TerraMosaic Daily Digest: August 12, 2026
Daily Summary
Direct hazard papers sharpen the separation of where hazards are primed, when they are triggered, and how exposure amplifies losses. In the Himalaya, free-satellite analyses show antecedent weather carries most of the short-term signal for glacial-lake bursts, rainfall-triggered landslides, and small ice-related floods, while simple matched baselines remain competitive with deeper models; across Italy, historical flood-risk growth is attributed mainly to settlement expansion, but projections to 2100 shift the dominant contribution toward climate change. Coastal and climate-linked studies extend this driver accounting to wave-driven Scottish barrier impacts, the record 2025 European fire season, and concurrent coastal marine and terrestrial heat extremes, while Konya subsidence emerges as a spatially continuous deformation corridor rather than a scatter of isolated hotspots.
Solid-Earth and engineering studies likewise foreground hidden structure, material state, and coupled processes. Santorini tremors separate into swarm-like and deeper magmatic families linked to a mid-crustal reservoir; receiver-function and ambient-noise imaging around the Altyn Tagh fault and eastern Tibet indicate vertically partitioned shortening above ductile crustal flow; Mendocino tremor shutdown and recovery point to fluid-mediated modulation beyond simple Coulomb stress change; and P-wave attenuation beneath the Alaska Peninsula suggests that mantle melting alone does not explain along-arc volcanic behaviour. At smaller scales, field measurements after the Kahramanmaras sequence, crack-aware and physics-informed slope models, structure-constrained rockfall simulation, coupled scour-liquefaction analysis, automated rock-mass rating, vibration-based rock-strength estimation, and stochastic frozen-soil and permafrost studies all show that localized heterogeneity strongly governs failure timing, runout, and engineering response. Probabilistic post-earthquake building safety and road-network resilience frameworks extend the same logic into operational decision support.
Transferable papers extend the technical toolkit without collapsing methodological advance into geohazard validation. Remote-sensing studies improve segmentation, change description, hyperspectral fusion and classification, super-resolution, thin-cloud removal, aerial detection, GNSS-R rainfall retrieval, hypertemporal bathymetry, and vegetation or biomass mapping through stronger representations, multimodal constraints, and physics-aware supervision; parallel work on burned-area mapping, coseismic landslide segmentation, and supraglacial lake bathymetry illustrates where those design choices do intersect hazards directly. In modeling and forecasting, Bayesian reduced-order tsunami surrogates, topology-aware cascade optimization, structure-preserving uncertainty quantification, and CMIP7 evaluation and bias-correction workflows all push hazard-relevant inference toward calibrated, computationally efficient decision support.
Key Trends
August 12's selections converge on hazard inference that is more driver-resolved, mechanics-aware, and explicitly uncertainty-calibrated, while transferable sensing and AI studies prioritize structural fidelity over generic benchmark inflation.
- Hazard analyses are disentangling susceptibility, triggering, and exposure: The Himalayan multi-hazard study treats site susceptibility and trigger timing separately, Italy's flood reconstruction separates urban growth from climate forcing, and wildfire, humid-heat, and coastal-storm papers likewise attribute hazard escalation to specific meteorological or exposure pathways rather than single composite indices.
- Failure-scale mechanics are being embedded directly into assessments: Crack-aware unsaturated slope analysis, elastoplastic neural slope modeling, multidimensional slope fragility, fracture-network-constrained rockfall simulation, coupled scour-liquefaction modeling, and automated rock-mass and rock-strength estimation all move stability assessment away from coarse empirical screening and toward process-resolving diagnosis.
- Deep structure and fluids remain central to seismic-volcanic interpretation: Santorini tremor families, Altyn Tagh shortening, eastern Tibetan low-velocity flow paths, Mendocino tremor suppression, and Alaska attenuation collectively show that reservoir geometry, ductile layers, and fluid effects can dominate surface hazard expression beyond simple fault or magma-supply metrics.
- Observation pipelines favor robust representations over larger model stacks: Across wildfire susceptibility, coseismic landslide mapping, burned-area retrieval, supraglacial bathymetry, GNSS-R rainfall, bathymetry, and multiple remote-sensing restoration tasks, gains come mainly from stronger encoders, physics-aware supervision, multimodal constraints, and radiometric control rather than complexity for its own sake.
- Probabilistic calibration is moving upstream in forecasting and simulation: Bayesian pooled tsunami surrogates, dam-breach uncertainty propagation, probabilistic post-earthquake safety assessment, CMIP7 evaluation workflows, SDMBCv2 bias correction, and structure-preserving uncertainty quantification all treat uncertainty as part of model design rather than as a residual caveat.
Selected Papers
The selected papers split between direct geohazard studies and transferable sensing, mechanics, climate, and AI methods. Direct contributions cover mountain, flood, wildfire, seismic-volcanic, coastal, cryospheric, permafrost, and engineering hazards, whereas the transferable cohort offers observation, reconstruction, and uncertainty tools whose geohazard utility is prospective unless demonstrated in the paper itself.
1. Which Site, and When: A Free-Satellite-Data Test of Himalayan Glacial Lake Bursts, Landslides, and Ice Floods
Core Problem: Predict which Himalayan sites are susceptible and when triggers arrive for glacial-lake bursts, landslides, and small ice-related floods using only free satellite data.
Key Innovation: Uses matched controls and strict spatial cross-validation to show simple weather-plus-terrain models outperform deep models while producing a Nepal watchlist.
2. The Portrait of Flood Risk in Italy: Past, Present and Future, From 1870 to 2100
Core Problem: Quantify how urbanization and climate change have shaped Italy's population flood risk from 1870 to 2100.
Key Innovation: Long-horizon RESCUE-FR modeling separates the historical dominance of settlement growth from the future dominance of climate change.
3. Deep Seismic Tremors in the 2025 Santorini Seismo-Volcanic Crisis Highlight Magmatic Feeding From a Mid-Crustal Reservoir
Core Problem: Disentangle tremor sources and magma-migration processes during the 2025 Santorini seismo-volcanic crisis.
Key Innovation: Identifies two tremor families, including deeper magmatic tremor linked to a mid-crustal reservoir near Kolumbo.
4. Evaluating AlphaEarth Foundations Embeddings for Wildfire Susceptibility Mapping
Core Problem: Test whether geospatial foundation embeddings can replace handcrafted variables for wildfire susceptibility mapping.
Key Innovation: Shows AEF embeddings reconstruct standard predictors well, achieve ROC-AUC above 0.92, and transfer across regions far better than physical-variable models.
5. Reduced Order Modeling for Tsunami Forecasting with Bayesian Hierarchical Pooling
Core Problem: Fast surrogate tsunami forecasting must remain physically grounded, statistically calibrated, and valid beyond a single reduced trajectory.
Key Innovation: Combines corrected reduced-order modeling with Bayesian hierarchical pooling to generate probabilistic tsunami surrogates.
6. Attribution of the record breaking 2025 European fire season to climate change
Core Problem: Europe's record 2025 fire season needs region-specific attribution to understand whether extreme fire weather is emerging beyond natural variability.
Key Innovation: Links the season's exceptional wildfire activity across five regions mainly to rising vapour pressure deficit and more frequent hot-dry-windy conditions.
7. Remote Sensing, Vol. 18, Pages 2729: Severity-Based Mapping of Land-Subsidence Hazard Zones and Critical Hotspots Using SBAS-InSAR and Spatial Statistics: The Konya Metropolitan Area, Turkey
Core Problem: InSAR studies often monitor subsidence but stop short of producing robust, quantitative hazard classes and connected hotspot zones.
Key Innovation: Integrates SBAS-InSAR, spatial statistics, a composite severity index, and connectivity analysis to map subsidence hazard corridors and critical hotspots.
8. Remote Sensing, Vol. 18, Pages 2727: Encoder Choice Outweighs Modular Refinement in U-Net Architectures for Globally Distributed Coseismic Landslide Segmentation
Core Problem: Post-earthquake landslide segmentation lacks a clear quantitative hierarchy of how much encoder choice versus refinement modules affects performance.
Key Innovation: Uses controlled ablations on a globally distributed coseismic landslide dataset to show encoder backbone choice dominates downstream module gains.
9. Remote Sensing, Vol. 18, Pages 2723: Integrating Multi-Source Remote Sensing and Geospatial Data for Snow Disaster Risk Assessment in Northwestern China
Core Problem: Mapping combined regional risk from heavy snowfall, blowing snow, and avalanches
Key Innovation: Integrates multi-source remote sensing with entropy-weighted multi-hazard assessment validated against records
10. Coupling fracture extraction and network topology for structure-controlled rockfall hazard assessment in complex terrain
Core Problem: Identifying structure-controlled rockfall sources and scenario runout in complex terrain
Key Innovation: Couples AI fracture extraction, fracture-network topology, and MPDEM rockfall simulation
11. Uncertainty propagation and its impact on earth-fill dam failures
Core Problem: Quantifying how hydrologic and breach uncertainty change downstream flood hazard maps
Key Innovation: Links HEC-HMS uncertainty bounds to probabilistic breach simulations and scenario hazard zoning
12. Geotechnical field measurements and structural damage in Iskenderun following the 2023 Turkey-Kahramanmaras earthquake sequence
Core Problem: Explaining land-class-dependent ground damage and structural damage in Iskenderun after the 2023 earthquake sequence
Key Innovation: Combines historical land classification with PDCPT, MASW, HVSR, and damage observations to relate subsurface conditions to failures
13. An SPCL-Stacking ensemble learning framework for landslide susceptibility assessment
Core Problem: Reducing label noise and weak single-model performance in landslide susceptibility mapping
Key Innovation: Integrates self-paced curriculum learning for negative-sample optimization with heterogeneous stacking
14. A unified 3 D limit analysis framework for rainfall-induced tensile-shear stability of unsaturated cracked slopes
Core Problem: Capturing tensile cracks and suction evolution in rainfall-triggered failure of unsaturated slopes
Key Innovation: Unified 3D limit-analysis framework coupling infiltration-driven suction change with tensile-shear composite failure
15. Multidimensional seismic fragility evaluation of slopes with multiple performance indicators under stochastic seismic ground motions
Core Problem: Evaluating slope seismic fragility when several performance indicators capture different failure modes
Key Innovation: Builds multidimensional limit states from FOS, crest displacement, and Newmark displacement under stochastic ground motions
16. Wave-driven storm impacts on Scotland's sedimentary barriers intensify with climate change
Core Problem: Assessing how climate change intensifies wave-driven impacts on sedimentary barriers
Key Innovation: Frames climate-change amplification of barrier-storm impacts at the national scale
17. Seismo- and Tsunami-Traveling Ionospheric Disturbances of the 29 July 2025 M8.8 Kamchatka Earthquake Observed by CW-HF Doppler Sounding Systems and Ionosondes in Taiwan and Japan
Core Problem: Characterize ionospheric disturbances generated by the 29 July 2025 Kamchatka earthquake and tsunami.
Key Innovation: Combines HF Doppler sounding and ionosondes to distinguish fast seismic ionospheric waves from slower tsunami-tracking disturbances in 3D.
18. Numerical investigation of coupled scour and liquefaction mechanisms around a mono-pile
Core Problem: Scour and liquefaction interact nonlinearly around monopiles under wave and current loading, but their coupled evolution is poorly resolved.
Key Innovation: Builds a 3D OpenFOAM model that jointly simulates scour and liquefaction and shows how loading regime, seepage velocity, and current direction alter hazard timing and depth.
19. Remote Sensing, Vol. 18, Pages 2725: Long-Term Changes in Shelterbelt Stability Along the Taklimakan Desert Highway Revealed by Landsat Observations
Core Problem: Assessing long-term stability of a highway shelterbelt that protects against sand hazards
Key Innovation: Builds a Landsat-based Shelterbelt Stability Index combining vegetation cover, connectivity, area share, and fractal form
20. Integrating deep neural network with elastoplastic analysis: a hybrid approach for slope stability analysis
Core Problem: Reducing purely data-driven bias in slope stability prediction by embedding mechanics
Key Innovation: Physics-informed neural network that solves elastoplastic slope fields and couples them to slip-surface search and FOS calculation
21. Empirical calibration of facility wildfire vulnerability index using post-fire damage records and neural additive models
Core Problem: Calibrating facility wildfire vulnerability using observed post-fire damage
Key Innovation: Uses empirical damage records with neural additive models to tune a facility-level vulnerability index
22. An integrated framework for evaluating road network resilience: Wellington fault Mw 7.5 scenario
Core Problem: Evaluating road network resilience under a Wellington fault Mw 7.5 scenario
Key Innovation: Integrated framework linking a large earthquake scenario to transport-network resilience assessment
23. BAM: A physics-informed self-supervised framework for near-real-time wildfire burned area mapping from multi-source earth observation
Core Problem: Producing near-real-time burned-area maps without manual labels
Key Innovation: Physics-informed self-supervision using combustion stoichiometry and ATBI pseudo-labels within Google Earth Engine
24. Deep learning-based precipitation nowcasting integrating radar echoes and rain gauge data
Core Problem: Improving precipitation nowcasting by combining radar echoes with rain-gauge observations in a deep-learning framework.
Key Innovation: Fuses radar and gauge data in a deep-learning nowcasting setup aimed at more accurate short-term rainfall prediction.
25. A probabilistic framework for post-earthquake safety assessment of mainshock-damaged buildings
Core Problem: How to use probabilistic post-earthquake hazard information to support safety assessment and public-protection decisions after a mainshock.
Key Innovation: Frames operational earthquake forecasting and probabilistic hazard thresholds as a structured basis for post-event safety and alert decisions.
26. Crustal Shortening Mechanism Near the Central Altyn Tagh Fault in Northern Tibet
Core Problem: Resolve how crustal shortening is partitioned around the central Altyn Tagh fault system in northern Tibet.
Key Innovation: Dense receiver-function imaging reveals vertically partitioned shortening with brittle upper-crust deformation over ductile mid-lower crustal thickening.
27. Positive Association Between Coastal Sea Surface Temperatures and Humid Heat Stress on Nearby Land
Core Problem: Test whether warmer coastal seas intensify nearby land humid-heat stress globally.
Key Innovation: Global coastal analysis links marine heatwaves to more frequent concurrent humid terrestrial heat extremes.
28. Abrupt Suppression of Tectonic Tremors Outside the Stress Shadow of Mendocino Earthquakes
Core Problem: Explain why a Mendocino tremor cluster shuts down after large earthquakes despite positive Coulomb stress changes.
Key Innovation: Combines moment tensors, stress modeling, migration patterns, and tomography to implicate fluid-related suppression and recovery.
29. DiCoR: Decoupled Referent Disambiguation and Contour Recalibration for Efficient Referring Remote Sensing Image Segmentation
Core Problem: Referring remote sensing segmentation struggles to jointly resolve ambiguous targets and refine masks efficiently.
Key Innovation: Decouples referent disambiguation from contour recalibration within an efficient remote-sensing segmentation pipeline.
30. A Deep-Learning-Embedded Sparse Variational Optimization Method for Hyperspectral and Multispectral Image Fusion
Core Problem: Physically interpretable sparse variational HSI-MSI fusion struggles to capture complex spatial-spectral structure with hand-crafted priors alone.
Key Innovation: Embeds a deep implicit prior and explicit sparsity into a variational fusion model solved by a hybrid ADMM scheme with zero-shot initialization.
31. Dense-array ambient noise tomography of the eastern Tibetan Plateau: a high-resolution model of the crustal low-velocity zone and its tectonic evolution implications
Core Problem: The geometry and anisotropy of the Tibetan mid-lower crustal low-velocity zone need better resolution to test crustal-flow deformation models.
Key Innovation: Uses dense-array ambient noise tomography to map the low-velocity zone and its anisotropy, clarifying ductile flow pathways beneath eastern Tibet.
32. Along-Strike Variations in Sub-Arc Melting Beneath the Alaska Peninsula Revealed by P-Wave Attenuation
Core Problem: Explain along-strike differences in sub-arc melting and volcanism beneath the Alaska Peninsula.
Key Innovation: Uses P-wave attenuation to separate mantle-melting variations from trans-crustal controls on surface volcanism.
33. HIMEC: Directional Change Representation and Fixed-Interface Decoding for Remote Sensing Image Change Captioning
Core Problem: Improve semantic change captioning from bitemporal remote-sensing imagery.
Key Innovation: Separates directional change streams and uses a fixed-interface decoder to stabilize caption generation.
34. LocusGS: Spatially Grounded Tokens for Feed-Forward 3D Gaussian Splatting
Core Problem: Feed-forward 3D Gaussian splatting lacks spatially coherent query tokens, weakening scene structure recovery.
Key Innovation: Adds explicit 3D anchor states that guide evidence aggregation and local Gaussian decoding.
35. Fast Iterative Five point Relative Pose Estimation
Core Problem: Speed up calibrated five-point relative pose estimation without sacrificing precision.
Key Innovation: Replaces the standard solver with a Powell Dog Leg iterative method that is about twice as fast and extends naturally beyond five points.
36. TANGCO: Learning Topology-Aware Capacity Allocation for Overload-driven Cascading Failures
Core Problem: Allocate limited node capacity to reduce overload-driven cascading failures on networks.
Key Innovation: Trains a topology-aware graph neural policy with policy gradients and heuristic anchoring to outperform hand-designed allocation rules.
37. History-informed Lagrangian Neural Networks
Core Problem: Forecast mechanical-system evolution from position-only histories while inferring hidden velocities and varying physical parameters.
Key Innovation: Uses a recurrent history encoder to reconstruct latent dynamics and modulate a structured Lagrangian model end to end.
38. Shortest-Path Decomposition for Foliage-Robust 3D Tree Modeling and Above-Ground Biomass Estimation from Point Clouds
Core Problem: Leaf-on foliage severely degrades quantitative structural modeling and biomass estimation from point clouds.
Key Innovation: Recovers branching hierarchy through shortest-path traversal frequencies, avoiding explicit leaf-wood separation.
39. Ocean Rainfall Identification and Intensity Retrieval From GNSS-R Observations
Core Problem: Ocean rainfall detection and intensity retrieval from GNSS-R remain challenging across varying wind and precipitation conditions.
Key Innovation: Benchmarks optimized ML classifiers and regressors on matched GNSS-R, IMERG, and ERA5 data and improves rainfall identification and retrieval.
40. Multistep Tunable Super-Resolution for Remote Sensing Image Clarity Enhancement
Core Problem: Remote-sensing super-resolution methods either oversmooth details or introduce artifacts and unstable outputs.
Key Innovation: Uses a multistep tunable super-resolution network that preserves richer texture while controlling artifact formation.
41. Thin Cloud Removal From Remote Sensing Imagery Using Difference Map Supervision and Band Balancing
Core Problem: Optical remote sensing loses usable hazard information when thin clouds partially corrupt pixels across bands and times.
Key Innovation: Adds a lightweight difference-map supervision and band-balancing module that uses an auxiliary image to recover thin-clouded scenes more robustly.
42. Dynamic Global-Local Selection Network for Hyperspectral Image Classification
Core Problem: Existing Mamba-based hyperspectral classifiers waste computation with fixed scan paths and weaken local spatial structure.
Key Innovation: Combines adaptive Mamba block selection with dynamic local branch selection across spatial, spectral, and frequency views.
43. Rapid Evaluation Framework for the CMIP7 Assessment Fast Track
Core Problem: CMIP7 needs rapid, reproducible, community-scale evaluation of Earth system model performance as outputs are published.
Key Innovation: Integrates multiple benchmarking tools through a containerized CMEC workflow for automated, portable diagnostics on new model output.
44. SDMBCv2 (v1.0): correcting systematic biases in RCM inputs for future projection
Core Problem: Regional climate simulations inherit GCM boundary-condition biases, and common surface-only corrections break physical consistency.
Key Innovation: Extends the SDMBC package to correct multivariate RCM inputs for future projections with interpolation and HPC-ready Python tooling.
45. Remote Sensing, Vol. 18, Pages 2726: Supraglacial Lake Bathymetry Retrieval from ICESat-2 Altimetry Data and Sentinel-2 Imagery Using Deep Learning Algorithms
Core Problem: Retrieving spatially continuous supraglacial lake depth from sparse lidar and multispectral imagery
Key Innovation: Uses ICESat-2-derived depth labels with CatBoost and residual dense networks to upscale lake bathymetry
46. Toward Automated and Refined Rock Mass Quality Evaluation: An AR-RMR System Driven by AI and Multi-source Heterogeneous Data
Core Problem: Making rock mass rating more objective and spatially representative
Key Innovation: Automates extraction and quantification of RMR parameters from multi-source heterogeneous data within a unified AR-RMR pipeline
47. Linking Fractal Fault Geometry to Stress Rotation and Fault Reactivation in the Haishiwan Coal Mine, Western China
Core Problem: Understanding how fractal fault geometry controls stress rotation and reactivation risk in a coal mine
Key Innovation: Integrates fractal fault metrics, stress inversion, and instability modeling into a Fractal-Stress-Hazard framework
48. Rock Strength Estimation Using Drilling Vibration Signal
Core Problem: Estimating rock UCS rapidly from drilling vibration signals
Key Innovation: Combines vibration denoising and XGBoost on time and frequency features for high-accuracy UCS prediction
49. RDA-DETR: Rotation-aware decoupled oriented object detection in aerial images with agent query
Core Problem: Improving oriented object detection in aerial imagery
Key Innovation: Rotation-aware decoupled DETR architecture with agent-query mechanism
50. A pore-scale hydraulic conductivity model for frozen saturated soils incorporating wall-first and center-first freezing mechanisms
Core Problem: Predicting hydraulic conductivity of saturated frozen soils across freezing regimes
Key Innovation: Pore-scale conductivity model linked to SFCCs, with Transformer-XGBoost estimation when measurements are unavailable
51. Pore-scale sedimentary architecture controls the emergence and constitutive scaling of asymptotic hydrodynamic dispersion
Core Problem: Determining how pore-scale sedimentary architecture governs the onset and scaling of asymptotic hydrodynamic dispersion.
Key Innovation: Shows that spatial architectural organization, not just heterogeneity magnitude, controls dispersion scaling and transition lengths, with architecture-specific constitutive relationships.
52. Structure-preserving uncertainty quantification for GENERIC dynamics
Core Problem: Add epistemic uncertainty to hard-constrained scientific ML models without breaking physical admissibility.
Key Innovation: Attaches lightweight epinets to constrained GENERIC components and calibrates outputs with conformal prediction while preserving thermodynamics.
53. Assessing the Effectiveness of Deep Embeddings for Tree Species Classification in the Dutch Forest Inventory
Core Problem: National forest inventory tree-species mapping remains labor-intensive and hard to scale with limited labels.
Key Innovation: Shows pretrained remote-sensing embeddings outperform handcrafted features for few-shot inventory classification.
54. BinMod1D v1.0.10: a Python package for explicitly simulating 1D collisional coalescence/breakup processes with corresponding polarimetric radar signatures
Core Problem: Researchers need efficient 1D simulations linking collisional cloud microphysics to polarimetric radar signatures.
Key Innovation: Provides a fast JIT-parallelized Python bin model that explicitly evolves hydrometeor spectra and forward-simulates radar variables.
55. Remote Sensing, Vol. 18, Pages 2731: An Adaptive Shooting and Bouncing Ray Method Based on Q-Learning for Efficient Synthetic Aperture Radar Imaging Simulation
Core Problem: Synthetic aperture radar image simulation remains slow because conventional SBR methods allocate too many rays inefficiently.
Key Innovation: Recasts ray allocation as a Q-learning problem so the simulator adaptively densifies complex regions and coarsens simple ones.
56. Remote Sensing, Vol. 18, Pages 2730: Analysis of the Variation Trends and Driving Forces of Growing-Season kNDVI in Hainan Island over the Past Three Decades
Core Problem: Long-term vegetation changes in Hainan need spatial attribution across topography, climate extremes, and human disturbance.
Key Innovation: Builds a 30-year Landsat kNDVI record and uses XGBoost plus SHAP to separate topographic, climatic, and anthropogenic controls.
57. Remote Sensing, Vol. 18, Pages 2728: Wind Direction Retrieval from X-Band Marine Radar Images Using 2D-DTCWT-CSC and Maximum-Energy Radial Rings
Core Problem: Wind-direction signals in X-band marine radar are easily obscured by wave textures, clutter, and blind-zone interference.
Key Innovation: Combines dual-tree complex wavelets, convolutional sparse coding, and maximum-energy radial-ring fitting to improve wind-direction retrieval accuracy.
58. Remote Sensing, Vol. 18, Pages 2722: High-Resolution Aboveground Biomass Estimates of Tropical Peatland Forest Based on Planet NICFI Imagery and Airborne LiDAR
Core Problem: Upscaling sparse airborne LiDAR into wall-to-wall peatland biomass maps
Key Innovation: Compares two Planet-plus-LiDAR deep learning workflows and shows partial-coverage LiDAR can drive high-resolution mapping
59. Community asset identification for natural hazards research: A view from the Pacific Northwest
Core Problem: Identifying community assets for natural hazards research
Key Innovation: Asset-identification framing for community-centered hazard studies
60. Hypertemporal Sentinel-2 images reveal deep seafloor features and significantly improve satellite-derived bathymetry
Core Problem: Improving satellite-derived bathymetry from multi-temporal Sentinel-2 imagery
Key Innovation: Hypertemporal preprocessing and radiometric normalization workflow for more accurate bathymetry
61. Impact of spatial variability of soil void ratio on the hydrothermal process in permafrost engineering: A case study of cast-in-place piles on the Qinghai-Tibet Plateau
Core Problem: Assessing how spatial variability in soil void ratio affects hydrothermal behavior in permafrost engineering around cast-in-place piles.
Key Innovation: Introduces spatially variable void-ratio effects into a permafrost pile hydrothermal case study on the Qinghai-Tibet Plateau.