TerraMosaic Daily Digest: August 4, 2026
Daily Summary
Wide-swath altimetry and stochastic inundation modelling expose two complementary components of tsunami risk. SWOT imaged dispersive sea-surface-height arches about 70 minutes after the 2025 Mw 8.8 Kamchatka earthquake; inversion places approximately 15 m of slip in the southern trench-adjacent rupture and reproduces nearby deep-ocean records. At Kahului Harbor, transport of 2,150 debris objects through 3,801 building footprints shows that stochastic post-collision deflection expands debris dispersion by 27% and identifies 15.6% more struck buildings than deterministic tracking. The first result constrains the offshore source from the propagating wave field; the second quantifies how object-scale uncertainty changes community-scale consequences.
Slope-hazard analysis is becoming both time dependent and geometry aware. UAV LiDAR and photogrammetry identify roughness anomalies before stockpile failure, while geometry-aware InSAR filtering avoids treating low line-of-sight deformation as proof of stability in landslide-susceptibility sampling. SBAS-InSAR displacement is also combined with environmental triggers for dynamic Himalayan risk assessment, while an open 29-epoch Sentinel-1 archive preserves the deformation history surrounding the 2025 Njintout failure in Cameroon. Failure-time radar rainfall and soil-water indices replace static climatologies in spatially validated susceptibility models. Three-dimensional point clouds, discontinuity kinematics and runout simulation extend this chain from regional screening to individual rockfall sources; reservoir monitoring links impoundment above 810 m to renewed landslide and tunnel deformation.
Hazard models and open data products increasingly expose the physical or observational assumptions that control their predictions. A global sandy-beach-slope product combines Sentinel-1/2 observations with environmental and human-influence variables, reporting independent-validation R-squared of 0.8879 and releasing the mapped slopes for coastal erosion and flood-vulnerability analysis. Three-dimensional simulations of metro inundation and spillway-driven riverbank loading translate flow geometry into evacuation and protection choices. Relocated Tonga earthquakes delineate deep seismic planes less than about 3 km wide, and a 273-node receiver-function survey resolves a Los Angeles basin approaching 10 km depth with a markedly uplifted Moho. Repairability-informed wharf fragilities propagate damage location into recovery, whereas a stress-normalized shear-wave-velocity model represents confining-pressure effects on liquefaction resistance. The WEXA archive supplies event-level heat-wave and cold-spell objects from 1979 onward; sub-daily streamflow analysis, radar-rainfall flood nowcasting and physics-guided typhoon downscaling show why temporal resolution and process constraints must be evaluated as part of the model rather than treated as fixed inputs.
Key Trends
Five shifts connect wave-field observation, time-dependent slope state, three-dimensional geometry, physically constrained prediction and reusable hazard data.
- Tsunami Observation Is Moving from Point Records to Wave-Field Imaging: SWOT's two-dimensional sea-surface-height field resolves dispersive structure and constrains megathrust slip, while stochastic debris transport translates inundation into building-scale impact distributions.
- Observed Slope State Is Replacing Static Stability Proxies: UAV topographic change, InSAR deformation, reservoir level, failure-time rainfall and soil water are incorporated directly; the Njintout release also makes a pre-failure Sentinel-1 displacement sequence independently inspectable.
- Three-Dimensional Geometry Connects Source, Path and Exposure: Fault-plane relocation, basin imaging, point-cloud rock-block reconstruction, tunnel-landslide interaction, metro inundation and riverbank-flow simulation make geometry an estimated state variable rather than a schematic input.
- Physical Structure Is Entering Learned Models at the Process Interface: Shallow-water residuals constrain flood synthesis, scale equivariance encodes unit-hydrograph behaviour, and typhoon downscaling and flood nowcasting are tested against independent events or contrasting regimes.
- Open Data Are Becoming Part of Hazard Inference: Global beach slopes, Njintout displacement histories, event-based weather extremes and station-scale flood simulations expose inputs and intermediate states that can be reanalysed rather than only presenting final risk classes.
Selected Papers
The issue opens at two scales of tsunami inference: SWOT wave-field imaging constrains megathrust slip, while stochastic debris collisions alter building-level impact across an inundated community. Slope studies then connect deformation, rainfall, soil water, surface roughness, discontinuity geometry and reservoir operation to observable changes in hazard state, with the Njintout archive preserving one pre-failure sequence for reanalysis. A global beach-slope product and station-scale flood simulations extend the same emphasis on mechanism, uncertainty, reusable evidence and decision-relevant resolution.
1. The slip distribution of the 2025 Kamchatka earthquake determined from SWOT satellite dispersive tsunami images
Core Problem: Conventional tsunami source inversions rely on spatially sparse gauges and pressure sensors, leaving the offshore wave field itself largely unobserved.
Key Innovation: SWOT captured dispersive sea-surface-height arches about 70 minutes after the Mw 8.8 Kamchatka earthquake; joint inversion resolves approximately 15 m of southern trench-adjacent slip and reproduces the nearest deep-ocean time series.
2. Stochastic Tsunami-Driven Debris Hazard in a Built Environment: Community-Scale Assessment at Kahului Harbor, Maui
Core Problem: Community-scale debris models usually advect objects deterministically, although building collisions redirect debris stochastically and can change which structures are hit.
Key Innovation: A stochastic post-collision model transports 2,150 objects through 3,801 Kahului building footprints, expanding debris dispersion by 27% and identifying 15.6% more impacted buildings than deterministic tracking.
3. Mapping global sandy beach slope using multi-source remote sensing data and ensemble learning
Core Problem: Coastal erosion and flood-vulnerability studies lack a globally consistent beach-slope layer derived under one observation and modelling framework.
Key Innovation: Sentinel-1/2, morphology, marine conditions, climate and human-influence variables are fused in a stacking ensemble to map global sandy-beach slope; the released product reports independent-validation R-squared of 0.8879 and RMSE of 0.0250.
4. Dynamic landslide risk assessment (DLRA) using SBAS-InSAR method, integrated with geo-environmental data as temporal triggers in the Northwestern Himalayas, Pakistan
Core Problem: Terrain-conditioned susceptibility does not reveal whether a mapped slope is actively moving or how triggering conditions evolve through time.
Key Innovation: SBAS-InSAR deformation is combined with geo-environmental temporal triggers to construct a dynamic landslide-risk framework for the northwestern Himalaya, with line-of-sight and regional-calibration limits retained explicitly.
5. Remote Sensing-Assisted Stockpile Landslide Monitoring Based on Change Detection Analysis and Identification of Topographical Failure Precursors
Core Problem: Heterogeneous engineered stockpiles can fail without dense instrumentation, and small precursor changes are difficult to separate from vegetation and survey noise.
Key Innovation: Registered UAV LiDAR and photogrammetric point clouds, multiscale vegetation filtering and level-of-detection thresholds recover displacement zones; anomalous roughness variability distinguishes terrain later affected by failure.
6. Geometry-Aware InSAR Feedback Purification Sampling for Negative Sample Selection in Landslide Susceptibility Assessment: A Case Study in the Shigatse Region
Core Problem: Low InSAR velocity cannot be equated with stable terrain when line-of-sight geometry makes a moving slope insensitive to the satellite look direction.
Key Innovation: Geometry-aware dual-orbit feedback purifies candidate negatives before susceptibility training, improving reliability over buffer and low-deformation intersection sampling in Shigatse.
7. Flood and landslide susceptibility assessment and multi hazard interaction mapping using machine learning and GIS for sustainable settlement planning in Nepal
Core Problem: Flood and landslide susceptibility are commonly mapped independently even where shared rainfall and terrain controls create overlapping constraints on settlement.
Key Innovation: Machine-learning susceptibility surfaces built from harmonized inventories are combined in GIS to identify flood-landslide interaction zones for planning, without presenting spatial overlap as a simulated cascade.
8. Imaging Active Fault Zones in the Deep Tonga Slab
Core Problem: Deep-focus earthquakes occur where pressure should inhibit ordinary brittle faulting, so their active structures and rupture mechanism remain poorly resolved.
Key Innovation: Relative relocation and focal-mechanism clustering delineate narrow, coplanar seismic zones in the Tonga slab, including a northern plane less than about 3 km wide that is consistent with recurrent thermal shear runaway.
9. Three‐Dimensional Structure of the Los Angeles Basin and Its Underlying Moho
Core Problem: Strong-motion prediction in Los Angeles depends on three-dimensional basin and crustal geometry that sparse receiver-function arrays usually resolve poorly.
Key Innovation: Receiver functions from 273 nodes are interpreted jointly with gravity, revealing an oblong basin approaching 10 km depth and Moho uplift of as much as 45% relative to the surrounding crust.
10. Probabilistic seismic resilience assessment of a pile-supported wharf in liquefiable ground using repairability-informed damage states
Core Problem: Conventional wharf fragility states describe damage severity but not whether pile damage is accessible and repairable, weakening recovery estimates after liquefaction.
Key Innovation: Damage location and repairability are embedded in modified fragility states and propagated through stochastic recovery; resilience falls from 0.96 to 0.68 as peak ground velocity rises from 0.1 to 1.4 m s⁻¹ in the analysed system.
11. The Weather Extremes Archive (WEXA)
Core Problem: Global heat-wave and cold-spell studies lack a consistent event object that preserves duration, footprint and intensity across the reanalysis era.
Key Innovation: WEXA applies seasonally varying percentile thresholds and three-dimensional connected components to ERA5 from 1979 onward, releasing event-level diagnostics through an open archive and interactive explorer.
12. Land Subsidence-Induced Horizontal Displacement Along the High-Speed Rail in Central Taiwan: An Integrated Multi-Temporal InSAR, GNSS, and Leveling Approach
Core Problem: Subsidence along Taiwan's high-speed railway also produces horizontal convergence, which vertical-only deformation products cannot characterize.
Key Innovation: Ascending and descending Sentinel-1 SBAS-InSAR are integrated with GNSS and levelling to resolve vertical rates above 60 mm yr⁻¹ and horizontal convergence of roughly 2-10 mm yr⁻¹ along the corridor.
13. An improved shear wave velocity-based characterization model for liquefaction resistance considering confining pressure
Core Problem: The relation between small-strain shear-wave velocity and cyclic liquefaction resistance changes with confining pressure and density state.
Key Innovation: Discrete-element monotonic and cyclic tests support stress normalization of small-strain modulus and a confining-pressure correction for cyclic resistance, yielding an improved field-oriented Vs-based characterization model.
14. Kinematic Characteristics and Risk Analysis of Potential Rockfall based on 3D Point Clouds
Core Problem: Potential rockfall blocks on high, steep slopes are still identified mainly by manual field interpretation, which is difficult to reproduce at corridor scale.
Key Innovation: A point-cloud workflow detects discontinuity-bounded planar, wedge and toppling blocks, estimates source volumes and propagates representative blocks through three-dimensional runout simulations for risk analysis.
15. Numerical and experimental study on the influence of debris morphology on sliding and accumulation characteristics of dry debris flow with three-dimensional sphere DDA
Core Problem: Spherical or single-shape debris representations suppress the influence of real particle morphology on dry-debris-flow mobility and deposition.
Key Innovation: Shape-controlled laboratory tests and three-dimensional sphere DDA compare collapse, flume transport and UAV-terrain runout, showing that morphology changes collision, rolling and accumulation behaviour.
16. Potential Impulse Wave Analysis for Sejiang Deforming Slope on the Near-Dam Reservoir Bank of Bala Hydropower Station of China
Core Problem: A deforming reservoir-bank slope can generate locally amplified impulse waves whose interaction with gorge bends and dam infrastructure is not represented by empirical peak formulas.
Key Innovation: Three-dimensional CFD predicts a 5.57 m opposite-bank wave and approximately 156% bend amplification; comparison with a conservative 13.60 m empirical estimate supports three spatial protection zones without predicted dam overtopping.
17. The effect of reservoir impoundment on the deformation of the tunnel inside a reactive landslide
Core Problem: Reservoir impoundment can reactivate an old landslide around an operating tunnel, but the hydraulic threshold and coupled deformation are difficult to isolate.
Key Innovation: Monitoring, numerical simulation and theory show acceleration after water level exceeded 810 m, surface displacement of 299.1-1146.8 mm and concentrated cracking and spalling in the tunnel section crossing the landslide.
18. Njintout landslide
Core Problem: Pre-failure motion at the 2025 Njintout landslide cannot be independently examined from a final velocity map alone.
Key Innovation: The release provides PyGMTSAR SBAS-InSAR products from 29 Sentinel-1A acquisitions, including velocity rasters, epoch-wise line-of-sight displacement, point velocities with goodness-of-fit values and an interactive map.
19. Evacuation assessment under metro flooding based on 3D modelling of flood inundation dynamics
Core Problem: Metro-flood evacuation decisions depend on how quickly water reaches platforms, stairs and refuge areas, yet station-scale three-dimensional models and route assumptions are rarely released together.
Key Innovation: The Figshare package exposes inundation simulations and evacuation artefacts for Shenzhou Road Station, where modelled platform depth reaches 0.5 m within roughly 200-400 seconds under the analysed scenarios.
20. Can Neural Networks Think Like Geomorphologists?
Core Problem: Understanding the connection between topography and the processes and properties that shape landscapes both helps us better understand landscape evolution, and use topography to infer information relevant for a wide range of human activities.
Key Innovation: As a first test, we train a convolutional neural network to invert topography generated from a landscape evolution model into the model parameters, and interpret the model's learning.
21. Diurnal Pulses in Tropical Cyclones as Convectively Coupled Gravity Waves
Core Problem: The study uses convection‐permitting simulations to examine the vertical extent, propagation, and nighttime preference of diurnal pulses (DPs) in tropical cyclones (TCs).
Key Innovation: Here we leverage ensemble case study simulations of Typhoon Haiyan (2013) and Hurricane Maria (2017) to analyze DP events and quantify their structure and propagation characteristics. Results show that DPs are deep, column‐spanning features visible in multiple atmospheric variables, including precipitation, temperature, vertical velocity, and radial wind.
22. Earth Embeddings
Core Problem: Earth observation is moving from foundation models that users must run themselves toward embedding products that package model feature outputs as reusable data without needing to download and process the imagery used to generate them.
Key Innovation: The authors review their use in land cover and crop mapping, ecological and hazard modeling, socioeconomic prediction, and semantic search, with evidence on when embeddings improve on conventional features and when pooling, fusion, or spatial transfer limit performance.
23. FlowForm: Synergizing Fluid Physics with Topological Consistency for Satellite Flood Synthesis
Core Problem: Developing robust flood assessment models requires high-quality paired satellite imagery, yet such data remain scarce for flood-specific image generation.
Key Innovation: The authors propose FlowForm, a framework for satellite flood synthesis that integrates SWE-inspired latent regularization with structure-aware conditioning. Across all reported comparisons, FlowForm achieves higher visual fidelity, greater similarity between paired images, and stronger consistency of flooded regions.
24. Spatio-temporal evolution of low-magnitude seismicity before the May 24, 2013, Sea of Okhotsk earthquake recovered by waveform cross correlation. Is it an earthquake prediction case?
Core Problem: According to the International Data Centre (IDC), the Sea of Okhotsk earthquake occurred at 05:44:49.7 on May 24, 2013, had coordinates 54.89{\deg}N,153.31{\deg}E, mb=6.27, and depth of 604 km.
Key Innovation: More than 200 events obeying the Event Definition Criteria adapted by the IDC were found between May 13 and the mainshock, with a sudden increase in their occurrence rate starting on the afternoon May 19.
25. Dynamic resilience modelling of offshore mooring system under sequential tropical cyclone hazards
Core Problem: Offshore systems are exposed to diverse and potentially interacting hazards throughout their service life.
Key Innovation: The study presents a Markov-based framework for assessing offshore system resilience modelling under hazard disruption, with emphasis on progressive mooring failure combined with sequential hazards. Recovery Markov models capture stochastic performance during recovery, incorporating sequential hazards and different recovery actions (failed unit repair and motion intervention action).
26. Snow-eater heat waves of the western United States
Core Problem: However, the characteristics (e.g., area, duration, and frequency), impacts, and trends of snow-eater heat waves have received little attention.
Key Innovation: To address this gap, we developed a method to identify snow-eater heat waves and estimate their melt potential using 20th Century Reanalysis version 3 air temperature data, the TempestExtremes algorithm, and an operational snowmelt model (SNOW-17) across 1850-2015. Since the 1850s, snow-eater heat waves have increased in area and frequency, decreased in duration, and shifted earlier in the melt season.
27. Super-resolution downscaling of WRF-based typhoon wind fields over complex coastlines using a decoupled physics-guided U-Net
Core Problem: Accurate kilometer-scale typhoon wind fields are important for coastal hazard assessment, but Weather Research and Forecasting (WRF)-based dynamical downscaling remains computationally expensive, while traditional statistical methods often fail to capture localized topographic effects.
Key Innovation: The study develops a decoupled physics-guided U-Net to emulate WRF-based downscaling of hourly maximum 10-m typhoon wind fields from 25 km to 3 km resolution.
28. Monitoring agricultural drought mitigation using MODIS data: a spatiotemporal study of Henan province, 2001–2022
Core Problem: Agricultural drought represents one of the most persistent and damaging natural hazards in Henan province, posing a severe threat to regional food security and socio-economic stability.
Key Innovation: The study evaluates the efficacy of four satellite-borne remote sensing drought indices, namely, the Crop Water Stress Index (CWSI), Vegetation Supply Water Index (VSWI), Normalized Difference Drought Index (NDDI), and Temperature Vegetation Dryness Index (TVDI), against the meteorological and physical measurement-based Self-Calibrating Palmer Drought Severity Index (SC-PDSI) to determine their applicability across the province’s diverse landscapes.
29. Public perception of AI-personalized disaster alerts: Evidence from a randomized controlled trial
Core Problem: However, for AI-personalized alerts, features related to personal context, such as having dependents and higher education levels (Master's or above), emerge as relatively more influential compared to standard alerts.
Key Innovation: These results highlight both the potential and the limitations of AI-based personalization in emergency communication, emphasizing the need for careful design, transparency, and further validation in real-world settings.
30. A two-dimensional framework for building-level heat risk assessment: Refining vulnerability identification and advancing equity
Core Problem: To bridge these gaps, this study develops a novel two-dimensional, building-level heat risk assessment framework.
Key Innovation: To address this disparity, we propose targeted interventions tailored to each profile, ranging from retrofitting subsidies or optimized resource reallocation.
31. Monitoring post-fire ecohydrological recovery through integrated remote sensing and ecological modeling
Core Problem: Wildfire simultaneously disrupts ecosystem carbon, water, and soil processes, yet these coupled effects and post-fire recovery trajectories cannot be reliably assessed from spectral vegetation indices alone.
Key Innovation: Here, we developed a fine-scale, multi-dimensional framework to quantify wildfire impacts and post-fire ecohydrological recovery, using the 2016 megafire in Great Smoky Mountains National Park in the eastern United States as a testbed. We found that GPP declined by 20.7%, ET by 20.0%, WY increased by 19.2%, and soil erosion increased 12.5 times in the first post-fire year.
32. HydEquivNet: bridging classical unit hydrograph theory and data-driven hydrology via scale-equivariant learning
Core Problem: Traditional data-driven Deep Learning (DL) hydrological models often suffer from poor physical consistency and weak extrapolation robustness.
Key Innovation: To address this, we propose HydEquivNet (HEN), a novel architecture that introduces Lie group equivariance into hydrological modeling. Extensive validation on 531 Catchment Attributes and Meteorology for Large-sample Studies (CAMELS) catchments shows that HEN achieves a median Nash-Sutcliffe Efficiency (NSE) of 0.724 and avoids negative NSE values in the evaluated catchments.
33. Dynamic Landslide Susceptibility Assessment Using Machine Learning Models
Core Problem: Landslide susceptibility assessments have traditionally used static rainfall statistics that do not reflect the actual meteorological conditions when slopes fail.
Key Innovation: The study develops a machine learning framework that aligns high-resolution radar rainfall (XRAIN, 250 m / 10 min) and modeled soil moisture (XSWI) with documented landslide occurrence times as dynamic triggering factors.
34. Adaptive flood hazard index: a probabilistic framework for flood risk assessment
Core Problem: Climate change and global warming are expected to worsen in the future, significantly increasing the severity and frequency of extreme events.
Key Innovation: In this regard, flood monitoring and forecasting play a vital role in developing effective flood mitigation policies.
35. Analysing Operational Aspects of River Ice Monitoring with Earth Observation for Flood Early Warning in Canada
Core Problem: River ice jams are a major contributor to flood risk in cold regions.
Key Innovation: The authors describe the setup designed and employed operationally for the relevant agencies of Alberta, the Northwest Territories and Yukon in Canada, as well as a framework to analyse various factors relevant for operational monitoring purposes.
36. Hybrid AI framework for comparative urban seismic vulnerability and risk assessment across multiple cities
Core Problem: The study presents a hybrid AI-analytical framework for comparative seismic vulnerability and risk assessment across three urban systems with contrasting characteristics: Pohang (South Korea), Jammu (India), and Muscat (Oman).
Key Innovation: A harmonized building inventory incorporating construction period, building height, structural material, and roof type was utilized to ensure cross-city consistency. The framework captured physical damage, human losses, economic losses, and debris generation.
37. Reliance on daily mean streamflow data biases inferred flood seasonality
Core Problem: Floods are commonly analyzed using daily mean streamflow data, an approach that can obscure short-lived, high-intensity floods generated by sub-daily processes.
Key Innovation: Here, we examine how temporal resolution influences the inferred flood seasonality of peak flows. Our results show clear spatial patterns in flood regimes as well as differences between daily mean and maximum regimes within groups.
38. Testing Characteristic Magnitude Distributions in Modern PSHA Models
Core Problem: Characteristic magnitude distributions have been commonly applied to faults in probabilistic seismic hazard analysis (PSHA), and in modern models they can emerge from the way short-term seismicity constraints are combined with long-term geologic and geodetic constraints.
Key Innovation: The authors test the characteristic magnitude distribution hypothesis by comparing the fault-based magnitude distributions from the 2023 update to the National Seismic Hazard Model (NSHM23) in the Western United States with observed seismicity over the past 93 yr.
39. TS-DInSAR tool: a temporal & spatial tool for the analysis of DInSAR data at the national scale
Core Problem: National deformation archives contain hundreds of thousands of InSAR measurement points that cannot be interpreted reliably by manual inspection.
Key Innovation: TS-DInSAR introduces an automatic, scalable and unsupervised procedure for spatial and temporal analysis of horizontal and vertical P-SBAS products.
40. Unsupervised Mapping of Flood-prone Areas in Ghana Using Sentinel-1 Time-Series
Core Problem: Flooding is one of the most persistent natural hazards in Ghana, causing recurrent damage to infrastructure, livelihoods, and local economies.
Key Innovation: The study addresses this spatial gap by integrating Earth Observation (EO) datasets to identify and characterise flood-prone areas across Ghana at a national scale. Results showed that flood is concentrated in the southern half of the country, particularly in Western, Western North and Eastern Regions, and hotspots around Kumasi in Ashanti and the Weija dam in Greater-Accra regions.
41. Towards a Digital Twin infrastructure for landslides: users and data requirements
Core Problem: Our findings highlight that most existing systems function as "Digital Shadows" characterised by unidirectional data flows and a topography gap, where dynamic sensor data is superimposed onto static, outdated 3D meshes.
Key Innovation: Based on these requirements, we propose a theoretical layered architectural framework for a Data Hub designed to bridge these gaps. The increasing frequency and magnitude of landslides necessitates a fundamental shift from reactive mitigation to proactive, predictive risk governance.
42. Vulnerability Analysis of House Construction on Cut Slopes Based on Coupled Data-Driven and Physics-Based Simulations
Core Problem: To assess the risk of landslide geohazards induced by House Construction on Cut Slopes (HCCS) in mountainous areas, this study used HCCSs in the granitic slope region of Yanling County, Hunan Province, China, as case samples.
Key Innovation: An intelligent vulnerability assessment framework was developed by coupling multiple machine learning algorithms with SPH–FEM numerical simulations. As TDS increases, the energy absorbed by the wall decreases progressively, and the disaster intensity is significantly reduced.
43. Vegetation Improves Hydro‐Mechanical Soil Resilience on Humid Tropical Slopes: Field Evidence for Sustainable Land Management
Core Problem: Vegetation is widely promoted as a nature‐based solution to improve near‐surface soil conditions; however, systematic field‐based hydro‐mechanical evidence from humid tropical environments remains limited.
Key Innovation: Therefore, this study investigates the effects of vegetation on soil shear strength, moisture content and percolation rate across seven representative slopes within the Universiti Malaya campus, providing a comparative field‐based assessment of vegetated and bare slopes under humid tropical conditions.
44. Efficacy of deep replacement (cement-stabilised sand column) length in mitigating liquefaction: shaking table tests under earthquake loading
Core Problem: Deep replacement with cement-stabilized sand columns can reduce liquefaction, but the replacement length needed to control earthquake-induced deformation remains uncertain.
Key Innovation: The work provides experimental guidance on how treatment depth governs mitigation performance instead of assuming that any partial replacement is equivalent.
45. Flood nowcasting based on deep learning and radar rainfall estimates: A reliable and efficient framework for diverse flood regimes
Core Problem: Flash floods in particular are among the most destructive flood hazards due to their rapid onset, short response times, and thus limited time for warning.
Key Innovation: Overall, this study presents a scalable, data-driven approach for real-time flood nowcasting, providing a practical tool to support early warning systems and inform emergency planning in vulnerable regions. However, physics-based or conceptual hydrological models often struggle to deliver reliable short lead-time predictions, particularly in small or fast-responding catchments where complex and rapidly evolving hydrological processes are at play.
46. Integrated hydrologic and hydrodynamic modeling for flood hazard assessment in Ajay River Basin, West Bengal, India
Core Problem: The Ajay River, originating from the Chotanagpur Plateau (Bihar) and flowing through Jharkhand into West Bengal before merging with the Bhagirathi‐Hugli near Katwa, is among the most flood‐prone rivers in the region.
Key Innovation: Here, an integrated hydrologic‐hydraulic approach was adopted to better understand flood hazards, a practice that is rarely applied in this basin.
47. Integrating multi-source data and large language models in flood vulnerability curve analysis for buildings in mainland China
Core Problem: Reliable flood vulnerability assessment is critical for risk mitigation, yet mainland China faces a persistent bottleneck of historical depth-damage data scarcity.
Key Innovation: The study proposes an innovative methodology leveraging Large Language Models (LLMs) integrated with Retrieval-Augmented Generation (RAG) to develop flood vulnerability curves for buildings in China. The results indicate that building-component vulnerability is heavily dictated by typology and the indoor property loss trajectories reveal distinct spatial dependencies for different types of buildings.
48. Landslide susceptibility modelling with solely DEM-derived indicators: a data-efficient and explainable framework
Core Problem: Landslide susceptibility models often depend on geological and environmental layers that are inconsistent or unavailable between regions.
Key Innovation: The study develops a data-efficient framework using indicators derived solely from a digital elevation model and applies interpretable machine learning to identify the terrain controls behind predictions.
49. Mechanical responses, failure behaviours and stability of soft–hard interbedded antidip slopes under various degrees of anisotropy
Core Problem: Purpose This study was aimed at investigating the mechanical responses and failure characteristics of soft–hard interbedded rock samples under uniaxial compression and Brazilian splitting tests with varying strength ratios between the hard and soft layers (i.e. anisotropy degrees, λ).
Key Innovation: In addition, different combinations of the anisotropy degree (λ = 1.5, 2, 4, 6, 8 and 10), layer dip angle (θ = 0°, 15°, 30°, 45°, 60 and 75°) and slope angle (β = 70°, 75°, 80°, 85°, 90°, 95° and 100°) were designed to investigate the effects of the degree of anisotropy on the stability of SHIADSs.
50. Risk Assessment of River-Channel Washout Disasters for Long-Distance Oil and Gas Pipelines Considering Storm-Induced Flood Scour Effects
Core Problem: Existing assessment methods for river-channel washout are effective for single river cross-sections or post-disaster field investigations; however, their application remains limited when dealing with long-distance pipeline systems characterized by numerous river- and gully-crossing sections and large spatial variability in upstream catchment conditions.
Key Innovation: To address this issue, this study investigates storm-flood discharge and scour-depth calculation methods suitable for river- and gully-crossing sections of long-distance oil and gas pipelines, and establishes a quantitative evaluation index system that considers river-channel washout susceptibility, pipeline vulnerability, and pipeline failure consequences.
51. Numerical investigation structural protection measures for riverbank due to flood flow-driven damage
Core Problem: Riverbank structures can reduce local erosion while redirecting flood momentum toward other exposed assets, so protection measures cannot be judged from bank velocity alone.
Key Innovation: Three-dimensional CFD compares submerged groynes, detention zones and concrete lining; low-submergence groynes increase maximum upper-slab impact force by about 3.5 times, whereas combined detention and lining reduce loading at the banks and bridge.
52. Guivve-A/Paper: MangroveShield — IEEE ETCM 2026 camera-ready artefacts
Core Problem: Real-time estuarine flood inference requires a reproducible chain from satellite and tide inputs to calibrated classification, latency measurement and deployment.
Key Innovation: MangroveShield packages Sentinel-1 and tide-calibration artefacts, classifier validation, latency benchmarks and cloud-deployment specifications for the Guayas Estuary, while remaining a case-specific research implementation.
53. Tracing river dynamics: Multi‐temporal planform evolution and morphological change analyses in a meandering reach using high‐resolution satellite imagery and UAV data
Core Problem: The study investigates the spatiotemporal dynamics of meander evolution and morphological change along a 7 km reach of the Atrak River ’s Chat -Gonbad corridor in northern Iran.
Key Innovation: These findings offer crucial guidance for developing adaptive, reach-specific monitoring and intervention strategies.
54. Two Distinct Growth Styles of Riedel Shear Zones in Cratonic Strike‐Slip Fault System, Tarim Basin, NW China
Core Problem: The kinematics and progressive evolution of Riedel shear zones have been extensively studied, yet the factors controlling their distinct evolutionary styles in natural settings remain debated.
Key Innovation: By integrating borehole and high‐resolution 3D seismic reflection data sets, we characterized two geometric types: the F I 7 and F I 17 fault zones are dominated by right‐stepping synthetic (R) shears with partially developed secondary synthetic (P) shears, whereas the F I 5 fault zone is characterized by left‐stepping R shears and low‐angle (Y) shears, corresponding to R‐P and R‐Y shear fracture systems, respectively.
55. Volcanic Eruptions Reorganize Climate Teleconnections: Detection via Riemannian Covariance Fingerprinting
Core Problem: Standard detection and attribution of volcanic forcing targets spatial temperature anomalies.
Key Innovation: All three forcings are reorganization‐dominated ( volcanic, 0.026 anthropogenic, 0.091 solar), but at the limited effective sample size these values lie within the null spread and are not individually significant; we present the decomposition as a physically motivated diagnostic, not a demonstrated inter‐forcing contrast.
56. Vertical Coherence Dominates Decadal Land Water Storage Change Despite Hydroclimatic Extremes and Groundwater Pumping
Core Problem: Whether surface and deeper water storage trends move in the same direction underpins assessments of freshwater availability.
Key Innovation: The authors find that vertical coherence dominates, with same‐sign surface and total storage trends across >70% of the analyzed area. Theory predicts vertical coherence, yet it is untested globally and questioned under climate extremes and human management.
57. Hydrogen‐Enhanced Grain Boundary Conductivity Explains High‐Conductivity Anomalies in Trans‐Lithospheric Shear Zones
Core Problem: Magnetotelluric (MT) surveys reveal high electrical conductivity anomalies along major trans‐lithospheric shear zones.
Key Innovation: Given that trans‐lithospheric shear zones are proposed conduits for deep‐Earth volatiles, we employ molecular dynamics simulations to quantify H + diffusivity and electrical conductivity at olivine grain boundaries. This enhanced grain boundary transport increases bulk olivine conductivity by approximately two orders of magnitude relative to anhydrous conditions.
58. Subsurface Vertical Connectivity Shapes Solute Transport to Montane Streams: Insights From Watershed‐Scale Geophysics‐Informed Modeling
Core Problem: Headwater catchments provide essential water and nutrients to downstream ecosystems.
Key Innovation: Here, we test controls of three‐dimensionally resolved subsurface structure on solute transport, which reflects source waters and water residence time. Our results reveal greater vertical connectivity elongates flow paths and enhances deep groundwater contributions to streams, therefore substantially influencing the timing and magnitude of solute transport.
59. Climatology of Storm Characteristics for Sub‐Daily Heavy Precipitation in the Greater Alpine Region
Core Problem: Characterizing storms that generate heavy precipitation across different timescales is essential for parameterizing and assessing stochastic weather generators, evaluating climate models, enhancing risk management, and understanding climate change impacts.
Key Innovation: The study develops a detailed climatology of precipitation and temperature time series for storms producing heavy rainfall at sub‐hourly to daily scales across the Greater Alpine Region. Results show that storm characteristics vary systematically with both duration and geography.
60. Land Surface Feedbacks Regulate Atmospheric Water Demand and Maintain Global Drying Stability
Core Problem: Atmospheric aridity is widely projected to intensify under climate warming, driven by rapidly increasing atmospheric water demand, as reflected in potential evaporation (PE).
Key Innovation: Yet global trends in precipitation, runoff, and vegetation productivity reveal a contrasting pattern of enhanced water fluxes and widespread greening, known as the “aridity paradox.” Here we show that this apparent contradiction stems from a conceptual flaw in conventional PE formulations under a changing climate, which treat vapor pressure deficit (VPD) as an external driver, despite its strong regulation by land–atmosphere coupling.
61. Multimodal Auto-regressive Transformer Surrogate for Modeling Variable Operations and Quantifying Uncertainty in Geological Carbon Storage
Core Problem: The use of variable well perforation and injection strategies can improve the efficiency of geological carbon storage operations.
Key Innovation: The authors develop a new multimodal auto-regressive transformer surrogate to model these operations under geological uncertainty. These are fused via self-attention in a transformer encoder, and a temporal decoder generates predictions auto-regressively through encoder-decoder cross-attention.
62. Standalone DINOv3 for Training-Free Open-Vocabulary Semantic Segmentation in Remote Sensing
Core Problem: Remote sensing semantic segmentation is hindered by costly pixel-level annotations, motivating training-free open-vocabulary methods.
Key Innovation: The authors propose DinoSplat-OV, a training-free framework that adapts DINOv3 to remote sensing without fine-tuning or additional pretraining. Its Text-aware Laplacian Propagation module de-noises patch-level predictions by combining textual semantic affinities with local visual similarity, improving regional consistency while preserving boundaries.
63. FaithIR: Rethinking Infrared Image Super-Resolution from Perceptual Sharpness to Task Relevant Fidelity
Core Problem: Infrared image super-resolution (IISR) is important for downstream tasks such as object detection and semantic segmentation.
Key Innovation: To address this issue, we propose FaithIR, a faithful infrared super-resolution framework for reliable machine perception. FaithIR consists of a patch-level conditioning branch that captures global thermal and structural information and a pixel-level restoration branch that performs dense local reconstruction under structural guidance.
64. CROSS: Cascaded Distillation and Dual-Constraint Grounding for Remote Sensing Referring Segmentation
Core Problem: However, this progress largely relies on strong pre-trained capabilities, while leaving two fundamental limitations insufficiently addressed: (1) Architectural Weak-Coupling, where the unidirectional flow forces reliance on coarse VLM prompts and wastes SAM's pixel-level structural guidance, causing localization drift; and (2) Object-Centric Semantic Bias, where models overemphasize dominant object semantics while remaining insensitive to spatial reasoning crucial for RRSIS.
Key Innovation: Motivated by these observations, we propose CROSS, a tightly integrated paradigm for RRSIS. Extensive experiments on RRSIS benchmarks demonstrate that CROSS achieves state-of-the-art performance and maintains precise localization even under severe spatial description perturbations, standing as a robust new paradigm for RRSIS.
65. Distilled Roads: Generalisable Road Network Extraction Across Sensors, Resolutions, and Region
Core Problem: Road network segmentation from satellite imagery remains challenging due to large geographic variation in road appearance, occlusions, and domain shifts introduced by differing resolutions and sensors.
Key Innovation: Our framework combines cross-resolution knowledge distillation across a resolution-decreasing curriculum, multi-sensor training, and topology-aware supervision, yielding a single model that generalises across 0.3-1.0 m imagery from multiple satellite platforms across continents.
66. Compass: Degradation-Simulated Reciprocal Learning with Lightweight Needle RWKV for Multimodal Crack Segmentation under Missing Modalities
Core Problem: In multimodal crack segmentation for industrial facilities, the key challenge is preventing missing modalities from degrading pixel-level performance while maintaining low computational cost.
Key Innovation: The authors propose Compass, a lightweight network for robust crack segmentation under arbitrary missing modalities. Even with 90% depth modality missing on CrackDepth, Compass achieves F1 of 0.8216 and mIoU of 0.8434 with only 2.58M parameters.
67. Geospatial-Prior Guidance for 3D Semantic Scene Completion
Core Problem: Inferring complete 3D geometry and semantics from onboard images remains challenging because occlusions and restricted fields of view leave large scene regions underconstrained.
Key Innovation: The authors present GeoScene, a geospatially guided framework that jointly uses satellite imagery and structured OpenStreetMap cues as soft priors for 3D semantic scene completion. Experiments on SemanticKITTI and SSCBench-KITTI-360 demonstrate that GeoScene consistently improves both geometric and semantic completion under the geospatial-prior-assisted setting, with the most pronounced benefits for large-scale static and geospatially structured classes.
68. Geo-Embed: Towards Unified Multimodal Embeddings for Urban Understanding
Core Problem: Geospatial and urban applications increasingly require models to compare heterogeneous evidence across street-view imagery, remote-sensing observations, text descriptions, region proposals, and temporal change cues.
Key Innovation: However, existing multimodal embedding models and benchmarks are still largely designed and evaluated around general-purpose image-text matching, leaving unclear whether unified embedding space can support heterogeneous geospatial tasks involving spatial relationships, fine-grained semantics, and temporal changes. On GeoMEB, Geo-Embed achieves the strongest overall performance among representative multimodal embedders, with a 15.3% relative improvement over the strongest baseline.
69. UniEvo-RS: Omni-Prompt Unified Remote Sensing Segmentation with Representative Exemplar-Driven Prototype Evolution
Core Problem: Moreover, existing unified paradigms primarily rely on intra-image specific prompts, lacking flexible task routing to adapt to multi-intent operational workflows.
Key Innovation: Motivated by this practice, we propose UniEvo-RS, an omni-prompt unified RS segmentation framework equipped with representative exemplar-driven prototype evolution. By contrasting manual annotations with initial predictions on exemplars, UniEvo-RS distills prediction errors into positive and negative prototypes.
70. PRISMA: Improving the Accuracy-Latency Frontier of Diffusion-based PDE Solvers Using Physics-Informed Spectral Attention
Core Problem: Diffusion-based solvers for partial differential equations (PDEs) are often bottle-necked by slow gradient-based test-time optimization routines that use PDE residuals for loss guidance.
Key Innovation: To address these limitations, we introduce PRISMA (PDE Residual Informed Spectral Modulation with Attention), a conditional diffusion neural operator that embeds PDE residuals directly into the model's architecture via an attention-inspired modulation mechanism in the spectral domain, enabling gradient-descent free inference.
71. Rethinking Uncertainty Quantification and Entanglement in Image Segmentation
Core Problem: Uncertainty quantification (UQ) is crucial in safety-critical applications such as medical image segmentation.
Key Innovation: The authors present a comprehensive empirical study covering a broad range of AU-EU model combinations, propose a metric to quantify uncertainty entanglement, and evaluate both across downstream UQ tasks.
72. On the limits of univariate deep learning for significant wave height forecasting
Core Problem: The study conducts a systematic hyperparameter search across five deep learning architectures - DLinear, LSTM, PatchTST, ResAttLstm, and Mamba2 - and nine context lengths (1-168 h) for single-station significant wave height ( H s ) forecasting on NDBC buoy 41009, followed by re-evaluation of the best configurations on a 47-buoy, 37-year corpus.
Key Innovation: By establishing a rigorous reference baseline for what univariate H s models can and cannot achieve, this study provides a benchmark against which future multivariate and physics-informed approaches can be calibrated, and offers practical guidance for lightweight buoy-level forecasting in mid-latitude storm-dominated and swell-mixed environments.
73. Numerical study of wave impact and overtopping of defence structures on reef flats
Core Problem: Existing studies have mainly focused on the evolution of waves with periods shorter than 10 s, whereas the impact and overtopping characteristics of medium-to long-period waves have received relatively limited attention.
Key Innovation: To address this gap, a numerical wave flume was established using the coupled Finite Difference-Smoothed Particle Hydrodynamics (FD-SPH) model proposed by Wu et al. (2026), and the numerical results were validated against the experimental measurements.
74. AguaTrack-ARCO-SA: A 30-year, high-resolution atmospheric moisture tracking dataset for South America
Core Problem: However, existing moisture tracking datasets are often limited by coarse spatial resolution, short temporal coverage, or high computational demands, constraining their use for subnational-scale and short-term event analyses.
Key Innovation: Here, we present AguaTrack-ARCO-SA, a 30-year (1990-2019), daily, 0.25° resolution atmospheric moisture backtracking dataset for South America, generated using the Eulerian WAM2layers v3 model driven by ERA5 reanalysis data.
75. The TIPMIP Earth system model experiment protocol: phase 1
Core Problem: The authors describe a new Earth system model (ESM) experiment protocol, as part of the international Tipping Points Modelling Intercomparison Project (TIPMIP) project.
Key Innovation: The authors propose this as a protocol for the Coupled Model Intercomparison Project 7 (CMIP7). The protocol requires ESMs to run in CO₂ -emission mode, with atmospheric CO₂ a predicted variable.
76. Impact of seasonal snow on the recharge of a mountain karst aquifer under climate change: the Dévoluy case study (Southern Alps, France)
Core Problem: Seasonal snow strongly influences groundwater recharge in mountain aquifers, yet its role in mid-altitude karst systems under climate warming remains poorly quantified.
Key Innovation: The authors investigated the Dévoluy karst aquifer (Southern French Alps) to assess how snow controls recharge and how spring discharge may respond to rising temperatures. The model was calibrated and validated over four contrasting years (two low-snow, one high-snow, and one very high-snow year).
77. Accelerating biomass loss from forest disturbances across Europe
Core Problem: However, declines in aboveground biomass from tree cover loss remain poorly quantified, limiting our understanding of the role of disturbances in global carbon dynamics.
Key Innovation: Here we present a spatially explicit estimate of aboveground biomass losses from natural disturbances and harvest across Europe’s 216 million hectares of forests, using satellite remote sensing. From 2018 onward, aboveground biomass losses increased by 46% reaching annual values unprecedented in the preceding four decades.
78. Himalayan rivers reflect rainfall patterns, not river piracy
Core Problem: Han et al. Nature Geoscience https://doi-org.proxy.lib.umich.edu/10.1038/s41561-024-01535-w (2024)
Key Innovation: Han et al. Nature Geoscience https://doi-org.proxy.lib.umich.edu/10.1038/s41561-024-01535-w (2024)
79. Shoreline Behavior at California Groin Fields from Satellite-Based Measurements
Core Problem: Satellite imagery has helped to provide decades of shoreline change data to coastal regions all over the world.
Key Innovation: Here we apply satellite-derived shoreline techniques to three areas of California: Ventura, Santa Monica, and Newport Beach, each site containing groin fields of varying number and length.
80. Temporal-Variation-Resistant Bidirectional Convolution-Transformer GAN for Remote Sensing Image Spatiotemporal Fusion
Core Problem: However, current STF approaches still suffer from severe prediction distortion under abrupt changes, long-interval temporal variations, and land-cover type transitions, as well as poor robustness against disturbances in prior data.
Key Innovation: First, the TRBG designs a temporal-variation-resistant bidirectional encoder to capture prior information and arbitrary time-varying local–global features, enhancing prediction robustness and representation capability for time-varying information.
81. Resilience of national freight rail networks as complex systems: Efficiency, vulnerability, and investment trade-offs, and spatial governance mismatches in South Korea
Core Problem: The study examines South Korea's national freight rail network as a complex system, showing that its efficiency-vulnerability trade-off in fact operates along three separable dimensions - efficiency, vulnerability, and investment cost - compounded by spatial governance mismatches.
Key Innovation: The authors propose a "Strategic Resilience Engineering" framework combining targeted hub protection, rich-club redundancy, and governance aligned with functional economic corridors.
82. Active Learning Accelerated Knowledge Fusion of Finite Element Models for Tunnel Blasting Risk Prediction
Core Problem: Current frameworks minimize high-fidelity data needs using extensive low-fidelity datasets; however, in blast-induced overbreak risk assessment, diverse borehole configurations make acquiring even low-fidelity data costly and heavily reliant on manual modeling efforts.
Key Innovation: Compared to other fusion methods, the proposed approach achieves accuracy improvement exceeding 9% while reducing modeling costs by 300 hours.
83. Evacuation efficiency analysis of rescue strategy combinations considering rescue and guidance behaviors: Based on an extended cellular automata model
Core Problem: In emergency evacuations where real-time access to comprehensive information on the incident environment and crowd conditions is limited and rescue resources are constrained, rescue strategy selection significantly influences evacuation efficiency.
Key Innovation: The study examine the following question: Under these constraints, should rescuers prioritize Nearest Response Rescue (NR), Maintain Rescue Mission (MM), or combined strategy to maximize overall efficiency? Simulation results indicate that evacuation efficiency under different strategy combinations depends on crowd characteristics and contexts (e.g., crowd size, injured pedestrians or guides' initial distribution).
84. Identifying and mitigating worst-case disruptions in critical infrastructure systems: Models and algorithms
Core Problem: Traditional hazard-specific models inadequately capture systemic complexities and deep uncertainties of CIS disruptions.
Key Innovation: The authors further assess solution algorithms by distinguishing between exact and approximate methods, including heuristic and machine learning-based approaches. Moreover, their validation is frequently hampered by limited historical data.
85. Intelligent emergency navigation for residents near natural gas well sites: Models, algorithms, and applications
Core Problem: Preplanned evacuation routes and static signs cannot respond when toxic-gas exposure, congestion, road access and sign availability change together during an emergency.
Key Innovation: A multi-objective routing model updates travel and exposure costs from real-time conditions, while road-failure resilience metrics and conformally calibrated LightGBM screening prioritize limited variable-message-sign deployments.
86. Wind-induced structural collapse of lattice hybrid-supported wind turbine structures with blade icing
Core Problem: However, high mean wind speeds, strong turbulence, and frequent blade icing caused by low temperatures can significantly alter aerodynamic loads and dynamic responses, posing challenges to structural safety.
Key Innovation: Based on the relative displacement between the tower top and mid-section and the additional bending moment induced by the concentrated mass at the tower top, a deformation-based failure index δ was proposed, and structural collapse is defined when δ > 9.34%. Results show that blade icing reduces the along-wind aerodynamic load on blades by up to 25.41%.
87. iS3-based digital twin system for engineering geological information along rock tunnel axis
Core Problem: With the extensive development of tunnel engineering in complex geological environments, traditional approaches to geological information acquisition and management can no longer meet the demands of refined management and real-time decision-making.
Key Innovation: Based on the iS3 intelligent tunnel construction platform, this paper proposes a digital-twin-based geological information management method for tunnel alignments. The results show that, when the tunnel axis is divided into 20 m intervals for surrounding rock classification decision-making, the accuracy reaches 91.1%.
88. Physics-informed prediction of tunnel excavation-induced ground response in layered soils considering geostatic equilibrium
Core Problem: Accurate prediction of excavation-induced ground response in layered soils remains challenging because layered ground contains abrupt stiffness contrasts and gravity-induced geostatic stress gradients.
Key Innovation: To address these limitations, this study proposes a gravity-consistent multi-domain PINN framework (g-MPINN) for tunnel excavation in multilayered soils. Compared with conventional single-network PINNs, the g-MPINN improves the accuracy of stress, strain, and displacement predictions.
89. Impact of loading rate profiles and loading system stiffness contrasts on dynamic rock collapse
Core Problem: In situ, however, excavation boundaries deform under spatially non-uniform displacement fields, with axial displacement gradients that decay into the rock mass, while the loading system stiffness (LSS) of the surrounding mine environment may become increasingly compliant as extraction progresses.
Key Innovation: The response is evaluated using axial and lateral stress-strain behaviour, fracture development, specimen-scale strainburst intensity (the average ejection of breakout formations), severity (ejected mass at test termination), bulking response and rupture duration. Results show that LSS primarily controls the magnitude of dynamic instability, whereas non-uniform loading governs the spatial and temporal evolution of collapse.
90. How emergent-constrained precipitation projections reshape watershed hydrology?
Core Problem: Intensifying climate change is expected to substantially reshape watershed hydrology, yet large discrepancies among global climate models (GCMs)-particularly in precipitation projections-remain a major obstacle to reliable runoff forecasting.
Key Innovation: Here, using the Xijiang River Basin as a testbed, we develop and apply a novel model-selection-based emergent constraint (EC) framework to constrain CMIP6 precipitation projections and reduce inter-model spread by identifying models that exhibit physically consistent historical-future relationships. The EC increases projected precipitation growth rates by 7-84% relative to the raw CMIP6 ensemble while reducing their uncertainty by 6.8-15.2%.
91. A boundary adaptive optimization-based meshfree large deformation method for asphalt concrete/earth core wall rockfill dams on deep overburdens
Core Problem: A core wall rockfill dam is an important dam type on the deep overburden foundation, and its upper core wall is directly connected to the cutoff wall, which may cause the local large deformation due to the significant difference of stiffness.
Key Innovation: In this paper, a boundary adaptive optimization-based meshfree large deformation method (MFLDM) is developed, which implements adaptive truncation optimization of background meshes near model boundaries and enables more reasonable numerical integration. A typical numerical example is presented to verify the precision improvement achieved by the proposed method.
92. A novel SSI-based optimization framework for non-traditional TMDI system under near- and far-fault ground motions
Core Problem: However, neglecting soil-structure interaction (SSI) in practical applications may lead to significant design inaccuracies and reduced control performance of the N-TMDI system.
Key Innovation: To overcome this limitation, a new design framework is proposed for the tuning of N-TMDI systems installed in primary structures subjected to ground motion by explicitly considering SSI effects. The performance of the N-TMDI system is subsequently validated using real earthquake records, with a specific focus on the comparative effects of near-fault and far-fault ground motions under diverse soil conditions.
93. Centrifuge modeling of rocking foundations on sand improved with soil-cement columns
Core Problem: The study presents a centrifuge modeling investigation of the seismic performance of shallow, rocking-dominated foundations supported by soil-cement columns intended to limit detrimental settlement and rotation, while preserving the beneficial energy dissipation afforded by a rocking foundation.
Key Innovation: A baseline case was developed for a rocking footing without ground improvement by subjecting the soil-foundation system to shaking at a level that caused excessive settlement (more than 5% of its length in this study) and permanent rotation.
94. Wave attenuation and profile evolution of artificial mangrove-protected coastal highway embankments
Core Problem: Tidal inundations, storm surges, and wave-induced flooding pose a significant threat to coastal transportation infrastructure, including highway embankments.
Key Innovation: The study was conducted using large-scale laboratory experimentation. Test results demonstrated that the artificial mangrove-protected system achieved wave height reductions of about 16 to 45%, whereas this reduction was found to be only 8 to 18% in the unprotected configuration.
95. Controlling large deformation of deep soft rocks
Core Problem: This challenge significantly impedes the safe and efficient extraction of deep coal resources.
Key Innovation: To address this challenge, based on the engineering geological conditions and mining characteristics of deep coal mines under dynamic pressure, this study proposes a novel stress compensation and pressure relief (SCPR) synergetic control technology. The results demonstrate that the SCPR control technology enables multi-scale synergetic regulation of strength compensation, stress compensation, and pressure-relief compensation for the roadway surrounding rocks.
96. TMSA‐Net: Transformer‐Based Multi‐Scale Attention U‐Net for Flood Image Segmentation
Core Problem: However, the flood region segmentation is challenging due to the complex background and occlusions with debris and the effect of external adverse factors.
Key Innovation: In contrast to existing models that focus on satellite and remote sensing images, for real‐time applications, this study proposes a new Transformer‐based Multi‐Scale‐Attention U‐Net (TMSA‐Net) model for flood region segmentation in images with cluttered backgrounds. Experimental results show that TMSA‐Net achieves a test IoU of 92.01%, with improvements of 2.82% and 5.95% IoU over DeepLabV3+ and TransUNet, respectively, demonstrating the effectiveness of the proposed approach.
97. Directional seismic performance of reinforced concrete buildings under near-fault ground motions
Core Problem: Near-fault ground motions are strongly directional, whereas reinforced-concrete buildings are often assessed using record orientation as a secondary choice.
Key Innovation: Nonlinear structural analyses rotate near-fault records and compare demand across building axes to quantify orientation-dependent drift and damage.
98. Modal parameter determination and structural assessment of historical minaret under earthquake and wind loads
Core Problem: Historic masonry minarets are vulnerable to both earthquakes and wind, but their dynamic properties are often poorly constrained.
Key Innovation: Ambient-vibration measurements identify modal frequencies and shapes, which are used to calibrate a structural model and evaluate earthquake and wind response.
99. Seismic behaviour of aggregated reinforced masonry dwellings considering the combined influence of diaphragm stiffness and height irregularity
Core Problem: Aggregate dwellings are a common construction configuration in urban environments in the Americas, where space constraints favour terraced construction.
Key Innovation: The study analyses the seismic behaviour of clusters of terraced dwellings built with reinforced masonry using partially grouted concrete blocks (PG-RCM) through non-linear dynamic analysis.
100. Seismic fragility models of highway reinforced concrete bridges between cities considering the effects of aging and time-varying degradation
Core Problem: Regional bridge-loss estimates require fragility curves that evolve as reinforced-concrete bridges age, rather than treating capacity as constant over their service life.
Key Innovation: The study develops seismic fragility models for intercity highway bridges while representing deterioration as a time-varying process.
101. Sensitivity Analysis of Peak Rate Factors for Floods Assessment in the Wadi Ibrahim Watershed
Core Problem: The study investigates flood behavior in the Wadi Ibrahim watershed by evaluating the sensitivity of flood estimates to different Peak Rate Factor (PRF) values, with a focus on their influence on peak discharge (Qp), time to peak (tp), volume (V) and inundation depth.
Key Innovation: The study highlights that PRF selection and model choice significantly influence flood predictions, emphasizing the importance of using arid-zone-adapted models to ensure reliable flood risk assessment and resilient infrastructure planning.
102. A synthetic flow network decision-making tool to identify potential downstream contamination risk in permafrost-bound Alaskan hazard sites
Core Problem: Permafrost degradation can mobilize contaminants from remote Alaskan hazard sites, yet sparse observations make downstream pathways difficult to prioritize.
Key Innovation: The output is a screening instrument for monitoring and remediation, not a substitute for site-specific contaminant transport measurements.
103. Earthquake and Multi-Hazard Resilience: Community-Level Insights and AI/ML Applications
Core Problem: Earthquake resilience is increasingly a problem of interactions across hazards, assets, infrastructure systems, and recovery processes [...]
Key Innovation: Earthquake resilience is increasingly a problem of interactions across hazards, assets, infrastructure systems, and recovery processes [...]
104. Flood and drought risk management across the hydroclimatic spectrum in the most affected regions under changing water resources conditions
Core Problem: Floods and droughts are the most damaging and costly water-related hazards globally, and their frequency and severity are accelerating under changing climatic conditions.
Key Innovation: This perspective examines the evolving landscape of flood and drought risk management in the most affected areas, notably South Asia, Sub-Saharan Africa, East Asia, Southeast Asia, Europe, and Latin America.
105. Real Estate Exposure to Seismic and Subsurface Risks
Core Problem: Evidence connecting seismic and subsurface hazards to built-environment and real-estate exposure is fragmented across geotechnical, structural, planning and valuation research.
Key Innovation: A systematic review organizes 55 studies into seismic risk, building vulnerability, subsurface hazards, urban and heritage vulnerability, and mitigation and resilience, exposing integration and data gaps across these domains.
106. The Post-Landslide Collapse in Small Italian Villages
Core Problem: Small settlements damaged by repeated landslides need long-term spatial and heritage planning after emergency response, but transferable recovery methods remain limited.
Key Innovation: The Monacilioni case integrates landslide history, territorial interpretation, cultural heritage and landscape design into a post-disaster planning method intended for other declining small villages.
107. Groundwater Variability Is Strongly Linked to Soil Temperature Across Global Aquifers
Core Problem: Groundwater plays a critical role in land‐atmosphere interactions by regulating soil moisture and surface energy fluxes, yet its contribution to land‐surface processes and heat extremes remains poorly quantified globally.
Key Innovation: Here, we combine observation‐based aquifer records with gridded groundwater table depth (GWD) model outputs (2000–2015) to assess how GWD annual fluctuations covary with soil moisture, evaporation, soil temperature, and heatwave frequency across 1,658 aquifers worldwide.
108. Multidecadal Changes in ENSO Drive a Substantial Decline in West Antarctic Sea Ice Predictability
Core Problem: ENSO teleconnections are a key source of Antarctic sea‐ice predictability, particularly in the Antarctic Dipole (ADP) region, but whether this predictability remains stable under the recent transition toward Central Pacific (CP) El Niño is unclear.
Key Innovation: Here, using a Markov model, we reveal an 83% decrease in sea ice concentration (SIC) predictability, measured by the anomaly correlation coefficient (ACC), during austral winter and spring since 2002.
109. Time‐Varying Hydroclimatic and Oceanic Controls on Arctic Submarine Groundwater Discharge Inferred Using Explainable AI
Core Problem: Submarine groundwater discharge (SGD), an important component of coastal water and nutrient budgets, is challenging to monitor and predict in the Arctic given the remoteness and harsh conditions.
Key Innovation: Here, we used explainable artificial intelligence to quantify the time‐varying importance of hydroclimatic and oceanic drivers of SGD at an Arctic beach from the early thaw period to late summer. Deep learning models were trained on in situ observations and reanalysis data, and feature contributions to model prediction were quantified using SHAP (Shapley Additive Explanations).
110. Pore Size Effect on Melting Point in Saturated Frozen Soil Systems: A Pressure Perspective at Molecular Scale
Core Problem: Existing models remain insufficient for describing water phase change in smaller pores (i.e., nano pore), where surface effects become significant and water molecules are inhomogeneously distributed near solid surface.
Key Innovation: The study investigates the pore size effect on melting point from a molecular scale perspective and introduces a local pressure variable, namely lateral pressure. Further analysis shows that lateral pressure increases as pore size decreases, corresponding to lower melting points.
111. Qwen-3D: A Generalist 3D Vision-Language Model for Spatial Understanding
Core Problem: Large Multimodal Models (LMMs) have achieved remarkable success on images and short videos, yet scaling them to long videos remains challenging due to frame-centric tokenization and limited context windows.
Key Innovation: The authors argue that a key limitation is geometry-aware decoding: existing methods communicate 3D predictions through language tokens, proposal selection, or lightweight grounding queries, creating a bottleneck between language reasoning and dense geometric prediction.
112. LiteMVS: Efficient Multi-View Stereo with Foundation Distillation and Expert Aggregation
Core Problem: Existing multi-view stereo (MVS) methods primarily rely on geometric correspondences, which often fail in textureless or repetitive regions, while monocular depth models leverage strong image-level priors but lack robust multi-view geometric constraints.
Key Innovation: To obtain visual representations with stronger structural awareness and greater potential for spatiotemporal extension, we present LiteMVS, a lightweight multi-view depth estimation model that integrates plane-sweep geometric reasoning with strong monocular semantic and structural priors.
113. HyVIC: A Metric-Driven Spatio-Spectral Hyperspectral Image Compression Architecture Based on Variational Autoencoders
Core Problem: However, existing methods typically adapt variational image compression models designed for natural images, without adequately accounting for the distinct spatio-spectral redundancies inherent in HSIs.
Key Innovation: Based on our results, we offer insights and derive practical guidelines to guide future research directions in learning-based variational HSI compression in RS. To achieve this, we introduce spatio-spectral variational hyperspectral image compression architecture (HyVIC), a configurable variational autoencoder (VAE) for HSI compression.
114. Tarot-SAM3: Training-free SAM3 for Any Referring Expression Segmentation
Core Problem: Existing RES methods, however, rely heavily on large annotated datasets and are limited to either explicit or implicit expressions, hindering their ability to generalize to any referring expression.
Key Innovation: To this end, we present Tarot-SAM3, a novel training-free framework that can accurately segment from any referring expression. Extensive experiments demonstrate that Tarot-SAM3 achieves strong performance on both explicit and implicit RES benchmarks, as well as open-world scenarios.
115. From Forest to Future Capital: Tracking Land Cover Change in Ibu Kota Nusantara (IKN) from 2021 to 2026 with PlanetScope Imagery
Core Problem: Indonesia's relocation of its political and administrative capital from Jakarta to Ibu Kota Nusantara (IKN) has been framed around a Forest City vision, yet rapid construction within the Core Government Area (KIPP) raises concerns over land conversion, vegetation loss, and carbon stock decline.
Key Innovation: The study applies remote sensing techniques to systematically assess land use and vegetation cover change in KIPP from 2021 to 2026 using PlanetScope SuperDove satellite imagery. Results show substantial environmental transformation, with mean NDVI declined by 17.1%, total carbon stock decreased by 0.28%, developed land expanded by 672%, and total vegetation declined by 18.1%.
116. GLBD-FED: a global first-hand in-situ daily temperature dataset preferentially with a 00:00–24:00 UTC 24 h window (1981–2024)
Core Problem: However, converting these data into a global daily temperature dataset with a uniform definition - especially for daily maximum ( T max ) and minimum ( T min ) temperatures - has proven challenging due to the independent observation schedules across the world.
Key Innovation: To address this issue, we developed a new method that decomposes sub-daily T max and T min records from the Integrated Surface Database (ISD) into finer intervals, subsequently reaggregating them into daily T max and T min based on a prospective 00:00-24:00 UTC dateline.
117. DReaMIT: a dynamical reanalysis framework for modelling surface-based temperature inversions in cold environments
Core Problem: Surface-based temperature inversions (SBIs) are critical to high-latitude mountain climatology, shaping permafrost stability and near-surface thermal regimes.
Key Innovation: The study develops and evaluates a new surface-based inversion model, DReaMIT (Dynamical Reanalysis Model for Inversions of Temperature), that extends the framework of Pozsgay and Gruber ( 2025 ) by spatializing the inversion strength parameter ( α ) using hypsometric position rather than absolute elevation. The reformulated approach enables a unified calibration across two contrasting Yukon valleys (WS01 and WS02), improving model transferability and reducing site-specific bias.
118. New classes of climate model emulators to improve paleoclimate reconstructions
Core Problem: Reconstructing spatial climate variability from proxy records requires forward models “emulators” that capture the dynamical structure of the climate system while remaining computationally efficient.
Key Innovation: Here we develop and evaluate a hierarchy of CMIP-class climate model emulators, for annual surface air temperature field emulation, that integrate autoencoder-based dimensionality reduction with nonlinear prediction architectures, including Reservoir Computing (RC) and Recurrent Neural Networks (RNNs).
119. Near real-time estimation of daytime and nighttime evapotranspiration using GOES-R observations and machine learning models
Core Problem: However, most satellite-derived ET products are limited to daily or coarser temporal resolutions, despite the strong diurnal variability of ET processes.
Key Innovation: The authors test Gradient Boosting Regression (GBR) and Long Short-Term Memory (LSTM) models to assess their ability to estimate ET variations across the diurnal cycle. GBR captures daytime ET with an R-squared of 0.74 (normalized RMSE of 0.91) while maintaining low computational cost.
120. Hysteresis between groundwater and surface water levels indicates the states of hydrological turnover affecting solute transport and redox processes
Core Problem: Small streams are highly sensitive to variations in discharge, a sensitivity predicted to increase in future climate scenarios, impacting ecological health of streams and water management practices.
Key Innovation: The study examines the relationship between hydrological turnover (HT) and stream-stage-groundwater-level hysteresis patterns under various system states in a third-order tributary of the River Mosel in Trier, Germany, using high-resolution stream-stage and groundwater-level data (GW1, GW2) together with complementary chemical observations collected over two years. Our results reveal distinct seasonal dynamics in GW-SW exchange.
121. FDM-Net: A Multi-Level Feature Aggregation Network Based on Frequency-Decomposition for Hyperspectral Image Classification
Core Problem: However, existing Mamba–CNN hybrid frameworks typically adopt a parallel-branch architecture where identical spectral–spatial information is fed into both branches, failing to rectify the inherent frequency-specific bias: Mamba tends to prioritize low-frequency information, while CNNs excel at capturing high-frequency details.
Key Innovation: To handle these limitations, a novel multi-level Aggregation Network based on Frequency Decomposition (FDM-Net) is proposed. Meanwhile, a Multi-Level Feature Aggregation Module (MLFA) leverages multi-level depthwise convolutions and gated aggregation to capture complex multi-level interactions in high-frequency features.
122. Agreement–Disagreement Guided Knowledge Transfer for Cross-Scene Hyperspectral Imaging
Core Problem: However, existing studies often overlook the challenges of gradient conflicts and dominant gradients that arise during the optimization of shared parameters.
Key Innovation: To address these issues, we propose an Agreement–Disagreement Guided Knowledge Transfer (ADGKT) framework that jointly models optimization consistency and representation diversity for heterogeneous cross-scene HSI classification.
123. PCFD-Net: A Parallel Collaborative Fusion-Detection Network for SAR and Optical Imagery
Core Problem: However, effectively integrating these two tasks within a unified training framework remains challenging.
Key Innovation: To address these issues, we propose PCFD-Net (Parallel Collaborative Fusion-Detection Network), which consists of a fusion branch, a detection branch, and a bidirectional interaction branch, and unifies fused image generation and oriented object detection within a single training framework through explicit bidirectional interaction.
124. Airborne Point Cloud Fusion with Local Plane Constraints for Advanced Semantic Consistency
Core Problem: However, integrating airborne hybrid sensors or cross-source point clouds remains challenging due to variations in geometric accuracy, data precision, gaps, and sensor attributes.
Key Innovation: The authors propose a new point cloud fusion algorithm that leverages local plane constraints to achieve advanced semantic consistency.
125. SPFMamba: A Mamba-Based Network with Semantic Prompt and Frequency-Adaptive Fusion for Remote Sensing Image Semantic Segmentation
Core Problem: Modern remote sensing images (RSIs) provide increasingly fine spatial detail, making pronounced scale variations and complex spatial distributions of land-cover classes more apparent and thereby increasing the difficulty of semantic segmentation.
Key Innovation: Accordingly, we develop SPFMamba, an architecture built around Mamba that combines semantic prompting with frequency-adaptive fusion, thereby improving global context modeling and fine-detail representation.
126. Penetration Depth Investigation of L-Band and S-Band SAR Signals into Soils and Hard Ground Surfaces
Core Problem: However, direct measurements of penetration depth under controlled material conditions remain limited, especially for comparisons across radar bands, soil water content, sand, and hard ground surfaces.
Key Innovation: The study provides direct laboratory measurements of L-band and S-band SAR signal penetration using a ground-based SAR system in a microwave anechoic chamber. These results show that SAR penetration depends strongly on wavelength, water content, material type, and compaction/surface condition.
127. “It's how you say it”: Text-based measures of social capital and how it shapes public evaluations of environmental risk perceptions
Core Problem: The study examines how different dimensions of social capital shape individuals' perceptions of environmental health risks.
Key Innovation: The authors introduce an approach that leverages text-based measures to capture individuals’ emotional attachment to their communities.
128. The perils of risk communication in a context of uncertainty: The long dispute over contamination after the Grenfell Tower fire
Core Problem: Disasters, in disrupting societies' normal coping capacities, produce uncertainties about secondary risks and authorities' capacities to protect communities.
Key Innovation: The authors present a case study of the long dispute over the potential health-damaging contamination of air and soil after the Grenfell Tower fire in London, UK, drawing on ethnography and document analysis.
129. Development of a dynamic construction site fire occurrence likelihood index based on fire probability and cumulative frequency using meteorological data
Core Problem: Construction sites face high fire vulnerabilities due to dynamic environmental exposures but current assessments rely on subjective static checklists.
Key Innovation: The study proposes an objective dynamic fire occurrence likelihood index utilizing public meteorological data to overcome these limitations. The methodology synthesizes fire occurrence probability risk derived from an XGBoost algorithm using four weather variables and cumulative fire frequency risk calculated via an optimized 54-day rolling sum to capture historical environmental stress.
130. Fire evacuation design optimization of public buildings based on building information modeling semantic enrichment for fire simulation requirements
Core Problem: The complex structures and dense, diverse populations of public buildings create severe fire evacuation challenges, necessitating design optimization to improve safety.
Key Innovation: Based on the fire simulation results generated from the semantically enriched BIM model, an evacuation simulation model is constructed. The results show that the framework can support BIM-based fire simulation, evacuation performance assessment, and layout modification under the defined case-study conditions.
131. Innovative machine learning-based prediction of time-to-failure and domino fire escalation in chemical storage tanks
Core Problem: Detailed simulation of interacting storage-tank fires is too slow for rapid estimates of time to failure and escalation probability across many domino scenarios.
Key Innovation: A physics-informed fire-interaction model generates training scenarios for Tab-Transformer, CatBoost and neural-network surrogates; uncertainty intervals accompany rapid predictions, although validation remains simulation based.
132. Extending the MAIAC algorithm for hyperspectral atmospheric correction of TROPOMI measurements: A case study over vegetated regions
Core Problem: Hyperspectral surface reflectance (SR) is a critical input for retrieving atmospheric composition (e.g., aerosols and trace gases) and supports a wide range of land applications.
Key Innovation: The authors present a Multi-Angle Implementation of Atmospheric Correction (MAIAC)-based simultaneous retrieval framework for aerosol optical depth (AOD) and hyperspectral SR over land from the TROPOspheric Monitoring Instrument (TROPOMI) aboard Sentinel-5 Precursor. TROPOMI AOD shows good agreement with AERONET at most sites (with correlation coefficients typically exceeding 0.7), with larger discrepancies observed over optically bright and heterogeneous surfaces in the western United States.
133. Beyond RMSE: The GEDI imputed waveform product minimizes artificial homogeneity evident in existing global forest height maps
Core Problem: However, when predictor data do not explain all height variability in the training sample, this objective function leads to prediction toward the mean, distorting population-level prediction of height range and variability.
Key Innovation: Several existing forest height maps have used training data from NASA's Global Ecosystem Dynamics Investigation (GEDI) lidar mission, and while the mission's retrievals are subject to measurement error, we used GEDI to: 1) evaluate population-level errors of three prominent global height maps; and 2) produce new maps designed to better represent the full height distribution. Validation shows that L4D top height RMSE values were up to approximately 1.1 m greater than other maps.
134. Retrieval of sea surface salinity from SMAP L-band radiometry under rainfall conditions
Core Problem: Sea surface salinity (SSS) retrieval during precipitation remains challenging for L-band radiometry because rainfall alters both the atmospheric radiative transfer and sea surface emission, resulting in rain-related fresh biases.
Key Innovation: The study presents a physically based correction framework for SMAP-derived SSS. Across three independent near-surface validation datasets (drifters, Surface Salinity Snake, and Wave Gliders), the approach consistently reduces rain-induced fresh bias to within ∼ ± 0.16 psu and decreases RMSE by ∼15-55% relative to the uncorrected product.
135. A loop-outlier-aware Riemannian pose-graph optimization framework for robust visual localization
Core Problem: Robust visual localization is a fundamental component of image-based mapping, SLAM, and 3-D scene understanding, especially when long-range loop closures, poor initialization, and inconsistent measurements affect global geometric consistency.
Key Innovation: The study presents RiLO-PGO, a lightweight robust pose-graph optimization framework that improves the backend stage of visual localization without redesigning the full front-end pipeline.
136. Graph-Regularized Point-to-Trajectory Association for RS-AIS fusion toward maritime moving-ship identification and spatial refinement
Core Problem: Remote sensing (RS) satellites provide wide-area observations but lack explicit vessel identity information, while the Automatic Identification System (AIS) offers rich static and dynamic attributes but suffers from incomplete availability, reporting latency and noise.
Key Innovation: The study proposes a Graph-Regularized Point-to-Trajectory Association (GR-PTA) framework that formulates RS-AIS fusion as a point-to-trajectory association problem and introduces graph regularization to enforce trajectory-neighborhood coherence. Experiments across five representative maritime regions demonstrate that GR-PTA achieves an average Precision of 0.89, Recall of 0.96, and F1-score of 0.92, consistently outperforming strong baselines.
137. V-BUILD: A graph Transformer-based decision framework for selective updating of vector building data
Core Problem: Keeping vector building databases current requires deciding which database operations should be applied, rather than merely detecting binary geometric differences.
Key Innovation: The authors present V-BUILD, a hybrid decision framework that assigns each matched building unit one of four operations: addition, deletion, replacement, or preservation.
138. Urban digital twin for environmentally sensitive mobility planning: conceptual framework and application in pilot region Leipzig
Core Problem: The study presents the development and pilot implementation of a conceptual, practice-oriented Urban Digital Twin (UDT) used for environmentally sensitive mobility management within the pilot region of Leipzig (Germany).
Key Innovation: The results show the potential of UDTs to support responsive, data-driven decision-making using the developed workflow manager.
139. Structural damage and stiffness degradation model for artificially frozen silty clay considering unfrozen water content
Core Problem: However, the complex mechanical behavior of frozen soil, influenced by factors such as temperature and stress history, remains a significant research focus.
Key Innovation: The study establishes a damage model capable of accurately describing the cementation failure and strength degradation of frozen soil during the loading process. The results indicate that: (1) As temperature decreases, the unfrozen water content decreases, following an approximate power-law relationship with temperature; lowering the temperature significantly increases the initial cementation area.
140. Growth mechanism and prediction method of ice accretion on bridge deck pavements in low-temperature rainfall environments
Core Problem: In low-temperature rainfall environments (air temperatures near 0 °C with rainfall), bridge deck surfaces are more prone to icing than adjacent conventional road sections due to the lack of heat storage from the base and subgrade, coupled with rapid heat dissipation.
Key Innovation: Based on surface free energy and heat transfer theories, the coupled spreading and heat transfer process of raindrops was analyzed to elucidate the factors influencing ice accretion growth and to establish an ice coverage thickness prediction model.
141. Fracture water migration mechanism in single-fractured rock mass under low-temperature
Core Problem: Cold region projects are confronted with the frequent challenge of freeze-thaw disasters, and the root cause lies in the insufficient understanding of the water migration characteristics of fractured rock masses.
Key Innovation: Based on experimental data and the theory of water migration in frozen soil, a hydraulic aperture evolution equation for low-temperature rock masses and a single-fracture seepage model, both considering the influences of temperature, phase change, and mechanics, are established. The research results can contribute to the improvement of the hydro-thermo-mechanical coupling theory for low-temperature fractured rock masses.
142. Evacuation behavior and safety egress analysis in tunnel fires: a virtual reality simulation study
Core Problem: Tunnel evacuation models often prescribe reaction times and exit choices, leaving limited evidence for how smoke, warning timing and proximity to fire alter actual decisions.
Key Innovation: A full-scale virtual-reality tunnel experiment coupled with CFD varies smoke, alarm timing and fire distance, showing that poor visibility can cause occupants to bypass a nearby exit despite relying on exit signs.
143. An anisotropic Mogi failure criterion for transversely isotropic rocks
Core Problem: Accurately predicting the true triaxial strength of transversely isotropic rocks (TIRs) remains a fundamental challenge in rock mechanics.
Key Innovation: To address these limitations, this study proposes a continuous anisotropic Mogi failure criterion. A dynamic nonlinear parameter dependent on hydrostatic pressure is also added to capture the transition from low-pressure weak-plane sliding to high-pressure cross-weak-plane failure.
144. How to overcome hydrologic data scarcity in assessing potential groundwater risk to subterranean infrastructure: a finite-difference modelling approach
Core Problem: Groundwater infiltration (GWI) in sewer networks is an increasing challenge under climate change and human pressures, threatening the sustainability of sewer infrastructure.
Key Innovation: The study develops a finite-difference groundwater flow model (MODFLOW) to identify GWI susceptibility hotspots in non-pressurised sewer networks in Dawlish, southwest United Kingdom. The model was calibrated for October 2007 to September 2020 and validated over the following four years using monthly time steps, achieving a mean error of − 0.05 m, a mean absolute error of 0.94 m, and a root mean square error of 1.62 m.
145. Quantifying hydrological responses to climate change and associated canopy variability using an integrated LSTM-SWAT modeling framework
Core Problem: Climate change and climate-driven canopy variability jointly influence terrestrial hydrological processes, while their combined effects on individual water-balance components across complex terrains remain insufficiently quantified.
Key Innovation: The study integrated a modeling framework by coupling a Long Short-Term Memory (LSTM) network with the Soil and Water Assessment Tool (SWAT) to assess scenario-based hydrological responses in the Weihe River Basin (WRB) through 2100.
146. Evaluating the impact of climate change on water supply system reliability – a robust assessment framework using stochastic data
Core Problem: The study presents a stochastic data-based framework that enables dynamic and uncertainty-aware evaluation of water supply system performance under climate change and climate variability, and supports the determination of optimal timing for system augmentation.
Key Innovation: Under an extremely hot and dry climate scenario, the proposed augmentation, framed as a 20% reduction in annual demand, helped to alleviate system restrictions, and early augmentation was more effective than late augmentation.
147. Physics-constrained neural network modeling of soil thermal conductivity curve across full saturation
Core Problem: Unlike point-wise estimates at selected water contents or at the saturated endpoint, a continuous λ(Sr) curve over the full saturation range is needed to represent soil heat-transfer behavior from dry to saturated conditions.
Key Innovation: To address these limitations, this study introduces a Physics-Constrained Neural Network (PCNN) to learn a continuous, soil-specific λ(S r) curve over 0 ≤ S r ≤ 1 while applying established physical constraints. The MLP achieves lower error on the training dataset and a slightly lower testing RMSE, whereas the PCNN provides comparable testing accuracy while producing smoother and more physically consistent derivatives behavior.
148. Modeling and prediction of progressive salinization and deformation of unsaturated subgrades in saline seasonally frozen regions
Core Problem: Progressive salinization is a concealed deterioration mechanism for unsaturated highway subgrades in saline and seasonally frozen regions, where initially non-saline or slightly saline fills may gradually accumulate salts during service.
Key Innovation: The study develops a coupled thermo-hydro-salt-mechanical numerical model to investigate the spatiotemporal evolution, deformation response, and long-term prediction of progressive subgrade salinization under seasonal temperature variations.
149. Micromechanical modeling of elastoplastic damage behavior in rock-like materials with multiscale heterogeneity
Core Problem: Heterogeneous rock-like materials with a dual-scale “porous matrix + mineral grain” microstructure are widespread in rock engineering, where pore dimensions are far smaller than grain size.
Key Innovation: To quantitatively link microstructural features to macroscopic mechanical behavior, this paper develops a novel micromechanics-based elastoplastic damage constitutive model via a two-step homogenization scheme, addressing the limitation of traditional phenomenological models that cannot explicitly incorporate material compositional information.
150. A bedded rock stress corrosion model
Core Problem: A bedded rock stress corrosion (BRSC) model is proposed to capture the anisotropic time-dependent mechanical behavior and progressive damage of bedded rock, such as shale.
Key Innovation: The BRSC model is verified and then employed to reproduce the anisotropic creep behavior of shale, leading to the following conclusion: shale samples with horizontal bedding planes exhibit significantly more creep resistance, creep life, and damage tolerance than samples with vertical bedding planes.
151. Transient response of a circular tunnel with eccentric lining under plane P waves
Core Problem: However, most existing studies have been focused on the steady-state response of concentric lined tunnels under harmonic wave incidence, whereas the dynamic stress concentration mechanism of eccentric lined tunnels under arbitrary transient loading has not yet been systematically understood.
Key Innovation: Therefore, an analytical model based on the wave function expansion method is established for the dynamic response of a circular tunnel with an eccentric annular lining subjected to incident plane P waves. The results show that the local transient response of the eccentric lining can be amplified or attenuated by the high-frequency spectral components of the incident wave.
152. Dynamic behavior of sand-rubber mixtures: A CT image-based DEM study
Core Problem: The dynamic behavior of two sand-rubber mixture (SRM) types (irregularly shaped vs. spherical grains) under cyclic triaxial testing was investigated, focusing on the effects of rubber content, particle shape, and shear strain amplitude.
Key Innovation: A fitting model linking SRM damping ratio to that of high-stiffness (sand-sand) and low-stiffness contacts was established, which reflects how specimen damping ratio varies with rubber content and shear strain amplitude, and further elucidates particle shape effects. An X-ray computed tomography (CT) image-based discrete element model (DEM) was used for the investigation, which was calibrated and validated using real SRM specimens subjected to triaxial compression with real-time CT scanning.
153. Freeze-thaw damage evolution and crack propagation failure mechanism of red sandstone under non-uniform water distribution conditions
Core Problem: To reveal the freeze-thaw damage evolution and failure mode transformation mechanism of red sandstone under non-uniform water distribution conditions, repeated freeze-thaw cycle tests were conducted at different immersion heights.
Key Innovation: The spatial distribution and migration process of water were characterized using nuclear magnetic resonance (NMR) with the GR-HSE sequence and magnetic resonance imaging. The results show that the immersion height dominates the internal spatial water distribution of red sandstone, and the lower the immersion height, the more non-uniform the axial spatial water distribution.
154. Thermo–hydro–mechanical simulation of frost-heave mitigation measures for railway culverts
Core Problem: In Norway, railway culverts are essential for drainage and track continuity but also promote airflow and heat exchange that increase frost penetration in high-latitude regions.
Key Innovation: The study first validated THM simulations of frost heave in railway subgrades with and without culverts using field observations and measurement data from cold regions in Norway, Sweden, China, and Canada.
155. Weak-link minerals in crystalline rock: How mica cleavage orchestrates fracture propagation?
Core Problem: However, when numerically modeling mica-bearing rocks, conventional grain-based models (GBMs) tend to oversimplify mica as isotropic grains similar to quartz and feldspar, thereby failing to capture cleavage-controlled propagation and crack deflection.
Key Innovation: To bridge this gap, we propose a computed tomography (CT)-derived cleavage grain-based modeling (CT-CGBM) method that integrates high-resolution X-ray CT with nanoindentation to explicitly reconstruct the laminar cleavage structure of mica. The proposed model accurately reproduces the experimental stress-strain response (error <5%) and failure morphology, while effectively capturing complex mica-related cracking.
156. Interlayer effect on the rock failure and structural ring characteristics of tunnels in high geostress
Core Problem: Unfavorable geological conditions are one of the major concerns causing the tunnel failure, among which the overlying interlayer is a common discontinuity.
Key Innovation: Thereafter, the structural ring concept was introduced to quantify the load-bearing effect of the surrounding rock. The proposed method was validated by the model test results on homogeneous rock, based on which the interlayer effect on the spatial distribution and evolution characteristics of the structural ring was further revealed.
157. Stochastic tunnel convergence in spatially variable medium using rotated and non-rotated anisotropic random fields with adaptive surrogate model
Core Problem: Deterministic analysis of tunnel convergence completely neglects the uncertainties present in geotechnical parameters, while conventional probabilistic approaches treat inputs as random variables but often ignore the spatial variability of soil.
Key Innovation: The study presents a probabilistic framework for tunnel convergence analysis that incorporates spatial soil variability via an adaptive sparse polynomial chaos expansion (SPCE) surrogate model, further enhanced by global sensitivity analysis (GSA) to improve computational efficiency.
158. High-resolution land cover mapping with GeoAI: instance segmentation for land cover analysis
Core Problem: Land cover classification has become a key method for understanding natural and ecological resources, as well as for sustainable land-use planning and management.
Key Innovation: The study investigates the potential of instance-based segmentation within a GeoAI workflow for high-resolution land cover classification in the area of San Vito di Cadore (Veneto), a UNESCO mountain region with high ecological heterogeneity. The results show an overall precision of 0.847 and an overall recall of 0.575, with mAP@0.5 greater than 0.65.
159. Identifying hydrological indices derived with datasets from remotely sensed products to monitor climate change and variability impacts on water resources: a systematic review
Core Problem: Hydrological indices are widely used to assess climate change and climate variability impacts on water resources; however, their application is often constrained by limited in-situ observations, particularly in data-scarce regions.
Key Innovation: Advances in remote sensing have enabled the development of hydrological indices derived from satellite and reanalysis datasets, providing new opportunities for monitoring droughts, water storage dynamics, and hydrological extremes. The findings indicate that optical sensors (Landsat, MODIS, Sentinel-2), precipitation products (CHIRPS, TRMM, IMERG), gravity missions (GRACE and GRACE-FO), and reanalysis datasets (ERA5-Land, GLDAS, FLDAS) were the most frequently used products.
160. Wildfire Hazard for Seveso Installations
Core Problem: Wildfire can initiate technological accidents when flames, heat or embers reach installations regulated under the Seveso framework.
Key Innovation: The study examines wildfire exposure of these industrial sites and organizes the factors needed to screen escalation potential at the wildland-industrial interface.
161. AI Decision Support for Urban Fire Risk Management: A Framework for Validation, Governance, and Bounded Deployment
Core Problem: AI-based decision support is moving into fire practice and governance, where it is used to prioritise inspections, analyse building and community risk, examine station coverage, support evacuation planning, interpret warnings, and explore fire scenarios.
Key Innovation: These tools can extend analytical capacity, but they also create a decision role migration problem: an output developed for prediction, prioritisation, warning, simulation, or planning may later be treated as clearance, justification, or authority.
162. FedRings: A Scalable and Topology-Aware Federated Learning Framework for LEO Satellite Constellations
Core Problem: Federated learning over low Earth orbit (LEO) satellite networks is limited by frequent link changes, short contact times, and a highly dynamic topology, making centralized or synchronized training inefficient and hard to scale.
Key Innovation: To address this, we propose FedRings, a decentralized framework that organizes satellites into ring-based communication structures. By combining topology-aware routing, communication scheduling, and efficient aggregation, FedRings enables stable and efficient learning in dynamic LEO networks while reducing communication cost, and experiments show it consistently outperforms existing methods in realistic settings.
163. Efficient unsupervised domain adaptation via self-supervised vision transformer and synergistic cross-domain alignment
Core Problem: Despite recent advances, existing methods often rely on fine-tuning large backbone models, which leads to high computational cost and limits scalability in resource-constrained environments.
Key Innovation: To address this question, we propose Efficient Unsupervised Domain Adaptation (EUDA), a parameter-efficient framework that leverages a frozen DINOv2 backbone as a feature extractor and updates only a lightweight bottleneck and classification head. Experimental results on Office-Home, Office-31, VisDA-2017, and DomainNet demonstrate that EUDA achieves competitive performance across diverse domain complexities, while reducing the number of trainable parameters by 42 to 99.7%.
164. UMS-DETR: A lightweight Transformer-based framework for real-time ship detection in complex remote sensing scenes
Core Problem: Real-time ship detection in optical remote sensing images remains challenging due to large scale variation, elongated target geometry, dense ship distribution, weak object boundaries, and complex maritime backgrounds.
Key Innovation: To address these issues, this paper proposes UMS-DETR, a lightweight end-to-end Transformer-based detector for real-time ship detection in remote sensing images. In addition, UMS-DETR adopts the uncertainty-minimal query selection strategy from RT-DETR to initialize object queries from reliable and object-relevant feature tokens, thereby improving decoder prediction stability.
165. CMRWNet: Cross-Modal RWKV Network With Window-Guided Alignment for High-Fidelity Pansharpening
Core Problem: A key challenge lies in achieving spectrally faithful PAN injection while enabling scalable global feature interaction, without incurring the quadratic cost of full-image attention.
Key Innovation: The study proposes Cross-Modal RWKV Network (CMRWNet), a three-level encoder-fusion-decoder built from stacked Cross-Modal RWKV Fusion Blocks (CMRFBs).
166. PSG-RTDETR: Towards Stable Cross-Scale Feature Fusion for Small Object Detection
Core Problem: Small object detection in UAV remote sensing remains challenging due to fine-grained feature loss during downsampling and unbalanced cross-scale feature aggregation.
Key Innovation: To address these issues, we propose PSG-RTDETR, a P2-aware Softplus-Gated RT-DETR framework tailored for UAV-based dense small-object detection. Extensive experiments on VisDrone-DET demonstrate that, compared with the RT-DETR-ResNet18 baseline under the same ablation setting, PSG-RTDETR improves mAP50–95 and APsmall by 4.1 and 4.7 percentage points, respectively.
167. A Real-Time Subband SAR Imaging Algorithm Based on an Approximate Echo Signal Model
Core Problem: Conventional synthetic aperture radar (SAR) imaging algorithms establish the echo signal model based on the exact hyperbolic range equation and compensate for the range-azimuth coupling through complicated approximations to obtain high-resolution SAR images.
Key Innovation: The study proposes an approximate SAR echo signal model based on two-dimensional block partitioning.
168. TunVECM: vector graph structured extraction and intelligent checking for tunnel portal
Core Problem: Existing automated tools target structured IFC formats or execute rule-based logic, lacking adaptability to unstructured DXF vector graphs prevalent in railway engineering.
Key Innovation: To address this, this paper proposes TunVECM, a data-driven intelligent checking framework for tunnel portal vector graphs.
169. Improving reference evapotranspiration estimation by attention-enhanced autoencoders and hybrid neural networks under noisy chaotic conditions
Core Problem: The present study established a climate-adaptive Multi-Layer Perceptron-Self Attention (MLP-SA) framework for robust estimation of reference evapotranspiration (ETo) under noise-distorted conditions under different climatic conditions.
Key Innovation: The proposed model redefined noise-resilient ETo estimation through a climate-physics-informed two-stage architecture design: i) a denoising autoencoder (DAE) that first purifies noisy input data (meteorological variables, e.g. air temperature) by convolutional encoding-decoding, and ii) an MLP enhanced with a climate-adaptive self-attention mechanism, which dynamically recalibrated feature importance based on climate-specific physics.
170. Vibration isolation performance of periodic piles under actual soil conditions for mitigating train-induced building vibrations
Core Problem: However, in practical engineering, imperfect pile-soil contact and spatial variability of soil properties may influence wave propagation, highlighting the need for field experiments to evaluate the performance of periodic piles under actual soil conditions.
Key Innovation: Furthermore, the effect of periodic piles on building vibrations has rarely been reported in numerical studies. Results demonstrate that periodic piles provide effective isolation within the theoretical frequency band gap of 32-46 Hz, with attenuation coefficients below 0.5 at most measurement points.
171. Long-term creep and damage modeling of EICP-stabilized saline soil with FGD gypsum under freeze–thaw cycles
Core Problem: To mitigate the excessive settlement of saline soils in seasonally frozen regions, this study evaluates a sustainable stabilization method utilizing soybean urease-induced carbonate precipitation and industrial flue gas desulfurization (FGD) gypsum.
Key Innovation: A confining pressure sensitivity coefficient (Kp) was proposed, revealing a non-monotonic evolution law governed by frost damage accumulation. Furthermore, an improved Burgers model incorporating a time-dependent damage variable was established.
172. Micro-nano silica-based grouting materials for improving the mechanical behavior of fractured rock masses: performance optimization, laboratory and field validation
Core Problem: Cement-based grouting is widely used for reinforcement, yet poor slurry-rock interfacial bonding in fractured rock under deep underground conditions remains insufficiently addressed.
Key Innovation: The study develops a micro-nano silica-based grouting material incorporating micro-silica fume (MSF), nano-silica (NS), and graphene oxide (GO) to enhance ultrafine Portland cement (UPC) for fractured rock reinforcement. Compared with conventional slurry, it increased the peak strength of fractured sandstone by up to 16.95% and enhanced the energy dissipation capacity of the rock mass.
173. Intelligent Classification of Subsurface Road Defects Based on 3D GPR and Deep Learning
Core Problem: Three-dimensional ground-penetrating radar (3D GPR) produces high-density voxel data with complex spatial structures, making automated subsurface road defect recognition challenging.
Key Innovation: To address these limitations, this study proposes 3D-Vam, a hybrid architecture integrating multi-scale dilated convolution, Transformer self-attention, and Mamba-based state space modeling. Experimental results show that 3D-Vam achieves an overall classification accuracy of 90.06%, outperforming competing architectures.
174. Stability and fracture surface morphology evolution in granite subjected to high-order thermal cycles – Implications for enhanced geothermal systems
Core Problem: Granitoid-based enhanced geothermal systems (EGS) undergo cyclic thermal loading during fracture stimulation, increasing fracture complexity and modifying fracture morphology, with direct implications for geothermal productivity.
Key Innovation: In the present study, the impact of high-order thermal cycles (1, 10, 30, and 50) at temperatures of 100 °C, 200 °C, 300 °C, 400 °C, and 500 °C on the physical and mechanical properties, as well as variations in the fracture surface morphology of the granite, is examined.