TerraMosaic Daily Digest: Jan 14, 2026
Daily Summary
Today's landslide research landscape is dominated by the integration of advanced technologies, particularly AI and remote sensing, for improved hazard assessment, prediction, and mitigation. A significant portion of the new papers focus on leveraging foundation models, deep learning, and multi-sensor data fusion to enhance the accuracy and efficiency of landslide detection, mapping, and monitoring. These AI-driven approaches address challenges such as limited labeled data, computational constraints, and the need for real-time analysis.
Beyond AI, there's a strong emphasis on understanding the underlying physical processes that trigger landslides, including rainfall thresholds, soil properties, and the impact of human activities. Several studies explore the use of numerical modeling and experimental techniques to investigate slope stability, soil erosion, and the effects of environmental factors like freeze-thaw cycles and water content. A growing number of papers address specific applications, such as early warning systems, infrastructure resilience, and the impact of landslides on critical infrastructure like tunnels and bridges. Finally, a cluster of papers focus on flood risk management, including urban flooding, flash flood preparedness, and the development of more accurate flood forecasting models.
Key Trends
- AI-Powered Remote Sensing: A surge in papers utilizing AI, particularly deep learning and foundation models, to analyze remote sensing data (optical, SAR, LiDAR) for landslide detection, mapping, and risk assessment. This includes innovative approaches for few-shot learning, domain adaptation, and multi-sensor data fusion.
- Focus on Soil Properties and Hydrological Processes: Continued emphasis on understanding the role of soil properties (e.g., spatial variability, water content, freeze-thaw effects) and hydrological processes (e.g., rainfall infiltration, seepage erosion) in triggering landslides.
- Infrastructure Resilience: A growing number of studies address the impact of landslides and other hazards on critical infrastructure, such as tunnels, bridges, and subway systems, with a focus on developing resilience strategies and early warning systems.
Selected Papers
This digest features 167 selected papers from 2,479 papers analyzed across multiple journals. Each paper has been evaluated for its relevance to landslide research and includes links to the original publications.
1. Compressing Vision Transformers in Geospatial Transfer Learning with Manifold-Constrained Optimization
Core Problem: Deploying large geospatial foundation models on resource-constrained edge devices is challenging due to their size and compression-induced accuracy loss.
Key Innovation: Uses manifold-constrained optimization (DLRT) to compress vision transformer-based geospatial foundation models during transfer learning, preserving task-specific accuracy with substantial parameter reduction.
2. Thermo-LIO: A Novel Multi-Sensor Integrated System for Structural Health Monitoring
Core Problem: Traditional thermography is limited in assessing complex geometries, inaccessible areas, and subsurface defects in structural health monitoring.
Key Innovation: Fuses thermal imaging with high-resolution LiDAR and LiDAR-Inertial Odometry (LIO) to create accurate temperature distribution representations for detailed monitoring of temperature variations and defect detection across large-scale structures.
3. Vision Foundation Models for Domain Generalisable Cross-View Localisation in Planetary Ground-Aerial Robotic Teams
Core Problem: Localizing rovers in a local aerial map using ground-view images is challenging due to the scarcity of real-world training data with ground-truth position labels.
Key Innovation: Uses cross-view-localizing dual-encoder deep neural networks, leveraging semantic segmentation with vision foundation models and high volume synthetic data to bridge the domain gap to real images for rover localization.
4. Small but Mighty: Dynamic Wavelet Expert-Guided Fine-Tuning of Large-Scale Models for Optical Remote Sensing Object Segmentation
Core Problem: Fine-tuning large-scale foundation models for optical remote sensing image (ORSI) segmentation is difficult due to excessive GPU memory consumption and high computational costs.
Key Innovation: Proposes a dynamic wavelet expert-guided fine-tuning paradigm (WEFT) with fewer trainable parameters, adapting large-scale foundation models to ORSI segmentation tasks by leveraging the guidance of wavelet experts.
5. SAM-Aug: Leveraging SAM Priors for Few-Shot Parcel Segmentation in Satellite Time Series
Core Problem: Few-shot semantic segmentation of time-series remote sensing images is challenging due to limited labeled data.
Key Innovation: Leverages the Segment Anything Model (SAM) to generate geometry-aware mask priors, integrated into training through a RegionSmoothLoss function, improving few-shot land cover mapping.
6. LPCAN: Lightweight Pyramid Cross-Attention Network for Rail Surface Defect Detection Using RGB-D Data
Core Problem: Current vision-based rail defect detection methods have high computational complexity, excessive parameter counts, and suboptimal accuracy.
Key Innovation: Proposes a Lightweight Pyramid Cross-Attention Network (LPCANet) that leverages RGB-D data for efficient and accurate defect identification, integrating MobileNetv2, a lightweight pyramid module, a cross-attention mechanism, and a spatial feature extractor.
7. Magnifying change: Rapid burn scar mapping with multi-resolution, multi-source satellite imagery
Core Problem: Timely and detailed burn scar mapping after wildfires is limited by the trade-off between spatial resolution and temporal revisit frequency of current satellite systems.
Key Innovation: A novel deep learning model (BAM-MRCD) employing multi-resolution, multi-source satellite imagery (MODIS and Sentinel-2) for timely production of detailed burnt area maps with high spatial and temporal resolution.
8. DeTracker: Motion-decoupled Vehicle Detection and Tracking in Unstabilized Satellite Videos
Core Problem: Multi-object tracking in unstabilized satellite videos is challenging due to platform jitter and weak object appearance.
Key Innovation: A joint detection-and-tracking framework (DeTracker) with a Global-Local Motion Decoupling module and a Temporal Dependency Feature Pyramid module for improved trajectory stability and object representation.
9. Searth Transformer: A Transformer Architecture Incorporating Earth's Geospheric Physical Priors for Global Mid-Range Weather Forecasting
Core Problem: Accurate global medium-range weather forecasting is fundamental but existing Transformer-based forecasting models neglect the Earth's spherical geometry and zonal periodicity.
Key Innovation: Proposes the Shifted Earth Transformer (Searth Transformer), a physics-informed architecture that incorporates zonal periodicity and meridional boundaries into window-based self-attention for physically consistent global information exchange.
10. High-fidelity lunar topographic reconstruction across diverse terrain and illumination environments using deep learning
Core Problem: Limited availability of meter-scale lunar topographic data constrains detailed planetary investigations.
Key Innovation: A deep learning framework incorporating a robust scale recovery scheme and extending the model to polar regions under low solar illumination conditions for high-fidelity topographic reconstruction.
11. Multimodal Signal Processing For Thermo-Visible-Lidar Fusion In Real-time 3D Semantic Mapping
Core Problem: Autonomous robot navigation and environmental perception in complex environments require enhanced 3D point cloud maps with semantic understanding.
Key Innovation: A method for semantically enhancing 3D point cloud maps with thermal information by performing pixel-level fusion of visible and infrared images and segmenting heat source features in the thermal channel.
12. Tipping Point Forecasting in Non-Stationary Dynamics on Function Spaces
Core Problem: Tipping points are abrupt, drastic, and often irreversible changes in the evolution of non-stationary and chaotic dynamical systems, such as climate.
Key Innovation: Learn the evolution of such non-stationary dynamical systems using a novel recurrent neural operator (RNO), which learns mappings between function spaces, and detect future tipping points using an uncertainty-based approach.
13. SlumpGuard: An AI-Powered Real-Time System for Automated Concrete Slump Prediction via Video Analysis
Core Problem: Traditional concrete slump testing is manual, time-consuming, and operator-dependent, making it unsuitable for continuous or real-time monitoring.
Key Innovation: An AI-powered vision system (SlumpGuard) that analyzes the natural discharge flow from a mixer-truck chute using a single fixed camera for automated slump prediction.
14. CuMoLoS-MAE: A Masked Autoencoder for Remote Sensing Data Reconstruction
Core Problem: Accurate atmospheric profiles from remote sensing instruments are frequently corrupted, and traditional gap filling blurs fine-scale structures, while deep models lack confidence estimates.
Key Innovation: A Curriculum-Guided Monte Carlo Stochastic Ensemble Masked Autoencoder (CuMoLoS-MAE) designed to restore fine-scale features, learn a data-driven prior over atmospheric fields, and quantify pixel-wise uncertainty.
15. AQ-PCDSys: An Adaptive Quantized Planetary Crater Detection System for Autonomous Space Exploration
Core Problem: Deploying sophisticated detection architectures on space-qualified hardware is challenging due to limited memory and computation power.
Key Innovation: An Adaptive Quantized Planetary Crater Detection System (AQ-PCDSys) that integrates a Quantized Neural Network with an Adaptive Multi-Sensor Fusion module for efficient and precise real-time crater detection.
16. LC4-DViT: Land-cover Creation for Land-cover Classification with Deformable Vision Transformer
Core Problem: Scarce and imbalanced annotations and geometric distortions hinder remote sensing-based land-cover classification.
Key Innovation: Combines generative data creation with a deformation-aware Vision Transformer (DViT) using GPT-4o-generated scene descriptions and super-resolved exemplars to synthesize class-balanced, high-fidelity training images.
17. Scaling Remote Sensing Foundation Models: Data Domain Tradeoffs at the Peta-Scale
Core Problem: Understanding scaling behaviors of AI for training foundation models on high-resolution electro-optical (EO) datasets in remote sensing.
Key Innovation: Trains progressively larger vision transformer (ViT) backbones using over a quadrillion pixels of commercial satellite EO data, analyzing successes and failure modes at peta-scale to inform data collection strategies and optimization schedules.
18. Controlling Ensemble Variance in Diffusion Models: An Application for Reanalyses Downscaling
Core Problem: Generating ensemble members in meteorology with controlled variance for reanalysis downscaling.
Key Innovation: Using a Denoising Diffusion Implicit Model (DDIM) to control ensemble variance by varying the number of diffusion steps, applied to wind speed downscaling.
19. Can AI Dream of Unseen Galaxies? Conditional Diffusion Model for Galaxy Morphology Augmentation
Core Problem: Improving generalization of machine learning models in astronomy due to limited labeled datasets.
Key Innovation: Using a conditional diffusion model (GalaxySD) to synthesize realistic galaxy images for augmenting ML training data, improving morphology classification and rare object detection.
20. AURASeg: Attention Guided Upsampling with Residual Boundary-Assistive Refinement for Drivable-Area Segmentation
Core Problem: Improving fine-grained feature segmentation for autonomous robots in indoor environments.
Key Innovation: Introducing Attention-Guided Upsampling with Residual Boundary-Assistive Refinement (AURASeg) to enhance edge delineation and preserve region accuracy in ground-plane semantic segmentation.
21. Roles of different controls influencing the intensity–duration rainfall thresholds for triggering landslides: an intercomparison of published thresholds
Core Problem: Rainfall thresholds for triggering landslides are influenced by site-specific conditions, determination methodologies, and targeted scales.
Key Innovation: Intercomparison of published rainfall thresholds to determine the effects of different controlling factors, showing data quality and quantity are crucial.
22. Coupling hydrological, geotechnical and machine learning models to enhance landslide prediction for an early warning system: application to Upper Garonne River Basin, Pyrenees, Spain
Core Problem: Landslide Early Warning Systems (LEWS) accuracy is often limited by simplified approaches that do not fully capture subsurface hydrological processes.
Key Innovation: The SCLAM model integrates snowmelt, hydrological, and landslide models with a random forest approach, using baseflow excess as a subsurface hydrological condition indicator. This is the Editor's Choice paper for its comprehensive integration of physical and data-driven approaches to enhance LEWS performance.
23. Interaction between anti-dip rock slope and anchor system under unloading effect: an experimental investigation
Core Problem: Understanding the anchoring and deformation characteristics of slope-anchor systems is crucial for assessing slope stability and optimizing anchorage.
Key Innovation: Experimentally explores the interaction between an anti-dip rock slope and anchor system under unloading conditions, simulating excavation to analyze stress and displacement.
24. Energy absorption behavior of mild steel tube-core sandwich structures for rockfall protection
Core Problem: Sandwich structures have limited application in high-energy rockfall protection due to low energy absorption efficiency and complex fabrication.
Key Innovation: Proposes a novel sandwich structure with mild steel tubes as core energy absorbers, using finite element modeling to optimize design variables for rockfall impact mitigation.
25. Mechanical behavior and damage constitutive model of silty mudstone under thermo-hydro-mechanical interactions
Core Problem: Understanding the deterioration behaviors and mechanisms of rocks under thermo-hydromechanical (THM) interactions is crucial for mitigating slope instability.
Key Innovation: Investigates the physicomechanical properties of silty mudstone under THM interactions using triaxial tests and develops a new constitutive model considering hydrothermal-expansion strain.
26. Predicting and mapping soil behavior types of dredged sediments using CPTU data and a hybrid geostatistics–data transformation method
Core Problem: Characterizing spatial distribution of soil types in subsurface environments.
Key Innovation: Integrates geostatistics with data transformation based on Piezocone Penetration Tests to evaluate spatial variability of soil behavior types.
27. Size effects of desiccation cracking behavior in clayey soil
Core Problem: Understanding the size effect on desiccation cracking in clayey soils to link laboratory findings to field conditions.
Key Innovation: Identifies critical size parameters to optimize laboratory tests for reflecting field conditions.
28. Analytical determination of the pressure exerted by a swelling clay soil on a tunnel
Core Problem: Determining soil pressures acting on a tunnel lining affected by swelling clay.
Key Innovation: In-situ measurement of swelling pressure on a tunnel and its upper bound.
29. Slurry pressure transmission in saturated sand during shield tunneling: expanding from infiltration tests to realistic scale simulation
Core Problem: Investigating slurry pressure transmission in saturated sand during shield tunneling.
Key Innovation: Combines laboratory infiltration tests and 3D numerical simulations to understand pressure transmission, sand concentration, and cutterhead operation interactions.
30. Physics-informed neural networks for back-analysis and consolidation settlement prediction using field measurements
Core Problem: Predicting consolidation settlement under staged loading.
Key Innovation: A physics-informed neural network (PINN) framework that integrates field-measured settlement data with governing equations from consolidation theories.
31. Macro–Microscopic Investigation of the Shear Behavior of the Grout–Rock Interface in Red-Bed Mudstone and Sandstone Under Different Water Contents
Core Problem: Investigating the effect of water content on the shear behavior of grout–rock interface.
Key Innovation: Systematic study of macroscopic and microscopic mechanical characteristics of interfaces between grout and red-bed mudstone and sandstone under varying moisture conditions.
32. Progressive Failure Characteristics and Damage Evolution of Anisotropic Rocks with Oriented Arrangement of Minerals Under Triaxial Cyclic Loading and Unloading
Core Problem: Investigating the progressive failure characteristics and damage evolution of anisotropic rocks under cyclic disturbances.
Key Innovation: Analysis of failure modes, crack propagation, and damage evolution using CT scanning and numerical simulations.
33. Seismic Stability Assessment of a Nuclear Reactor Mat Foundation Using Numerical Modeling
Core Problem: Evaluating the settlement and tilting of a nuclear reactor mat foundation under seismic loading.
Key Innovation: Numerical analysis in Plaxis 3D incorporating structural and seismic load parameters from the Rooppur Nuclear Power Plant Project.
34. Predictive Models for Estimating Swelling Potential of Expansive Soils Based on Geotechnical Properties
Core Problem: Estimating swelling potential of expansive soils using routine index properties.
Key Innovation: Development of new regression equations using only routine index properties to improve reliability of expansive-soil screening.
35. Influence of Soil Fabric and Gradation on Excess Pore Water Pressure Development Based on Laboratory PWP Tester
Core Problem: Investigating the role of soil fabric and granulometric properties in the generation of excess pore water pressure.
Key Innovation: Use of a newly developed PWP Tester to conduct experiments on natural and crushed sands with varying grain size distributions, grain shapes, and particle orientations.
36. Effect of spatial variability of soil properties on liquefaction behaviour – a probabilistic approach
Core Problem: Assessing soil liquefaction hazard realistically requires modeling subsurface spatial variability and seismic hazard uncertainty.
Key Innovation: A high-resolution, site-specific probabilistic framework using Gaussian Random Field modeling and Monte Carlo simulation to evaluate liquefaction potential, incorporating both soil spatial variability and seismic hazard uncertainty.
37. Probabilistic seismic liquefaction triggering assessment of gravelly soils
Core Problem: Gravelly soils, often overlooked in liquefaction assessments, have been shown to undergo significant strength and stiffness reductions during earthquakes.
Key Innovation: Development of probability-based liquefaction triggering predictive models using a database of 215 gravelly soil case histories, incorporating adjustments for earthquake duration, vertical effective stress, and median grain size.
38. Seismic performance of RC staircases in the 2023 Kahramanmaraş earthquakes: field observation, damage patterns and evacuation risk
Core Problem: Staircases, critical for evacuation, are often overlooked in post-earthquake assessments, leading to potential life-safety risks.
Key Innovation: Introduction of a new Evacuation Difficulty Score (EDS) that combines building height and staircase damage severity to estimate the probability of critical evacuation difficulty, bridging structural damage and functional usability.
39. Seismic vulnerability assessment and retrofitting of soft-storey buildings with different types of masonry infill walls
Core Problem: Soft-storey configurations in RC buildings pose a significant seismic risk due to stiffness irregularity.
Key Innovation: Characterizing the seismic behavior of a four-story building with different infill wall distributions and evaluating the effectiveness of retrofitting solutions like shear walls and column jacketing.
40. Paleo-landslide analysis reveals underestimated seismic hazards in the outer Western Carpathians
Core Problem: Paleoseismic methods often underestimate seismic hazards in stable continental interiors due to the rarity of surface-rupturing earthquakes.
Key Innovation: An integrated approach transforming paleolandslides into quantitative paleoseismic indicators by combining morphometric/structural analyses with dynamic back analysis and velocity-dependent friction laws, revealing evidence for M7+ paleoearthquakes.
41. Automatic characterization of rock blocks in jointed exposures using 3D point clouds
Core Problem: Extracting rock mass structural information from point cloud data is crucial for rockfall source analysis and engineering assessment.
Key Innovation: An automated framework for extracting discontinuity orientations, traces, surface areas, and block volumes from point cloud data, using a growth algorithm based on orientation buckets and depth-first search to identify complex blocks.
42. Forecasting CO2 injection-induced fault reactivation: A hybrid approach and its application to the Illinois Basin–Decatur Project
Core Problem: The risks of injection-induced fault reactivation during geological CO2 storage require comprehensive assessment to ensure long-term and effective CO2 storage.
Key Innovation: An integrated assessment combining physics-based modeling and probabilistic forecasting to evaluate fault reactivation risks, analyzing fault slip tendency indices, Coulomb failure stress, seismogenic index, and magnitude probability distributions.
43. Seismic site characterization using satellite-derived terrain morphometry and geological data: A machine learning approach for predominant frequency prediction
Core Problem: Characterizing predominant frequency (fo) across large seismically active regions is challenging due to limited field measurements and cost constraints.
Key Innovation: A DEM-based machine learning methodology for regional-scale fo prediction, using stacked ensemble models trained on terrain morphometric parameters, geological classifications, and bedrock depth information.
44. Sediment yield assessment of a small ungauged montane catchment in the North Caucasus
Core Problem: Assessing sediment yield in ungauged mountain catchments with limited long-term gauging data.
Key Innovation: Application of multiple independent methods (lake sedimentation, geomorphic process assessment, RUSLE model adaptation) to estimate erosion rates and sediment yield, highlighting the contribution of rockfalls, soil creep and solifluction.
45. Evaluating long-term geomorphic responses and sediment budget impacts of threshold-based floods and sediment replenishment in the Naka River, Japan
Core Problem: Quantifying the geomorphic effectiveness of sediment replenishment (SR) in restoring riverine sediment continuity under variable flow regimes in dam-regulated systems.
Key Innovation: Integration of threshold-based flood frequency analysis (POT), sediment grain size measurements, and multitemporal geomorphic change detection (GCD) to investigate downstream sediment dynamics and the impact of sediment replenishment.
46. Late-Quaternary activity of parallel normal faults along the southern margin of the Yuguang Basin in the Shanxi Rift, China and its seismogeological implications
Core Problem: Assessing the co-seismic slip history and seismogenic potential of the Yu-Guang Basin southern marginal fault (YBSM Fault) in the Shanxi Rift, China.
Key Innovation: Combining quantitative morphological analysis, aerial, and field surveys with terrestrial Light Detection and Ranging (t-LiDAR) to analyze fault scarps and identify paleoearthquakes, demonstrating the seismic risk of both bedrock and sedimentary fault branches.
47. Effects of sediment connectivity changes on channel evolutionary trajectory: the case study of the Taro and Ceno rivers in the Northern Apennines (Italy)
Core Problem: Investigating the post-1950s evolutionary trajectories of the Taro and Ceno rivers, focusing on the correlation between channel adjustments and sediment alterations.
Key Innovation: Quantifying changes in active channel width and bed level using multi-temporal orthophotos and topographical cross-sections, and assessing changes in structural sediment connectivity by applying the Index of Connectivity (IC).
48. Identifying gravel-sand transition zones in alluvial plains using normalized steepness index
Core Problem: Identifying gravel-sand transition (GST) location, where riverbed sediments shift from coarse gravel to fine sandy-muddy materials, for assessing natural disaster risk.
Key Innovation: Analyzing the longitudinal valley slope (S) and upstream drainage area (A) of rivers using LiDAR-derived digital terrain models to determine the normalized steepness index (ksn) threshold for GST zones.
49. A novel approach for utilizing UAV imaging and deep learning techniques for river channel measurement and flood simulation
Core Problem: Improving the efficiency and accuracy of river hydraulic analysis and flood simulation by reducing reliance on labor-intensive field observations.
Key Innovation: Integrating image sensing techniques and unmanned aerial vehicles (UAVs) with deep learning to collect spatial information on river channels, automatically detect gravel size distribution, and estimate Manning's roughness coefficients for hydraulic analysis and flood simulation.
50. Barriers to effective flood risk management in India: A case of 2021 Chiplun flooding
Core Problem: Understanding the narratives surrounding flood risk and how they influence government decisions and policy changes for flood risk reduction in Chiplun, Maharashtra, after the 2021 flooding.
Key Innovation: Integrates the Narrative Policy Framework (NPF) with the Pressure and Release Framework (PAR) to reveal how post-disaster narratives direct attention to certain dimensions of risk while overlooking others, impacting effective flood risk management.
51. Communicating safety: The impact of warning signs and messages on reducing risky driving in flood conditions
Core Problem: Understanding how flood-related road signs and messages affect drivers' decisions to turn around after encountering flood warning signs, aiming to reduce vehicle-related flood fatalities.
Key Innovation: Examines the impact of past exposure to warning signs, comprehension of those signs, and exposure to flood-related risk messages on drivers' decisions, highlighting the critical role of static road signage and risk messaging in shaping driver behavior.
52. Impact of urban gray infrastructure on urban flooding: a city-scale drainage and surface water modeling framework
Core Problem: Addressing urban floods by dynamically integrating the role of gray infrastructure and real-time storage dynamics in urban drainage networks.
Key Innovation: Introduces an advanced city-scale urban flood modeling framework with innovative dual-mode flow equations for inlet systems and a dynamic storage model, validated against historical flood records and satellite observations in Hong Kong.
53. The Effect of Evacuation Decisions on Flash Flood Preparedness in Fujairah, UAE: When the Waters Rise Are We Ready in Desert Country?
Core Problem: Examining households’ decisions to evacuate in response to flash flooding in the UAE and their intentions to prepare for future floods, addressing the increasing vulnerabilities of the population to this hazard.
Key Innovation: Applies the Protective Action Decision Model (PADM) to identify significant predictors of flash flood evacuation, including risk perceptions, affective responses, and receipt of warnings, providing insights for emergency management agencies.
54. A Stochastic Rain-on-Grid Framework for Handling Spatio-Temporal Rainfall Uncertainty in Impact-based Flood Nowcasting
Core Problem: Predicting flash flood impacts by addressing the intrinsic uncertainty in rainfall spatial-temporal structure and its propagation through hydrological and hydrodynamic processes.
Key Innovation: Introduces a Stochastic Rain-on-Grid framework that couples a high-resolution stochastic rainfall generator with a 2D hydrodynamic model, assessing how rainfall spatio-temporal variability influences catchment response and street-level flood impacts.
55. Nuisance flood risk: Defining a new horizon in urban flood risk management through hydrodynamic flood hazard modelling and indicator-based vulnerability assessment
Core Problem: Addressing the underrepresentation of Nuisance Flooding (NF) in existing literature by proposing a flood risk assessment framework tailored to NF conditions in Mumbai.
Key Innovation: Employs a sophisticated 1D–2D coupled hydrodynamic model to simulate high-resolution flood hazards and assesses vulnerability using a Shannon Entropy-cum-TOPSIS framework, producing spatially explicit NF risk maps.
56. Neighborhood-scale assessment of urban flood impacts on transportation network resilience: A case study of Mavişehir, İzmir
Core Problem: Assessing neighborhood-scale flood resilience by evaluating the vulnerability and resilience of the road transportation network in Mavişehir, İzmir, under flood conditions.
Key Innovation: Uses GIS-based network modeling and quasi-real-time traffic data to measure accessibility disruptions, increased travel times, and the network's capacity to support emergency operations during inundation.
57. Waterlogging susceptibility assessment in developed urban area using explainable machine learning methods with different negative sampling strategies
Core Problem: Assessing urban waterlogging susceptibility and identifying high-risk areas using explainable machine learning (ML) methods, addressing data limitation and the compromise of model classification performance.
Key Innovation: Conducts a catchment unit-based framework based on morphological characteristics and proposes three data-driven negative sample sampling strategies, significantly outperforming conventional random sampling.
58. Drought and flood hazards in inland valleys of Benin: severities, adaptation strategies and associated factors across the Sudano-Guinean and Guinean zones
Core Problem: Inland valleys in Sub-Saharan Africa are vulnerable to drought and flooding, requiring effective adaptation strategies.
Key Innovation: Spatiotemporal analysis of drought and wet conditions, identification of associated climatic and biophysical factors, and determination of socio-economic and environmental factors influencing farmers' adaptation responses in Benin.
59. Multiple levels of human instability due to urban overland flow within the 21st century: An urban Catchment study case in Brazil
Core Problem: Extreme rainfall events in urban environments pose a risk to pedestrian stability due to overland flow forces.
Key Innovation: Hydrodynamic modeling to assess human instability with multiple levels of vulnerability based on age, gender, weight, and height under climate change conditions in a catchment in São Paulo city, Brazil.
60. Time-Vertex machine learning for optimal sensor placement in temporal graph signals: Applications in structural health monitoring
Core Problem: Efficient sensor placement in structural health monitoring (SHM) to reduce costs without compromising monitoring quality, considering temporal dynamics.
Key Innovation: Time-Vertex Machine Learning (TVML) framework integrating Graph Signal Processing (GSP), time-domain analysis, and machine learning for interpretable and efficient sensor placement.
61. Rapid post-earthquake functionality prediction of subway systems based on graph neural networks and attentive transfer learning
Core Problem: Predicting the post-earthquake functionality of subway systems rapidly and accurately to aid in disaster response and recovery.
Key Innovation: A graph neural network (FuncGNN) with a Hierarchical Gate-Query Attention (HGQA) mechanism for enhanced cross-domain transferability in predicting subway system functionality after earthquakes.
62. Connectivity-based seismic design strategy for bridge networks by controlling fragility correlation among individual bridges
Core Problem: Enhancing the seismic connectivity of bridge networks, crucial for post-earthquake relief and reconstruction, by addressing the fragility correlation among bridges.
Key Innovation: A connectivity-based seismic design strategy that adjusts design parameters to enhance and reduce fragility correlation on the same and different paths, respectively, while ensuring bridge safety against the design seismic action.
63. Probabilistic risk and resilience assessment of high-rise concrete structures under mainshock-aftershock sequences: A case study
Core Problem: Assessing the probabilistic risk and resilience of high-rise concrete structures under mainshock-aftershock sequences, considering the cumulative damage effects of aftershocks.
Key Innovation: A comprehensive analytical method using the Copula function to construct a joint seismic hazard surface for mainshock-aftershock sequences, revealing the amplification effect of aftershocks on structural risk.
64. Seismic resilience evaluation for substations considering diversity of earthquake sources
Core Problem: Improving the seismic resilience assessment of substations by incorporating parameters of earthquake sources (PES) for site-specific hazard-consistent input.
Key Innovation: An improved evaluation framework integrated with parameters of earthquake source (PES), delivering site-specific, full-probability lifecycle resilience for given substations.
65. Reliability-informed inverse design of dual tunnels with deep evidential regression
Core Problem: Evaluating stability and performing inverse design of dual tunnel systems in cohesive–frictional soils under surcharge loading.
Key Innovation: Coupling finite element limit analysis, Deep Evidential Regression, and reliability-informed optimization for interpretable, probabilistic, and cost-efficient tunnel designs.
66. Modeling and performance analysis of emergency intelligence service process based on stochastic petri nets: A case study of the "7·20″ zhengzhou rainstorm in China
Core Problem: Analyzing the performance of Emergency Intelligence Service (EIS) processes in disaster response.
Key Innovation: An integrated framework based on Fault Tree Analysis (FTA) and Stochastic Petri Nets (SPN) for modeling and performance analysis of EIS processes.
67. Adaptive capability-based functionality assessment for post-disaster response in urban regional healthcare system incorporating demand uncertainty
Core Problem: Assessing the functionality of urban healthcare systems exposed to disasters, considering the uncertainty and dynamics of healthcare demand.
Key Innovation: An adaptive capability-based assessment method combined with a stochastic computing approach for dynamic demand generation and a multi-phase programming-based approach with a pseudo dispatch mechanism.
68. Introduction to a 45-year (1979–2023) global daily snow cover fraction product from multiple AVHRR satellites with accuracy assessment
Core Problem: Accurate monitoring of snow cover dynamics is essential for climate change attribution and water resource management, but long-term consistent datasets are lacking.
Key Innovation: A 45-year global daily snow cover fraction product (AVHRR10C1.V4) is generated by integrating data from 16 AVHRR sensors, addressing orbital drift and inter-sensor inconsistencies.
69. Estimating the upper depth of subsurface water on the Greenland Ice Sheet using multi-frequency passive microwave remote sensing, radiative transfer modeling, and machine learning
Core Problem: Monitoring subsurface water depth on the Greenland Ice Sheet is important for accurate runoff estimations, but is challenging.
Key Innovation: Multi-frequency microwave brightness temperatures and machine learning are combined with radiative transfer modeling to estimate the upper depth of liquid water on the ice sheet.
70. Probabilistic mapping of high-intensity forest fire potential via time series machine learning and remote sensing-informed fire spread simulations
Core Problem: Accurate probabilistic risk assessment of high-intensity forest fires is challenging, limiting the effectiveness of spatial estimations.
Key Innovation: Fire spread simulations and machine learning are integrated to enhance high-intensity forest fire potential estimations through multi-step time-series forecasting.
71. A global intercomparison of SWOT and traditional nadir radar altimetry for monitoring river water surface elevation
Core Problem: The water surface elevation (WSE) of rivers serves as fundamental data for various hydrological research and applications, but global validation of WSE is needed.
Key Innovation: SWOT WSE are compared with virtual stations derived from Sentinel-3 and Sentinel-6 missions, across five different node quality categories.
72. A novel UAV lidar-derived shrub structural index for estimating above-ground biomass
Core Problem: Accurate estimation of shrub AGB in arid regions is crucial but limited by sparse and low-quality UAV LiDAR point clouds.
Key Innovation: A novel Shrub Structure Index (SSI) is proposed, integrating UAV-based multispectral and LiDAR data to reconstruct the three-dimensional shrub structure under sparse point cloud conditions, improving AGB estimation accuracy.
73. Performance assessment of canopy gap fraction retrieval using multiple airborne LiDAR instrument configurations
Core Problem: Inconsistency in canopy gap fraction (GF) retrieval from airborne LiDAR due to different instruments or scanning configurations.
Key Innovation: Assessment of canopy GF retrieval performance from small-footprint airborne LiDAR under different collection scenarios using NEON data, encompassing 30 sites across 16 eco-regions in the United States from 2016 to 2022.
74. Satellite monitoring of Greenland wintertime buried lake drainage and potential ice flow response
Core Problem: Limited understanding of spatiotemporal dynamics of buried lake drainages (BLDs) across the Greenland Ice Sheet (GrIS) and their influence on ice flow dynamics.
Key Innovation: Detection of pan-GrIS wintertime BLDs by integrating Sentinel-1 and -2 satellite imagery and ArcticDEM data, revealing that BLDs have a more profound effect on ice flow dynamics than previously assumed.
75. Combined use of R-VSPI and VSPI for enhanced quantification of fire severity in south-eastern Australian forests
Core Problem: Need for accurate and near-real-time fire severity mapping to improve emergency response and post-fire recovery strategies.
Key Innovation: Introduces Decision-Based Hierarchical Learning (DBHL), a novel multi-sensor fire severity classification model that integrates Synthetic Aperture Radar (SAR; Sentinel-1 backscatter) and optical (Sentinel-2 reflectance) data.
76. Change tensor: Estimating complex topographic changes from point clouds using Riemann manifold surfaces
Core Problem: Estimating complex 3D topographic surface changes including rigid spatial movement and non-rigid morphological deformation from multi-temporal 3D point clouds.
Key Innovation: Considers the dynamic evolution of topographic surfaces as the geometric changes of Riemann manifold surfaces, building Euclidean and non-Euclidean coordinate systems to solve for rigid transformation and non-rigid deformation.
77. A Spatially Masked Adaptive Gated Network for multimodal post-flood water extent mapping using SAR and incomplete multispectral data
Core Problem: Mapping water extent during a flood event is essential for effective disaster management, but the adaptive integration of partially available MSI data into the SAR-based post-flood water extent mapping process remains underexplored.
Key Innovation: Proposes the Spatially Masked Adaptive Gated Network (SMAGNet), a multimodal deep learning model that utilizes SAR data as the primary input and integrates complementary MSI data through feature fusion.
78. National mapping of wetland vegetation leaf area index in China using hybrid model with Sentinel-2 and Landsat-8 data
Core Problem: Accurately mapping LAI at a broad scale is essential for conservation and rehabilitation of wetland ecosystem, but retrieving LAI of wetlands remains a challenging task with significant uncertainty.
Key Innovation: Proposes a hybrid strategy that incorporated active learning (AL) technique, physically-based PROSAIL-5B model, and Random Forest machine learning algorithm to map wetland biomes LAI across China from Sentinel-2 and Landsat-8 imagery.
79. Mapping aboveground tree biomass and uncertainty using an upscaling approach: A case study of the larch forests in northeastern China using UAV laser scanning data
Core Problem: Forest biomass mapping and monitoring are vital for understanding carbon cycling, but a remote sensing-based framework that upscales biomass estimation from the individual-tree level remains underdeveloped, especially for rigorously quantifying the propagation of associated uncertainties.
Key Innovation: An upscaling framework for aboveground biomass (AGB) mapping and prediction uncertainty estimation using unmanned aerial vehicle laser scanning (UAVLS) data, including an analytical framework to characterize the AGB prediction uncertainty considering error propagation throughout the whole upscaling workflow.
80. Toward noise-resilient retrieval of land surface temperature and emissivity using airborne thermal infrared hyperspectral imagery
Core Problem: Effective retrieval of surface parameters from thermal infrared remote sensing is fundamentally challenging due to degraded spectral quality caused by narrow bandwidths of thermal infrared hyperspectral imagers, atmospheric line absorption interference, and limitations in sensor manufacturing.
Key Innovation: A Noise-Resilient Atmospheric Compensation with Temperature and Emissivity Separation (NRAC-TES) method, where the noise-resistant capability is mainly achieved through the NRAC module during the atmospheric compensation (AC) stage.
81. Dual-domain representation alignment for unsupervised height estimation from cross-resolution remote sensing images
Core Problem: Existing unsupervised domain adaptation methods for single-view remote sensing images often neglect the discrepancy in spatial resolution between the source and target domains, restricting their ability to generalize from low ground sample distance (GSD) to fine GSD images effectively.
Key Innovation: A cross-resolution unsupervised height estimation framework with Dual-Domain Representation Alignment (DDRA) to address both challenges: (1) How to capture resolution-invariant representations for better unsupervised domain adaptation; (2) How to maintain geometric integrity and spatial layout across domains.
82. ARSGaussian: 3D Gaussian Splatting with LiDAR for aerial remote sensing novel view synthesis
Core Problem: Novel View Synthesis (NVS) can reconstruct scenes from multi-view images, but large distances and sparse viewing angles during aerial remote sensing collection can cause the model to easily produce floaters and overgrowth issues due to geometric estimation errors.
Key Innovation: ARSGaussian, an innovative novel view synthesis (NVS) method for aerial remote sensing that incorporates LiDAR point cloud as constraints into the 3D Gaussian Splatting approach, adaptively guiding the Gaussians to grow and split along geometric benchmarks, thereby addressing the overgrowth and floaters issues.
83. Information transmission: Inferring change area from change moment in time series remote sensing images
Core Problem: Deep learning approaches to time series change detection treat change area detection and change moment identification as distinct tasks, despite the intrinsic relationship between them.
Key Innovation: A time series change detection network, named CAIM-Net (Change Area Inference from Moment Network), to ensure consistency between change area and change moment results by inferring change area from change moment.
84. Denoising VIIRS and Sentinel-2 MSI ocean color imagery for improved floating algae monitoring using noise-simulation-aided deep learning
Core Problem: Floating algae index (FAI) images derived from VIIRS and Sentinel-2 MSI often suffer from complex and variable noise, and adapting deep learning methods to sensor-specific noise variations remains challenging due to the need for large-volume and high-quality training data.
Key Innovation: A two-step denoising process: 1) simulating noise using spatial frequency domain information to generate customized representative training data from limited samples, and 2) training the state-of-the-art Multi-scale Image Restoration Network (MIRNet) for optimal performance.
85. A differentiable method for novel view SAR image generation via 3D Gaussian Splatting
Core Problem: SAR target images often suffer from insufficient view coverage due to the constraints of observation geometry, which poses challenges for data-driven SAR target classification and recognition methods.
Key Innovation: Integration of SAR imaging mechanisms with the concept of 3D Gaussian Splatting (3DGS), proposing an advanced differentiable method for novel view SAR image generation.
86. MHFNet: Multimodal hybrid fusion framework for misaligned SAR-Optical ship detection
Core Problem: Most existing fusion-based methods for multimodal ship detection assume spatially aligned optical–SAR pairs, an assumption that rarely holds in practice due to differences in sensor geometry, acquisition timing, and registration errors.
Key Innovation: MHFNet, an alignment-aware hybrid fusion framework that systematically integrates improvements at the feature, loss, and decision levels to address the challenge of misaligned SAR-Optical imagery.
87. Towards resolution-arbitrary remote sensing change detection with Spatial-frequency dual domain learning
Core Problem: Deep learning-based change detection (CD) is widely used for same-resolution images, but high-resolution remote sensing images are often not continuously available over time, necessitating the ability to handle images with arbitrary resolution differences.
Key Innovation: A novel resolution-arbitrary CD network that enables CD with input images of arbitrary resolution differences by solving resolution mismatch and boosting accuracy through high- and low-level tasks learning from each other.
88. SARCLIP: a multimodal foundation framework for SAR imagery via contrastive language-image pre-training
Core Problem: The SAR domain lacks a general-purpose multimodal foundation model, and conventional models suffer from limitations in semantic understanding and generalization capabilities.
Key Innovation: A Contrastive Language-Image Pre-Training framework tailored for SAR data (SARCLIP) that integrates two SAR-specific modules: Noise-Robust Encoding (NRE) and Hierarchical Prompt Learning (HPL).
89. Complex-valued mix transformer for SAR ship detection
Core Problem: Transformer-based object detection algorithms neglect the importance of phase data in capturing fine structural details of SAR targets, making existing methods vulnerable to speckle noise and failing to capture high-frequency structural information.
Key Innovation: A complex-valued SAR target detection framework based on a “dual-branch feature learning–feature fusion–feature selection” strategy, named Complex-valued Mix Transformer (CVMT), which incorporates both SAR amplitude and phase information.
90. Beyond spectral signals: Geographic features drive bathymetric accuracy in the turbid Sancha Lake using machine learning
Core Problem: Accurate bathymetric mapping in inland water bodies presents significant challenges for conventional optical remote sensing due to complex water quality conditions and variable bottom types.
Key Innovation: A novel Spectral-Geospatial XGBoost Regression (SG-XGBoost) model that revolutionizes depth estimation by integrating comprehensive spectral transformations with explicit geographic coordinates through gradient boosting methodology.
91. Altimetry river water level retrieval over complex environments: assessment and diagnosis of different strategies
Core Problem: There remain many challenges to retrieving accurate river water levels from satellite altimetry, especially for rivers surrounded by various water bodies.
Key Innovation: Investigation of six retrackers in the Yangtze River and a proposed strategy combining FFSAR and MWaPP+ to enhance accuracy and the number of observations.
92. A spatio-temporal machine learning method for estimating high-resolution XCO2 in China
Core Problem: Satellite XCO2 data exhibit significant spatial discontinuity, making it difficult to meet the needs of research at small spatial scales, and mainstream machine learning methods are mostly data-driven, limiting the accuracy and generalization ability of the models.
Key Innovation: A new spatiotemporal XGBoost model (XGBKT) to generate high-resolution XCO2 dataset covering the entire territory of China, focusing on spatial correlation, temporal heterogeneity, and temporal periodicity.
93. Mapping hidden heritage: Self-supervised pre-training on high-resolution LiDAR DEM derivatives for archaeological stone wall detection
Core Problem: Automated mapping of undocumented dry-stone walls in remote, vegetated landscapes is challenging due to occlusion and limited labeled training data.
Key Innovation: A self-supervised cross-view pre-training framework (DINO-CV) based on knowledge distillation, using LiDAR DEM derivatives for accurate and data-efficient mapping of dry-stone walls.
94. An integrated object-based-deep learning approach applied for mapping armed conflict impacts and land scars
Core Problem: Efficient, cost-effective, and transferable methods are needed for mapping environmental impacts and land scars from wars and armed conflicts.
Key Innovation: An integrated object-based image analysis (OBIA) and deep learning convolutional neural networks (DL-CNNs) approach for mapping war and armed conflict scars and impacts, validated across multiple conflict zones.
95. Near real-time monitoring reveals extensive recent forest disturbance in Ghana's protected areas
Core Problem: Undocumented extent, rate, and locations of forest disturbances in Ghana's Protected Areas (PAs) due to recent policy changes facilitating logging and mining.
Key Innovation: Application of the fusion near real-time (FNRT) algorithm using Landsat, Sentinel-1, and Sentinel-2 data to monitor and quantify forest loss in Ghana's PAs.
96. Improving soil moisture estimation in wet soils using L-band Synthetic Aperture Radar (SAR) through polarization and filtering optimization
Core Problem: The efficacy of microwave remote sensing diminishes in regions with high soil water content, limiting accurate soil moisture mapping for irrigation management and landslide risk assessment.
Key Innovation: Optimizing polarization, despeckling filter application, and ground truth data determination to improve soil moisture estimation in areas with volumetric water content exceeding 0.3 m3/m3 using L-band SAR.
97. Sensitivity to soil moisture of Capella X-band high-resolution SAR data over forests
Core Problem: Monitoring soil moisture in forests is challenging due to tree interference. Current methods lack the resolution and agility for effective monitoring.
Key Innovation: Demonstrates the strong response of Capella X-band SAR data to soil moisture changes in forest openings, enabling high-resolution soil moisture mapping for applications like landslide prediction and wildfire risk assessment.
98. 2023 activity of Nyamulagira volcano monitored by SAR interferometric coherence
Core Problem: Monitoring volcanic activity, particularly lava flows, in remote areas like Nyamulagira volcano is challenging due to limited accessibility and the need for timely information for hazard management.
Key Innovation: Utilizes multi-temporal InSAR coherence from Sentinel-1 data to track lava effusion, flow extent, and solidification phases during the 2023 Nyamulagira eruption, providing valuable data for the GVO local monitoring team.
99. Terrace abandonment enhances rather than diminishes hydrological functioning in karst terraces by reshaping soil pore structure
Core Problem: Terrace abandonment impacts karst soil hydrology, but the pore-scale mechanisms are unclear.
Key Innovation: NMR analysis shows vegetation recovery fosters finer, better-connected pore networks, enhancing water transport capacity in abandoned karst terraces.
100. Linking runoff and sediment characteristics to karst development degree in small watershed
Core Problem: Surface/subsurface structure in karst areas complicates soil erosion and sediment transport processes.
Key Innovation: Relates sediment transport to karst development using hysteresis analysis, linking sediment sources to watershed characteristics.
101. Can Sentinel-1 reliably provide regional-scale information on avalanche activity
Core Problem: Validating SAR-based avalanche detections is challenging due to limited ground truth data and uncertainties in detection algorithms.
Key Innovation: Automated Sentinel-1 SAR algorithm to map avalanche debris, validated against diverse datasets, demonstrating consistent agreement and strong alignment with avalanche hazard levels.
102. Experimental study of test methods for the acoustic properties of snow
Core Problem: Automated monitoring of snow acoustic attenuation remains highly challenging due to dynamic changes in snow properties.
Key Innovation: Designed acoustic testing system using the transmission method to conduct acoustic measurements on natural snow samples for avalanche early warning.
103. A “Proactive-Prevention and Post-Resistant” support method for alleviating rockburst in deep-buried large-section tunnels
Core Problem: Rockbursts challenge the safe construction of large-section hard rock tunnels under high geostress environments, and conventional active support strategies have limitations.
Key Innovation: Proposes a 'Proactive-Prevention and Post-Resistant' support method involving initial tunnel excavation with pre-stressed rockbolts to establish an advanced pressure arch, followed by tunnel expansion and supplementary reinforcement.
104. Experimental and numerical study on the anti-impact performance of fluted tapered tube
Core Problem: Insufficient anti-impact performance of hydraulic supports during rockbursts.
Key Innovation: Proposes a fluted tapered tube energy absorption component connected to the column to improve energy absorption capacity during rockbursts.
105. Precursor response mechanism of tunnel water surge in water-rich fault fracture zones based on similar physical model
Core Problem: Tunnel water surge in fault fracture zones poses a threat to tunnel engineering safety, requiring understanding of the water surge mechanism during excavation.
Key Innovation: Constructs a tunnel similar physical model and conducts water surge tests during excavation to analyze disaster precursors response characteristics, proposing a critical anti-outburst thickness calculation model.
106. A computational framework for dynamic quantitative assessment of surrounding rock damage based on failure approaching index in underground construction
Core Problem: Superimposed surrounding rock stress during ultra-large mining height working face passing through abandoned roadways induces intense dynamic disasters, and existing evaluation methods fail to achieve full-process dynamic quantification of damage.
Key Innovation: Constructs a dynamic evaluation system for surrounding rock damage based on the failure approaching index, verified through experiments, numerical simulations, and on-site engineering applications.
107. Prediction of mud cake formation on shield cutterheads based on multi-source monitoring data integrated with deep learning method
Core Problem: Accurate prediction of mud cake formation on shield cutterheads is critical for construction efficiency and operational safety, but traditional approaches are inadequate for dynamic and accurate prediction.
Key Innovation: An integrated approach combining a Transformer–LSTM deep learning model with real-time shield monitoring and advanced geological prediction to dynamically predict cutterhead mud cake rate.
108. Evaluation and analysis of the ductile dynamic response of mountain tunnels based on shaking table tests
Core Problem: Limited research on ductile displacements and deformation rate control indices in shaking table tests of mountain tunnels.
Key Innovation: Development of a method to directly measure ductile displacement time-history in tunnels and propose MDR thresholds for shallow, medium, and deep-buried tunnels.
109. Rock mass discontinuity trace mapping using a voxel-based morphology-topology framework
Core Problem: Existing point cloud-based methods for discontinuity trace mapping face challenges including insufficient trace connectivity and ambiguous topological relationships between traces.
Key Innovation: A voxel-based morphology-topology approach that enhances and refines spatial connectivity through morphological dilation and an improved secondary erosion process.
110. Structure-constrained semi-supervised segmentation of complex geomaterial CT images
Core Problem: Accurate segmentation of pore–fracture systems in rock CT images is challenging due to the scarcity of annotated data and the morphological complexity of heterogeneous geological materials.
Key Innovation: A semi-supervised segmentation framework that integrates limited manual annotations, pseudo-labels generated by the Segment Anything Model (SAM), and a multi-task learning strategy guided by physical structure constraints.
111. Developing a novel tricriteria-based reinforced belief rule Base optimization Algorithm to assess building tilt in underground tunnel construction
Core Problem: Assessing building tilt in underground tunnel construction.
Key Innovation: A novel tricriteria-based reinforced Belief Rule Base (BRB) optimization algorithm.
112. Failure behaviour simulation of transversely isotropic rocks considering realistic grain structure and bedding plane morphology
Core Problem: Accurately capturing rock failure behavior in transversely isotropic rocks with complex microstructures.
Key Innovation: A novel grain-based model with transverse isotropy (GBM-T) based on the bonded particle method is proposed to explore the intrinsic mechanism underlying the failure of transversely isotropic rocks at the grain scale.
113. Microstructure-driven prediction of undrained shear strength of deep-sea sediments: a multivariate approach bridging physical–mechanical properties
Core Problem: Predicting undrained shear strength in deep-sea sediments using conventional terrestrial soil models is inaccurate due to the unique properties of deep-sea environments.
Key Innovation: Developed predictive models incorporating buoyant unit weight, liquidity index, and sensitivity as key factors, achieving superior accuracy for deep-sea sediments.
114. Predictive modeling of pore pressure build-up in vibratory pile driving through machine learning
Core Problem: Understanding and predicting pore pressure build-up and liquefaction potential during vibratory pile driving in saturated sandy soils for large-scale infrastructure projects.
Key Innovation: Integrated 3D numerical modeling with machine learning (ANN and symbolic regression) to predict soil liquefaction, using SHAP values to assess feature importance and creating design charts for practical guidance.
115. Real-time separation of reservoir-induced and rainfall-induced seepage in earth-rock dams using data-driven methods
Core Problem: Conventional seepage monitoring systems for earth-rock dams struggle to dynamically separate reservoir-level-induced seepage from rainfall-induced seepage, hindering effective dam safety assessment.
Key Innovation: Developed a data-driven real-time separation framework using 1DGAN and a CGO-enhanced TCN-BiLSTM-Attention model to dynamically separate rainfall-reservoir coupled seepage components.
116. SMOTE-BN-FLA: enhanced Bayesian network for rainfall-induced flood loss estimation and mechanism decoding in data-scarce regions
Core Problem: Accurate estimation of rainfall-induced flood losses is challenging due to imbalanced data distributions and a lack of understanding of disaster mechanisms, especially in data-scarce regions.
Key Innovation: Proposed SMOTE-BN-FLA, an integrated framework coupling Synthetic Minority Oversampling Technique (SMOTE) with data-driven Bayesian Networks (BN) for flood loss assessment and mechanism decoding.
117. Freeze-thaw processes induce soil water and salt migration in farmland-ditch systems
Core Problem: Understanding the 2-D dynamics of water-salt-heat in farmland-ditch systems undergoing freeze-thaw processes in cold regions, where previous studies mainly focused on 1-D vertical dynamics.
Key Innovation: A 2-D model examining water-salt-heat dynamics in farmland-ditch systems undergoing freeze-thaw, revealing asymmetric recharge and drainage patterns and the impact of ditch water depth on soil salinity reduction.
118. Isotopic evidence unveils the regulation of biocrusts on shrub root water uptake strategy and water use efficiency in a semiarid ecosystem
Core Problem: Understanding how biocrusts influence vascular plant water use strategies in drylands, specifically focusing on shrub-biocrust interactions and their impact on ecohydrological processes.
Key Innovation: Demonstrates that biocrusts alter soil moisture profiles, leading to increased shallow water uptake by shrubs and enhanced water use efficiency, highlighting the importance of shrub-biocrust relationships in dryland ecosystems.
119. Beyond depth-direction segregation: Independent flow-direction mechanisms drive size segregation in granular flows
Core Problem: Granular flows exhibit size segregation along both the depth and flow-directions, yet the mechanisms driving segregation in the flow-direction remain poorly resolved.
Key Innovation: Identified two intrinsic flow-direction segregation mechanisms that operate independently of depth-direction stratification: forward kinetic sieving and shear-induced migration. Developed a continuum framework that couples depth- and flow-direction segregation into a unified model.
120. Study on the seepage characteristics and influencing factors of microbially stabilized gap-graded sand using coupled CFD-DEM
Core Problem: Traditional physical testing methods are significantly limited in characterizing the particle migration mechanisms and microstructural evolution of MICP-treated sand under seepage conditions.
Key Innovation: Employed a coupled computational fluid dynamics-discrete element method (CFD-DEM) numerical model to investigate the seepage characteristics and induced particle migration of MICP-treated gap-graded sand under hydraulic loading.
121. An explainable resilience-informed framework for surrogate modeling and multi-objective optimization of embankments under seismic loading
Core Problem: Ensuring seismic resilience of transportation earth structures is critical for maintaining lifeline functionality following major earthquakes, with interpretability and data-driven modeling being key to achieving maximum resilience and rapid demand prediction.
Key Innovation: Presents an explainable resilience-informed framework that integrates finite element (FE) simulations, probabilistic demand modeling, and machine learning-based surrogate modeling to rapidly assess and optimize the seismic resilience of earth embankments.
122. Advancing internal erosion analysis through three-dimensional FEM simulation: insights from the Agly river dike
Core Problem: Internal erosion is a primary cause of degradation and failure in hydraulic structures and natural deposits, but most existing studies are predominantly limited to two-dimensional (2D) cross-sectional analyses.
Key Innovation: Integration of geophysical imaging data to develop a large-scale three-dimensional (3D) geometry model and a transient FEM that couples multi-field and multi-phase suffusion processes to simulate seepage and fines migration within the dike foundation under realistic flood conditions.
123. Mathematical descriptions of grading linked with prediction of mechanical consequences of suffusion
Core Problem: Internal erosion results from water flowing through the embankments, removing particles from the soils forming the embankments. It changes a soil’s particle size distribution, increases its void ratio, shifts its critical state line upwards in the compression plane and alters its stress–strain behavior.
Key Innovation: New mathematical links and constitutive model ingredients to capture the effects of internal erosion. The evolution of the particle size distribution is characterised through a grading state index, defined in terms of geometrical properties which are fractal.
124. A Novel Continuous-Discontinuous Deformation Analysis for Full-Process Rock Fracture Modeling with Dynamic Contacts and Data-Driven Constitutive
Core Problem: Nonlinear behavior in continuum, brittle fracturing, and afterwards large displacements are ubiquitous in disaster prevention but remain significant challenges for numerical simulations, which entail accurate nonlinear constitutive, continuum-discontinuum transition, and dynamic contacts between discrete blocks.
Key Innovation: A novel continuous–discontinuous deformation analysis (CDDA) method for full-process rock fracture modeling. A data-driven double-minimization iterative solver is developed to replace the conventional linear elastic constitutive in DDA.
125. Mesoscopic mechanisms of anisotropic suffusion behaviors of gap-graded soil: Identifying preferential suffusion paths based on strong-force chains and anisotropic pore structures
Core Problem: The suffusion behavior of gap-graded soil can be significantly influenced by various factors, including hydraulic conditions, particle size distribution, and stress states. Among these factors, the effects of stress states are particularly complex.
Key Innovation: A series of CFD-DEM coupled simulations is conducted to examine the suffusion behavior of gap-graded sand under different initial stress conditions from both macroscopic and mesoscopic perspectives. A meso-scale analysis framework based on Voronoi tessellation was developed to identify preferential suffusion paths under anisotropic stress conditions.
126. Machine learning-based national Vs30 models and maps for Italy
Core Problem: This study introduces a machine learning XGBoost–SHAP framework to estimate time-averaged shear-wave velocity in the uppermost 30 m (Vs 30 ) across Italy using over 20,000 in-situ measurements.
Key Innovation: Two independent XGBoost models are trained and their residuals modelled, enabling the construction of a final combined Vs 30 map, at 300 m resolution, through inverse-error weighting. This integrated approach captures both geological and physiographical controls, while reducing local prediction errors.
127. Physical modeling of pile with different bending stiffness under lateral spreading with distinct ground conditions: A 1-g shaking table investigation
Core Problem: Liquefaction-induced lateral spreading causes damage to pile foundations, and the influence of pile bending stiffness on soil-pile interaction is not fully understood.
Key Innovation: A series of shaking table tests analyze the response of piles with varying bending stiffness to lateral forces under liquefaction conditions, considering ground slope and crust thickness.
128. Effects of soil heterogeneity on the numerical simulation of liquefiable soil deposits using a multi-surface plasticity model
Core Problem: Soil spatial variability impacts the response of liquefiable deposits under earthquake loading, but is often not accounted for in simulations.
Key Innovation: Stochastic numerical analyses using the Random Finite Element Method (RFEM) represent soil relative density as a spatially correlated Gaussian random field to predict soil response within a confidence interval.
129. Toward explainable pile buckling capacity prediction in liquefiable strata: Integrating a hybrid framework of AutoML and SHAP
Core Problem: Existing solutions for predicting pile buckling capacity in liquefiable strata rely on empirical assumptions and are computationally expensive.
Key Innovation: An explainable machine learning framework using AutoML and SHAP is proposed for rapid and accurate prediction of pile buckling capacity, quantifying the contributions of input features.
130. Investigation on evolution mechanism of air injection desaturation region: physical experiments and twophase flow simulation
Core Problem: Soil desaturation by air injection is effective for mitigating liquefaction, but the mechanism of air migration under different loads is not fully understood.
Key Innovation: Physical experiments and a two-phase flow model investigate air migration during air-injection desaturation, considering the influence of different loads on saturation distribution and air entrapment capacity.
131. Dynamic centrifuge test on the reliquefaction characteristics of saturated sand deposits subjected to multiple earthquakes
Core Problem: Previously liquefied sand deposits are vulnerable to reliquefaction in subsequent earthquakes, and the evolution of liquefaction resistance during multiple events needs further investigation.
Key Innovation: A dynamic centrifuge test investigates reliquefaction characteristics, revealing differences in liquefaction resistance evolution at various depths and proposing a generalized correlation between liquefaction resistance and shear wave velocity.
132. Effects of soil heterogeneity on the numerical simulation of liquefiable soil deposits using a multi-surface plasticity model
Core Problem: Soil spatial variability impacts the response of liquefiable deposits under earthquake loading, but is often not accounted for in simulations.
Key Innovation: Stochastic numerical analyses using the Random Finite Element Method (RFEM) represent soil relative density as a spatially correlated Gaussian random field to predict soil response within a confidence interval.
133. Influence of water table rise on the seismic performance of subway stations in clay
Core Problem: Rising water tables alter the load state of underground structures in clay, impacting their seismic performance.
Key Innovation: A 3D numerical model investigates the effect of water table rise on the internal forces and seismic response of a subway station in clay, using a single bounding surface constitutive model.
134. Seismic response of pile group foundations in deep saturated sand sites under strong earthquakes
Core Problem: The seismic performance of high-rise buildings on liquefiable sites with pile group foundations needs further investigation.
Key Innovation: Centrifuge shaking table tests examine coupled interactions among ground acceleration, pore water pressure, and structural dynamic response in a scaled model of saturated sand, a pile-raft foundation, and a detailed superstructure.
135. Seismic behavior of caisson quay walls with fiber-reinforced calcareous sand backfill
Core Problem: Calcareous sand backfill in caisson quay walls is prone to liquefaction under earthquake action, causing instability.
Key Innovation: Cyclic simple shear tests and shaking table tests evaluate the seismic performance of calcareous sand caisson quay walls with fiber reinforcement, focusing on fiber content and shaking intensity.
136. Incorporating layer-specific spatial variability in CPT-based liquefaction assessment: A comparative study of random field and conventional probabilistic approaches
Core Problem: Assessment of soil liquefaction potential using Cone Penetration Test (CPT) data is an important aspect of modern geotechnical seismic analysis. Probabilistic models are increasingly adopted, as they better capture inherent soil uncertainties compared with deterministic approaches
Key Innovation: A novel CPT-based methodology that automatically detects soil layer boundaries and characteristics from raw CPT measurements, assigning each layer stochastic variability parameters according to its thickness and heterogeneity.
137. Enhanced CRR prediction for liquefaction analysis using advanced machine learning techniques
Core Problem: Accurate prediction of the cyclic resistance ratio (CRR) in granular soils is crucial for assessing the liquefaction potential of soils subjected to seismic loading.
Key Innovation: An interpretable hybrid machine learning framework is developed for direct CRR prediction from a large and diverse cyclic triaxial (CTX) dataset, integrating ensemble learning, model-agnostic explainability, and symbolic regression.
138. Prediction for freeze–thaw cycles induced degradation of dynamic shear modulus in Qinghai loess: structural damage concept-based model-Case study
Core Problem: Freeze-thaw cycles and earthquakes cause foundation deformation and instability in loess regions, leading to catastrophic failures like landslides and settlement. Freeze-thaw cycles degrade loess stiffness by causing microstructural damage.
Key Innovation: A dynamic shear modulus prediction model is developed, effectively accounting for structural disturbance by coupling microstructural damage metrics with dynamic stiffness, offering a novel tool for slope-stability and subgrade-settlement assessments in cold region loess.
139. DEM investigation on granular soil arching with emphasis on particle size distribution effect
Core Problem: Soil arching, a common load transfer mechanism in geotechnical engineering, is significantly influenced by soil particle size distribution (PSD). Existing studies have not fully understood the PSD effect on the arching evolution and critical height.
Key Innovation: This study investigates the PSD effect on the evolution of soil arching using the discrete element method, revealing that coarser and better-graded granular soils promote a more rapid development of soil arching effect, thereby enhancing the initial load-transfer efficiency.
140. Investigation on the triaxial creep behavior of red-stratum mudstone soil-rock mixture with different rock contents
Core Problem: Red-stratum mudstone, widely used in high-fill subgrade projects in western China, exhibits low strength and susceptibility to fracturing, making creep deformation a prominent issue.
Key Innovation: Triaxial creep tests on red-stratum mudstone soil-rock mixture clarify the influence of rock content and stress state on creep behaviors, revealing that creep behaviors depend not only on particle breakage but also on the internal pore structure of the mixture.
141. Erosion mechanism of interlayer soils under different seepage directions: a CFD-DEM perspective
Core Problem: Research on the erosion-induced deformation of interlayer soils under seepage action remains at the macroscopic level, with insufficient understanding of the underlying microscopic mechanisms behind the observed macroscopic deformations.
Key Innovation: This study investigates the macroscopic deformation process and the evolution of the microscopic contact mechanics of interlayer soils during seepage, considering the effects of different seepage directions and hydraulic gradients using CFD-DEM.
142. A DEM creep contact model with damage evolution for frozen soil
Core Problem: Frozen soil creep is a key factor in the settlement of cold region subgrades. Clarifying its macro and micromechanical deformation and damage mechanisms is essential for mitigating subgrade distress.
Key Innovation: A new discrete element creep contact model that incorporates damage evolution is proposed to accurately simulate the non-attenuating creep behavior of frozen soil, capturing the third-stage creep behavior of frozen soil.
143. “Brazil nut effect”–based gas–vibration system for reinforcing SRM subgrades/foundations: evidence from In–situ tests
Core Problem: Soil–rock mixtures (SRM) are widely distributed globally, but their high heterogeneity, collapsibility, and dynamic degradation often make thick SRM deposits unsuitable for subgrades or foundations.
Key Innovation: A novel gas–vibration coupled reinforcement method for SRM, theoretically derived from the “Brazil nut effect” (BNE), is proposed. Through gas–vibration coupling, a quasi-fluidized environment was established to amplify the BNE, thereby enhancing particle stratification and migration.
144. Sustainable development of critical filling height for an embankment constructed on unsaturated soils
Core Problem: The soils beneath an embankment often remain unsaturated, and the strength of unsaturated soils is influenced by the intermediate principal stress.
Key Innovation: Analytical solutions of critical filling height for an embankment constructed on unsaturated soils are derived under uniform and linear suction profiles, incorporating comprehensive influences from the intermediate principal stress, unsaturated soil natures, an actual in-situ stress state, and the presence of stiff crust.
145. CFD-DEM investigation into multi-mode evolutionary mechanisms of underground seepage erosion
Core Problem: In the context of seepage erosion induced by damage to underground structures, the transformation from suffusion to leakage is a particularly typical and complex phenomenon that warrants increased attention.
Key Innovation: This study adopts the well-validated CFD-DEM method to simulate a series of erosion unit tests, revealing three common erosional modes: stable suffusion, catastrophic suffusion, and continuous leakage, and determines the governing equations to enable a closed-loop description of these erosional modes.
146. Liquefaction Characteristics and Damage Evolution of Rapid and Long-traveling Landslides
Core Problem: Understanding the causes of rapid and long-traveling landslides, which pose significant threats due to their high velocities and large scales.
Key Innovation: Investigation of the shear mechanical behavior and liquefaction mechanism of landslide soil through undrained ring shear tests, and development of a statistical damage constitutive model using damage theory and the Weibull distribution function to simulate landslide initiation, transportation, and accumulation.
147. Acousto-optic characteristics of dry and wet damaged rock under cyclic impact loads
Core Problem: Quantifying the impact of dynamic disturbances and water content on rock deformation behavior for understanding long-term response during excavation.
Key Innovation: Using acoustic emission and digital image correlation to quantify internal damage and surface crack propagation in sandstone samples under cyclic impact and wet-dry conditions.
148. Microseismic signal processing and rockburst disaster identification: A multi-task deep learning and machine learning approach
Core Problem: Rapid and reliable interpretation of microseismic signals for timely identification of rockbursts in underground engineering.
Key Innovation: A multi-task deep learning model for automated processing of microseismic signals, extracting key parameters for rockburst disaster recognition with high accuracy and efficiency.
149. Statistical method for quantifying the strain localization process in Beishan granite under multi-creep triaxial compression based on distributed optical fiber sensing
Core Problem: Investigating damage evolution and strain localization in granite under creep conditions to predict failure.
Key Innovation: Combining distributed optical fiber sensing and X-ray CT to quantify strain localization using Gini and skewness coefficients, providing early warning of failure.
150. Effect of initial fracture angle on the failure pattern and gas flow channel of sandstone under multistage loading
Core Problem: Understanding how initial fracture angles in overlying rock strata affect failure modes and gas flow channels after coal seam mining.
Key Innovation: Using multistage loading experiments, numerical simulations, and 3D reconstruction to analyze the fragmentation process of rocks with different initial fracture angles.
151. Early warning system for risk assessment in geotechnical engineering using Kolmogorov-Arnold networks
Core Problem: Developing a comprehensive and fair evaluation method for risk assessment in geotechnical engineering.
Key Innovation: Using Kolmogorov-Arnold networks (KAN) for geohazard assessment and comparing its performance with other classification models across different datasets.
152. Injection-induced slip of splitting granite fracture
Core Problem: Understanding the effect of fluid pressure on injection-induced slip behavior in fractured granite, relevant to reservoir stimulation.
Key Innovation: Laboratory friction experiments on split-cutting granite fracture to investigate the effect of fluid pressure on slip behavior, linking mechanical responses to microstructural evolution.
153. Mechanical behavior and energy dissipation characteristics of coal under coupled 3D static and graded cyclic impact loading
Core Problem: Simulating the stress environment of surrounding rock under impact ground pressure caused by cyclic disturbances in coal mines.
Key Innovation: Developing a coupled loading method combining 3D static loading with graded cyclic impacts to study the mechanical behavior and energy dissipation of coal.
154. Experimental and model investigation on fatigue properties of sandstone after fully coupled thermo-hydro-mechanical cycles
Core Problem: Understanding the impact of thermo-hydro-mechanical (THM) cycles on the fatigue properties of sandstone and its implications for long-term slope stability.
Key Innovation: A fractional order-based damage fatigue model is introduced to quantitatively describe the rock viscoelastic parameters after different initial damage treatments, offering potential applications for slope stability assessment.
155. Invited perspectives: Redefining disaster risk – the convergence of natural hazards and health crises
Core Problem: Disaster risk assessments often overlook the cascading effects of natural hazards on public health, leading to incomplete risk management strategies.
Key Innovation: Advocates for integrating health metrics into disaster risk frameworks and using compound risk models to better understand the interplay between natural disasters and disease outbreaks.
156. FLEMOflash – Flood Loss Estimation MOdels for companies and households affected by flash floods
Core Problem: Accurate estimation of flash flood losses for companies and households is crucial for effective disaster management and mitigation strategies.
Key Innovation: Development of probabilistic models (FLEMO flash) to estimate flash flood losses, identifying key drivers such as emergency measures for companies and knowledge about emergency action for households.
157. Enabling real-time high-resolution flood forecasting for the entire state of Berlin through multi-GPU accelerated physics-based modeling
Core Problem: Urban pluvial flooding is intensifying due to climate change and urbanization, necessitating faster and more accurate forecasting capabilities.
Key Innovation: Introduction of RIM2D, a multi-GPU accelerated 2D flood model, capable of simulating high-resolution flood events across Berlin, demonstrating the operational viability of GPU-based early warning systems for urban-scale flooding.
158. Evaluation of microphysics and boundary layer schemes for simulating extreme rainfall events over Saudi Arabia using WRF-ARW
Core Problem: Accurate simulation of extreme rainfall events is crucial for flood prediction and water resource management, especially in arid regions like Saudi Arabia.
Key Innovation: Systematic testing of 36 combinations of microphysics and boundary layer schemes in the WRF model to identify the optimal configuration for simulating extreme rainfall events in Saudi Arabia.
159. China's three major cereal crops exposed to compound drought and extreme rainfall events
Core Problem: Compound drought and extreme rainfall events pose a significant threat to food security in China, requiring a better understanding of their frequency and impact on crop areas.
Key Innovation: Analysis of the spatiotemporal patterns of compound drought and extreme rainfall events in China, linking event frequency with crop areas to assess exposure risk and guide disaster planning.
160. Low-frequency earthquakes track the motion of a captured slab fragment
Core Problem: Uncertain plate configuration at the Mendocino triple junction affects seismic hazard assessment.
Key Innovation: Analysis of low-frequency earthquakes reveals a captured slab fragment translating northward, challenging existing interpretations and highlighting a potential unaccounted earthquake hazard.
161. Complex mesoscale landscapes beneath Antarctica mapped from space
Core Problem: Limited knowledge of Antarctica's subglacial landscape due to spatial biases in geophysical surveys.
Key Innovation: A continental-scale elevation map of Antarctica’s subglacial topography is produced by applying the physics of ice flow to ice surface maps, enriching understanding of mesoscale subglacial landforms.
162. Remote Sensing, Vol. 18, Pages 283: An Automatic Identification Method for Large-Scale Landslide Hazard Potential Integrating InSAR and CRF-Faster RCNN: A Case Study of Ahai Reservoir Area in Jinsha River Basin
Core Problem: Manual delineation of landslide anomalies from InSAR data is labor-intensive and time-consuming.
Key Innovation: An automated approach for large-scale landslide identification by integrating InSAR technology with an improved Faster Regional Convolutional Neural Network (Faster R-CNN).
163. Remote Sensing, Vol. 18, Pages 286: Cross-Domain Landslide Mapping in Remote Sensing Images Based on Unsupervised Domain Adaptation Framework
Core Problem: Deep learning-based landslide mapping relies on extensive annotated labels and is sensitive to domain shifts, leading to poor generalization.
Key Innovation: An unsupervised domain adaptation framework (LandsDANet) for cross-domain landslide identification using adversarial learning, image style transformation, and contrastive loss.
164. Remote Sensing, Vol. 18, Pages 290: Revealing Spatiotemporal Characteristics of Global Seismic Thermal Anomalies: Framework Based on Annual Energy Balance and Geospatial Constraints
Core Problem: Limited understanding of the polarity evolution of thermal anomalies related to earthquakes.
Key Innovation: A dynamic spatiotemporal adaptive framework to quantify global thermal anomaly responses using annual energy balance and geospatial constraints.
165. Remote Sensing, Vol. 18, Pages 234: Spatiotemporal Prediction of Ground Surface Deformation Using TPE-Optimized Deep Learning
Core Problem: Surface deformation prediction from resource extraction is complex, with hyperparameter tuning and interpretability issues.
Key Innovation: Combines Tree-structured Parzen Estimator (TPE) Bayesian optimization with ensemble inference for improved deep learning-based surface deformation prediction, enhancing accuracy and interpretability.
166. Remote Sensing, Vol. 18, Pages 239: Coseismic Slip and Early Postseismic Deformation Characteristics of the 2025 Mw 7.0 Dingri Earthquake
Core Problem: Understanding coseismic slip and postseismic deformation after a major earthquake is crucial for assessing regional tectonic activity and potential hazards.
Key Innovation: Uses Lutan-1 and Sentinel-1 data with DInSAR to analyze coseismic and postseismic deformation, inverting coseismic slip distribution and time-series deformation characteristics.
167. Remote Sensing, Vol. 18, Pages 211: Deep Learning Applied to Spaceborne SAR Interferometry for Detecting Sinkhole-Induced Land Subsidence Along the Dead Sea
Core Problem: Automating the detection of land subsidence areas associated with sinkhole formation along the Dead Sea using InSAR data is time-consuming and prone to human error.
Key Innovation: A deep learning segmentation model is developed and applied to InSAR data to detect land subsidence areas, learning interferometric phase patterns to capture subtle ground deformations.