TerraMosaic Daily Digest: Jan 22, 2026
Daily Summary
This digest synthesizes 75 selected papers and focuses on infrastructure-focused hazard performance, flood generation, routing, and hydroclimatic forcing, high-resolution remote-sensing monitoring workflows. Top-ranked studies examine satellite and LiDAR-based deformation monitoring, wildfire hazard and adaptation, and earthquake-triggered slope response and liquefaction.
Across the full set, evidence converges on mechanism-constrained analysis with operational relevance, especially for risk, fragility, and resilience quantification and wildfire hazard dynamics and adaptation. The strongest contributions pair interpretable process evidence with monitoring or forecasting workflows that support warning design and risk prioritization.
Key Trends
- Infrastructure-facing outputs are increasingly decision-ready: Asset performance is evaluated with uncertainty-aware frameworks to support mitigation and maintenance prioritization.
- Flood analyses are becoming event-specific and process-based: Papers emphasize precipitation structure, antecedent wetness, and catchment controls rather than static hazard descriptors.
- Monitoring workflows rely on integrated remote-sensing products: Multi-source satellite and airborne observations are used for deformation retrieval, change detection, and rapid post-event mapping.
- Risk studies move beyond hazard mapping to consequence pathways: Vulnerability, fragility, exposure, and recovery metrics are integrated to compare interventions under compound hazards.
- Wildfire research is integrated with broader geohazard management: Physical drivers, landscape controls, and operational planning are analyzed together to evaluate cascade risk.
Selected Papers
This digest features 75 selected papers from 2420 papers analyzed across multiple journals. Each paper has been evaluated for its relevance to landslide research and includes links to the original publications.
1. Self-Supervised Score-Based Despeckling for SAR Imagery via Log-Domain Transformation
Core Problem: Speckle noise in SAR imagery degrades image quality and complicates analysis.
Key Innovation: A self-supervised framework for SAR image despeckling based on score-based generative models operating in the transformed log domain, converting speckle noise residuals into an approximately additive Gaussian distribution.
2. Real-Time Wildfire Localization on the NASA Autonomous Modular Sensor using Deep Learning
Core Problem: Automating the human-intensive process of fire perimeter determination from high-altitude, multi-spectral aerial imagery.
Key Innovation: A deep-learning model combining image classification and pixel-level segmentation to efficiently localize active wildfire in real-time, leveraging a multi-spectral dataset with SWIR, IR, and thermal bands.
3. Forest-Chat: Adapting Vision-Language Agents for Interactive Forest Change Analysis
Core Problem: Pixel-level change detection and semantic change interpretation for complex forest dynamics using satellite imagery.
Key Innovation: An LLM-driven agent (Forest-Chat) for integrated forest change analysis, supporting multiple RSICI tasks and incorporating zero-shot change detection with an interactive point-prompt interface.
4. UniRoute: Unified Routing Mixture-of-Experts for Modality-Adaptive Remote Sensing Change Detection
Core Problem: Scalability of remote sensing change detection (CD) methods toward modality-adaptive Earth observation.
Key Innovation: A unified framework (UniRoute) for modality-adaptive learning by reformulating feature extraction and fusion as conditional routing problems, using Adaptive Receptive Field Routing MoE (AR2-MoE) and Modality-Aware Difference Routing MoE (MDR-MoE) modules.
5. Filtered 2D Contour-Based Reconstruction of 3D STL Model from CT-DICOM Images
Core Problem: Reconstructing accurate 3D models from 2D medical images is challenging due to imperfections and noise.
Key Innovation: Filtering 2D contour data points to improve the geometry of the reconstructed 3D STL model from CT-DICOM images.
6. Three-dimensional visualization of X-ray micro-CT with large-scale datasets: Efficiency and accuracy for real-time interaction
Core Problem: Balancing accuracy and efficiency in 3D visualization of micro-CT data for real-time defect detection.
Key Innovation: Reviewing and analyzing CT reconstruction and volume rendering methods for efficient and accurate 3D characterization of microscopic features.
7. Tracing 3D Anatomy in 2D Strokes: A Multi-Stage Projection Driven Approach to Cervical Spine Fracture Identification
Core Problem: Efficient and precise detection of cervical spine fractures in 3D CT volumes.
Key Innovation: Using 2D projection-based vertebra segmentation for vertebra-level fracture detection in 3D CT volumes, reducing computational complexity.
8. Geo-Registration of Terrestrial LiDAR Point Clouds with Satellite Images without GNSS
Core Problem: Accurate geo-registration of LiDAR point clouds in urban environments where GNSS signals are unreliable.
Key Innovation: A post-hoc geo-registration method aligning LiDAR point clouds with satellite images using road segmentation, skeleton extraction, and RBF interpolation, validated on KITTI and a new Perth dataset.
9. GAIA: A Global, Multi-modal, Multi-scale Vision-Language Dataset for Remote Sensing Image Analysis
Core Problem: Existing Vision-Language Models (VLMs) are not performing well on RS-specific tasks, as commonly used datasets often lack detailed, scientifically accurate textual descriptions.
Key Innovation: A novel dataset designed for multi-scale, multi-sensor, and multi-modal RS image analysis. GAIA comprises of 201,005 meticulously curated RS image-text pairs, representing a diverse range of RS modalities associated to different spatial resolutions.
10. Socio-economic vulnerability assessment to climate change-induced disasters among the agriculture-based Indigenous communities: empirical evidence from Indian Sundarban biosphere reserve
Core Problem: Assessing socio-economic vulnerability to climate change in the Sundarban Biosphere Reserve, focusing on Indigenous agrarian communities.
Key Innovation: Site-specific socio-economic vulnerability index using IPCC framework and mixed-method approach to identify priority regions for mitigation.
11. Recent GLOFs in the Fitz Roy Area, Patagonia, Argentina
Core Problem: Documenting and characterizing glacial lake outburst floods (GLOFs) in the Fitz Roy area to understand their recurrence and potential impacts.
Key Innovation: Multidisciplinary approach combining historical records, remote sensing, field evidence, and hydraulic modeling to reconstruct past GLOF events and assess risk.
12. Influence of accumulated geotechnical deterioration (AGD) on detailed scale landslide phenomena: Cortinas Sector, Toledo, Colombia
Core Problem: Understanding the influence of Accumulated Geotechnical Deterioration (AGD) on landslide occurrence at a detailed scale.
Key Innovation: Demonstrating the role of AGD, where conditioning factors vary over time, in predicting future instability events through monitoring resistance parameters.
13. Region-specific assessment of flood disaster risk and contributing factors, based on historical data and machine learning
Core Problem: Assessing flood disaster risk globally and identifying contributing factors using historical data and machine learning.
Key Innovation: Integration of multi-source data and machine learning to reveal regional heterogeneity in flood risk and identify key determinants based on climate and socio-economic factors.
14. Spatiotemporal performance and error analysis of satellite precipitation products over a topographically complex semi-arid region in Iran
Core Problem: Accurate precipitation estimation is critical for water resource management and climate risk monitoring in semi-arid, topographically complex areas, but satellite precipitation products have limitations.
Key Innovation: Detailed evaluation of CHIRPS, PERSIANN-CDR, IMERG-F v07, and GSMaP satellite precipitation products against rain gauges in a semi-arid region, revealing elevation-dependent biases and product performance variations.
15. Establishment and Verification of an XGBoost–SHAP Interpretable Model for Rock Failure Mode Identification Based on Acoustic Emission Signals
Core Problem: Accurate identification of rock failure modes is crucial for safety assessment and disaster early warning in rock engineering, but traditional acoustic emission (AE) analysis methods have limitations in complex signal processing and model interpretability.
Key Innovation: An interpretable machine learning framework combining XGBoost and SHAP values to identify rock failure modes from acoustic emission signals, providing high accuracy and transparent analysis of decision-making processes.
16. Emerging trends in coefficient of consolidation determination: a statistical meta-analysis and nonparametric agreement assessment
Core Problem: Different methods for determining the coefficient of consolidation (cv) yield varying results, making it difficult to assess the compatibility of methods and identify systematic differences.
Key Innovation: A comprehensive comparative analysis of seven methods for determining cv using a large dataset and nonparametric statistical methods, identifying systematic differences and providing recommendations for assessing method compatibility.
17. A Comparative Evaluation of Scale-of-Fluctuation Estimation Methods for Small-Sample SWCC Data in Unsaturated Soils
Core Problem: Estimating the scale of fluctuation (SOF) for soil properties is challenging with limited data, which is common in laboratory-measured unsaturated soil properties.
Key Innovation: Evaluation of the autocorrelation function method (ACM) and an improved variance reduction function method (iVRM) for SOF estimation under small-sample conditions, demonstrating iVRM's superior stability and accuracy.
18. Soil characterization through shear wave velocity analysis of Lucknow city in the Indo-Gangetic plain of India
Core Problem: Understanding soil deformation under dynamic loading during earthquake shaking is essential for geotechnical engineering applications in Lucknow, Uttar Pradesh, a fast-growing urban city on the banks of Gomati and Sai rivers in the central Indo-Gangetic Plain (IGP).
Key Innovation: Multichannel Analysis of Surface Waves (MASW) surveys to characterise geotechnical parameters of the shallow subsurface soil in Lucknow, Uttar Pradesh, a fast-growing urban city on the banks of Gomati and Sai rivers in the central Indo-Gangetic Plain (IGP), and lies to the south of the central seismic gap region in the Himalayan collision zone.
19. Static and Dynamic Mixed Mode I–II Fracture of Anisotropic Shale: Numerical Analysis, Experimental Characterization and Theoretical Prediction
Core Problem: Revealing the fracture mechanism and establishing fracture criteria for shale rocks are fundamental issues in shale reservoir fracturing engineering.
Key Innovation: A comprehensive numerical, experimental, and theoretical study is conducted to investigate the anisotropic mixed mode I–II fracture behaviors of the Lushan shale under static and dynamic loading conditions.
20. Road Transport Resilience under Extreme Rainfall: Integrating Multiple Impact Factors and Delay Propagation
Core Problem: Assessing road network resilience under extreme rainfall events considering flood, visibility, and traffic signal failures, and delay propagation.
Key Innovation: A dynamic traffic assignment framework based on the link transmission model to represent network-wide delay propagation under rainfall, and a delay-based resilience index.
21. Probabilistic assessment of dynamic urban evacuation-sheltering functionality under typhoons based on interdependent road-shelter network
Core Problem: Assessing the functionality of urban evacuation-sheltering systems (UESS) during typhoons, considering the interdependent operational states of road networks and shelters.
Key Innovation: A holistic functionality metric capturing temporal variations in evacuation timeliness and shelter availability, combined with a probabilistic assessment framework for UESS.
22. Measuring the resilience of urban healthcare service availability using metro-bus double-layer network against extreme disturbances
Core Problem: Assessing the resilience of urban healthcare service availability under disturbances in public transportation systems, particularly metro and bus networks.
Key Innovation: A multi-layer network framework integrating community population, hospital capacity, and public transport network characteristics to simulate disturbance scenarios and analyze cascading risk effects.
23. A three-stage framework for stand-level automated stem volume estimation in temperate forests using Mobile laser scanning
Core Problem: Automated stem volume estimation in temperate natural forests from dense Mobile Laser Scanning (MLS) point clouds is challenging due to understory vegetation and complex stem structures.
Key Innovation: A three-stage method using deep learning for understory removal, bidirectional section growing (BSG) for stem detection/segmentation, and sector median points (SMP) for stem reconstruction, enabling accurate volume estimation.
24. Characterizing mangrove forest succession in Suriname using GEDI waveform metrics
Core Problem: Monitoring mangrove succession and structural development in dynamic coastal settings is difficult, especially in inaccessible regions.
Key Innovation: Using GEDI spaceborne LiDAR data to derive structural metrics and integrate them with Landsat-derived chronosequences to monitor mangrove successional trajectories and structural complexity.
25. A large-scale framework for deriving tidal flat topography from SWOT data
Core Problem: Accurate and spatially consistent large-scale monitoring of tidal flat topography is challenging.
Key Innovation: A novel framework leveraging satellite altimetry from the Surface Water and Ocean Topography (SWOT) mission, combined with a tide-constrained adaptive best-quantile reconstruction strategy, to derive high-accuracy tidal flat topography.
26. High-resolution annual desertification mapping in northern China using a novel comprehensive desertification index and unsupervised algorithm
Core Problem: Current desertification monitoring methods lack clear remote sensing mechanisms, robust extraction methods, and high-resolution large-scale desertification products.
Key Innovation: A comprehensive desertification index (CDI) integrating multisource remote sensing data (Sentinel-1/2) with a Gaussian mixture model (GMM) for automated desertification mapping, yielding a 10 m-resolution annual desertification dataset.
27. Advances and challenges in predicting wave runup in coastal regions: A scoping review of empirical, numerical, and AI-based approaches
Core Problem: Coastal regions are increasingly vulnerable to short-wave runup-induced hazards, and accurately predicting short-wave runup is essential for effective coastal management.
Key Innovation: A comprehensive scoping review of short-wave runup prediction methodologies, systematically evaluating empirical formulas, numerical models, and artificial intelligence (AI)-based approaches.
28. Leveraging additional VIIRS information to improve wildfire tracking in the western US
Core Problem: Current satellite fire monitoring has limitations due to infrequent coverage, smoke/cloud obscuration, omission of small fires, and atmospheric attenuation, hindering accurate quantification of fire behavior and emissions.
Key Innovation: Augmenting the standard VIIRS active fire product with additional candidate fire pixels to improve fire tracking, fill data gaps, and enhance the representation of smoldering fire activity.
29. Response of soil conservation functions to vegetation dynamics and restoration potential on the Loess Plateau
Core Problem: Quantifying the impact of vegetation change on soil conservation in the Loess Plateau, considering the diminishing marginal benefits of restoration.
Key Innovation: Integration of the InVEST model with multi-scale geographically weighted regression (MGWR) to elucidate the spatiotemporal heterogeneity of soil conservation driving mechanisms.
30. Investigating the effect of microbial inoculation on raindrop splash erosion: Indoor simulated rainfall and Hairsine-Rose model analysis
Core Problem: Understanding the process and mechanism of raindrop splash erosion under microbial application conditions for soil erosion control.
Key Innovation: Application of different types of microbial inoculants to silty clay loam black soil under simulated heavy rainfall, combined with COMSOL Multiphysics and the Hairsine-Rose soil erosion model.
31. Kernel density change: A new bitemporal lidar metric for directly mapping wildland fire fuel consumption
Core Problem: Scaling fuel load and consumption estimates in wildland fires requires remote sensing, but current methods have limitations.
Key Innovation: Introducing Kernel Density Change (KDC), a new modeling approach that directly predicts fuel consumption from bitemporal point cloud structural change metrics derived from LiDAR data.
32. Impact of dams on river regime and extreme flow events in MIÑO–SIL river basin (NW of the IBERIAN peninsula)
Core Problem: Assessing the long-term effects of river regulation by dams on flow regimes and extreme flood events relative to climate and land-use variability.
Key Innovation: Analyzing hydro-meteorological records spanning 1944–2023 to evaluate the roles of precipitation variability, land-use change, and dam regulation on river flow.
33. Hydrological alterations induced lakeward expansion of wetland vegetation in Dongting Lake, China's second-largest lake
Core Problem: Understanding the effects of hydrological alterations on wetland vegetation dynamics in Dongting Lake, considering both natural and human factors.
Key Innovation: Combining long-term hydrological data with high-resolution remote sensing imagery to examine the impacts of hydrological alterations before and after the construction of the Three Gorges Dam.
34. Mapping of anomalous C-band backscatter signals caused by subsurface scattering and their correlations with land surface characteristics over the Tibetan Plateau
Core Problem: Detecting and mapping backscatter anomalies and subsurface scattering on the Tibetan Plateau to improve soil moisture retrievals.
Key Innovation: Using C-band radar data (ASCAT, Sentinel-1A) to identify areas with anomalous backscatter signals related to subsurface scattering, and correlating these with land surface characteristics.
35. Integrating convolutional neural networks and explainable AI for enhanced winter road surface conditions classification using stationary RWIS imagery
Core Problem: Adverse winter weather compromises driving safety. Automated road surface condition (RSC) classification from RWIS imagery is challenging due to complex scenes.
Key Innovation: CNN applied to full stationary RWIS imagery, optimized with explainable AI (XAI) for camera angles, and evaluated for image resolution and data quantity trade-offs.
36. Advance prediction of rock mass classification in tunneling using improved D-S fusion and hybrid machine learning
Core Problem: Accurate rock mass classification is crucial for tunnel safety and cost. Single classification models lack robustness.
Key Innovation: Multi-model fusion framework using D-S evidence theory and LSTM for ahead-of-face rock class prediction, enhancing classification reliability and enabling proactive support design.
37. Safety performance evaluation of tunnel with void behind lining using an artificial neural network
Core Problem: Voids behind tunnel linings impact safety. Accurate and quick evaluation of tunnel safety is needed for maintenance.
Key Innovation: Using an artificial neural network to predict the safety performance of a tunnel with voids, using the minimum relative safety coefficient as the output variable.
38. Data-driven causal factor analysis of metro construction incidents using complex network theory
Core Problem: Near misses and accidents in metro construction pose safety threats. Identifying potential hazards and formulating prevention measures is difficult due to many factors and relationships.
Key Innovation: Applying complex network theory to examine interrelationships among causal factors of metro construction incidents, using near-miss and accident reports to build a Metro Construction Incident Network (MCIN).
39. Operational safety assessment of cracked tunnel linings reinforced with polypropylene fiber: investigation, field validation and numerical simulation
Core Problem: Crack control is essential for tunnel safety. Polypropylene fiber-reinforced concrete (PFRC) is investigated as a superior alternative for rehabilitating deteriorated linings.
Key Innovation: Combining field investigations and numerical modeling to analyze the mechanical properties and structural performance of PFRC linings, demonstrating enhanced load-bearing capacity and crack resistance.
40. Intelligent prediction of surface settlement troughs induced by twin shields tunnelling: Insights from a numerical modelling-empirical formulation-interpretable automated machine learning fusion method
Core Problem: Accurately predicting surface settlement induced by twin-shield tunneling is crucial for risk mitigation in urban areas.
Key Innovation: Integrating numerical modeling, empirical formulas, and automated machine learning (AutoML) to predict surface settlement troughs, with improved accuracy and interpretability using SHAP analysis.
41. New sensing-inversion integrated method for mechanical behavior analysis of shield tunnels during heavy rainfall
Core Problem: Understanding the mechanical response of shield tunnels under heavy rainfall conditions.
Key Innovation: Integrating displacement monitoring, distributed fiber optic sensing, and a strain–displacement-internal force recursive inversion method to analyze tunnel deformation, load development, and internal force distribution.
42. Data-physics integration model for predicting tunnel convergence subject to water level fluctuations and lining structure degradation
Core Problem: Monitoring data from underwater tunnels are often corrupted by noise and degradation is difficult to extract, limiting data utility.
Key Innovation: A data-physics integration model combining improved STL for seasonal components and a probabilistic degradation model with dynamic Bayesian networks to predict tunnel convergence.
43. A knowledge-data dual-driven framework for predicting surrounding rock classification in TBM tunneling and generalization analysis
Core Problem: Accurate prediction of surrounding rock classification (SRC) in TBM operations is critical, but machine learning generalization performance is lacking.
Key Innovation: A knowledge-data dual-driven SRC prediction framework (KD-SRC) combining the CSM model and APCA method with a TPE-LightGBM voting model, improving accuracy and cross-project applicability.
44. Physics-informed dictionary learning of time-varying 3D settlements from sparse monitoring data and 2D numerical models with consideration of complex stratigraphy
Core Problem: Computational challenges in predicting 3D settlements in land reclamations with complex stratigraphy, especially with limited monitoring data.
Key Innovation: Physics-informed dictionary learning approach integrating 2D numerical models and sparse monitoring data to efficiently predict time-varying 3D settlement.
45. Aquifer characterization and salinization origin using unsupervised machine learning and 3D gravity inversion modeling, Siwa Oasis, Egypt
Core Problem: Groundwater salinization in arid oasis environments, specifically the complex hydrogeological conditions in Siwa Oasis, Egypt, and the need for comprehensive characterization.
Key Innovation: Integrated approach combining machine learning clustering with gravity data analysis to characterize aquifer systems and identify structural controls on groundwater flow.
46. Physics-guided deep learning for global sea surface temperature forecasting: Balancing accuracy and stability across timescales
Core Problem: Accurate sea surface temperature (SST) forecasting across multiple timescales remains challenging. Daily forecasting frequently relies on autoregressive models prone to instability and over-smoothing, whereas monthly forecasting suffers from sparse data and the complex dynamics of ocean systems. Existing deep learning methods struggle to address these diverse challenges simultaneously.
Key Innovation: We introduce SSTFormer, a novel physics-guided deep learning framework that achieves leading results, with root mean squared error of 0.17 °C for daily forecasts and 0.60 °C for monthly forecasts, yielding lower bias and improved spatial coherence.
47. Physics-constrained neural network for daily pan evaporation forecasting in hyper-arid climates optimized by the Bat Algorithm
Core Problem: Accurate measurement of atmospheric water loss in hyper-arid regions, where evaporation rates are high and freshwater resources are scarce.
Key Innovation: A hybrid framework integrating a Physics-Constrained Neural Network (PCNN) and the Bat Algorithm (BA) to predict daily pan evaporation, incorporating physical constraints into the loss function.
48. Improving precipitation nowcasting via multiphysical parameter fusion in radar echo extrapolation
Core Problem: Existing modeling methodologies typically simplify precipitation nowcasting to a task of spatiotemporal sequence prediction based on radar echo reflectivity data. However, the reliance on unimodal reflectivity data including intensity-only information restricts the model’s ability to characterize the phase evolution and dynamic processes of hydrometeor particles, ultimately leading to insufficient extrapolation accuracy.
Key Innovation: We integrate radar echo reflectivity and four additional physical parameters of hydrometeor particles into a deep learning framework and propose a novel Physics-Informed Multimodal Echo Extrapolation neural network (PIEE).
49. High-resolution daily surface soil moisture mapping over the Qinghai–Tibet Plateau via predictors fusion and machine learning
Core Problem: The need for high-accuracy and high-spatiotemporal-resolution soil moisture (SM) products for understanding regional water–carbon–energy cycles and climate change over the Qinghai–Tibet Plateau (QTP), despite limitations of current microwave remote sensing.
Key Innovation: Development of a hyperparameter-optimized Random Forest (RF) downscaling framework that learns the nonlinear relationships between coarse-scale SMAP retrievals and regionally calibrated high-resolution predictors, producing a seamless 500 m daily SM product spanning 2015–2023.
50. Sorting carbonate clay-type lithium ores using a deep learning model with adaptive spectral-texture feature fusion
Core Problem: Beneficiation challenges of carbonate clay-type lithium ore due to its coexistence with karst-type bauxite, resulting in mixed, low-lithium ores.
Key Innovation: Joint modeling of spectral–texture features derived from hyperspectral imaging (HSI) with a deep learning framework to explore the feasibility of pre-sorting carbonate clay-type lithium ores.
51. A causal-aware artificial intelligence framework for mineral prospectivity mapping
Core Problem: Mineral prospectivity mapping (MPM) plays a vital role in locating potential mineral resources and informing exploration planning. Although data-driven machine learning (ML) techniques have demonstrated effectiveness in handling complex geological patterns, they often depend purely on statistical dependencies, ignoring the underlying causal mechanisms responsible for mineralization. This shortcoming can introduce spurious links and reduce the robustness of the model.
Key Innovation: In this study, we introduce a new causal-aware artificial intelligence (CausalAI) framework tailored for MPM, which incorporates both traditional metallogenic features and their causal interdependencies as input data.
52. RemoteWaterNet: a lightweight and efficient algorithm for remote river surface velocimetry
Core Problem: Accurate measurement of river surface velocity is essential for hydrological research, hydraulic engineering, flood forecasting, and hydrological monitoring. Although non-contact imaging technologies offer promising alternatives to traditional contact-based methods, existing approaches often lack robustness under diverse environmental conditions, especially with unstable results under different flow velocities and varying densities of tracer features.
Key Innovation: To address these challenges, this study proposes a novel RemoteWaterNet, a lightweight deep learning framework for robust and efficient remote river surface velocimetry.
53. Physics-constrained neural network for daily pan evaporation forecasting in hyper-arid climates optimized by the Bat Algorithm
Core Problem: Accurate measurement of atmospheric water loss in hyper-arid regions, where evaporation rates are high and freshwater resources are scarce.
Key Innovation: A hybrid framework integrating a Physics-Constrained Neural Network (PCNN) and the Bat Algorithm (BA) to predict daily pan evaporation, incorporating physical constraints into the loss function.
54. Joint inversion of hydraulic and thermal parameter fields using a CNN-based framework and transient hydraulic tomography
Core Problem: Accurate reconstruction of hydraulic and thermal parameter fields is essential for understanding coupled groundwater flow and heat transport.
Key Innovation: This study develops a Thermal–Hydraulic Tomography Neural Network (THT-NN), a DenseNet-based convolutional encoder–decoder that jointly estimates permeability (kp), porosity (n), thermal conductivity (kt), and specific heat capacity (Sp) from transient hydraulic and thermal tomography data.
55. Improved flash drought forecasting and attribution: A spatial-temporal causality-aware deep learning approach
Core Problem: Flash droughts pose significant challenges to water resource management and agricultural sustainability, making it imperative to improve their predictability to mitigate potential risks.
Key Innovation: This study presents a novel deep learning framework that integrates a spatial–temporal causality-aware (STC) module into a CNN-LSTM hybrid architecture to enhance flash drought prediction in China’s Greater Bay Area (GBA).
56. Synthetic geological porous media generation using deep learning: a comprehensive review
Core Problem: Challenges in characterizing porous media, which are fundamental to hydrology, geoscience, energy, and environmental systems, due to limitations in scalability, tunability, and accessibility of imaging and simulation techniques.
Key Innovation: A domain-specific synthesis of deep generative models (DGMs) for synthetic porous media (SPM) generation, evaluating how Variational Autoencoders (VAEs), diffusion models, and multimodal 3D pipelines extend beyond conventional GAN-based approaches.
57. Spatiotemporal variations of rainfall interception before and after implementation of Grain for Green Program in the Loess Plateau
Core Problem: Quantifying the impact of the Grain for Green Program (GFGP) on rainfall interception (RI) in the Loess Plateau, considering climate variability and ecological restoration.
Key Innovation: Quantified the individual contributions of climate change and vegetation restoration to changes in rainfall interception using the Gash sparse model and scenario simulations.
58. Effects of moss cover patterns on hydrodynamic parameters and particle size selectivity during karst erosion under rock surface flow
Core Problem: Understanding the impact of epilithic mosses on rock surfaces in regulating rock surface flow and downstream soil erosion in karst rocky desertification areas.
Key Innovation: Quantified the combined effects of moss cover on rock and slope surfaces on runoff generation, flow velocity, and sediment loss, revealing the dual role of epilithic mosses in erosion regulation.
59. Lagged streamflow depletion due to pumping-induced stream drying: Incorporation into analytical streamflow depletion estimation methods
Core Problem: Improving the estimation of streamflow depletion caused by groundwater pumping by incorporating stream drying into analytical depletion functions (ADFs).
Key Innovation: Developed an approach to incorporate stream drying into ADFs, simulating the temporal shift in streamflow depletion that occurs when summer stream drying causes stream network disconnections.
60. High-precision bedrock motion inversion from surface records using decoder-only convolutional attention Transformer
Core Problem: Conventional bedrock motion inversion methods often fail to accurately capture the real nonlinear behavior of soils under seismic loading.
Key Innovation: A decoder-only convolutional attention Transformer (D-Conv-Transformer) model for directly inverting bedrock ground motion time histories from surface-recorded data.
61. Experimental investigation of the seismic response and ultimate lateral resistance of pile–superstructure systems considering ground settlement due to pore water dissipation
Core Problem: Pile damage due to liquefaction-induced ground settlement and consolidation under repeated seismic loading.
Key Innovation: Centrifuge model tests reproducing progressive pile failure under repeated seismic shaking, clarifying ground settlement and pile head exposure effects.
62. Seismic damage assessment using frequency characteristics of acceleration records in centrifugal vibration table tests for CSG dam
Core Problem: Understanding the seismic performance characteristics of Cemented Sand and Gravel (CSG) dams, which are primarily found in seismically active regions.
Key Innovation: Centrifuge shaking table test on a CSG dam, examining frequency characteristics of acceleration records to assess seismic damage.
63. Influence of earthquakes on landslide activity in the permafrost regions of the Qilian Mountains, China
Core Problem: Quantifying the impact of earthquakes on landslide activity in permafrost regions, especially given increasing activity due to thawing.
Key Innovation: InSAR analysis reveals post-seismic acceleration of landslides, with deformation linked to fault proximity, altitude, and ground temperature, highlighting seismic impacts on permafrost slope stability.
64. Relationship between geomorphological characteristics, environmental settings and activity of transitional rock glaciers: Insights from a statistical analysis in the French Alps
Core Problem: Understanding the processes and geomorphic responses associated with transitional rock glaciers (TRGs) under degrading permafrost conditions.
Key Innovation: DInSAR analysis correlated with topo-climatic and geomorphic characteristics reveals that fast-moving TRGs are linked to higher latitudes, elevations, steep slopes, and convex morphologies, contrasting with slow-moving TRGs.
65. Autonomous and efficient large-scale snow avalanche monitoring with an Unmanned Aerial System (UAS)
Core Problem: Avalanche monitoring is difficult due to remote and dangerous locations.
Key Innovation: An unmanned aerial system (UAS) capable of autonomously navigating and mapping avalanches in steep mountainous terrain.
66. Assessing the spatial correlation of potential compound flooding in the United States
Core Problem: Assessing the likelihood of widespread compound flooding along the U.S. coastline.
Key Innovation: Generated plausible events preserving observed dependence to find that nearly half of compound floods on the West coast affect multiple sites.
67. A unified 3D geological model for Germany and adjacent areas
Core Problem: Creating 3D finite element models of geological structures is time-consuming.
Key Innovation: A unified 3D geological underground model of Germany and neighboring countries, combining 27 individual models with 146 units at 1x1 km² resolution.
68. Mapping and spatial distribution of relict charcoal hearths across Poland
Core Problem: Mapping and classifying relict charcoal hearths for research on forest history and human-environment interactions.
Key Innovation: A national inventory of >634,000 relict charcoal hearths in Poland using LiDAR data, classified by size, form, slope, and environment.
69. SWIIFT v0.10: a numerical model of wave-induced sea ice breakup with an energy criterion
Core Problem: Understanding and modeling wave-induced sea ice breakup, which is not well understood due to harsh field conditions.
Key Innovation: A novel criterion parameterizing ice fracture incorporated into a numerical model that simulates wave propagation, relating results to lab experiments.
70. Reentry and disintegration dynamics of space debris tracked using seismic data
Core Problem: Uncontrolled reentry of space debris poses a growing risk, but tracking debris during atmospheric burn-up is difficult, leading to poor prediction of fallout locations.
Key Innovation: A minimum-gradient fit seismic inversion methodology allows determining debris trajectory, speed, altitude, size, and fragmentation pattern, improving space situational awareness and debris hazard mitigation.
71. Safeguarding tailings dams at the climate brink
Core Problem: Tailings dams are vulnerable to climate change impacts, posing significant environmental and safety risks.
Key Innovation: This paper likely discusses strategies and recommendations for improving the safety and resilience of tailings dams in the face of climate change, though the abstract is unavailable.
72. Progressive eastward rupture of the Main Marmara fault toward Istanbul
Core Problem: The Main Marmara fault (MMF) poses a high seismic risk to Istanbul, but its rupture dynamics are not fully understood.
Key Innovation: Integrated observations reveal a series of eastward-propagating M > 5 events and a gradual eastward partial rupture of the MMF over the past ~15 years, highlighting the necessity of real-time monitoring of this part of the MMF.
73. Slip-Surface Depth Inversion and Influencing Factor Analysis Based on the Integration of InSAR and GeoDetector: A Case Study of Typical Creep Landslide Groups in Li County
Core Problem: Deep-seated slip surfaces of creeping landslides are difficult and costly to detect using traditional methods.
Key Innovation: Integrates SBAS-InSAR with GeoDetector to invert slip-surface depths and quantitatively evaluate the dominant factors controlling landslide depth, validated with borehole data.
74. Combining moored observations and SAR images in validating compound flood models
Core Problem: Validating flood inundation models with robust spatiotemporal data is challenging.
Key Innovation: Combines in situ measurements and Sentinel-1 SAR data to calibrate and validate a hydrodynamic model for simulating compound flood events in estuaries.
75. Remote Sensing, Vol. 18, Pages 280: YOLOv11-MDD: YOLOv11 in an Encoder–Decoder Architecture for Multi-Label Post-Wildfire Damage Detection—A Case Study of the 2023 US and Canada Wildfires
Core Problem: Fast and accurate post-wildfire damage detection is crucial for rescue planning, but traditional methods are time-consuming and may lack accuracy.
Key Innovation: An encoder-decoder architecture using YOLOv11 blocks and Modified UNet blocks for multi-label post-wildfire damage detection from remote sensing data, achieving high accuracy without pre-event imagery.