drought vulnerability in central Morocco reflects interacting climate, terrain, vegetation, water and population factors that single-index maps cannot represent.
Key Innovation: A GIS-AHP model combines ten environmental and socioeconomic variables and classifies 43.98% of Beni Mellal-Khenifra as high vulnerability and 42.78% as medium vulnerability.
51. Effects of pulse-type near-fault ground motions on the nonlinear seismic response of steel frames with inverted-V CBF and EBF systems
Core Problem: Near-fault velocity pulses can drive different nonlinear demands in concentrically and eccentrically braced steel frames, but system-level comparisons remain limited.
Key Innovation: Three-dimensional frame models compare inverted-V CBF and EBF systems under pulse and non-pulse records, separating the influence of bracing mechanism on drift, force and damage response.
52. Energy‐Based Characterisation of Earthquake Damage in Masonry Infills Through Combined In‐Plane and Out‐of‐Plane Experimental Tests
Core Problem: Deformation thresholds alone can miss cumulative earthquake damage in masonry infills when long shaking, soft soils or event sequences produce low-cycle fatigue.
Key Innovation: Combined in-plane and out-of-plane experiments test hysteretic energy as a complementary damage measure and derive energy-based limits tied to observed infill states.
53. Experiencing wildfire, understanding risk in mountain regions of Switzerland: a case study of the 2023 Riederhorn wildfire
Core Problem: Technical wildfire maps do not capture how direct experience changes risk perception, preparedness and agency in mountain communities newly exposed to fire.
Key Innovation: Narrative analysis of twelve resident and professional interviews after the 2023 Riederhorn fire links sensory and emotional memory with community cohesion and practical understanding of future risk.
54. Experimental Investigation on Seismic Performance of the Masonry Structure with Reinforced Concrete Walls and Large Openings at Its Bottom Floor
Core Problem: Large ground-floor openings weaken masonry buildings, and replacing selected walls with reinforced concrete may alter stiffness, force transfer and failure mode.
Key Innovation: A one-quarter-scale four-storey specimen uses stiffness-equivalent concrete walls and pseudo-static loading to test whether the hybrid system restores integrated, ductile seismic behaviour.
55. Flash drought characteristics based on three identification methods in the North China Plain, China
Core Problem: Flash-drought climatologies depend strongly on whether rapid intensification is defined by atmospheric demand, evaporative stress or soil moisture.
Key Innovation: A mean-scaled evaporative stress ratio is evaluated against two established definitions over the North China Plain for 1981-2022, exposing method-dependent differences in onset, duration and spatial pattern.
56. Large-scale circulation-conditioned attribution of the historic 2024 Spanish extreme precipitation
Core Problem: Attribution of Spain's record 2024 rainfall must separate the contribution of unusual atmospheric circulation from thermodynamic amplification by warming.
Key Innovation: Circulation-conditioned analogues estimate that pre-trough dynamics explain about 31% of the event's maximum daily precipitation, allowing the remaining amplification to be assessed against a dynamically comparable baseline.
57. Machine Learning-Based Landslide Susceptibility Mapping and Environmental Controls in Hanang District, Northern Tanzania
Core Problem: The 2023 Hanang disaster exposed the absence of an inventory-based landslide-susceptibility map for northern Tanzania and uncertainty about its dominant environmental controls.
Key Innovation: A Random Forest trained with 429 landslide and 430 non-landslide locations combines remote sensing and GIS predictors to map susceptibility and rank the factors associated with observed failures.
58. MSDT v1.0: An Open-source Multi-method Snow Drought Toolbox
Core Problem: Snow-drought methods are implemented in disconnected, region-specific scripts, preventing reproducible comparison of event definitions.
Key Innovation: MSDT packages data acquisition, multiple identification methods, visualization and cross-method comparison into an open-source Python workflow with a graphical interface.
59. Performance assessment of real-time earthquake localization algorithms for Earthquake Early Warning Systems in the Ibero-Maghrebian Region
Core Problem: Earthquake early warning in the Ibero-Maghrebian region needs a tested localization method that remains accurate with heterogeneous station geometry and limited initial arrivals.
Key Innovation: Real-time replay compares RTLoc with MParLoc, which adds P-wave amplitudes and back azimuth to arrival times, using regional M4-and-larger earthquakes recorded during 2011-2026.
60. Priority-aware independent path-driven multi-agent reinforcement learning for optimal post-earthquake road network recovery
Core Problem: Post-earthquake road restoration must prioritize unequal emergency demand while searching a combinatorial repair space across complex networks.
Key Innovation: Priority-Aware Independent Path decomposition accelerates a multi-agent reinforcement-learning optimizer that sequences repairs around emergency accessibility rather than network connectivity alone.
61. Reliability-qualified Nighttime Lights for Disaster Impact and Recovery in cloud-impacted tropical Regions
Core Problem: Persistent cloud and irregular sampling can make tropical nighttime-light changes look like power disruption or recovery when they are only gaps in observation.
Key Innovation: A reliability-qualified workflow conditions disaster inference on observational support and urban-dominant signal extraction, retaining daily light changes only where the satellite record can sustain them.
62. Seismic response and damage characteristics of an underground metro station in composite strata under near-fault strong vertical ground motions
Core Problem: Near-fault vertical shaking can exceed conventional design assumptions, but its interaction with burial depth and composite strata around underground stations is poorly quantified.
Key Innovation: A three-dimensional nonlinear soil-structure model varies vertical-to-horizontal acceleration ratio and burial depth, locating amplified contact forces and concentrated damage at base-slab, sidewall and central-column joints.
63. Spatio-temporal distributions of compound drought and heat events (CDHEs) with different risk levels in Shaanxi Province, China
Core Problem: Compound drought-heat monitoring often ignores daily duration and the sensitivity of event frequency to alternative combinations of heat and drought thresholds.
Key Innovation: Daily relative thresholds define multiple risk levels across Shaanxi for 1961-2023, revealing more persistent events after the late 1990s, rapid growth since 2020 and larger affected areas for events lasting more than five days.
64. Evaluating the skill of medium-range machine learning-based convective hazard forecasts derived from convection-allowing ensembles
Core Problem: Medium-range convection-allowing ensembles extend to day eight, but their raw output does not establish whether local severe-weather probabilities remain skillful at those leads.
Key Innovation: Neural networks trained on ensemble fields and storm reports evaluate daily tornado, hail and wind probabilities from two experimental systems over the 2023-2024 spring seasons.
65. Structure and Origin of an Active Hydrothermal System in NW Svalbard: A Combined Magnetotelluric and Gas‐Water Isotope Study
Core Problem: The plumbing and heat source of Svalbard’s remote Trollkjeldene-Jotunkjeldene hydrothermal system are poorly constrained by surface observations.
Key Innovation: A twelve-site three-dimensional magnetotelluric survey is interpreted jointly with water and gas isotopes to connect conductive structures with deep fluid pathways.
66. Entrainment‐Based Framework for Understanding Extratropical Moist Heat Extremes
Core Problem: Moist-neutrality arguments underestimate extratropical moist-heat extremes when dry environmental air is entrained into convection.
Key Innovation: Reanalysis, climate-model perturbations and a zero-buoyancy plume model isolate entrainment, saturation deficit and free-tropospheric energy as controls on present and future extremes.
67. Unprecedented Extreme Dust Storm Triggered by Stratosphere‐Troposphere Interactions
Core Problem: The April 2025 southern China dust storm reached hundreds of millions of people far south of the region’s usual dust corridor, outside standard lower-tropospheric explanations.
Key Innovation: Trajectory diagnosis and numerical perturbations attribute predictability to coupled tropospheric wave-energy dispersion and descending stratospheric signals.
68. Demographically-Informed Heat-Mortality Risk Curves via Risk Graph Neural Networks
Core Problem: Conventional heat-mortality curves preserve nonlinear and lagged exposure effects but omit the demographic and geographic structure that governs local vulnerability.
Key Innovation: A hierarchical risk graph neural network maps granular census attributes to distributed-lag nonlinear model coefficients, retaining interpretable risk curves while improving point accuracy and uncertainty coverage across ten regions during the 2022 heatwave.
69. Peak-trough season-partitioned differences in tropical cyclone genesis and related waves
Core Problem: Global tropical-cyclone trends are dominated by peak-season activity, leaving the genesis environments and wave context of less frequent trough-season storms poorly resolved.
Key Innovation: A 42-year hemispheric, season-partitioned analysis separates thermodynamic controls at higher latitudes from dynamic controls at lower latitudes and identifies northward peak-season and westward trough-season shifts in Northern Hemisphere genesis.
70. Use of nonlinear principal components of CHIRPS precipitation data and ocean–atmospheric variables for streamflow forecasting in an area of scarce data – case study: Tocaría river basin – Orinoquia Colombiana
Core Problem: Streamflow forecasting in sparsely gauged tropical basins must represent local rainfall nonlinearity and remote ocean-atmosphere forcing without long calibration records.
Key Innovation: A SARIMAX framework combines nonlinear CHIRPS precipitation components with macroclimate indices and maintains skill for peak flows and forecasts extending to 24 months.
71. GSIM-PLUS: A Gap-Filled Global Monthly Streamflow Dataset for 1995–2015
Core Problem: Global streamflow archives contain long and uneven gaps that prevent consistent monthly hydrological analysis across basins.
Key Innovation: GSIM-PLUS reconstructs monthly discharge for 1995-2015 while retaining uncertainty and expands the spatially complete basis for drought, flood and water-resource comparisons.
72. Future tropical cyclone rainfall constrained by increased atmospheric dryness
Core Problem: Warmer atmospheres hold more moisture, but future tropical-cyclone rainfall also depends on environmental dryness and cannot be inferred from thermodynamic scaling alone.
Key Innovation: The study constrains cyclone rainfall change with the competing evolution of atmospheric moisture and dryness, refining projections of storm-total precipitation.
73. Modeling and analysis of material stocks in areas susceptible to flooding: case study in novo Hamburgo-RS
Core Problem: Flood-prone cities lack spatial estimates of the building materials that would become demolition waste and constrain post-disaster management.
Key Innovation: Georeferenced buildings, flood-susceptibility maps and material-intensity factors identify 22,767 exposed buildings and about 1.6 million tonnes of concrete in Novo Hamburgo.
74. Spatiotemporal drought assessment using remote sensing indicators and land-use change analysis: implications for loss and damage in pastoral communities of Simanjiro District, Tanzania
Core Problem: Pastoral drought loss in semi-arid East Africa is difficult to track where field records are sparse and land-use change alters vegetation-based signals.
Key Innovation: Multiple satellite drought indicators and land-use trajectories map persistent hotspots and connect worsening severity to pastoral loss-and-damage planning in Simanjiro.
75. Intensification and expansion of concurrent droughts in Northern China via a copula-based drought index
Core Problem: Separate meteorological, agricultural and hydrological drought indices cannot measure their concurrent onset, lag structure and expansion across northern China.
Key Innovation: A multivariate copula index combines the three drought types, selects regional dependence structures and maps where concurrent droughts are becoming more frequent and severe.
76. An Approach to Determine the Hoek–Brown Rock Mass Disturbance Factor in Open-Pit Mining Through Near-Field Peak Particle Velocity
Core Problem: The Hoek-Brown disturbance factor is difficult to assign in open-pit highwalls where blasting, weak seams, joint opening and bedding shear act together.
Key Innovation: Near-field peak particle velocity and the Heelan-Blair vibration model are translated into field guidelines for estimating disturbance in surface-coal-mine slopes.
77. RoDiAn: A Tool for Kinematic and Dynamic Stability Analysis of Discrete Blocks in Masonry and Engineered Rock Structures
Core Problem: Large-scale rock-slope software does not readily represent the small, discrete blocks and boundary geometry of masonry and engineered rock structures.
Key Innovation: RoDiAn derives block orientation and geometry from three-dimensional point clouds and unifies kinematic screening with dynamic stability analysis in a MATLAB tool.
78. A Moisture-Accelerated Cross-Scale Damage Cascade in Paleogene Red-Bed Mudstone Under Freeze–Thaw Cycling, NE Tibetan Plateau
Core Problem: Freeze-thaw deterioration of red-bed mudstone cannot be predicted from cycle count alone because moisture links pore-scale damage to macroscopic failure.
Key Innovation: Tests across twelve cycles and two moisture states connect microstructure, wave response and strength loss into a moisture-accelerated cross-scale damage cascade.
79. Simultaneous influence of spatially varying ground motions and soil support on a rockable bridge with unequal slenderness
Core Problem: Rocking-bridge response is commonly evaluated with uniform excitation or ideal support, omitting the joint effects of unequal piers, soil compliance and spatially variable shaking.
Key Innovation: Shake-table tests on a two-segment bridge isolate the simultaneous influence of support condition, pier slenderness and differential ground motion.
80. Enhancement of enzyme-induced carbonate precipitation (EICP) for sustainable soil stabilization in loess slopes using soybean-derived urease and organic nucleating agents
Core Problem: EICP stabilization of cracked loess is constrained by enzyme cost, sparse nucleation and spatially uneven carbonate cementation.
Key Innovation: Soybean-derived urease, alternative calcium salts and organic nucleating agents are screened together; sucrose-enhanced treatment improves strength, water stability and freeze-thaw durability.
81. A Framework to Mitigate Wind-induced Damages to Electric Power System Components and Reduce the Risk of Wildfire Incidents Caused by Hot and Dry Seasonal Winds
Core Problem: Power-line wildfire mitigation lacks a quantitative link between wind-driven tree failure, component damage and ignition likelihood.
Key Innovation: A probabilistic framework evaluates tree-removal programmes across vegetation and seasonal wind conditions and estimates their reduction of infrastructure-caused ignitions.
82. Probabilistic Wildfire Risk Assessment: A Computationally Efficient Framework for Systematic Uncertainty Propagation – Hazard Analysis
Core Problem: Wildfire risk estimates propagate uncertain fuel, weather and response variables inconsistently, making tail risk difficult to compare across scenarios.
Key Innovation: A computationally efficient probabilistic framework carries parameter uncertainty through hazard analysis rather than reporting a single deterministic fire-risk estimate.
83. A Novel and Efficient Methodology for Full-Field Structural Reliability Analysis of High Earth and Rockfill Dams
Core Problem: Dam reliability studies usually concentrate on global slope failure and do not resolve how reliability varies among individual elements across the full structure.
Key Innovation: Third-moment screening, zonal representative elements and a LASSO-assisted surrogate recover the reliability field of a 314 m earth-rockfill dam about six times faster than the reference calculation, with roughly 10% mean relative error.
84. Reliability and Sensitivity Analysis of Global Failure Modes of Concrete Buttress Dam Monoliths
Core Problem: Reliability evidence for concrete dams is concentrated on gravity and arch designs, leaving the competing global failure modes of buttress monoliths poorly quantified.
Key Innovation: A population of Swedish-derived geometries supports reliability and sensitivity analysis across sliding and overturning, identifying sliding as the governing mode.
85. Parameterization of ground ice dynamics in permafrost models: Advances, challenges, and future directions
Core Problem: Permafrost models simplify ground-ice gain and loss even though excess ice controls thaw settlement, hydrology and infrastructure response.
Key Innovation: The review compares current parameterizations and identifies the observations and process couplings needed to represent dynamic ground ice across scales.
86. Pixel-wise change detection for tunnel linings from multi-temporal inspections
Core Problem: Single-date tunnel-lining segmentation records defects but cannot determine whether an individual crack or damaged region is growing.
Key Innovation: A bi-temporal pixel-wise framework registers repeated inspections and directly segments lining change under difficult underground illumination and surface conditions.
87. Simultaneous aridification of soil moisture driven by anthropogenic activities in China’s loess critical zone
Core Problem: Short instrumental records cannot establish whether soil-moisture drying across China’s Loess Plateau is recent, synchronous or primarily anthropogenic.
Key Innovation: Tree-ring oxygen isotopes reconstruct two regional series back to 1556 and 1675; CMIP6 attribution links their post-1850 coherence and stronger extremes chiefly to human forcing.
88. Drivers, dynamics, and cascading effects of phased drought-flood abrupt alternation on basin water quality
Core Problem: Drought-flood abrupt alternation is usually treated as two endpoints, obscuring how phased transitions cascade into river-water-quality deterioration.
Key Innovation: Events are partitioned into drought development, drought recovery and flood phases to trace dynamic changes in pH, oxygen demand, dissolved oxygen and ammonia nitrogen.
89. Impact of reservoir storage on propagation from meteorological to hydrological drought
Core Problem: Reservoirs interrupt the propagation from meteorological to hydrological drought, but storage and network position are rarely quantified together.
Key Innovation: Four drought states across eleven Krishna Basin reservoirs are linked with a downstreamness metric to separate storage buffering from upstream-downstream connectivity.
90. Spatiotemporal evolution and propagation of GNSS-derived PWV-based meteorological and soil moisture drought: a case study of the contiguous United States (2003–2022)
Core Problem: Conventional drought indices lack the all-weather, high-frequency atmospheric moisture observations needed to connect meteorological deficits with soil response.
Key Innovation: GNSS precipitable-water-vapour anomalies are integrated into a 2003-2022 propagation analysis and compared across timescales with established meteorological and soil-moisture indices.
91. Improved design method for the nuclear island periodic raft-pile foundation hybrid isolation system considering soil structure interaction
Core Problem: Nuclear pile-raft foundations on non-bedrock sites require broadband seismic isolation without losing a tractable representation of soil-structure interaction.
Key Innovation: The design combines row piles with a periodic raft that filters seismic frequency bands, then uses spatially decoupled modelling and partitioned soil-structure response to reduce foundation-to-ground motion across a broad band.
92. Intensity measure–damage relationship in instrumented school and nearby buildings under near-fault pulse-type ground motions
Core Problem: Empirical near-fault damage relations rarely pair building-specific strong-motion records with observed performance across multiple damage states.
Key Innovation: Records from 26 instrumented schools and nearby housing during the 2023 Kahramanmaraş sequence directly relate pulse-sensitive intensity measures to collapse, moderate and slight damage.
93. Seismic performance of post-tensioned segmental precast concrete shear walls integrated with pinned steel frames
Core Problem: Rapidly erected hybrid frames need seismic resistance without sacrificing the self-centering behaviour of post-tensioned precast walls.
Key Innovation: Validated nonlinear models of 4-, 8- and 12-storey wall-pinned-frame systems test drift, recentering and member demand under static and dynamic loading.
94. Cyclic experiments on novel self-centering beam-column connections with replaceable reduced steel plate energy dissipation
Core Problem: Self-centering steel connections must dissipate earthquake energy while confining inelastic damage to components that can be replaced after an event.
Key Innovation: Full-scale cyclic tests combine post-tensioned strands with replaceable cut or drilled fuse plates, producing stable recentering, largely elastic beams and columns, and equivalent viscous damping ratios of 17.8-18.9%.
95. Mitigation of higher-mode flexural demands in controlled rocking walls using outriggers
Core Problem: Controlled rocking walls reduce residual drift but can amplify higher-mode flexural demand in the upper stories of tall buildings.
Key Innovation: A shake-table-validated OpenSees model shows that damped outriggers reduce upper-story demand and base rotation; moderate yielding distributes demand more effectively than maximizing outrigger strength.
96. Evaluation of the Hydrothermal Response of a Permafrost Highway Embankment in permafrost regions of Qinghai-Tibet Plateau to CMIP6 Model: A Case Study from the Beiluhe Area
Core Problem: Permafrost embankment projections often omit rainfall sensible heat and position-dependent energy exchange, biasing long-term thaw response.
Key Innovation: A validated coupled hydrothermal and surface-energy model projects distinct 30-year SSP2-4.5 responses across positions of the Beiluhe highway embankment.
97. Deriving continuous deglaciation chronology from infrequent historical outlines
Core Problem: Sparse historical glacier outlines provide dated boundaries but leave the timing of deglaciation unresolved for most terrain between them.
Key Innovation: A semi-automatic open-source interpolation converts irregular outlines into continuous year-of-retreat rasters, making exposure duration available for every proglacial pixel.
98. Maximum Temperature Forecasts in the Yangtze River Valley Using a Predictor‐Corrector Lightweight Deep Learning
Core Problem: Numerical weather prediction methods often face challenges from atmospheric chaos, limited physical understanding, and external forces, resulting in uncertainties and biases.
Key Innovation: Because individual predictors, whether convolutional or transformer‐based networks, can fall short, this study introduces a lightweight forecasting framework that combines a ResNet‐based predictor with a multilayer‐perceptron‐based corrector, achieving up to 7.8% accuracy improvement.
99. Stiffened sand compaction piles for coastal applications: A numerical assessment
Core Problem: The seismic and service-load performance of stiffened sand compaction piles in layered marine soils has not been established for coastal infrastructure.
Key Innovation: Numerical models validated against full-scale observations and a centrifuge test show that the composite piles can triple vertical capacity, raise lateral capacity by 80% and reduce earthquake-induced force, settlement and tilt, with gains normalized by embodied carbon.
100. Multi-source data-driven and interpretable hybrid prediction framework for the dynamic shear modulus of marine clays
Core Problem: Dynamic shear-modulus estimates for marine clay remain uncertain across sites, limiting offshore seismic-response and foundation analyses.
Key Innovation: A 1,265-sample multi-region database, anomaly screening and hybrid XGBoost optimization yield an interpretable predictor with a reported test R-squared of 0.9983; SHAP, ICE and LIME consistently identify effective stress and shear strain as the dominant controls.
101. Bias Correction of Satellite Precipitation Products Near the Data-Scarce Region of Al-Ahsa, Saudi Arabia
Core Problem: This requirement is further intensified in data-scarce arid regions where ground-based observations are limited and sparsely distributed.
Key Innovation: The study aims to establish an effective bias correction algorithm in arid regions by evaluating the performance of three recommended machine learning models, namely, random forest, XGBoost, and ConvLSTM, over the area surrounding Al-Ahsa oasis in Saudi Arabia. CMORPH and IMERG datasets, along with climate variables (minimum temperature, relative humidity, and cloud fraction) from ERA-5, were compared with ground observations for model calibration and validation.
102. Wind-Field-Integrated Deep Learning for Marine Oil Spill Detection: Suppressing False Positives in SAR Imagery
Core Problem: However, SAR images frequently contain dark patches induced by low-wind-speed areas (LWSAs) and the leeward sides of islands (LSIs), which can cause persistent false detections in conventional models.
Key Innovation: To address this challenge, we propose a wind-field-integrated deep learning approach built on the UNet++ architecture with a spatial and channel squeeze-and-excitation (SCSE) attention module. The proposed SAR-UV configuration based on UNet++ with SCSE attention achieved a Precision of 89.4%, Recall of 83.4%, F1-score of 85.7%, and false-positive rate (FPR) of 2.6%.
103. Integrating SBAS and a Farneback Iterative Optical Flow Model for Large-Gradient Deformation Estimation in Mining Areas
Core Problem: Large-gradient mining deformation breaks phase unwrapping, while offset tracking alone is too noisy and computationally costly for dense time series.
Key Innovation: SBAS-FOM combines small-baseline error suppression with Farneback optical flow and is validated on eleven TerraSAR-X acquisitions over the Daliuta mine.
104. ST-Mamba-LDM: An Attention-Mamba Hybrid Spatio-Temporal Latent Diffusion Model for Radar Echo-Based Precipitation Nowcasting
Core Problem: Existing deep learning-based methods for radar echo prediction in precipitation nowcasting often exhibit limited capability in modeling complex spatio-temporal dynamics, which leads to over-smoothed forecasts and inadequate representation of irregular variations and fine-scale structures in precipitation systems.
Key Innovation: To address these limitations, we propose ST-Mamba-latent diffusion model (LDM), an Attention-Mamba hybrid spatio-temporal latent diffusion model for radar echo-based precipitation nowcasting. The SHM integrates spatial attention with spatial Mamba to jointly model global dependencies and local details, while the THM combines temporal attention with bidirectional temporal Mamba to capture temporal evolution patterns more effectively.
105. A Robust and Fast InSAR Phase Denoising Algorithm Based on Coarse-to-Fine Adaptive Residual Iteration
Core Problem: Mainstream denoising methods face a key constraint: it is challenging to strike an optimal balance between denoising power, detail preservation, and computational efficiency.
Key Innovation: To address this, this article presents a robust and fast InSAR phase denoising algorithm based on coarse-to-fine adaptive residual iteration, with its core being a progressive denoising architecture. By fully leveraging the simple yet efficient EFAF module, the method iteratively restores detail information submerged in noise, achieving superior overall performance with a concise architecture.
106. Mapping three decades of urban growth in China: a 30 m annual building height dataset (1990–2019)
Core Problem: Long-term building height data are critical for analyzing urban morphological evolution and renewal processes, yet such datasets at fine spatial resolutions remain scarce for large geographical regions.
Key Innovation: The study proposes a framework to generate continuous annual building height maps for China at 30 m spatial resolution from 1990 to 2019, integrating multi-source remote sensing data (Landsat, Sentinel-1/2, etc.) through the eXtreme Gradient Boosting (XGBoost) model.
107. Insight into the Equivalent Shear Strength of Rock Masses with Multiple Nonpersistent Joints based on Laboratory Tests
Core Problem: Equivalent shear-strength models for jointed rock do not adequately represent how joint geometry and progressive rock-bridge degradation interact.
Key Innovation: Direct-shear experiments spanning joint persistence, inclination and number support an improved Jennings criterion that incorporates both geometry and compressive-shear degradation and is more accurate than the original formulation.
108. Discontinuum out-of-plane failure analysis of a stone masonry façade under different modelling assumptions
Core Problem: Out-of-plane assessment of historic masonry facades is highly sensitive to incomplete geometry, material assumptions and boundary representation.
Key Innovation: Distinct-element pushover analyses isolate these modelling effects and show when simplified masonry remains adequate if restraints are represented correctly.
109. Optimal design and performance evaluation of outrigger systems augmented with tuned viscous mass dampers
Core Problem: Conventional outriggers reduce drift but do not independently control displacement, acceleration and shear demands in tall buildings.
Key Innovation: Two tuned-viscous-mass-damper placements exploit amplified core-column motion; optimization separates a displacement-oriented design from one that better limits acceleration, drift and shear.
110. Fragility functions for corroded intermediate short reinforced concrete columns
Core Problem: Column fragility models seldom represent how reinforcement corrosion shifts damage and near-collapse thresholds.
Key Innovation: A database of 156 tests from 22 campaigns supports drift-based fragility functions by corrosion level, showing the largest median-capacity losses in severe and near-collapse states.
111. Probability-conserving decoupling for global dynamic reliability with multiple limit states: Theory, algorithms, and numerical results
Core Problem: Dynamic structural reliability with interacting limit states requires high-dimensional probability integration and becomes computationally prohibitive for nonlinear seismic response.
Key Innovation: The method decouples the problem into one-dimensional density-evolution equations, solves them with a finite-volume WENO scheme and recombines Boolean failure events, retaining errors below 1% in benchmark and seismic-frame tests.
112. Post-disaster recovery of infrastructure system with spatial zonation based on multi-agent deep reinforcement learning
Core Problem: Infrastructure recovery schedules waste scarce crew time when spatial travel and work-zone constraints are omitted from component prioritization.
Key Innovation: A zoned multi-agent reinforcement-learning model co-optimizes repair sequence and crew territory, reducing travel distance by 48.24-58.45%.
113. Substantial improvement of geostationary satellite-derived shortwave radiation product through correcting geolocation displacement
Core Problem: Terrain parallax displaces geostationary-satellite pixels and systematically distorts ridge-valley shortwave-radiation contrasts in mountainous regions.
Key Innovation: A terrain-based geolocation correction raises agreement with MODIS by 0.13 in spatial correlation and lowers root-mean-square difference by 30.8 W m-2 over the Tibetan Plateau, with independent validation at 130 stations.
114. HeightCLIP: A vision-language network for height estimation from monocular remote sensing imagery
Core Problem: However, existing vision-based methods rely solely on visual features, which limits their ability to fully exploit height-related semantic information.
Key Innovation: To overcome this limitation, we propose HeightCLIP, a novel vision-language framework tailored for height estimation from monocular remote sensing imagery. In addition, we incorporate a reciprocal cross-attention fusion mechanism to facilitate bidirectional interaction between visual and textual features, thereby improving cross-modal feature integration.
115. Unveiling shear-induced non-uniform deformation of segmented tunnels through fully instrumented model tests
Core Problem: The transfer of localized ground shear into discontinuous deformation and joint opening along segmented tunnels is not resolved by global response measurements.
Key Innovation: A fully instrumented large-scale test and validated three-dimensional model jointly track soil localization, distributed tunnel strain and discrete joint release, revealing an abrupt, finite-range multi-peak deformation pattern.
116. A physics-constrained data-driven method for predicting the damage zone extent of surrounding rock under coupled dynamic-static loading
Core Problem: Damage-zone extent around dynamically disturbed roadways is difficult to infer from coupled rock strength, in-situ stress and transient loading.
Key Innovation: An automated FLAC3D workflow generates 1,000 physics-constrained cases for six surrogate models; CatBoost reaches reported R-squared values of 0.852 and 0.953, while SHAP separates the controls on roof and sidewall damage.
117. Physics-consistent multi-module and multi-branch neural networks for hydro-mechanical coupling in soils
Core Problem: However, for complex multi-field coupling problems, such as seepage-consolidation coupling in soils, conventional PINN architectures may be insufficient to capture the strongly nonlinear and bidirectional interactions among different physical quantities.
Key Innovation: To address this limitation, a multi-module and multi-branch physics-informed neural network (MM-PINN) is proposed within the conventional PINN framework for reconstructing hydro-mechanical coupling responses in soils. Numerical examples demonstrate that MM-PINN can accurately predict key hydro-mechanical responses, including pore water pressure, displacement, and effective stress, across different stages.
118. Contribution of Aloft Gravitational Sorting and Lower‐Part Coalescence to Precipitation Over the Southeastern Tibetan Plateau
Core Problem: Observational constraints on the vertical microphysical evolution of mixed‐phase precipitation across the melting layer over the high‐altitude of southeastern Tibetan Plateau remain scarce.
Key Innovation: Using collocated two‐dimensional video disdrometer and Micro Rain Radar measurements in rainy seasons of 2022–2025, we characterize the vertical microphysical evolution of mixed‐phase precipitation across the melting layer.
119. WAT3R: Feedforward Underwater 3D Reconstruction
Core Problem: Reliable feedforward underwater 3D reconstruction remains challenging due to severe light attenuation and backscattering, which degrade visual quality and disrupt feature consistency across views, leading to inaccurate multi-view geometry.
Key Innovation: To address this issue, we propose WAT3R, a feed-forward framework for reconstructing 3D scenes directly from underwater images. By leveraging degradation adaptation as a geometry-constrained process, WAT3R integrates a lightweight neural adaptation module to flexibly account for these underwater imaging effects, thereby improving multi-view reconstruction quality.
120. Water-VGGT: Leveraging visual geometry and water-optics for robust 3D reconstruction in underwater environments
Core Problem: Underwater 3D reconstruction is severely challenged by complex optical distortions caused by refraction at camera interfaces, particle scattering, and wavelength-dependent attenuation, which violate the imaging assumptions of conventional multi-view reconstruction methods.
Key Innovation: To address this problem, we propose Water-VGGT, a physics-conditioned network for underwater 3D reconstruction. A set of auxiliary condition-aware prediction heads is incorporated to support camera, dense geometry, and tracking-related outputs under severe optical degradation and particle interference.
121. Source-Free Adaptation for Cross-Sensor SAR Target Detection via Prototype Anchoring and Foreground-Aware Manifold Constraint
Core Problem: Cross-sensor synthetic aperture radar (SAR) target detection faces significant domain shift challenges.
Key Innovation: To overcome these challenges, we propose a novel SFDA framework for cross-sensor SAR target detection by deeply mining the intrinsic physical invariants of SAR targets in the deep feature space. Experiments on cross-sensor tasks demonstrate that our framework reaches improvements of about 10%-25% in mean average precision and F1 score without accessing any source data, significantly enhancing the model’s generalization capability in unknown scenarios.
122. Unsupervised SAR Change Detection of Small Objects via Superpixel Classification
Core Problem: Unsupervised change detection for small objects using synthetic aperture radar imagery is particularly challenging due to the limited number of changed samples, the low signal to clutter ratio, and strong speckle noise, which often degrade the reliability of conventional change detection methods.
Key Innovation: In this article, we propose an unsupervised SAR change detection framework for small object detection via superpixel classification. Experimental results on multiple real SAR datasets show that the proposed method consistently improves detection performance compared to several unsupervised approaches.
123. Lightweight and Efficient Spatial–Frequency Network for Remote Sensing Change Detection
Core Problem: Although Transformer-based methods have demonstrated strong global modeling capability, they often suffer from limited local feature perception and quadratic computational complexity, which hinders their deployment for high-resolution remote sensing images.
Key Innovation: To address these limitations, we propose a lightweight and efficient spatial-frequency network (LENet) for RSCD. First, a lightweight local-global feature integration module is designed in the encoder to effectively capture fine-grained local details while efficiently modeling global structural information.
124. Assessing the Transferability of L-Band SAR Soil Moisture Retrieval Parameters Between Agricultural Crop Types
Core Problem: Field-scale soil moisture (SM) retrieval using synthetic aperture radar (SAR) data under heterogeneous vegetation conditions remains challenging due to complex SAR backscatter responses influenced by vegetation, surface roughness, and soil texture.
Key Innovation: The study examines whether the parameter transferability of the water cloud model (WCM) extends to different agricultural crops-corn, soybean, wheat, and canola-using SMAPVEX12 L-band UAVSAR data. Application of WCM parameters trained for a specific crop to other types of crop resulted in increased SM retrieval error (ubRMSE) by 26%-55% for VV-pol and 37%-63% for VH-pol.
125. DSCT: Dual-Sparse Coherent Transformer With Cross-Feature Alignment for Remote Sensing Image Super-Resolution
Core Problem: Remote-sensing super-resolution must recover fine local structure without the quadratic cost and oversmoothing of dense Transformer attention.
Key Innovation: DSCT couples adaptive sparse attention with cross-feature coherence, improving reconstruction across several benchmarks at competitive computational cost.
126. DMHF-Net: Dual-Modal Hierarchical Fusion Network for RGB–IR Joint Detection of Small-Scale Targets From UAV Perspective
Core Problem: However, under complex environments and varying illumination conditions, single-modal processing of small targets often suffers from weak feature representation, making effective feature extraction a formidable challenge.
Key Innovation: To address these challenges, we propose a novel multimodal remote sensing detector, namely dual-modal hierarchical fusion network (DMHF-Net). After detection, critical samples are weighted in the intersection-over-union-aware weighted loss and participate in backpropagation to improve the network.
127. Geometric Correction and Accuracy Verification of the Geology-1 Hyperspectral Satellite
Core Problem: A new hyperspectral satellite requires geometric calibration and orthorectification accuracy to be established before its imagery can support geological mapping.
Key Innovation: A rigorous imaging model, on-orbit calibration and Level-2 processing chain are evaluated under ground-control-point-free conditions for Geology-1.
128. Global Prototype-Driven Contrastive Partial Adaptation in Hyperspectral Image Classification
Core Problem: Cross-scene hyperspectral adaptation becomes unstable when the target contains only a subset of source classes and no target labels.
Key Innovation: Global prototypes, contrastive separation and partial-domain weighting suppress irrelevant source classes while preserving transferable spectral structure.
129. SMU-Net: A Mamba-Based Siamese U-Shaped Network for Unsupervised Heterogeneous Remote Sensing Change Detection
Core Problem: Unsupervised heterogeneous change detection (UHCD) remains challenging due to the significant modality discrepancies between multisource remote sensing data.
Key Innovation: To address these issues, we propose SMU-Net, a novel Siamese U-shaped network built upon the recently developed Mamba architecture. Furthermore, we introduce a cross-modal representation consistency loss to guide the network toward modality-invariant feature learning, and design a demosaicing-inspired decoder to improve local structural continuity and suppress artifacts.
130. ICP-Net: Exploiting Iterative Consistency of Pseudolabels for Cross-Domain Object Detection in Remote Sensing
Core Problem: Domain shifts caused by varying remote sensing sensors severely degrade the model performance of object detection when deployed in target domains.
Key Innovation: To address both domain discrepancy and pseudolabel degradation, we propose ICPNet, an Iterative Consistency Pseudo-Labeling Network that integrates self-supervised learning with pseudolabel refinement to enhance domain robustness and label reliability. In addition, we introduce an Iterative Consistency Pseudolabel Filter (ICPF) that leverages a memory of historical predictions.
131. GeoMamba: Geometric-Prior-Infused Multi-Scale Deformable Visual Mamba for Crack Semantic Segmentation
Core Problem: Accurate pavement crack segmentation is critical for road infrastructure assessment, yet it remains challenging due to complex background noise and highly variable crack topologies.
Key Innovation: To address this limitation, we propose GeoMamba, a geometric-prior-infused multi-scale deformable visual Mamba network for road crack semantic segmentation. Comprehensive experiments on DeepCrack and Concrete3K datasets demonstrate that GeoMamba outperforms nine state-of-the-art methods.
132. A Scene Attention-Based Semantic Segmentation Method for Urban Real-Scene 3-D Models
Core Problem: Urban three-dimensional meshes preserve geometry but lack the semantic labels needed for exposure inventories and object-level analysis.
Key Innovation: Scene attention adds contextual reasoning to mesh segmentation and reaches 73.2% mIoU on the Helsinki SUM benchmark.
133. Automatic Extraction Method of Road Intersection Center Points From Spaceborne SAR Images
Core Problem: However, automatic detection and accurate center-point extraction of road intersections from L1-level SAR images remain challenging owing to severe speckle noise, geometric distortions, and blurred structural boundaries.
Key Innovation: To address these issues, this study proposed a two-stage automatic method for road intersection detection and center-point coordinate extraction from spaceborne SAR imagery. In the detection stage, an enhanced YOLOv8-based model was developed by applying multiscale contextual perception and attention mechanisms to improve intersection recognition under complex SAR imaging conditions.
134. Radiometric Calibration of MODIS Using a Hyperspectral Radiometer on the Near-Space Platform
Core Problem: Traceable on-orbit radiometric calibration is difficult when satellite overpasses and ground references differ in atmosphere, scale and spectral response.
Key Innovation: A balloon-borne hyperspectral radiometer provides a near-space reference for MODIS bands 1-7, reducing atmospheric path length in a controlled Qinghai experiment.
135. Spatial Semantic Information Interaction for Lightweight Salient Object Detection in Optical Remote Sensing Images
Core Problem: However, existing approaches often struggle to exploit high-resolution spatial details and deep semantic information simultaneously.
Key Innovation: To overcome these issues, this study introduces a lightweight ORSI-SOD network named the spatial semantic information interaction network (SSINet). Extensive experiments on three ORSI benchmark datasets show that SSINet achieves leading performance across twelve evaluation metrics with only 3.17 M parameters and 1.96 G FLOPs.