TerraMosaic Daily Digest: September 22, 2026
Daily Summary
Mass-movement studies isolate why apparently similar slopes diverge. Two Mackenzie Valley failures define permafrost detachment slides as deep, rapid translational failures driven by bottom-up thaw; the mapped examples displaced 4.5 and 18 million m³, and hundreds of analogues occur in warm, thin permafrost. A comparison of burned Californian catchments shows that post-fire grain-size reduction raises Shields stress by two orders of magnitude, far exceeding the up-to-threefold runoff effect and explaining the presence or absence of debris flows under intense rainfall. At the Aru glaciers, elevation redistribution and velocity acceleration track evolving instability, whereas temperature and albedo do not provide collapse-specific anomalies and SAR backscatter responds differently between the two glaciers.
Earthquake and volcanic studies combine cross-system mechanics with event-spanning observations. In the Aegean, geodetic and seismic evidence reconstructs an elastic-stress relay from Santorini reservoir pressurization to a Kolumbo dike and a Mw 6.2 aseismic Amorgos-fault transient. Digitized records from 52 Pacific tide gauges indicate that the 1952 and 2025 Kamchatka earthquakes ruptured nearly colocated megathrust areas only 73 years apart, challenging conventional slip-budget accounting. Complementary geotechnical studies connect liquefaction triggering to consequences, distinguish extension and compression rheology, and test cyclic response across sand, marine clay and offshore foundations.
Hazard assessment is moving from static susceptibility toward process and decisions. A regional debris-flow index integrates simulated depth, velocity and affected area for 47 gullies and changes six classifications relative to a conventional scheme. Rainfall-conditioned geohazard warning in Nujiang retains a 76.5% retrospective detection rate while reducing the warned footprint, and a flood-evacuation model transfers multi-agent simulation labels from 4,278 residences to 724,845 unsimulated locations. Uncertainty-aware flood-risk modeling, donor-gage drought prediction and national Brazilian storm profiles further connect physical variability to mitigation choice, ungauged-basin planning and design discharge.
Earth-observation methods increasingly expose the conditions under which models fail. Terrain-aware multimodal landslide detection aligns optical, SAR and topographic inputs; DEM-resolution analysis treats susceptibility performance and explainability as scale-dependent; and Colombian InSAR-GNSS mapping explicitly separates deformation screening from validated landslide warning. Annual Earth-observation embeddings simplify burned-area mapping across continents but retain seasonal and landscape-specific failure modes. Geographic implicit representations, InSAR phase reconstruction, ice-bed radar extraction and foundation-model comparisons extend transferable capability while keeping spatial scale, sensor physics and observation geometry explicit.
Key Trends
The strongest papers replace single indicators and static maps with mechanism-specific observations, explicit uncertainty and decision-linked outputs.
- Hazard occurrence is being explained through material state: Basal permafrost thaw, channel grain size, debris-flow dynamics and rock-ice mixture behavior determine whether forcing becomes failure, runout or structural loading.
- Precursors are becoming indicator-specific: Glacier elevation, velocity and SAR backscatter respond to different parts of the instability process, arguing against a universal anomaly threshold and for complementary observations.
- Cascading deformation is resolved across systems: Caldera collapse, reservoir-dike-fault stress transfer, repeated megathrust rupture and InSAR deformation mapping reconstruct linked processes that single-event or single-geometry observations miss.
- Models are being evaluated against operational decisions: Flood mitigation, evacuation timing, drought transfer and rainfall-triggered warning are judged by consequence, spatial selectivity and uncertainty rather than predictive accuracy alone.
- Scale and protocol are treated as sources of model variability: DEM resolution, evaluation protocol, spatial holdout and sensor fusion are becoming explicit experimental variables in susceptibility, forecasting and remote-sensing systems.
Selected Papers
The 22 September selection is led by a newly characterized permafrost-landslide process, a mechanistic explanation for contrasting post-fire debris-flow occurrence, elastic stress relay across the Santorini-Kolumbo-Amorgos system and tsunami evidence for repeated Kamchatka megathrust rupture. Companion studies address glacier-collapse precursors, multimodal landslide detection, dynamic debris-flow classification, rainfall-conditioned geohazard warning, uncertainty-aware flood decisions, drought transfer, liquefaction and scale-sensitive Earth-observation models.
1. Permafrost Detachment Slides: A Distinct High-Magnitude Permafrost Mass Wasting Process
Core Problem: Warming permafrost is generating deep, rapid mass movements not captured by familiar active-layer detachment models.
Key Innovation: Defines permafrost detachment slides from two 4.5–18 million m³ failures, attributes initiation to bottom-up thaw, and identifies hundreds of analogues in warm, thin Mackenzie Valley permafrost.
2. Grain size and sediment storage explain why postfire debris flows are absent in some mountainous landscapes
Core Problem: Comparable post-fire rainfall can produce abundant debris flows in one mountain landscape and none in another.
Key Innovation: Field comparison and numerical experiments show that fire-driven grain-size reduction raises Shields stress by two orders of magnitude, dominating the up-to-threefold runoff effect.
3. Cascading magmatic unrest and aseismic faulting in the Aegean Sea enabled by elastic stress relay
Core Problem: Unrest can migrate among clustered magma reservoirs and fault systems, but the mechanical relay is rarely observed across an entire crisis.
Key Innovation: Geodetic and seismic constraints reconstruct stress transfer from Santorini pressurization to a Kolumbo dike and then a Mw 6.2 aseismic Amorgos-fault transient.
4. Pacific tsunami records suggest similar rupture locations of two great Kamchatka earthquakes only 73 years apart
Core Problem: The nearly consecutive 1952 and 2025 Kamchatka great earthquakes challenge conventional megathrust slip-budget expectations.
Key Innovation: Digitized records from 52 Pacific tide gauges and tsunami modeling indicate nearly colocated ruptures only 73 years apart, motivating fast viscoelastic energy accumulation.
5. MHT-MambaNet: a multi-source heterogeneous terrain-aware Mamba network for landslide detection in remote sensing imagery
Core Problem: Optical, SAR and terrain inputs differ in representation and make boundary-preserving landslide mapping difficult in rugged mountains.
Key Innovation: Aligns heterogeneous modalities, uses direction-adaptive Mamba interactions and object-level change decoding to improve segmentation and boundary delineation.
6. Debris-Flow Hazard Assessment Considering Dynamic Characteristics Index with High-Resolution Topography in Duoxiongla Valley (China)
Core Problem: Regional gully rankings seldom include flow depth, velocity and inundation extent from process simulations.
Key Innovation: Constructs a dynamic characteristics index for 47 gullies from 0.5 m imagery, 5 m LiDAR and numerical simulations, changing six hazard classifications relative to a conventional assessment.
7. Multi-Source Remote Sensing Data Reveal the Instability Evolution and Precursory Signals Before the Collapse of the Aru Glaciers on the Tibetan Plateau
Core Problem: Candidate remote-sensing precursors to glacier collapse have not been compared within a common anomaly framework.
Key Innovation: A 1990–2016 synthesis separates mass-redistribution and velocity signals from weaker temperature/albedo indicators and identifies localized SAR-backscatter sensitivity to crevassing at Aru 53.
8. A fluvial flood risk model for quantifying the benefit of mitigation measures under uncertainty
Core Problem: Flood-mitigation options must be compared while separating natural variability from model uncertainty.
Key Innovation: Builds an efficient risk framework for combined mitigation measures and demonstrates it on a flood-detention basin along the Bavarian Danube.
9. Predicting streamflow drought in the conterminous United States using machine learning and a donor-gage approach, 1982-2020
Core Problem: Streamflow-drought controls vary regionally, while many catchments lack continuous gauge observations.
Key Innovation: Combines machine learning with donor-gage transfer across the conterminous United States to predict drought in ungauged basins and resolve regional roles of temperature, snow and rainfall.
10. Meteorological Hazard Assessment for Risk Early Warning of Geohazards Based on Causal-Heuristic Coupling: A Case Study of Nujiang Prefecture, China
Core Problem: Static susceptibility overlays do not represent spatially heterogeneous rainfall responses.
Key Innovation: Couples susceptibility, rainfall thresholds and a causal-forest-derived heuristic modulation factor; retrospective Nujiang tests retain 76.5% detection while sharply narrowing warning footprints.
11. Mapping Active Surface Deformation in Fusagasugá and Silvania, Colombia, Using Sentinel-1 SBAS-InSAR and Campaign GNSS Observations
Core Problem: Mountain communities in Colombia lack spatially continuous deformation measurements for prioritizing geological-hazard investigation.
Key Innovation: Maps coherent Sentinel-1 SBAS-InSAR deformation and compares it with campaign GNSS, explicitly treating the result as a screening product rather than validated landslide warning.
12. Development and internal validation of a machine-learning framework for spatial screening using simulation-derived flood evacuation planning typologies
Core Problem: Detailed inundation and agent-based evacuation simulations cannot cover every residence in a basin.
Key Innovation: Learns six planning typologies from 4,278 simulated residences and screens 724,845 additional locations, with nested-cross-validation balanced accuracy of 0.720 for the combined typology.
13. DEM investigation of dry rock-ice avalanche impacts on slit dams: effects of ice content, slit size and flow inertia
Core Problem: Title-level focus: identifies uncertainty in how ice fraction, slit geometry and flow inertia control avalanche impact on slit dams.
Key Innovation: Title-signalled approach or contribution: The title reports a DEM investigation of coupled material and barrier-design controls; public metadata did not expose methods or quantitative results at review time. Methods, data and results could not be assessed because no reliable abstract was available.
14. A probabilistic energy-based framework for performance-based liquefaction hazard assessment: From triggering to consequence
Core Problem: Title-level focus: identifies the need to connect liquefaction triggering probability to consequences within one hazard framework.
Key Innovation: Title-signalled approach or contribution: The title reports a probabilistic energy-based triggering-to-consequence formulation; the public Crossref record did not expose validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
15. DEM resolution drives AI variability: Trade-offs between Performance and Explainability in Landslide Susceptibility Prediction
Core Problem: Title-level focus: identifies DEM resolution as a source of variability in both predictive performance and explanation of AI landslide-susceptibility models.
Key Innovation: Title-signalled approach or contribution: The title frames performance–explainability trade-offs across terrain resolutions; public Crossref metadata did not expose datasets, methods or numerical results. Methods, data and results could not be assessed because no reliable abstract was available.
16. Time Distribution of Heavy Rainfall in Brazil: Empirical Huff Curves from 290,164 Sub-Daily Storm Events
Core Problem: Imported temporal storm profiles can bias hydrological design where national sub-daily observations are sparse.
Key Innovation: Derives empirical Huff curves from 290,164 Brazilian storm events and shows that local curves raise median design peak discharge by 8%, with open regional parameters.
17. Annual Earth-observation embeddings encode wildfire disturbance and support simplified burned area mapping
Core Problem: Operational burned-area mapping usually depends on curated fire imagery or dense image time series.
Key Innovation: Shows that annual Tessera embeddings support simple burned-area classifiers, cross-continental transfer and ignition-date recovery, while documenting end-of-year and landscape-specific failure modes.
18. FAST-ML: A Hybrid Physics-Machine Learning Framework for Tropical Cyclone Intensity Forecasting
Core Problem: Tropical-cyclone intensity forecasts must reconcile sparse observations with physically constrained rapid intensification.
Key Innovation: Combines physical constraints with machine learning to forecast cyclone intensity; its value is a direct hazard application rather than a field-wide forecasting breakthrough.
19. Infra-Net: a robust parallel decision-making network for discriminating natural hazards and anthropogenic infrasound events via multi-view feature learning
Core Problem: Infrasound monitoring must distinguish natural hazards from anthropogenic signals under variable propagation conditions.
Key Innovation: Uses parallel multi-view feature learning to discriminate natural-hazard and anthropogenic infrasound events for monitoring applications.
20. Extending daily river discharge records across China using satellite-derived river widths
Core Problem: Short and discontinuous gauge records limit hydrological analysis across China.
Key Innovation: Uses satellite-derived river widths to extend daily discharge records nationally, increasing temporal coverage for flood and drought analysis.
21. PEGNet: A Peridynamics-Inspired and Emergent-Feature-Conditioned Spatio-Temporal Graph Neural Network for Land Subsidence Modeling
Core Problem: Sequence-only models struggle to represent spatial coupling and local physical consistency in InSAR deformation forecasts.
Key Innovation: Combines GRU temporal encoding, graph attention and peridynamics-inspired local consistency; diagnostics show improved spatial consistency without overstating aggregate accuracy gains.
22. Intersecting Hazards and Insecurity: Disaster Governance, Displacement, and Conflict Risk in the Philippines
Core Problem: Weak governance can transform hazard shocks into displacement, contested recovery and conflict, but the intervening mechanisms are underspecified.
Key Innovation: A mechanism-focused Philippine case analysis identifies displacement and delayed recovery as the hinge connecting preparedness, distributional decisions, legitimacy and conflict risk.
23. Unified Fluid Approach Based on Extension-Compression Apparent Viscosity for Liquefaction Triggering
Core Problem: Liquefaction-triggering models need to represent different apparent viscosities in extension and compression.
Key Innovation: Develops a unified-fluid formulation with extension–compression apparent viscosity for triggering analysis.
24. Liquefaction behavior of clean sand in CSSTs and evaluation of cyclic responses: Effects of particle size
Core Problem: Title-level focus: identifies particle-size effects on cyclic sand response in cyclic simple shear tests.
Key Innovation: Title-signalled approach or contribution: The title reports controlled liquefaction testing and cyclic-response evaluation; public Crossref metadata did not provide an abstract at review time. Methods, data and results could not be assessed because no reliable abstract was available.
25. A physically constrained multivariate bias correction framework for projecting compound drought and heatwave risk in the Yangtze River Basin
Core Problem: Title-level focus: identifies the need to preserve multivariate physical dependence when bias-correcting compound drought–heat projections.
Key Innovation: Title-signalled approach or contribution: The title reports a physically constrained multivariate correction framework for Yangtze Basin compound risk; quantitative performance was unavailable in public metadata. Methods, data and results could not be assessed because no reliable abstract was available.
26. Numerical investigation of seismic dynamic response and liquefaction characteristics of suction bucket foundation
Core Problem: Title-level focus: identifies coupled foundation response and soil liquefaction under seismic loading.
Key Innovation: Title-signalled approach or contribution: The title reports numerical analysis of suction-bucket foundation dynamics and liquefaction; quantitative evidence was unavailable in public metadata. Methods, data and results could not be assessed because no reliable abstract was available.
27. MIND the Gap: A Geographic Implicit Neural Representation with Adjustable Spatial Scale
Core Problem: However, generalization to distant regions remains largely unexplored, despite its importance for remote sensing applications.
Key Innovation: We introduce Matryoshka Implicit Neural Distillation (MIND), which distills embeddings from specialist pretrained geospatial models into a single generalist coordinate embedding with adjustable spatial granularity.
28. Agentic Building-Aware Satellite Gaussian Splatting for Auditable Urban DSM Reconstruction
Core Problem: Urban-scale 3D reconstruction from satellite imagery supports disaster response, city monitoring, and geospatial digital twins, yet neural rendering methods typically optimize average visual fidelity rather than the structures that analysts inspect first: buildings.
Key Innovation: We present an agentic building-aware satellite Gaussian Splatting workflow that uses Segment Anything-derived building masks as semantic priors and an Agentic Reconstruction Controller to select, verify, and record DSM reconstruction policies.
29. Hi-OPD: Hierarchy-Aware Open-Prompt Detection for Remote Sensing Images
Core Problem: Hi-OPD addresses a failure mode left uncontrolled by flat open-prompt training: descendant retrieval need not persist under ancestor queries when multi-source remote sensing annotations exhibit inconsistent granularity and missing labels.
Key Innovation: We propose Hi-OPD, a hierarchy-aware open-prompt detector, and construct RS153-HierOPD from 175,644 retained training image/tile records and 3.48M boxes mapped to 153 atomic categories with sparse hierarchy and alias relations.
30. Shallow-to-deep velocity model building via diffusion models-Part I: Method and Proof of concept
Core Problem: Traditional methods demand high-quality starting models and, also, remain limited in resolution in coverage and computationally intensive.
Key Innovation: To address this issue, we propose a depth-progressive diffusion framework that constructs velocity models incrementally from shallow to deep by propagating prior information.
31. Shallow-to-deep velocity model building via diffusion models-Part II: Realistic scenarios
Core Problem: Full-waveform inversion (FWI) requires accurate initial velocity models to avoid cycle-skipping, but constructing such models remains challenging in practice.
Key Innovation: Building on the depth-progressive diffusion framework introduced in Part~I, which relied on idealized reflectivity constraints, this work adapts the methodology to realistic exploration scenarios.
32. FREESIA: Covariance-Aware Posterior Transport for Expressive and Scalable Data Assimilation
Core Problem: However, accurate inference of the multimodal posteriors induced by nonlinear or non-injective observation operators remains a key challenge under high-dimensional and sparse observation conditions.
Key Innovation: To address these issues, we propose a training-free, asymptotically exact posterior transport method.
33. Hydrological Constraints on Temperature Sensitivities of Precipitation Frequency and Intensity
Core Problem: To improve understanding of how precipitation frequency and intensity change with warming, we combine the coupled land--atmosphere water balance with a stochastic hydrological model that retains explicit dependence on precipitation frequency and event depth.
Key Innovation: To improve understanding of how precipitation frequency and intensity change with warming, we combine the coupled land--atmosphere water balance with a stochastic hydrological model that retains explicit dependence on precipitation frequency and event depth.
34. Effect of initial water content on undrained cyclic behaviour of reconstituted marine soft clay
Core Problem: Initial water content alters pore-pressure accumulation, stiffness degradation and cyclic strength in marine soft clay.
Key Innovation: Monotonic and cyclic triaxial tests show that normalizing cyclic stress by monotonic shear strength collapses results from three initial water contents onto a common curve under the tested conditions.
35. AGPC: an Annual 500 m Gridded Population (1990-2020) for China incorporating 3D building volume dynamics
Core Problem: AGPC: an Annual 500 m Gridded Population (1990–2020) for China incorporating 3D building volume dynamics Xiaocong Xu, Shiyu He, Jinpei Ou, Yan Zhou, and Xiaoping Liu Earth Syst.
Key Innovation: Data, 18, 6995–7019, https://doi.org/10.5194/essd-18-6995-2026, 2026 We developed annual population maps for mainland China from 1990 to 2020 at 500-metre detail.
36. A National Depth-Resolved Soil-Moisture-to-Electromagnetic Proxy Database for Hydrogeophysical Monitoring
Core Problem: Subsurface soil moisture is central to drought, irrigation and recharge but is difficult to monitor consistently below the surface.
Key Innovation: Links measurements from 117 stations and six depths with soil properties and expected electromagnetic responses, identifying a mid-depth buffering zone and publishing an open monitoring reference.
37. The North American CORDEX-CMIP6 WRF evaluation run: comparing historical simulations from 25 km to convection-permitting scales
Core Problem: Regional climate simulations must improve precipitation timing and extremes without always requiring convection-permitting computational cost.
Key Innovation: Evaluates a 12 km CORDEX-CMIP6 WRF run against 25 km and 4 km configurations, finding reduced biases and extreme-precipitation behavior close to the convection-permitting simulation.
38. Characterising runoff processes for Australia: insights from a parsimonious rainfall-runoff event identification method
Core Problem: Characterising runoff processes for Australia: insights from a parsimonious rainfall-runoff event identification method Mohammad Masoud Mohammadpour Khoie, Danlu Guo, and Conrad Wasko Hydrol.
Key Innovation: We introduce RVEIM, a parsimonious method for event identification that improves both robustness and physical plausibility.
39. An eastward-propagating coupled mode amplifying the early growth of extreme El Niño events
Core Problem: Extreme El Niño events are among the most consequential climate phenomena, yet their early prediction remains challenging.
Key Innovation: Here, by analyzing the 1982/1983, 1997/1998, and 2015/2016 events together with a large ensemble of climate model simulations, we identify an unusually early and reproducible ocean-atmosphere precursor to extreme El Niño: a slow, eastward-propagating coupled mode that develops in the western-central equatorial Pacific during boreal winter and spring.
40. TCSF-Net: Transformer-Cloth Simulation Filtering Fusion Network for Semantic Segmentation of Densely Vegetated River Levee LiDAR Point Clouds
Core Problem: Accurate semantic segmentation of LiDAR point clouds from urban river levees is critical for flood hazard assessment and infrastructure maintenance, yet dense vegetation cover severely occludes the terrain and introduces ambiguous return signatures, challenging conventional filtering and learning-based methods.
Key Innovation: To address this, we propose TCSF-Net, a hybrid framework that synergistically integrates a Transformer backbone with a cloth simulation filtering (CSF) prior.
41. Experimental and Statistical Assessment of Variability and Uncertainty in Sandstone Failure Criteria
Core Problem: The inherent heterogeneity of rocks introduces significant variability in strength characteristics, which is often neglected in conventional deterministic failure analyses, leading to potentially unreliable predictions.
Key Innovation: This study develops an integrated experimental–statistical framework to quantify and propagate this variability into the uncertainty assessment of rock failure criteria.
42. Remote Detection of Streambank Erosion and Channel Incision Using High-Resolution, Repeat Lidar-Derived Metrics in Raleigh, North Carolina
Core Problem: However, unexplained variability remained, demonstrating the complexity of stream stability in urban watersheds.
Key Innovation: The study included five components.
43. Mapping Glacier Bed Topography in Crevassed Regions: A New Passive Seismic Reflection Method Applied to Isunnguata Sermia, West Greenland
Core Problem: However, heavily crevassed regions are challenging to survey using conventional methods, leading to data gaps.
Key Innovation: However, heavily crevassed regions are challenging to survey using conventional methods, leading to data gaps.
44. Basic friction angle of rock surfaces under varying water saturation, low-temperature, and freeze-thaw conditions
Core Problem: Title-level focus: identifies the problem signaled by: Basic friction angle of rock surfaces under varying water saturation, low-temperature, and freeze-thaw conditions.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
45. A grid-informed physics-guided graph deep learning framework for interpretable distributed daily streamflow prediction
Core Problem: Title-level focus: identifies the problem signaled by: A grid-informed physics-guided graph deep learning framework for interpretable distributed daily streamflow prediction.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
46. Continuous Phase Reconstruction in Two Stages and Global Graph Optimization for InSAR Deformation Monitoring over Large Areas
Core Problem: Title-level focus: identifies the problem signaled by: Continuous Phase Reconstruction in Two Stages and Global Graph Optimization for InSAR Deformation Monitoring over Large Areas.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
47. RGB Optical Remote Sensing in the Deep Learning Era: A Systematic Review of Methods, Models, Datasets, Challenges, and Future Research Directions
Core Problem: Title-level focus: identifies the problem signaled by: RGB Optical Remote Sensing in the Deep Learning Era: A Systematic Review of Methods, Models, Datasets, Challenges, and Future Research Directions.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
48. An Automatic Frequency-Spatial Deep Learning Framework for Continuous Ice-Bed Interface Extraction from Ice-Penetrating Radar Data
Core Problem: Title-level focus: identifies the problem signaled by: An Automatic Frequency–Spatial Deep Learning Framework for Continuous Ice–Bed Interface Extraction from Ice-Penetrating Radar Data.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
49. Nonuniform Mantle Suture Beneath Northern Tibet Revealed by Helium Isotopes and Attention-Based Machine Learning: Implications for Crustal Rheology and Growth of the Tibetan Plateau
Core Problem: However, inconsistent geophysical observations have led to long‐standing controversies over whether the Asian lithosphere beneath northern Tibet is undergoing underthrusting or southward subduction.
Key Innovation: The results show that attention‐based machine learning (ML) can reveal the spatial pattern of helium isotope ratios across the plateau and delineate a helium boundary in northern Tibet ( R C / R A > 0.1 R A ).
50. Evaluating Accuracy and Probabilistic Reliability of Zero-Shot Time Series Foundation Models
Core Problem: However, existing works often overlook trade-offs between predictive accuracy and probabilistic calibration.
Key Innovation: However, existing works often overlook trade-offs between predictive accuracy and probabilistic calibration.
51. SPEANet: Structural Prior Enhanced Attention Network for Parameter-Efficient Remote Sensing Object Detection
Core Problem: However, directly injecting these responses can amplify content-irrelevant textures, while applying a uniform operator design across the hierarchy may be poorly matched to stage-specific representation requirements.
Key Innovation: We propose the Structural Prior Enhanced Attention Network (SPEANet), a parameter-efficient RSOD backbone that integrates fixed operators through stage-specific prior extraction and context-conditioned response modulation.
52. MGRL-RSCC: Multi-Granularity Reward Reinforcement Learning for Fine-Grained Remote Sensing Change Captioning
Core Problem: Remote Sensing Change Captioning (RSCC), which aims to generate accurate and detailed linguistic descriptions of ground object variations from bi-temporal remote sensing images, is a critical and challenging task in intelligent remote sensing interpretation.
Key Innovation: The mainstream autoregressive training paradigm faces severe exposure bias and train-test distribution mismatch, resulting in cumulative generation errors.
53. Quantifying Protocol-Induced Uncertainty in Comparative Predictive-Model Evaluation: Evidence from Large-Scale Daily PM10 Forecasting
Core Problem: We formalize this problem as protocol-induced ranking uncertainty and introduce a framework that compares ranking displacement caused by switching protocols with displacement produced by conventional choices within a fixed protocol.
Key Innovation: These results show that model-selection conclusions can depend materially on legitimate evaluation choices.
54. Distributed Proximal Stein Variational Gradient Descent Algorithm for Large-scale Bayesian Inference in Traveltime Tomography
Core Problem: This architecture eliminates the need for the explicit construction of large-scale sensitivity matrices, significantly reducing the memory footprint for 3D surveys.
Key Innovation: We present a distributed framework for large-scale Bayesian inverse problems governed by the eikonal equation, with a specific focus on seismic traveltime tomography.
55. A Multi-Timestep LSTM Ensemble regressor for Enhanced Short-Term Runoff Prediction
Core Problem: Accurately forecasting river runoff is key to managing water resources, controlling floods, and planning agriculture.
Key Innovation: We introduce a daily runoff prediction model based on Long Short-Term Memory (LSTM) networks.
56. A dataset of high-spatiotemporal-resolution dust and non-dust aerosol mass concentration profiles from ground-based remote sensing in central China (2018-2024)
Core Problem: The vertical distribution of dust and non-dust aerosol over central China is poorly constrained by surface observations alone.
Key Innovation: Uses continuous ground-based lidar from 2018–2024 to resolve aerosol mass profiles and seasonal dust/non-dust structure for satellite and model evaluation.
57. Reduced Asian monsoon influence on Mediterranean summers in a warmer climate
Core Problem: Future Mediterranean summers depend partly on how warming alters the remote influence of the South Asian monsoon.
Key Innovation: Climate-model projections under a high-emissions scenario indicate that circulation shifts weaken the monsoon's influence on Mediterranean summer weather.
58. A strong Pacific Walker Circulation during the Little Ice Age
Core Problem: However, the strength of the east-west tropical Pacific atmospheric circulation (the Pacific Walker Circulation) during the Little Ice Age (LIA; approximately 1450 to 1850 CE), our most recent cool period, remains a subject of debate.
Key Innovation: We present a third line of evidence, the surface-to-thermocline temperature difference in the Indonesian Seas, which was smaller during the LIA than during the Medieval Climate Anomaly (approximately 950 to 1250 CE), suggesting a deeper thermocline, in agreement with hydroclimate evidence for a stronger LIA Walker Circulation.
59. DLMBMT: A Deep Learning Method Based on the Multi-Scale Technique for Hyperspectral Image Classification with Limited Training Samples
Core Problem: Hyperspectral classifiers must combine local, global and multiscale spectral–spatial information when only limited training samples are available.
Key Innovation: Combines spectral–spatial extraction, an improved 3D Swin Transformer and Mamba-assisted multiscale residual processing; evaluation spans four datasets.
60. Downscaling SMAP Soil Moisture to 250 m Using Deep Learning Models and Multi-Source Environmental Variables over the Loess Plateau
Core Problem: Coarse soil-moisture products cannot resolve heterogeneous terrain, but nominally finer predictions do not necessarily improve field agreement.
Key Innovation: Downscales SMAP L4 from 9 km to 250 m with SE-ResNet and eight predictors, then shows that in-situ performance varies by station and does not consistently exceed the original product, exposing scale-dependent uncertainty.
61. Surface Roughness and Chemical Weathering of Fluvial Geomorphic Surfaces: A Case Study from Qingyi River, Eastern Tibetan Plateau
Core Problem: River terraces and alluvial fans preserve important records of landscape evolution, yet establishing their relative chronology is often difficult where suitable materials for numerical dating are scarce.
Key Innovation: We combine digital-elevation-model-based roughness analysis with major-element geochemistry and previously established age constraints for a sequence of river terraces and paleo-alluvial fans developed under humid and strongly erosive conditions.
62. Climatic and Tectonic Controls on Late Quaternary Fluvial Terrace Development in the Turpan Fold-Thrust Belt, Eastern Tianshan
Core Problem: Fluvial terraces provide important archives for understanding the relative roles of tectonics and climate in landscape evolution.
Key Innovation: Our results show that terrace aggradation occurred mainly during marine isotope stage (MIS) 3c (ca.
63. Mitigation of mud cake adhesion by a humate-based dispersant and its compatibility with slurry performance in slurry shield tunnelling
Core Problem: Mud cake adhesion in high-plasticity clay strata can increase cutterhead torque, impair slurry circulation, and cause stoppage-related risks during slurry shield tunnelling.
Key Innovation: The results show that 0.8%–1.5% HH maintained a reasonable balance among slurry viscosity, filtration–mud-film formation, and spoil-carrying capacity, whereas dosages above 2.0% caused excessive viscosity reduction, increased filtrate volume, and reduced spoil-carrying performance.
64. Dynamic response of pile driven by different impact energies in clay-mudstone layers: field understanding based on fiber optic sensing
Core Problem: Pile driving in layered clay–mudstone frequently encounters refusal, excessive rebound, and stress-induced damage, highlighting the need to optimize hammer energy for safe and efficient installation.
Key Innovation: To address this gap, this study presents one of the few full-scale field investigations of driven closed-end PHC piles in layered clay–mudstone profiles using two hammer energies (101 kJ and 81 kJ).
65. High-fidelity displacement reconstruction method for offshore wind turbines under extreme sea conditions and high noise
Core Problem: Title-level focus: identifies displacement reconstruction under extreme sea states and high measurement noise as the target problem.
Key Innovation: Title-signalled approach or contribution: The title reports a high-fidelity reconstruction method for offshore wind turbines; public metadata did not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
66. Study on uniaxial compression mechanical properties and mesoscopic damage mechanism of granite under freeze-thaw cycles
Core Problem: Title-level focus: identifies how repeated freezing and thawing modify granite strength and mesoscopic damage.
Key Innovation: Title-signalled approach or contribution: The title reports uniaxial-compression and mesoscopic-damage analysis; public metadata did not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
67. Effects of freeze-thaw cycling and intermediate principal stress on the deformation and three-dimensional strength of sodium sulfate saline soil
Core Problem: Title-level focus: identifies coupled effects of freeze–thaw cycling and intermediate principal stress on saline-soil deformation and strength.
Key Innovation: Title-signalled approach or contribution: The title reports a three-dimensional strength investigation of sodium-sulfate saline soil; public metadata did not expose methods or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
68. A freezing-index-based criterion for passive frost protection design in cold-region tunnels
Core Problem: Title-level focus: identifies the need for a climate-linked design threshold for passive tunnel frost protection.
Key Innovation: Title-signalled approach or contribution: The title reports a freezing-index-based design criterion; public metadata did not expose derivation, validation or quantitative thresholds. Methods, data and results could not be assessed because no reliable abstract was available.
69. Depth-dependent response functions and rainfall transition zones of soil-profile activation under varying antecedent moisture in the Shandian River Basin, northern China
Core Problem: Title-level focus: identifies the problem signaled by: Depth-dependent response functions and rainfall transition zones of soil-profile activation under varying antecedent moisture in the Shandian River Basin, northern China.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
70. Numerical investigation of segment uplift induced by fresh grout in shield tunneling: an immersed boundary-based fluid-structure interaction model
Core Problem: Title-level focus: identifies the problem signaled by: Numerical investigation of segment uplift induced by fresh grout in shield tunneling: an immersed boundary-based fluid–structure interaction model.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
71. Balancing subsurface complexity and model simplicity: How much spatially distributed field data is needed to simulate streamflow in a subalpine critical zone
Core Problem: Title-level focus: identifies the problem signaled by: Balancing subsurface complexity and model simplicity: How much spatially distributed field data is needed to simulate streamflow in a subalpine critical zone.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
72. A hybrid physics and machine learning framework for trajectory prediction and uncertainty quantification of undrained behaviour in gassy clay
Core Problem: Title-level focus: identifies the problem signaled by: A hybrid physics and machine learning framework for trajectory prediction and uncertainty quantification of undrained behaviour in gassy clay.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
73. A new method for calculating rock mass pressure in shallow burdens tunnels: theoretical derivation, experimental validation, and analysis of influencing factors
Core Problem: Title-level focus: identifies the problem signaled by: A new method for calculating rock mass pressure in shallow burdens tunnels: theoretical derivation, experimental validation, and analysis of influencing factors.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
74. Analytical method for synergistic cyclic degradation of base and side resistances on large diameter monopiles
Core Problem: Title-level focus: identifies the problem signaled by: Analytical method for synergistic cyclic degradation of base and side resistances on large diameter monopiles.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
75. A Physically Informed Transformation of Sentinel-1 SAR Observations for Streamflow Inference in Small and Complex Rivers
Core Problem: Title-level focus: identifies the problem signaled by: A Physically Informed Transformation of Sentinel-1 SAR Observations for Streamflow Inference in Small and Complex Rivers.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
76. Mamba-LiteDSN: Lightweight Dual-Stream Network for Efficient Remote Sensing Change Detection
Core Problem: Title-level focus: identifies the problem signaled by: Mamba-LiteDSN: Lightweight Dual-Stream Network for Efficient Remote Sensing Change Detection.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
77. A Hierarchical Feature Assembly and Adaptive Enhancement Network for Remote Sensing Image Change Detection
Core Problem: Title-level focus: identifies the problem signaled by: A Hierarchical Feature Assembly and Adaptive Enhancement Network for Remote Sensing Image Change Detection.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
78. Enhancing High Spatial Resolution SAR-to-Optical Image Translation via Wavelet-Derived Multiscale Features over Croplands
Core Problem: Title-level focus: identifies the problem signaled by: Enhancing High Spatial Resolution SAR-to-Optical Image Translation via Wavelet-Derived Multiscale Features over Croplands.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
79. Near Real-time GNSS-IR Sea Level Retrieval And Prediction Using Deep Learning
Core Problem: Title-level focus: identifies the problem signaled by: Near Real-time GNSS-IR Sea Level Retrieval And Prediction Using Deep Learning.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
80. Evaluating the Potential of Multi-Source Snow Depth for Improving ICESat-2 Arctic Sea Ice Thickness Retrieval
Core Problem: Title-level focus: identifies the problem signaled by: Evaluating the Potential of Multi-Source Snow Depth for Improving ICESat-2 Arctic Sea Ice Thickness Retrieval.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
81. TSMRNet: A Method for High-Resolution Soil Moisture Mapping with Limited Labels by Integrating Self-Supervised and Semi-Supervised Learning
Core Problem: Title-level focus: identifies the problem signaled by: TSMRNet: A Method for High-Resolution Soil Moisture Mapping with Limited Labels by Integrating Self-Supervised and Semi-Supervised Learning.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
82. Short-Time DOA-Based 3-D Imaging of Complex Subsurface Targets Using Array-Based Borehole Radar Observations
Core Problem: Title-level focus: identifies the problem signaled by: Short-Time DOA-Based 3-D Imaging of Complex Subsurface Targets Using Array-Based Borehole Radar Observations.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
83. Spatial-Temporal-Spectral Unified Modeling for Remote Sensing Dense Prediction
Core Problem: Title-level focus: identifies the problem signaled by: Spatial-Temporal-Spectral Unified Modeling for Remote Sensing Dense Prediction.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
84. Quantifying the Effect of Topographic Normalization on Airborne LiDAR-Derived Individual Tree Parameters: A Unified Theoretical-Experimental Framework
Core Problem: Title-level focus: identifies the problem signaled by: Quantifying the Effect of Topographic Normalization on Airborne LiDAR-Derived Individual Tree Parameters: A Unified Theoretical-Experimental Framework.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
85. A Physics-Informed Ensemble Learning Framework for Transferable Shallow Water Bathymetry Inversion in Data-Scarce Regions
Core Problem: Title-level focus: identifies the problem signaled by: A Physics-Informed Ensemble Learning Framework for Transferable Shallow Water Bathymetry Inversion in Data-Scarce Regions.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
86. A Structural knowledge-guided deep learning for retrieving precipitable water vapor from satellite remote sensing observations
Core Problem: Title-level focus: identifies the problem signaled by: A Structural knowledge-guided deep learning for retrieving precipitable water vapor from satellite remote sensing observations.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
87. ITC-OVAD: An Image-Text Collaborative Strategy for Open-Vocabulary Aerial Detection
Core Problem: Title-level focus: identifies the problem signaled by: ITC-OVAD: An Image-Text Collaborative Strategy for Open-Vocabulary Aerial Detection.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
88. Adaptive-Dictionary-Regularized 3-D Direct Inversion of Surface-Wave Dispersion Data
Core Problem: Title-level focus: identifies the problem signaled by: Adaptive-Dictionary-Regularized 3-D Direct Inversion of Surface-Wave Dispersion Data.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
89. Residual Multiple Suppression Using a Physics-Informed Plug-in Network With Residual-Multiple-Level Prior
Core Problem: Title-level focus: identifies the problem signaled by: Residual Multiple Suppression Using a Physics-Informed Plug-in Network With Residual-Multiple-Level Prior.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
90. Latent Commonality Expectation-Maximisation for Box-supervised Tree Crown Instance Segmentation
Core Problem: However, existing models are predominantly trained on dense canopy forest imagery and degrade in savannah and drylands, where tree crowns are sparse, of variable appearance, and underrepresented in annotated benchmarks.
Key Innovation: We introduce LACE (LAtent Commonality Expectation-maximisation), a box-supervised instance segmentation model, evaluated on 0.1 m/px aerial RGB tree crown imagery.
91. Label-Efficient Learning for Ground-Based Sky-Image Classification: A Benchmark of Transfer Learning, Active Learning, and Pseudo-Labeling on GCD
Core Problem: However, reliable sky-image annotation is time-consuming, especially when cloud types are visually similar or mixed.
Key Innovation: However, reliable sky-image annotation is time-consuming, especially when cloud types are visually similar or mixed.
92. LLM-Driven Training-free Location-Attribute Synergic Fusion: A Closed-Loop Paradigm for Dual-source Encrypted POIs and LULC Mapping
Core Problem: Dual-source encrypted points of interest (DSEP), POIs from two encrypted coordinate systems, suffer from intertwined location and attribute uncertainties, including nonlinear systematic misalignment and naming inconsistency, hindering land-use/land-cover (LULC) mapping.
Key Innovation: To the best of our knowledge, this paper is the first to propose an LLM-driven, training-free location-attribute synergic closed-loop optimization paradigm for DSEP fusion.
93. Beyond the Flat Seafloor: A Closed-Form Two-View Constraint to Aid Sidescan Sonar Reconstruction
Core Problem: As sidescan sonar is limited to a 1D range measurement, many approximations are frequently used, including the long-standing assumption of a flat seafloor.
Key Innovation: Our results show that locus length is strongly governed by elevation misalignment, and peaks at a moderate oblique crossing angle of approximately 20 degrees, with minimal locus lengths obtained at near parallel and anti-parallel passes.
94. One Domain, Many Tongues: Composing Domain and Language LoRAs for Cross-Lingual Remote-Sensing MLLMs without Paired Data
Core Problem: Remote-sensing (RS) multimodal large language models (MLLMs) are trained and evaluated only in English, while text-only instruction data covers over 100 languages.
Key Innovation: We propose MODL (Mutually Orthogonal Domain-Language composition), a recipe that adds new languages to an English RS MLLM without a single multilingual RS example: a domain LoRA trained on English RS imagery and a language LoRA trained on text alone are learned jointly, under one loss term that keeps the two updates mutually orthogonal at every layer throughout training.
95. Bayesian Fusion of Active Contour Models and ConvNet Priors for Standing Dead Tree Segmentation
Core Problem: We address this problem and propose a novel instance segmentation method geared towards NRM imagery.
Key Innovation: Moreover, we introduce a novel prior for contour shape, namely, a class of Deep Shape Models based on architectures from Generative Adversarial Networks (GANs).
96. RGB-to-Multispectral Reconstruction with Spectral Attention and Non-Adversarial Perceptual Loss
Core Problem: RGB imagery lacks the red-edge, near-infrared and short-wave-infrared information needed by many environmental products, making spectral reconstruction ill posed.
Key Innovation: A residual U-Net with spectral attention and non-adversarial L1–LPIPS training reconstructs ten Sentinel-2 bands, outperforms tested GAN variants and retains cross-dataset performance without fine-tuning.
97. RASM-Net: A Reliability-Aware Structural and Multiscale Modeling Network for RGB-T Object Detection in Complex Environments
Core Problem: However, existing methods generally treat the contribution of each modality as fixed or globally consistent, overlooking that modality reliability varies dynamically across scenes and spatial regions.
Key Innovation: To address these issues, we propose RASM-Net, a Reliability-Aware Structural and Multiscale Modeling Network that coordinates three complementary stages of the feature-propagation pathway: cross-modal fusion, intra-scale structural modeling, and cross-scale feature aggregation.
98. CoGLoR: Collaborative Global-Local Representation Learning for Remote Sensing Scene Classification
Core Problem: However, existing self-supervised methods often emphasize either global semantic invariance or local structural modeling, while the interaction between global scene context and local patch-level representations remains insufficiently explored.
Key Innovation: To address this issue, we propose CoGLoR, a collaborative global–local representation learning framework for self-supervised remote sensing scene classification.
99. Transition to Double-Cell Mock Walker Circulations With Surface Warming Explained by Periodic Convection
Core Problem: Abstract Idealized mock Walker simulations are widely used to study the interactions between overturning circulation and convection in the tropics.
Key Innovation: Previous studies documented a transition from a single‐cell to a double‐cell mock Walker circulation when the average sea surface temperature exceeds 300 K.
100. Assessing the viability of electrical resistivity tomography for cirque sediment studies: New data from the Slovenian Alps
Core Problem: Title-level focus: identifies whether electrical resistivity tomography can resolve sediment architecture within alpine cirques.
Key Innovation: Title-signalled approach or contribution: The title reports new Slovenian Alps observations used to test ERT viability; public metadata did not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
101. Extending ocean color product to three dimensions assisted by high spectral resolution lidar data
Core Problem: Title-level focus: identifies the problem signaled by: Extending ocean color product to three dimensions assisted by high spectral resolution lidar data.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
102. Bridging cloud-induced gaps in MODIS leaf area index products using high-frequency geostationary satellite observations
Core Problem: Title-level focus: identifies the problem signaled by: Bridging cloud-induced gaps in MODIS leaf area index products using high-frequency geostationary satellite observations.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
103. Quaternary sediment provenance in the Qiaojia Basin, southeastern Tibetan plateau, and its implications for drainage reorganisation in the middle and lower reaches of the Jinsha River
Core Problem: Title-level focus: identifies the problem signaled by: Quaternary sediment provenance in the Qiaojia Basin, southeastern Tibetan plateau, and its implications for drainage reorganisation in the middle and lower reaches of the Jinsha River.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
104. Development of a high-pressure, temperature-controlled calibration chamber for cone penetration tests in hydrate-bearing sand
Core Problem: Title-level focus: identifies the problem signaled by: Development of a high-pressure, temperature-controlled calibration chamber for cone penetration tests in hydrate-bearing sand.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
105. Improving Empirical Pressure Estimation over Mainland China and Surrounding Regions Using Machine Learning
Core Problem: Title-level focus: identifies the problem signaled by: Improving Empirical Pressure Estimation over Mainland China and Surrounding Regions Using Machine Learning.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
106. Taylor-linearized ADMM Deep Unfolding Network with Classification Prior Guidance for Tomographic Inversion of Urban Areas
Core Problem: Title-level focus: identifies the problem signaled by: Taylor-linearized ADMM Deep Unfolding Network with Classification Prior Guidance for Tomographic Inversion of Urban Areas.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
107. MSGST-Net: A Multiscale Global Spatiotemporal Neural Network for Month-Scale Prediction of Coastal Sea Surface Temperature
Core Problem: Title-level focus: identifies the problem signaled by: MSGST-Net: A Multiscale Global Spatiotemporal Neural Network for Month-Scale Prediction of Coastal Sea Surface Temperature.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
108. Comparing a Vision Foundation Model (DINOv3) and a Task-Specific U-Net for Mapping Emergent Aquatic Vegetation From Fused UAV Multispectral and LiDAR Data
Core Problem: Title-level focus: identifies the problem signaled by: Comparing a Vision Foundation Model (DINOv3) and a Task-Specific U-Net for Mapping Emergent Aquatic Vegetation From Fused UAV Multispectral and LiDAR Data.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
109. AetherNet: Bridging Frequency Enhancement and Uncertainty Guided Refinement for Building Extraction
Core Problem: Title-level focus: identifies the problem signaled by: AetherNet: Bridging Frequency Enhancement and Uncertainty Guided Refinement for Building Extraction.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
110. Neural Network-Based Depth Extension of Absorption-Coefficient Profiles Retrieved from Water-Body LiDAR
Core Problem: Title-level focus: identifies the problem signaled by: Neural Network-Based Depth Extension of Absorption-Coefficient Profiles Retrieved from Water-Body LiDAR.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
111. Leveraging High-Resolution LISS IV Data and Machine Learning for Land Use Land Cover Mapping: A Comprehensive Workflow and Model Interpretation
Core Problem: Title-level focus: identifies the problem signaled by: Leveraging High-Resolution LISS IV Data and Machine Learning for Land Use Land Cover Mapping: A Comprehensive Workflow and Model Interpretation.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
112. CAEG-UNet: A Class-Imbalance-Aware Enhanced UNet with Edge Enhancement and Graph Reasoning for GaoFen-2 Image Semantic Segmentation
Core Problem: Title-level focus: identifies the problem signaled by: CAEG-UNet: A Class-Imbalance-Aware Enhanced UNet with Edge Enhancement and Graph Reasoning for GaoFen-2 Image Semantic Segmentation.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
113. An Efficient Processing Framework for Spaceborne GNSS-R Raw Intermediate-Frequency Data Based on NVSLSM and GPU Parallel Computing and Its Application to Sea Surface Altimetry
Core Problem: Title-level focus: identifies the problem signaled by: An Efficient Processing Framework for Spaceborne GNSS-R Raw Intermediate-Frequency Data Based on NVSLSM and GPU Parallel Computing and Its Application to Sea Surface Altimetry.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
114. Dual-Branch Spectral Group Embedding and FFT-Guided Feature Aggregation Network for Hyperspectral Image Classification
Core Problem: Title-level focus: identifies the problem signaled by: Dual-Branch Spectral Group Embedding and FFT-Guided Feature Aggregation Network for Hyperspectral Image Classification.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
115. A Deep Learning-Based Approach for Mapping Shrubs in Arctic Tundra from Very high-Resolution Imagery
Core Problem: Title-level focus: identifies the problem signaled by: A Deep Learning-Based Approach for Mapping Shrubs in Arctic Tundra from Very high-Resolution Imagery.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
116. Geometry-Driven Analysis of Sea-Surface Signatures in Monostatic and Long-Baseline Along-Track Bistatic SAR Configurations: Insights from the PLT-1 Mission
Core Problem: Title-level focus: identifies the problem signaled by: Geometry-Driven Analysis of Sea-Surface Signatures in Monostatic and Long-Baseline Along-Track Bistatic SAR Configurations: Insights from the PLT-1 Mission.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
117. Soil Knowledge-Guided Depth-Autoregressive Transformer for Predicting Soil Organic Carbon at Multiple Depths
Core Problem: Title-level focus: identifies the problem signaled by: Soil Knowledge-Guided Depth-Autoregressive Transformer for Predicting Soil Organic Carbon at Multiple Depths.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.
118. Reconstructing Multi-Angular Field HCRF Spectra from Indoor Measurements via a Spectral-Frequency-Global Network
Core Problem: Title-level focus: identifies the problem signaled by: Reconstructing Multi-Angular Field HCRF Spectra from Indoor Measurements via a Spectral-Frequency-Global Network.
Key Innovation: Title-signalled approach or contribution: The title identifies the study direction, but the available public metadata do not expose methods, validation or quantitative results. Methods, data and results could not be assessed because no reliable abstract was available.