TerraMosaic Daily Digest: September 10, 2026
Daily Summary
Two advances connect hazard simulation more directly to inference and cascading consequences. DiffSWE2d makes a two-dimensional shallow-water solver differentiable end to end, allowing flood and tsunami parameters to be optimized through the governing equations rather than through repeated black-box runs. In Chongqing, sequential 3D FEM-DEM modelling resolves how the Sibuhe landslide loads a bridge, transfers forces through structural components and develops cascading damage. Central Italian rock-slope analysis complements this process view with probabilistic peak-ground-velocity thresholds for coseismic detachment, turning regional shaking into spatially explicit rockfall source screening.
Mountain hazards are increasingly evaluated under changing forcing regimes. A 1950-2100 reconstruction in the French Alps quantifies how avalanche activity responds to evolving snow and climate conditions. In the Wumeng Mountains, machine-learning susceptibility projections are coupled to Shared Socioeconomic Pathways to track not only where slope failure may intensify but how relative exposure shifts as population ages and declines. HEC-RAS modelling of Dankanongba glacial lake constrains downstream outburst hazards along the Sichuan-Tibet corridor, while a spatially cross-validated random forest study on the Lhasa-Dingri corridor addresses the optimism introduced by geographically non-independent validation.
The Hunga eruption provides a geophysical bridge between source dynamics and far-field hazard: rapid submarine caldera collapse is identified as a mechanism capable of amplifying tsunami generation. Related work maps sinkhole susceptibility by combining remote sensing and field evidence, derives multihazard coastal risk surfaces in Bangladesh, and develops fire-potential and agricultural-drought monitoring from satellite time series. Across the methodological literature, the strongest common direction is physical structure inside learning systems-from differentiable conservation laws and physics-informed geotechnical inversion to geometry-aware remote-sensing detection and uncertainty-aware critical-infrastructure benchmarks.
Key Trends
The day's work embeds physical constraints and validation geography more deeply into hazard prediction and monitoring.
- Differentiable physics is becoming an inverse-modeling engine: DiffSWE2d exposes flood and tsunami dynamics to gradient-based calibration, while geotechnical PINNs and reduced-order seepage models use governing structure to constrain otherwise data-hungry inference.
- Landslide assessment is expanding from slopes to interacting infrastructure: The Sibuhe analysis follows forces and failure across a landslide-bridge system, and coseismic fragility studies convert shaking and slope detachment into transportation performance rather than treating each asset in isolation.
- Validation is becoming explicitly spatial: The Lhasa-Dingri susceptibility model uses spatial cross-validation, and regional exposure and multihazard studies retain geographic dependence instead of relying on random splits that mix neighboring terrain.
- Climate scenarios are being joined to changing exposure: Avalanche projections, Wumeng landslide scenarios, drought monitoring and flood-risk studies combine evolving forcing with population, land-use or ecosystem response.
- Multimodal sensing targets operational ambiguity: Remote-sensing methods fuse SAR, optical, hyperspectral, LiDAR and field constraints to distinguish true hazards from observational artifacts and to support deployment under incomplete coverage.
Selected Papers
The September 10 collection is led by differentiable flood and tsunami modelling, rapid submarine caldera collapse, landslide-bridge cascading failure, probabilistic coseismic rock-slope thresholds, long-term avalanche projections and scenario-based landslide exposure. Direct applications span GLOFs, sinkholes, spatially validated susceptibility, wildfire and drought monitoring, paleoseismology and liquefaction. The wider methods layer advances physics-informed inversion, multimodal Earth observation, geotechnical constitutive modelling and infrastructure resilience without extending into unrelated marine control or biomedical AI.
1. DiffSWE2d: a differentiable Shallow Water Equations solver for end-to-end flood and tsunami modelling
Core Problem: Inverse flood and tsunami problems require expensive repeated simulations when gradients through the governing solver are unavailable.
Key Innovation: Implements a differentiable 2D shallow-water solver for end-to-end parameter inference and optimization while retaining conservation-law structure.
2. Rapid submarine caldera collapse during the 2022 climactic eruption of Hunga volcano (Tonga)
Core Problem: The source process linking Hunga's submarine eruption to exceptional tsunami generation remains incompletely resolved.
Key Innovation: Links rapid caldera collapse to enhanced tsunami generation using the observed geomorphic evolution of the 2022 Hunga eruption.
3. Mechanisms of landslide-bridge interaction and cascading failure: A sequential coupled 3D FEM-DEM investigation of the Sibuhe Landslide, Chongqing, China
Core Problem: Bridge risk cannot be inferred from slope runout alone because impact loads propagate through interacting structural components.
Key Innovation: Couples 3D FEM and DEM sequentially to reconstruct landslide motion, bridge loading and cascading structural failure for the Sibuhe case.
4. 1950-2100 climate trends in avalanche activity in Haute-Maurienne, French Alps
Core Problem: Avalanche activity must be interpreted across historical and future snow-climate regimes rather than a short stationary record.
Key Innovation: Reconstructs and projects avalanche activity from 1950 to 2100 in Haute-Maurienne, linking hazard evolution to changing climate forcing.
5. Multi-scenario assessment of future landslide susceptibility and relative exposure burden under Shared Socioeconomic Pathways in the Wumeng Mountains Region, Southwest China
Core Problem: Future risk depends jointly on changing susceptibility and the demographic composition remaining in exposed mountain communities.
Key Innovation: Couples XGBoost susceptibility with Shared Socioeconomic Pathways to estimate future spatial hazard and relative exposure burden in the Wumeng Mountains.
6. Hazard assessment of the dankanongba glacial lake outburst flood based on HEC-RAS in Bomi, Tibet
Core Problem: A potentially unstable Tibetan glacial lake threatens a strategic mountain transportation corridor without well-constrained downstream inundation.
Key Innovation: Uses HEC-RAS to simulate Dankanongba outburst scenarios and map downstream hazard along the Sichuan-Tibet corridor.
7. Characterization of geotechnical and geological features with time-lapse infrared thermography and geo-spatial technologies-a case study
Core Problem: Near-surface thermal signatures are rarely integrated with geological and geospatial evidence for slope characterization.
Key Innovation: Combines time-lapse infrared thermography with geospatial technologies in a field case to characterize geotechnical and geological heterogeneity.
8. Optimizing landslide susceptibility mapping using spatially cross-validated random forest: Lhasa-Dingri corridor, Tibetan Plateau
Core Problem: Random validation can inflate susceptibility accuracy when neighboring samples share terrain and environmental structure.
Key Innovation: Optimizes a random-forest model under spatial cross-validation for the Lhasa-Dingri transport corridor, producing a more defensible estimate of regional transfer.
9. Tsunami amplification and time-delay mechanism; an insight into the case study of the 1945 tsunami event, Makran coast, Pakistan
Core Problem: The delayed arrival and exceptional height of the 1945 Makran tsunami are not explained by earthquake magnitude alone.
Key Innovation: Tests amplification and delay mechanisms for the Pasni coast to reconcile observed timing and wave height with source and coastal processes.
10. Probabilistic PGV-slope detachment thresholds for coseismic rock-slope instabilities in Central Italy
Core Problem: Regional rockfall screening needs probabilistic links between shaking intensity and slope detachment rather than deterministic cutoffs.
Key Innovation: Derives PGV-based detachment thresholds for Central Italian rock slopes, translating seismic motion into spatially variable source probability.
11. Byzantine-Robust Federated Fire Detection with a Rotating Coordinator
Core Problem: Such fire-detection systems use edge cameras that record sensitive footage which cannot easily be collected at a central server.
Key Innovation: Such fire-detection systems use edge cameras that record sensitive footage which cannot easily be collected at a central server.
12. A Dataset and Model for Imputing Water Surface Elevation on a Large and Extremely Sparse Spatiotemporal Graph
Core Problem: Satellite altimetry has the potential to alleviate this problem but its use is currently hindered by sparse temporal coverage.
Key Innovation: Yet, the scarcity of in situ gauges across much of the globe constrains the development of reliable modeling frameworks. We show that prior spatiotemporal graph imputation methods are not adapted to this topology, scale and sparsity, and propose a simple bidirectional selective state space model that outperforms them by sampling connected subgraphs and flattening space and time into a single token sequence with topology-aware.
13. Stress-Testing Dynamical and Generative Downscaling Using Subseasonal Extreme Precipitation Forecasts
Core Problem: Downscaling with either dynamical or deep generative models can overcome this issue, but the comparative performance of these models for extremes across different atmospheric regimes remains poorly understood.
Key Innovation: Downscaling with either dynamical or deep generative models can overcome this issue, but the comparative performance of these models for extremes across different atmospheric regimes remains poorly understood. WRF achieves the highest probabilistic skill for a multicell, non-stationary event.
14. Targeting tumor-associated macrophages using mRNA lipid nanoparticles for cytotoxic T lymphocyte-mediated cancer immunotherapy
Core Problem: The immunosuppressive tumor microenvironment (TME) remains one of the main obstacles that limit responsiveness to immunotherapy.
Key Innovation: Recently, lipid nanoparticles carrying messenger RNA (mRNA) have emerged as a promising strategy to modulate the immunosuppressive TME, with the ultimate goal of sustaining anticancer immunity of cytotoxic T lymphocytes (CTLs). We demonstrate that codelivery of a Toll-like receptor agonist and CXCL9-encoding mRNA encapsulated in our Ab-LNP successfully ameliorates immunosuppression and improves tumor infiltration and activity.
15. From heat risk to heat adaptation power: attribution and trade-off study of spatial differentiation of urban heat in Wuhan based on the perspective of spatial equity
Core Problem: Climate change and rapid urbanization have intensified urban heat risks and exacerbated their spatial inequities.
Key Innovation: This study investigates the spatial differentiation and equity of urban heat exposure through the interactions among natural conditions, the built environment, and social factors. Using Wuhan as a case study, remote-sensing data are integrated with self-organizing map neural networks and K-means clustering to identify heat-risk patterns and evaluate spatial equity.The results reveal a pronounced core-periphery gradient of.
16. Multi-hazard risk mapping using geospatial techniques in coastal Bangladesh: implications for resilience
Core Problem: Coastal areas play a vital role in supporting livelihoods, resources, and economic activities, yet in Bangladesh they are highly vulnerable to climate-induced hazards such as cyclones, floods, salinity intrusion, and erosion.
Key Innovation: This study focuses on Kalapara Upazila in the southern coastal region of Bangladesh and develops an integrated multi-hazard risk map using geospatial techniques. The findings reveal that 16.08% of the area falls within very high-risk zones and 34.03% within high-risk zones, mainly concentrated in the southern and southwestern parts, including Lalua, Khaprabhanga, Baliatali, Latachapli, and Mithaganj.
17. Spatiotemporal characteristics of tropical cyclone precipitation in China from 1960 to 2019 based on machine learning
Core Problem: However, existing research on TC precipitation is often constrained by sparse spatial coverage, limited temporal records, and coarse data resolution.
Key Innovation: To address these limitations, we developed a machine learning-based model, which reconstructs TC-induced precipitation by integrating key cyclone characteristics and environmental factors. Results show that from 1960 to 2019, the Average Annual Tropical Cyclone Precipitation, the Mean Tropical Cyclone Precipitation, and the Mean Maximum 3-hour Tropical Cyclone Precipitation exhibited a fluctuating downward trend, whereas the.
18. Assessing sinkhole hazard at the blue and white golf course, state college, using a combined remote-sensing and field methods approach
Core Problem: Sinkholes are a common problem faced by golf courses throughout the U.S. due to the large amount of water cycled through and transported beneath them by natural and anthropogenic means.
Key Innovation: In State College Pennsylvania, the Blue and White Golf Course has experienced sinkholes around the course including a large one by the Blue Course’s Hole 15 in the mid 2000s.
19. Paleoseismic record on the Menshi segment of the Karakorum fault zone, southwestern Tibetan Plateau and its tectonic implications
Core Problem: Despite its high activity since the Late Quaternary, the recurrence behavior of large earthquakes on this fault system remains poorly constrained.
Key Innovation: Through integrated field investigations, offset geomorphic mapping, trench excavation, and radiocarbon dating, this study identifies paleoseismic events along different branches of this segment.
20. An ensemble version of fire potential index (eFPI) integrating seasonal variability in sub-pixel fuel composition
Core Problem: However, anthropogenic climate change over the past century has catalyzed a shift in wildfire regimes, particularly in California, as evidenced by increases in the extent, frequency and severity of fires.
Key Innovation: The broad social-economic consequences of extreme wildfire events necessitate systematic frameworks that incorporate multi-source observations to support risk assessments and pre-fire interventions. Our results indicate the eFPI, through spatially-explicit integration of sub-pixel fuel characteristics, demonstrated a promising capacity to discriminate between burned and unburned regions across Southern California (χ2 = 3247.
21. A hierarchical metaheuristic framework for CPT-based soil liquefaction prediction with Kolmogorov-Arnold Network variants
Core Problem: Soil liquefaction is a critical geotechnical hazard that threatens the safety of civil infrastructure under seismic loading, necessitating robust and interpretable prediction models.
Key Innovation: This study proposes a hierarchical framework for CPT-based soil liquefaction prediction using meta-heuristic-optimized Kolmogorov-Arnold Network (KAN) variants. Statistical comparisons based on the Wilcoxon signed-rank test demonstrate that the Colony-Based Search algorithm (CSA) achieves lower objective-function values than the competing meta-heuristics ( p 0.01).
22. Meteoclimatic drivers of rock mass plasticity before failures: Insights from an artificial neural network trained on monitoring data and weather forecasts
Core Problem: Machine-learning approaches, such as artificial neural networks (ANNs), can support the interpretation of long-term monitoring time series and help to identify conditions that lead to rock mass failure.
Key Innovation: This study uses data from a multi-parametric monitoring system, local meteorological observations, and weather forecasting to train ANNs to evaluate the role of meteoclimatic stressors in preparing rock masses for failure. Based on 5 years of monitoring and local weather data, ANN-1 achieved a mean area under the curve (AUC) of 0.93 on class-balanced test sets.
23. Emergent Headwater Types for Watershed Monitoring, Research, and Planning
Core Problem: Headwaters-that is headwater streams and the watersheds they drain-strongly influence water quality, ecosystem services, and hydrologic connectivity, yet remain poorly mapped, monitored, and understood at large spatial scales.
Key Innovation: Our findings offer a transferable framework for classifying, modeling, and managing these critical yet understudied components of the hydrologic network. Our analysis (a) identified seven dominant headwater types with distinct combinations of biophysical and climatic attributes and (b) revealed systematic contrasts with downstream watersheds, with strongest differences in attributes representing potential flowpath depth.
24. Toward Interpretable Multimodal Fusion: Heat Conduction Modeling for Hyperspectral and LiDAR Joint Classification
Core Problem: However, existing multimodal fusion methods still struggle to model long-range dependencies and complex anisotropic interactions while maintaining computational efficiency.
Key Innovation: This paper introduces M2Heat, a physics-inspired framework that investigates multimodal fusion through the lens of heat conduction. M2Heat achieves competitive overall performance on three benchmarks, i.e., Trento, Houston2013, and Augsburg, while providing an interpretable heat-conduction-guided perspective for multimodal feature fusion.
25. R4Tun: LLM-guided adaptive segmental tunnel lining segmentation in point clouds
Core Problem: Automated inspection of segmental tunnel linings requires adaptive segmentation from 3D point clouds, yet expert-tuned pipelines often degrade when tunnel conditions vary.
Key Innovation: This paper presents R4Tun, a large language model (LLM)-driven adaptation framework that extends an expert-designed pipeline (SAM4Tun) with bounded parameter tuning informed by structured context: memory (m), state (s), and knowledge (k). Across 270 (30 tunnels × 3 different LLMs × 3 context settings) runs, the LLMs showed similar parameter-adjustment trends (with overlapping 95% CIs on mean gains) and consistently adjusted a.
26. A variational physics-informed graph neural network for heterogeneous solid mechanics
Core Problem: Stress localization in heterogeneous solids is governed by the bimaterial interface, where the displacement field remains C⁰-continuous, while in-plane stresses jump due to the stiffness mismatch.
Key Innovation: Coordinate-based physics-informed neural networks (PINNs) represent this jump via a prescribed regularization width or a weighted interface penalty, making their accuracy sensitive to how phase-contrast changes are handled.
27. Geospatial Foundation Models Capture Health-Relevant Dimensions of Place Beyond Conventional Social Risk Indices
Core Problem: Area-based social risk indices summarize residents' socioeconomic conditions but incompletely capture physical features of place that may affect health.
Key Innovation: We evaluated whether numerical representations of physical place produced by four geospatial foundation model families from 2022 satellite data explained residual variance in tract-level associations between the Area Deprivation Index, Social Deprivation Index, and Social Vulnerability Index with health outcomes.
28. Measuring Browser Webcam Gaze Honestly: A Capture-Clock Methodology and Open Reference Implementation
Core Problem: Browser-based webcam gaze trackers are increasingly used for crowd-scale data collection and in clinical settings where lab eye trackers are impractical, but the reported latency numbers may not represent real world functionality.
Key Innovation: The common practice of timestamping each gaze sample when it is emitted, rather than when its source frame was captured, makes the measured inference latency read about 0ms no matter how slow the engine really is.
29. Is elastic wave velocity a proxy for fabric evolution under shear in quartz sand? An experimental study using X-ray tomography and ultrasonic transducers
Core Problem: Understanding the link between the evolution of fabric and elastic wave velocity in sand subjected to shear is of critical significance in engineering practice.
Key Innovation: This study provides experimental evidence for a strong correlation between normalised wave velocity and selected fabric descriptors, offering a potential basis for improving the calibration and validation of discrete-element simulations of sheared granular media. To achieve this objective, an innovative multiscale experimental campaign was conducted by integrating triaxial compression with simultaneous wave velocity.
30. An energy-driven model for granular materials with crushing, contraction and dilation
Core Problem: Crushable granular materials exhibit complex mechanical behaviour due to the interplay between crushing, contraction and dilation.
Key Innovation: To address these gaps, this study presents a novel energy-based constitutive model for addressing the crushing-contraction-dilation coupling issues of crushable granular materials.
31. The non-monotonic shape of the loading-collapse curve for partially saturated soils
Core Problem: Canadian Geotechnical Journal, Volume 63, Issue, Page 1-13, January 2026.
Key Innovation: This study experimentally investigates the shape of the loading-collapse (LC) curve of unsaturated soils, focusing on the evolution of the yield stress with suction across the full range of degree of saturation.
32. Vortex control for mitigating bridge pier scour: A hypothesis test using a denticle-shaped device
Core Problem: This study proposes a novel scour countermeasure using a biomimetic denticle device inspired by vortex-trapping mechanisms observed in shark skin and aerospace applications.
Key Innovation: This study proposes a novel scour countermeasure using a biomimetic denticle device inspired by vortex-trapping mechanisms observed in shark skin and aerospace applications. Results show that the denticle significantly reduces near-bed TKE and RSS by disrupting downflow impingement and trapping a stable vortex beneath the device.
33. Air cavity effect on the impact pressure of dam-break induced bore overtopping on a dike crest
Core Problem: This study investigates the air cavity feature and its influences on the bore overtopping impact pressure on the dike crest of a vertical seawall through dam-break experiments and a simplified theoretical model.
Key Innovation: This study investigates the air cavity feature and its influences on the bore overtopping impact pressure on the dike crest of a vertical seawall through dam-break experiments and a simplified theoretical model. Measured overtopping impact pressures reveal distinct characteristics in different crest regions with diverse mechanisms.
34. A benchmark dataset for half-hourly evapotranspiration estimation in China from 2000 to 2024
Core Problem: Although flux observations based on the eddy covariance technique are widely regarded as essential benchmark data for evapotranspiration estimation, existing ChinaFlux observations are generally limited by short observation periods and extensive data gaps, which substantially constrain their applicability in long-term change analyses and multi-scale studies.
Key Innovation: To address these limitations, we developed a gap-filling and temporal prolongation framework specifically designed for half-hourly LE and established a continuous ground-based benchmark dataset covering China for the period 2000-2024 based on observations from 50 ChinaFlux sites.
35. A convective-scale reanalysis for the ‘Swabian MOSES 2023’ field campaign
Core Problem: Abstract.
Key Innovation: Here, we present the pioneering, 3-months convective-scale campaign reanalysis of the 'Swabian MOSES 2023' campaign, consisting of a control (CTRL) dataset without and a campaign (CMPG) dataset with additional field campaign observations from the mobile atmospheric measurement platform KITcube during June, July and August 2023.
36. Task aggregation as a strategy to optimize Earth System Model workflows in HPC: assessing real scenarios with EC-Earth
Core Problem: Abstract.
Key Innovation: To alleviate this issue, we propose achieving shorter times-to-response, which are the durations from the first submission to the completion of the final task, by applying task aggregation to reduce subsequent requests for resources and, consequently, reducing queue times.
37. The Geoengineering Model Intercomparison Project (GeoMIP) contribution to CMIP7 - description of new experimental protocols and preliminary results
Core Problem: Abstract.
Key Innovation: Here we present a suite of new climate model experiments designed for the Coupled Model Intercomparison Project Phase 7 (CMIP7), building on lessons learned from previous GeoMIP experiments, recent SRM research, and new simulations developed for CMIP7. Such experiments must both diagnose areas of model agreement and disagreement through the lens of climate science and provide results useful for understanding the potential.
38. Landscapes buried beneath large-volume ignimbrites reveal preeruptive uplift rates
Core Problem: Pyroclastic density currents from large explosive eruptions can generate expansive ignimbrite deposits with planar, low-angle surfaces.
Key Innovation: We demonstrate the utility of these surfaces for constraining preeruptive rock uplift rates from the relief of buried paleolandscapes.
39. Beyond empirical models: Discovering constitutive laws in solids with graph-based equation discovery
Core Problem: Constitutive models are fundamental to solid mechanics and materials science, underpinning the quantitative description of material behaviors.
Key Innovation: In this work, we propose a graph-based equation discovery framework for automated discovery of constitutive laws directly from multicase experimental data. The discovered models exhibit compact analytical structures and achieve higher accuracy than empirical models.
40. Class-dependent depositional onset and selective bypass in polydisperse turbidity currents
Core Problem: Turbidity currents commonly transport mixtures of clay, silt, and sand, but reduced theories of depositional onset often treat the suspended load as one effective class.
Key Innovation: Here we develop a class-dependent state-variable theory for depositional onset in polydisperse long-runout turbidity currents. Sensitivity tests show that absolute onset distances depend on closure parameters, but the coarse-to-fine ordering is stable for the tested one-parameter variations and for an alternative rational suspension-capacity penalty.
41. Inferring groundwater overdraft in data-scarce arid agro-ecosystems: a semi-empirical remote sensing framework applied to the Elfeija watershed, Morocco
Core Problem: Monitoring these diffuse withdrawals remains a critical challenge for water governance.
Key Innovation: Introduction Small-scale irrigated agriculture in arid regions relies heavily on unmetered groundwater, creating an “invisible pumping” threat to aquifer sustainability. Results The analysis reveals a distinct seasonal signature of irrigation.
42. Intelligent spatial prediction of shear failure zones on rock joint surfaces using 3D morphological features and Naive Bayes
Core Problem: This study combines direct shear tests with machine learning methods to analyze the influence of surface shape on the shear failure behavior.
Key Innovation: This study combines direct shear tests with machine learning methods to analyze the influence of surface shape on the shear failure behavior. The results showed that the inclination angle and bulge height significantly affected the shear mobility, crack development process, and expansion behavior.
43. Learning Thermospheric State Evolution: An Adaptive Neural Operator Framework Based on TIE-GCM Simulations
Core Problem: Reliable short-term prediction of thermospheric states is important for satellite drag applications but remains difficult because of nonlinear, multiscale variability.
Key Innovation: We developed a multivariable Adaptive Fourier Neural Operator (AFNO) surrogate using 24 years (2000-2023) of Thermosphere-Ionosphere Electrodynamics General Circulation Model (TIE-GCM) simulations.
44. Study on Nonlinear Driving Mechanisms of Spatiotemporal Evolution in Sanjiang Plain Wetlands Based on Explainable Learning Methods
Core Problem: However, wetland evolution is highly heterogeneous across space and time, and existing studies have generally paid insufficient attention to nonlinear effects and interactions among driving mechanisms, limiting a comprehensive understanding of its intrinsic processes.
Key Innovation: Understanding the mechanisms that govern wetland evolution is essential for the sustainable development of wetland ecosystems. Artificial wetlands expanded continuously, whereas marsh wetlands declined markedly before 2005 and showed partial recovery thereafter.
45. Boosting Multi-Class SAR Oriented Object Detection via Geo-Topology-Guided Diffusion Synthesis
Core Problem: Oriented object detection in Synthetic Aperture Radar (SAR) imagery plays an important role in remote sensing, but its performance is usually limited by the shortage of high-quality annotated samples.
Key Innovation: This problem is particularly prominent in multi-class scenarios, where different targets exhibit significantly different scattering characteristics, scale distributions, orientation variations, and background dependencies.
46. An Antecedent-Precipitation-Informed Soil Water Balance and Time-Aware Mamba-MoE Framework for Surface Soil Moisture Forecasting
Core Problem: Surface soil moisture forecasting is important for drought monitoring, irrigation management, and land-atmosphere process analysis but remains challenging because near-surface soil moisture is jointly influenced by antecedent precipitation, atmospheric drying, soil properties, vegetation conditions, and irregular multi-source observations.
Key Innovation: This study proposes an Antecedent-Precipitation-Informed Surface Soil Water Balance and Time-Aware Mamba-Mixture-of-Experts (API-SWB-Mamba-MoE) framework for forecasting in situ volumetric soil moisture at approximately 5 cm depth using only information available before the target time.
47. Developing an Approach to Agricultural Drought Monitoring for Timor-Leste Using Space-Based Observations
Core Problem: Drought is a natural hazard that poses significant threats to populations, economic sectors, and the environment.
Key Innovation: This study examines agricultural drought monitoring in Timor-Leste using two satellite-derived precipitation and vegetation health products: the Standardised Precipitation Evapotranspiration Index (SPEI) and the Vegetation Health Index (VHI). This study demonstrates that spaced-based drought monitoring can be a valuable tool for proactive drought management in Timor-Leste and the surrounding regions.
48. Improving the Spatial Resolution of GRACE-Derived GFZ G3P Groundwater Storage Anomaly Product Through Unsupervised Deep Learning Downscaling
Core Problem: While groundwater is the second largest freshwater reservoir on Earth, it remains difficult to monitor, relying mainly on sparse and often unavailable well measurements.
Key Innovation: In this study, an unsupervised deep learning approach was exploited to downscale groundwater storage anomalies (GWSA) from the GRACE-derived GFZ G3P product from 0.5° to 0.1° spatial resolution, covering from 2003 to 2023. Results show good agreement between the downscaled and the original G3P product, with average temporal and spatial correlations of r = 0.99.
49. Serpentinite as a strategic Earth material: geological controls, mineral reactivity, carbon mineralization, geologic hydrogen, and sustainable resource pathways
Core Problem: Serpentinite, the hydrated ultramafic lithology formed during serpentinization of mantle peridotites, is increasingly recognized as a strategic yet heterogeneous geomaterial for construction, carbon-management, and energy-related pathways.
Key Innovation: This review synthesizes knowledge by linking geological origin, serpentine polymorph assemblage, accessory phases, alteration degree, and structure-property-reactivity relationships to application performance.
50. Physics-informed convolutional neural network for soil-bedrock contact identification using integrated airborne transient electromagnetic survey and borehole data
Core Problem: Although traditional landslide investigation methods using borehole drilling can directly reveal present soil and bedrock layers and enable the use of soil and rock samples, their sampling points are usually sparse, they are costly, and they do not provide spatial data and sometimes cannot be conducted at desired locations.
Key Innovation: Here, we propose a physics-informed convolutional neural network that integrates borehole and ATEM data for stratigraphic classification. Although traditional landslide investigation methods using borehole drilling can directly reveal present soil and bedrock layers and enable the use of soil and rock samples, their sampling points are usually sparse, they are costly, and they do not provide spatial data and sometimes cannot.
51. Urban agriculture as a place-based resilience strategy in post-earthquake container settlements: Evidence from İskenderun, Türkiye
Core Problem: The 6 February 2023 Kahramanmaraş earthquakes led to large-scale displacement in southern Türkiye, resulting in the rapid emergence of container settlements as transitional living environments.
Key Innovation: This study examines urban agriculture in these settlements through the interconnected dimensions of food provisioning, social cohesion, wellbeing, place attachment, and post-disaster resilience. The findings indicate that practitioners engage in urban agriculture primarily as a subsistence-based, non-market practice and perceive cultivation as supporting household food access, social interaction, and wellbeing.
52. Assessing Fire Risk in High-Rise Buildings: A Hazard-Centric Façade Design Framework
Core Problem: Consequently, the overall performance of these complex systems often remains ill-defined, either overly conservative or lacking resilience.
Key Innovation: New requirements are introduced regularly, particularly those pertaining sustainability. An extensive list of major failures shows a lack of integrated design methodology, with the façade not always being considered an integral part of the building.
53. Monitoring global mangrove spatial evolution and functional degradation using BTSAD and CBPA methods over a 36-year dense Landsat and HLSv2.0 time series
Core Problem: Over recent decades, however, these ecosystems have confronted the dual pressures of spatial shrinkage and persistent functional deterioration.
Key Innovation: To address this critical gap, this study proposes a comprehensive remote sensing framework that captures mangrove spatiotemporal evolution across six representative global mangrove sites using a 36-year multispectral time series (Landsat and HLSv2.0, 1988-2024) on the Google Earth Engine (GEE) platform.
54. A mechanism-coupled split-window network for medium- to high-resolution land surface temperature retrieval
Core Problem: Land surface temperature (LST) retrieval from medium- to high-resolution thermal infrared (TIR) observations remains challenging under complex atmospheric and surface conditions.
Key Innovation: To address these issues, this study proposes a Parallel Component Decoupled Neural Network (PCD-Net) that reformulates SW retrieval as an adaptive learning problem of physical component parameters.
55. Mapping selective logging with Sentinel-1 image time series and a novel transformer-based model
Core Problem: Selective logging (SL) is a major driver of forest degradation, yet its small and short-lived canopy openings remain difficult to detect in persistently cloudy tropical regions.
Key Innovation: We present InterFormer, a Transformer-based architecture that jointly models feature interactions and temporal dynamics via a feature block with Multi-head Intercross Self Attention (MISA) and a temporal block with Multi-head Self Dissimilar Attention (MSDA). In the independent UPA validation, InterFormer localized 92.65% of tree-falling gaps in UPA areas and outperformed standard Transformer variants and Planet-based alerts.
56. Machine learning-based classification of laboratory fault peak slip-rate trends using acoustic emission catalogs
Core Problem: Across these experiments, we observe two distinct behaviors in the evolution of peak slip rate, which either increases or remains approximately constant during subsequent injection cycles.
Key Innovation: As a laboratory analogue of injection-induced earthquakes, we conducted four sets of fluid-driven fault slip experiments on cylindrical Bentheim sandstone samples containing pre-cut faults, subjected to different fluid injection rates and initial confining pressures. Results reveal that the model achieves a balanced classification accuracy of more than 99%, indicating that the peak slip-rate trend can be classified with a.
57. Assessing the reliability of DFN-generated block systems for DEM analyses: Insights from geometric and topological consistency
Core Problem: While the reliability of DFN construction and DEM simulation has been extensively investigated, the reliability of block-system generation remains largely unexplored.
Key Innovation: The discrete fracture network-discrete element method (DFN-DEM) framework is widely used to analyze the mechanical behavior of fractured rock masses and generally involves DFN construction, block-system generation, and DEM-based mechanical simulation. The proposed method is first applied to a hypothetical rock mass case, showing high structural consistency and low variability among block systems generated under identical DFN.
58. Three-branch adaptive physics-informed neural network for geotechnical stiffness inversion
Core Problem: Recovering spatially variable stiffness from displacement observations is fundamental to geotechnical model calibration, yet remains difficult when material interfaces are unknown.
Key Innovation: This study develops a Three-Branch Adaptive Physics-Informed Neural Network (TBA-PINN) for distributed inversion of bulk and shear modulus fields. Comparisons with controls that match parameter count, separate network roles, or alter weighting showed that neither additional network capacity nor loss balancing alone accounted for the improvement.
59. Engineering-scale rock fracture modeling by quasi-state-based peridynamics via coordinated GPU acceleration
Core Problem: At the engineering scale, however, the same nonlocal mechanism becomes a computational bottleneck: as damage evolution breaks bonds and creates damaged and free material points, the set of active interactions becomes increasingly irregular, rendering repeated updates highly inefficient.
Key Innovation: Peridynamics is inherently suited for multiscale rock fracture simulation as its nonlocal interactions can capture crack initiation, propagation, branching, coalescence, and block separation within a unified framework. Pure numerical benchmarks demonstrate a stable 16.5-17.1 × speedup over the representative peridynamic GPU implementations on both Windows and Linux, and ablation tests confirm that the performance gain stems.
60. A multilevel reduced-order method for rapid prediction of 3D saturated-unsaturated seepage in earth-core rockfill dams
Core Problem: Rapid and accurate characterization of 3D saturated-unsaturated seepage in earth-core rockfill dams is essential for dam safety assessment and real-time risk warning.
Key Innovation: This study proposes a multilevel reduced-order method for rapid prediction of seepage fields under unseen parameter combinations. Results show that the proposed method achieves an average relative error of 0.4421% under combined parameter conditions and reconstructs the 3D seepage field in approximately 5.07 ms.
61. Multifractal analysis of anisotropic creep failure mechanisms in shale under direct shear using acoustic emission
Core Problem: However, the effect of shale anisotropy on the micro-scale energy-driven creep failure mechanism under direct shear conditions requires further analysis.
Key Innovation: In this study, direct shear creep tests of shale with five bedding angles (0°, 30°, 45°, 60°, 90°) were conducted, along with acoustic emission (AE) and digital image correlation (DIC) analysis. The results showed that specimens with α ≤ 30° exhibited progressive creep slip instability along the bedding planes.
62. A Multidimensional Critical Regulation Approach to Water Resources Optimization
Core Problem: Conventional water allocation models focus on reducing water shortage, maximizing economic benefits, and controlling environmental impacts, but rarely incorporate the coordinated evolution of multidimensional water systems.
Key Innovation: To address this gap, an integrated prediction-simulation-optimization framework was developed for 62 prefecture-level cities in the Yellow River Basin, embedding multidimensional water-system order directly into the optimization objective. Support vector machine achieved the best prediction performance, with 98% of cities meeting the accuracy threshold.
63. Contrasting Impacts of Climate and Vegetation Changes on Water Yield Across Global and Regional Scales
Core Problem: Existing studies remain confined to the relative contribution and interacting pathways of multiple drivers, providing limited insights into heterogeneity across different regions.
Key Innovation: This study conducted a global spatiotemporal analysis of WY using multi-source remote sensing and meteorological data from 1982 to 2020, integrating the Random Forest model and Structural Equation Model to quantify the relative contributions and pathways of climate and vegetation changes to WY at global and regional scales (arid and humid regions).
64. Halo: Improving forecast accuracy through heteroscedastic estimation
Core Problem: Heteroscedastic forecasting, where a network estimates a scale parameter alongside a location parameter, is normally motivated by uncertainty quantification.
Key Innovation: This paper shows it also improves the point estimate, in contrast to reported negative results for heteroscedastic estimation outside time series.
65. GRADE: Single-Frame Generative Radar Depth Estimation Under Visual Degradation
Core Problem: Dense 3D depth perception fails under smoke, fog, and darkness because optical sensors cannot penetrate airborne particulates. mmWave radar remains usable and measures range accurately under these conditions, but its small aperture limits angular resolution.
Key Innovation: We present GRADE, which grounds a pretrained generative prior in single-frame radar geometry to estimate high-fidelity metric depth. Trained and evaluated on ~95K frames across 12 buildings with real smoke, GRADE achieves an MAE of 0.303 m in clear scenes and 0.313 m under smoke, outperforming existing baselines.
66. RiVaT-Fuse: Reliability-Calibrated Variational Tensor Fusion for Multimodal Prediction under Modality Uncertainty
Core Problem: Existing representation-level fusion methods typically choose an aggregation architecture, such as concatenation, gating, conditional modulation, or attention, without explicitly defining what the fused representation should mean under modality uncertainty.
Key Innovation: We propose RiVaT-Fuse, a reliability-calibrated variational tensor fusion framework that defines fusion as sample-wise latent-state estimation. On an image-level image-metadata prediction benchmark, RiVaT-Fuse achieves the strongest overall predictive rank among direct representation-level baselines while improving probability and label stability under perturbation.
67. Bidirectional Multimodal Fusion of Sky Images and Time-Series for Solar Forecasting with Large Language Models
Core Problem: Furthermore, forecast performance is highly sensitive to location and local observing conditions, creating a strong need for site-specific data that are often scarce.
Key Innovation: Effectively incorporating sky imagery into an LLM-based forecasting framework remains under-explored and an open challenge. Recently, large language models (LLMs) have demonstrated competitive performance and high data efficiency in time-series forecasting.
68. Improving Faint Object Detection for Space Situational Awareness with Variational Autoencoders
Core Problem: Detecting low signal-to-noise ratio (SNR) objects remains extremely challenging in optical observations, particularly in the cislunar (X-GEO) environment, where structured sky backgrounds, dense stellar fields, and scattered moonlight significantly degrade the performance of classical detection algorithms.
Key Innovation: We present a deep-learning pipeline for enhancing the detection of faint moving objects in optical space situational awareness (SSA) imagery through automated star removal and background reconstruction. Results demonstrate that the method reconstructs star-free backgrounds with high fidelity, while preserving moving targets and significantly enhancing detectability, thereby providing an effective data-driven preprocessing.
69. Estimating Inconsistency Response Surfaces under Uncertainty in Cyber-Physical System Development
Core Problem: During development, CPS consistency requires that shared model elements remain compatible across these models.
Key Innovation: Uncertainty, for example, due to sensor noise or model abstraction, changes the admissible values of model elements and can introduce inconsistencies, i.e., situations in which models can no longer be jointly satisfied. Experiments on 48 scenarios and 10 CPS domains show that the surrogate matches Monte Carlo estimates while reducing evaluation time from milliseconds to microseconds, enabling orders-of-magnitude more.
70. MC-DeTra: Motion-Consistent Joint Object Detection and Socially-Aware Trajectory Forecasting in Bird's-Eye-View Images
Core Problem: Their accuracy on dynamic, moving actors, however, remains the hardest part of the task, and the strongest such model, DeTra, has no public implementation.
Key Innovation: Their accuracy on dynamic, moving actors, however, remains the hardest part of the task, and the strongest such model, DeTra, has no public implementation.
71. CausalArena: Benchmarking Causal Discovery in the Foundation Model Era
Core Problem: The emergence of causal discovery foundation models (CDFMs) further complicates evaluation: performance may reflect not only causal discovery ability, but also overlap between pretraining environments and test SCMs, making results on fixed synthetic benchmarks difficult to interpret.
Key Innovation: We introduce CausalArena, a unified and evolvable benchmark for causal discovery under a common protocol. The emergence of causal discovery foundation models (CDFMs) further complicates evaluation: performance may reflect not only causal discovery ability, but also overlap between pretraining environments and test SCMs, making results on fixed synthetic benchmarks difficult to interpret.
72. A Station-Based Evaluation of Machine Learning-based Weather Forecasting Models in Northern Norway
Core Problem: However, their performance remains unclear in challenging environments such as Northern Norway, where narrow fjords and rapidly changing weather result in highly variable local wind conditions.
Key Innovation: However, their performance remains unclear in challenging environments such as Northern Norway, where narrow fjords and rapidly changing weather result in highly variable local wind conditions. Our results show that HRES slightly outperforms FCN3 and GraphCast, with an overall RMSE of 2.89 ms⁻¹, compared to 2.96 ms⁻¹ for FCN3 and 2.94 ms⁻¹ for GraphCast.
73. Temporal and Multimodal Deep Learning for Cyberattack Detection in LEO Satellite Systems
Core Problem: The growing reliance on Low-Earth Orbit (LEO) satellite communication systems has increased the need for intelligent methods capable of detecting cyberattacks across complex and dynamic space environments.
Key Innovation: In this work, we conduct a systematic study of deep-learning-based cyberattack detection using the recently introduced satellite-specific UNSW-IoTSAT dataset. Experimental results demonstrate the value of structured multimodal modeling and rigorous evaluation, with the hierarchical Transformer achieving up to 91.66% accuracy and 85.63% macro F1 under the leakage-resistant evaluation protocol.
74. Predicting Train Delays in Finland Using Machine Learning and Weather Data
Core Problem: Adverse weather, particularly in Arctic regions with extreme temperatures and heavy precipitation, remains a leading cause of train delays, yet most prediction approaches rely on raw meteorological inputs without exploiting domain-informed feature engineering.
Key Innovation: Adverse weather, particularly in Arctic regions with extreme temperatures and heavy precipitation, remains a leading cause of train delays, yet most prediction approaches rely on raw meteorological inputs without exploiting domain-informed feature engineering.
75. MindTopo: Can Foundation Models Reason in Topological Space?
Core Problem: Spatial reasoning depends not only on metric properties such as distance, angle, and shape, but also on topological relations that remain invariant under continuous deformation.
Key Innovation: We introduce MindTopo, a benchmark of topological intuition across five properties grounded in cognitive science and formal topology: continuity, separation, order, enclosure, and knots.
76. Learning Intrinsic Water-Quality Dynamics with Rainfall for Data-Driven Forecasting
Core Problem: Rainfall is an important environmental driver of water-quality variations through processes such as runoff, pollutant transport, dilution, and resuspension.
Key Innovation: RaiNet employs LocTrend to capture irregular water-quality dynamics, constructs station-oriented rainfall events from gridded precipitation, and introduces XGateFusion for conditional lag-aware fusion across scales. Experiments show that RaiNet outperforms general time-series, water quality, diffusion-based, and spatiotemporal models by over 20%, while component-wise analyses confirm the distinct contribution of each module.
77. Addressing A Posteriori Performance Degradation in Neural Network Subgrid Stress Models
Core Problem: Neural network subgrid stress models often have a priori performance that is far better than the a posteriori performance, leading to neural network models that look very promising a priori completely failing in a posteriori Large Eddy Simulations (LES).
Key Innovation: This performance gap can be decreased by combining two different methods, training data augmentation and reducing input complexity to the neural network.
78. Experimental and numerical analysis into the coupled effect of rubber granules and geogrids on the mechanical behaviour of railway ballast under impact loading
Core Problem: Ballast degradation under high-energy impact loads, such as that caused by wheel-rail irregularities and stiffness variations in railway tracks, significantly compromises track stability and exacerbates maintenance and repair costs.
Key Innovation: This study investigates the coupled effect of rubber granules and geogrid reinforcement on enhancing ballast performance under impact loading. The results show that when rubber granules are used in combination with geogrids, the two exhibit a synergistic effect, mitigating the deformation of ballast, optimising load transmission and striking a balance between deformation control and structural stability.
79. Comparative field evaluation of geogrid-stabilized flexible pavements under static and repetitive loading
Core Problem: Establishing a relationship between reinforcement benefits under static and repetitive loading remains a challenge for evaluating the field performance of geogrid-stabilized flexible pavements.
Key Innovation: This study examines the performance of conventional and geogrid-stabilized pavement sections constructed over different subgrade strengths using SPBTs and RPBTs. Static tests showed MIF values ranging from 2.10 for weaker subgrades to 1.72 for stronger subgrades.
80. Impact-load characteristics of a tail-rudder-equipped vehicle during oblique trans-media water entry
Core Problem: High-speed oblique water entry exposes the tail rudders of trans-media vehicles to secondary impacts that affect transient loading and attitude stability.
Key Innovation: This study numerically investigates a tail-rudder-equipped vehicle during high-speed oblique water entry, focusing on cavity evolution, rudder wetting, hydrodynamic loading, and pitch response. These results identify transient rudder wetting as the link between cavity evolution, load partitioning, and attitude stability.
81. Investigation of mooring anchor installation deviations and GA-based correction strategies for a 15 MW FOWT in shallow water
Core Problem: For floating offshore wind turbines (FOWTs) in shallow water, such deviations threaten mooring safety due to the pronounced nonlinearity of catenary mooring systems in limited water depths.
Key Innovation: This study analyzes the effects of anchor installation deviations for a 15 MW semi-submersible FOWT at 50 m water depth, and proposes a genetic algorithm (GA)-based correction method to mitigate these effects.
82. Seismic performance and wind-seismic response of steel-concrete hybrid wind turbine towers
Core Problem: The verified record establishes the scope of seismic performance and wind-seismic response of steel-concrete hybrid wind turbine towers, but the accessible metadata do not expose the study motivation or boundary conditions.
Key Innovation: The title, authors, source and publication identity are verified; methods and quantitative outcomes are not asserted because a reliable abstract was unavailable.
83. Global and national lime process emissions and carbonation sink from 1930 to 2024
Core Problem: Accurate quantification of lime process emissions and subsequent carbonation uptake is needed to represent the lime carbon cycle in global carbon-budget assessments.
Key Innovation: We developed a source-prioritized and internally cross-checked dataset for 1930-2024 by harmonizing USGS statistics, national statistical yearbooks, industrial records, and explicitly identified proxy-based reconstructions.
84. Improving seaweed cover estimation in high latitudes: a focus on kelp classification
Core Problem: Abstract.
Key Innovation: Satellite remote sensing offers the most practical means of monitoring Macrocystis pyrifera kelp forests across the remote Strait of Magellan, but standard atmospheric correction workflows assume a plane-parallel atmosphere that becomes inaccurate at the high solar zenith angles (SZA) characteristic of subantarctic latitudes, introducing wavelength-dependent radiometric biases into spectral unmixing products.
85. Explicit representation and calibration of different landscape units for a robust catchment DOC export model
Core Problem: However, lumped and landscape-explicit (separating upland and riparian zone) model structures are generally calibrated to stream DOC concentrations, while the internal DOC dynamics often do not receive sufficient attention.
Key Innovation: Here, we developed a flexible model with a lumped and landscape-explicit structure for four headwater catchments in the Harz Mountains, Germany. Data-driven studies revealed variable functioning of different landscape units (upland, riparian zone, and groundwater) in catchment DOC mobilization and export.
86. Heterogeneous slab mantle hydration controls the chemical architecture of the Kamchatka arc
Core Problem: Fluids released from subducting slabs play an important role in arc magmatism and global volatile cycles, yet resolving their lithological origins and spatial distribution remains a major challenge.
Key Innovation: Our results demonstrate that preexisting slab hydration exerts first-order control on subduction zone chemical architecture, providing a predictive framework for understanding global volatile cycling.
87. Spatiotemporal dynamics and driving mechanisms of net ecosystem productivity in the retrogressive thaw slump regions of the Qinghai-Tibetan Plateau
Core Problem: Previous studies have mainly focused on RTS dynamics, while the spatiotemporal variations of net ecosystem productivity (NEP) in RTS regions on the QTP remain insufficiently understood, mainly due to the limited comparative assessments of NEP dynamics between RTS and non-RTS regions, and the underlying driving mechanisms remain unclear.
Key Innovation: This study constructs an NEP dataset for the QTP using the Google Earth Engine cloud platform, and investigates the spatiotemporal variations of interannual and seasonal NEP in RTS regions from 2016 to 2022, while analyzing their response mechanisms. The results show that the NEP in RTS regions exhibited a slow overall increase and the spatial variation of NEP in most RTS regions ranged from −5 gC/m²·month to 5 gC/m²·month.
88. A thermodynamic-kinetic approach to equilibrium and quench speciation of volcanic mercury
Core Problem: Volcanic mercury cannot be interpreted from bulk flux alone because the atmospheric transport, deposition, and preservation depend on speciation.
Key Innovation: This study presents a thermodynamic-kinetic approach for constructing equilibrium and quench speciation diagrams for mercury-bearing volcanic gases. An Etna-type benchmark shows that the reduced equations recover the expected ordering and approximate thermal shift: hot magmatic gases favor H g 0, while cooling, oxidation, and HCl-rich conditions favor H g C l 2.
89. Thorpe Analysis of Atmospheric Turbulence in Parts of Inner Mongolia and Guangdong, China, Based on a Round-Trip Intelligent Sounding System
Core Problem: Thorpe analysis is a classic method for turbulence retrieval, but traditional observations are limited by low spatiotemporal resolution, the absence of a stratospheric turbulence inversion framework, and insufficient cross-layer comparisons between northern and southern China, restricting the understanding of turbulence modulation mechanisms.
Key Innovation: Thorpe analysis is a classic method for turbulence retrieval, but traditional observations are limited by low spatiotemporal resolution, the absence of a stratospheric turbulence inversion framework, and insufficient cross-layer comparisons between northern and southern China, restricting the understanding of turbulence modulation mechanisms.
90. Phenology-Guided Weakly Supervised Cropping Structure Mapping with Phenological Similarity Constraints
Core Problem: To address the challenges of cumulative pseudo-label noise and substantial phenological differences across regions in cropping structure mapping, this study develops a phenological knowledge-guided weakly supervised semantic segmentation framework, termed PhenoStruct-WSF.
Key Innovation: To address the challenges of cumulative pseudo-label noise and substantial phenological differences across regions in cropping structure mapping, this study develops a phenological knowledge-guided weakly supervised semantic segmentation framework, termed PhenoStruct-WSF.
91. Land-Use Change and Land-Cover-Based Ecological Quality Patterns in a Coal Resource-Based City: A Case Study of Ordos, China
Core Problem: This study investigated the long-term relationship between land-use change and land-cover-based ecological quality patterns in Ordos City, a typical coal resource-based city in northern China.
Key Innovation: This study investigated the long-term relationship between land-use change and land-cover-based ecological quality patterns in Ordos City, a typical coal resource-based city in northern China. In the internal semantic segmentation validation, the proposed model achieved a mean Intersection over Union (mIoU) of 69.51%, a mean accuracy (mAcc) of 81.65%, and a pixel-level overall accuracy (aAcc) of 83.16%.
92. Fly High or Fly Low? Selecting Time-Efficient UAV Search Strategies for High-Recall Aerial Detection
Core Problem: The use of uncrewed aerial vehicles (UAVs) for remote sensing continues to increase, and, in missions such as search and rescue (SAR) and landmine detection, a vision-based detector with high recall is critical.
Key Innovation: Flying at a higher altitude covers a larger area in each pass but enlarges the ground sample distance, reducing recall and precision. Experimental results indicate that Strategy 1 is more effective when targets are dense, and Strategy 2 when targets are sparse, provided its survey altitude is well above the verification altitude.
93. Dual-Stream Spatial-Spectral Network with Nested Attention for Hyperspectral Image Classification
Core Problem: CNN-based methods are effective for local spectral-spatial extraction, but their limited receptive fields can weaken broader context modelling.
Key Innovation: To address these limitations, this study proposes the Dual-Stream Spatial-Spectral Network with Nested Attention (DSSN), which separates local spectral-spatial feature extraction from multi-scale spatial-context modelling before adaptive fusion. Experiments on Indian Pines, Pavia University and Salinas show DSSN achieves overall accuracies of 98.11%%, 99.88% and 99.82%, respectively, outperforming other baselines.
94. Dual-Space Knowledge Distillation with Cross-Geometric Feature Interaction for Hyperspectral Image Classification
Core Problem: Hyperspectral image (HSI) classification is critical for remote sensing but faces challenges in balancing accuracy and inference efficiency.
Key Innovation: In this paper, we propose Dual-Space Knowledge Distillation (DSKD), a novel dual-student dual-space KD framework integrating Euclidean (GCN) and Hyperbolic (HGCN) teachers to jointly train a native MLP student and a native HNN student.
95. DPFS-YOLO: Missed-Detection Alleviation and False-Detection Risk Suppression for Small Objects in Complex UAV Aerial Scenes
Core Problem: To address these challenges, this paper proposes DPFS-YOLO, a small-object detection method designed for complex aerial scenes.
Key Innovation: Built upon YOLOv8n, the proposed method integrates detail enhancement, foreground selection, and semantic guidance into a collaborative optimization framework, aiming to improve small-object feature representation while suppressing spurious background responses. Experiments on the VisDrone2019-DET, HIT-UAV, and TinyPerson datasets demonstrate that the proposed method effectively alleviates missed detections of small objects.
96. Geometric Adaptive Matched Filtering on HPD Manifolds for Radar Target Detection
Core Problem: However, existing MIG detectors still adopt a single geometric distance to quantify the dissimilarity between target signals and clutter, overlooking the availability of target prior information.
Key Innovation: However, existing MIG detectors still adopt a single geometric distance to quantify the dissimilarity between target signals and clutter, overlooking the availability of target prior information. Experimental results on simulated and measured data demonstrate that GAMF achieves superior detection performance and robustness, with an average detection performance gain of 4.6 dB compared with conventional methods, while.
97. RSEI-Based Assessment of Ecological Quality in an Alpine Transitional Region Incorporating Climate Accumulation Effects and Multiscale Drivers
Core Problem: Conventional Remote Sensing Ecological Index (RSEI) assessments, however, rely on single-date imagery and often fail to account for cumulative climatic effects.
Key Innovation: Conventional Remote Sensing Ecological Index (RSEI) assessments, however, rely on single-date imagery and often fail to account for cumulative climatic effects. Results show an overall improvement: mean RSEI increased from 0.63 to 0.69, Good/Excellent areas expanded from 65.6% to 79.9%, and the interquartile range decreased from 0.19 to 0.14.
98. Applying the IOTA2 Chain for Automated 10 m Crop Map Production in a Mediterranean Environment
Core Problem: However, obtaining 10 m crop distribution maps remains challenging in Mediterranean regions, where data availability is often limited and landscapes are fragmented.
Key Innovation: This study addresses these limitations by implementing the IOTA2 automated chain in Sardinia (Italy), to create a large-scale crop map specifically targeting Mediterranean crops. Results indicate that the simplified nomenclature (N25) provided more robust performances, achieving an overall accuracy (OA) of 0.77 with full sampling, compared to 0.61 for the detailed version.
99. Residential Footings on Expansive Soils: Design Methods and Climate Soil Structure Interaction
Core Problem: Abstract Expansive soils continue to be one of the most difficult and costly ground conditions faced in residential construction.
Key Innovation: This study presents a critical review of residential footing design frameworks, focusing on two key aspects: the prediction of potential ground movement and the determination of footing design requirements. Results show that, although general trends in soil reactivity are consistently identified, substantial discrepancies exist in predicted footing requirements.
100. Improving effect of soybean urease induced calcium carbonate precipitation combined with xanthan gum on wind erosion control and its field application
Core Problem: Sandstorm activities in desert regions cause severe damage to local structures and ecosystems.
Key Innovation: This study investigated the effectiveness of combining soybean urease induced carbonate precipitation (SICP) with xanthan gum (XG) compared to pure SICP in enhancing the surface-targeted wind erosion resistance of aeolian sand. Laboratory results showed that the combined treatment reduced the wind erosion rate by 85% compared to single-cycle pure SICP treatment, achieving efficacy comparable to multiple SICP treatments of 2-3.
101. Uncertainty-Aware Digital Twins for Safe Operation and Active Fault-Tolerance in Degrading Heterogeneous Pumping Systems
Core Problem: Conventional deterministic controls often fail in fluid systems under mechanical degradation due to epistemic uncertainties, leading to constraint violations.
Key Innovation: To bridge this gap, this study proposes an Adaptive Probabilistic Digital Twin (APDT) framework. Crucially, trajectory analysis reveals an emergent “strategic offloading” behavior: the APDT autonomously shifts operational loads from degraded to healthy units.
102. WheatScoper: A lightweight organ-based framework for multi-view wheat phenotyping using time-series RGB images
Core Problem: However, RGB image-based phenotyping still suffers from expensive pixel-level annotation, unstable organ-level segmentation across growth stages, and limited multi-trait extraction under complex field conditions.
Key Innovation: To address these issues, a high-throughput phenotyping framework (WheatScoper) was proposed, enabling organ-level segmentation and plot-level multi-trait extraction. On the held-out test set from the same site and growing season, WheatScopeNet achieved an mIoU of 0.869 and an mDice of 0.930.
103. Spatiotemporal dynamics of riverine dissolved organic matter and its linkages to cryospheric changes in the Qilian mountains
Core Problem: However, the studies of the characteristics and migration of DOM in multiple water bodies are still scarce.
Key Innovation: Through extensive field sampling and measurements, this study systematically evaluates the spatiotemporal characteristics of dissolved organic carbon (DOC), total dissolved nitrogen (TDN), and dissolved organic nitrogen (DON) across six alpine rivers and two glacierized mountain sites in the high-altitude Qilian cryosphere. These findings are critically important for advancing our understanding of carbon and nitrogen cycling.
104. Numerical simulations of the release of liquid hydrogen from a freight train and explosion in a long railway tunnel: Consequences for passenger train safety
Core Problem: However, LH2 might cause safety concerns in railway tunnels, especially when passenger and freight trains are allowed to transit simultaneously in the same tube.
Key Innovation: However, LH2 might cause safety concerns in railway tunnels, especially when passenger and freight trains are allowed to transit simultaneously in the same tube.
105. Geoscience of the energy-climate-sustainability nexus: earth system processes, resources, and pathways to a sustainable future
Core Problem: The transition to sustainable energy systems depends fundamentally on Earth system processes, making geoscience central to addressing the interconnected challenges of energy security, climate change, and sustainable development.
Key Innovation: The study provides an interdisciplinary geoscience framework that supports more resilient and sustainable energy transitions. The analysis shows that sustainable energy transitions require an integrated understanding of interactions among the lithosphere, atmosphere, hydrosphere, and biosphere rather than treating these systems independently.
106. Spatiotemporal dynamics of reference evapotranspiration in China (1981-2020): Insights into its teleconnection with sea surface temperature patterns
Core Problem: Reference evapotranspiration (ET0) governs agricultural irrigation water demand, yet its teleconnection with sea surface temperature (SST) patterns remains poorly understood under climate change.
Key Innovation: This study revealed the teleconnection between China’s ET0 and SST patterns based on the Penman-Monteith equation, multi-source data, machine learning models, and wavelet coherence analysis.
107. Impacts of submarine groundwater discharge on inland and seaside basins of estuaries
Core Problem: Although the biogeochemical significance of submarine groundwater discharge (SGD) has been widely documented, the influence of groundwater-estuary exchange on estuarine hydrodynamics remains underrepresented in many surface-water modeling approaches.
Key Innovation: This study uses coupled surface water-groundwater modeling to quantify how a permeable-bed representation affects salinity, tidal storage, and transport across eight synthetic estuary types and three tidal amplitudes. Particle-tracking diagnostics further showed that groundwater-estuary exchange alters transport pathways and residence behavior, with effects that depend strongly on estuary morphology and basin connectivity.
108. Glacial meltwater-driven suppression of riverine methane emissions in the high-altitude cryospheric regions
Core Problem: Headwater streams are widely recognized as hotspots of riverine methane (CH4) emissions; however, this understanding remains to be clarified in high-altitude cryospheric regions by the overriding influence of glacial meltwater.
Key Innovation: This study proposes a new conceptual framework for cryosphere-regulated hydro-biogeochemical coupling, highlighting the necessity of integrating these unique mechanisms into carbon cycle models for cryospheric regions. The results indicated that the downstream riverine sections of the upper Heihe River basin were weak atmospheric CH4 sources (0.39 ± 0.02 t CH4 yr⁻¹).
109. Spatially distributed modeling of biofilm-induced hydrodynamic evolution in porous media
Core Problem: However, most models treat biofilm zones as hydraulically uniform, neglecting internal heterogeneity and potentially compromising flow-field predictions.
Key Innovation: Here, we establish a spatially distributed framework for modeling the evolving hydrodynamics of biofilm growth in porous media by translating time-lapse biofilm images into distributed porosity and permeability fields, and evaluate it against the conventional uniform strategy. The results show that geometric heterogeneity promotes preferential colonization in narrow throats and the formation of larger pore-spanning clusters.
110. A routing-based approach to rank LID locations for runoff management at the watershed scale
Core Problem: Low Impact Development (LID) is widely used to mitigate urban flood runoff, yet selecting suitable locations at the watershed-scale remains challenging because hydrologic benefit depends on both local runoff-generation potential and upstream-downstream routing (the latter is often overlooked in previous research).
Key Innovation: We propose the Routing-Based LID Distribution Approach (RLDA), a routing-informed screening framework that ranks candidate locations by combining a Hydrologic Index (HI) and a Routing-Based Index (RBI) into a Comprehensive LID Benefit Index (CLBI). These findings provide an efficient, connectivity-aware ranking framework that bridges index-based screening and process-based validation for watershed-scale LID planning.
111. Machine learning approaches for resilient modulus prediction in soils and unbound pavement materials: A scoping review
Core Problem: The resilient modulus (MR) is a central input in mechanistic-empirical pavement design, but the repeated load triaxial test used to obtain it is costly and time-consuming, driving sustained effort to predict it from more accessible material properties.
Key Innovation: This scoping review maps the methodological approaches developed for this purpose and characterizes the transition from mechanistic-empirical formulations to machine-learning and hybrid frameworks.
112. Fragment behavior of layered rock subjected to blasting: Energy dissipation characteristics
Core Problem: The drilling-and-blasting method is widely applied in the fields of tunnel excavation and mining engineering.
Key Innovation: This study combines laboratory blasting tests, image processing, and three-dimensional (3D) finite element modelling to investigate the fragmentation features and energy dissipation laws of schist with varying foliation orientations and specific charges. Bedded rock (e.g. schist) features an obvious foliation structure, which has been shown to dominate blast-induced crack propagation and energy transfer.
113. Bayesian inference of secant and tangent stiffness degradation in calcareous sand using noisy shear force-displacement data
Core Problem: Conventional methods typically require setting initial stiffness, and calculating discrete secant stiffness, which often introduces subjectivity and propagates uncertainties from measurement errors.
Key Innovation: This study proposes a Bayesian framework for directly deriving the secant and tangent stiffness degradation characteristics of calcareous sand from experimental shear force-displacement data. The accuracy and robustness of the proposed method are demonstrated using both real-life measurements of calcareous sand and simulated data.
114. TrajFusionNet+: Transformer-Based Prediction of Pedestrian Crossing Intention via Fusion of Trajectory Representations and Scene Graphs
Core Problem: The pedestrian crossing intention task involves predicting whether pedestrians are likely to cross the road from the point of view of an autonomous vehicle.
Key Innovation: We introduce TrajFusionNet+, a novel transformer-based model for pedestrian crossing intention prediction. TrajFusionNet+ achieves improved state-of-the-art performance on the two most widely used pedestrian crossing intention datasets, PIE and JAAD.
115. Local Robustness Quantification for Naive Bayes Classifiers and Generative Forests: a General Approach
Core Problem: We provide methods for calculating the robustness of the predictions of two types of generative classifiers whose underlying distribution is a Probabilistic Graphical Model (PGM): naive Bayes classifiers and generative forests (a probabilistic extension of random forests).
Key Innovation: Following the paradigm of robustness quantification, we define the robustness of a prediction as the extent to which the distribution of the classifier can be perturbed without changing this prediction. We test our methods on benchmark datasets, demonstrate that the robustness value of a prediction serves as an indicator for its trustworthiness and compare our approach with other such indicators.
116. Physics-guided underwater image enhancement via multimodal spatial distance-decay constraint
Core Problem: To alleviate the lack of training data, an unsupervised contrastive learning framework based on unpaired images is constructed, enabling clear-domain data to participate in model optimization directly.
Key Innovation: To alleviate the lack of training data, an unsupervised contrastive learning framework based on unpaired images is constructed, enabling clear-domain data to participate in model optimization directly. Finally, a physics-guided constraint based on the underwater imaging model is incorporated to ensure the physical plausibility of the enhanced results.
117. An equivalent constitutive model for random pitting corroded steel exposed to marine environments
Core Problem: A finite element (FE) model was then developed to reconstruct random pitting morphologies and was validated against the experimental results, showing good agreement in both the overall tensile response and local failure modes.
Key Innovation: To provide an equivalent material description for residual performance assessment of marine and offshore steel structures, this study investigates the degradation of the nominal stress-strain relationship of Q355B steel subjected to random pitting corrosion. A finite element (FE) model was then developed to reconstruct random pitting morphologies and was validated against the experimental results, showing good agreement in.
118. Wave attenuation characteristics of moored water ballast floating breakwaters with different internal configurations
Core Problem: A water ballast floating breakwater, which can be considered as a sloshing-based passive system for wave attenuation, was examined for the protection of floating photovoltaic and wind farms.
Key Innovation: The hydrodynamic response, pitch motion, and mooring force of breakwaters with different ballast configurations were investigated in an experimental flume.
119. TMA-SegRNN: An enhanced decomposition-based hybrid model leveraging information-theoretic optimization and SegRNN for offshore wind speed forecasting
Core Problem: Short-term offshore wind-speed forecasting is challenging because marine-atmosphere coupling produces nonlinear, non-stationary, and multi-scale dynamics.
Key Innovation: This study proposes TMA-SegRNN, a hybrid model integrating time-varying filter-based empirical mode decomposition (TVF-EMD), maximal information coefficient (MIC), artificial protozoa optimizer (APO), and segment recurrent neural network (SegRNN). Experiments on three offshore wind speed datasets from Northern Fujian show that TMA-SegRNN reduces average RMSE by 83.92% compared with SegRNN, by 33.70% compared with TMA-XGBoost.
120. Numerical investigation of regional-scale collector plume dynamics and particle transport in deep-sea mining
Core Problem: However, mining activities generate sediment plumes that pose significant environmental risks to marine ecosystems.
Key Innovation: This study investigates the regional-scale dynamics of near-bottom collector plumes using numerical simulations based on the TELEMAC-3D system. Key findings indicate that ambient ocean currents dominate the regional-scale plume dynamics and sediment transport, with plumes extending up to 30 km and high-concentration zones confined within 3-4 km.
121. Genotype-Aware Prediction of Soybean Seed Composition from Multimodal UAV Imagery of the Standing Crop
Core Problem: Geospatial artificial intelligence (GeoAI) integrates multimodal remote sensing with deep learning to model complex agricultural systems at scale.
Key Innovation: Within this framework, accurate and non-destructive prediction of seed composition from in-season standing crops is essential for breeding and precision agriculture. The highest accuracy was achieved for sucrose (R² = 0.80), followed by simple carbohydrate (R² = 0.70) and starch (R² = 0.55), with notable gains from GEN and DAS.
122. Multi-Timescale Variations in Cloud Water Resources and Their Relationships with Climatic and Environmental Factors over Northwest China
Core Problem: Northwest China is characterized by severe water scarcity, and cloud water represents an important potential supplement to regional water availability.
Key Innovation: In this study, ERA5 and JRA-3Q reanalysis datasets (1960-2025) and satellite cloud products from MODIS and Cloud_cci (2003-2016) were used to evaluate the consistency of multi-source datasets in capturing cloud water path variations. The results indicated these datasets generally agreed on the temporal variability of cloud water path, whereas differences were found in the absolute ice water path (IWP) and total cloud water.
123. Soil organic carbon content, stability and saturation in black soils of Northeast China
Core Problem: Black soils cover roughly 20% of cropland worldwide and constitute a critical agricultural resource for food security.
Key Innovation: Soil organic carbon (SOC) in black soil is related to arable land quality and soil fertility. Our results show that MAOC and POC contents decline progressively with depth, whereas SOC physical stability increases and the SOC humification degree intensifies.
124. Contrasting Fe and Al controls on mineral-associated organic carbon across a tropical wetland toposequence in palm swamps (Vereda)
Core Problem: Wetlands are major global natural carbon reservoirs, yet the mechanisms controlling organic carbon persistence in tropical systems remain poorly constrained.
Key Innovation: In the Brazilian Cerrado, Veredas (groundwater-fed palm swamp wetlands) are characterized by strong hydrological gradients, dynamic redox conditions, and highly weathered soils. Iron-associated carbon decreased along the gradient and showed no consistent relationship with organic carbon, reflecting redox-driven dissolution and mobilization.
125. Rapid non-destructive characterization of shallow concrete microdamage based on photoluminescence sensing technology
Core Problem: To enable early detection of microdamage in tunnel linings, this study presents a field-oriented, color-based extension of the contact sponge-fluorescence tracing technique for the qualitative identification of shallow microdamage in reinforced concrete cover zones.
Key Innovation: To enable early detection of microdamage in tunnel linings, this study presents a field-oriented, color-based extension of the contact sponge-fluorescence tracing technique for the qualitative identification of shallow microdamage in reinforced concrete cover zones. The feasibility of the approach is examined through laboratory four-point bending tests, linear-elastic stress-state analysis, and a proof-of-concept.