TerraMosaic Daily Digest: Feb 20, 2026
Daily Summary
This digest curates 81 papers at the interface of slope instability, seismic ground response, and hydro‑coastal hazards, with a clear move from static hazard layers toward workflows that are both fast and uncertainty-aware. The most compelling landslide studies push susceptibility assessment toward operational use: an ensemble deep-learning pipeline paired with a road vulnerability index for the Chittagong Hill Tracts, a probabilistic regional framework accelerated by geotechnical reliability updating, and a small-sample debris-flow workflow that uses SMOTE to stabilize Random Forest predictions.
Across themes, mechanistic evidence is being translated into decision metrics. Laboratory and shaking-table studies refine how fines, stress history, and soil type shape liquefaction resistance and site amplification, while validated finite-element simulations map soil–structure–tunnel interaction in liquefiable deposits into scalar and vector fragility curves. Monitoring is also broadening beyond conventional sensors: wide-swath SWOT altimetry and causal-attention models extract flood dynamics and water depth from remote and social imagery, pointing to near-real-time observations that can inform warning design and infrastructure planning.
Key Trends
- Uncertainty-aware susceptibility mapping is becoming the default: Regional workflows increasingly treat probability and model updating as first-class outputs—combining probabilistic formulations with reliability-based updating to cut compute costs while preserving accuracy, and pairing susceptibility with asset-specific vulnerability indices for road networks.
- Small-data strategies replace “bigger inventories” as the main bottleneck: Debris-flow and landslide studies lean on targeted augmentation and sampling design (e.g., SMOTE-assisted pipelines) to make learning stable when event inventories are sparse or biased.
- Liquefaction and site effects are re-framed through fabric and stress history: Critical-state and cyclic tests emphasize how fines content, depositional fabric, and prior stress paths control static liquefaction and ground-motion amplification—bridging laboratory mechanics to site-response practice.
- Monitoring is expanding to opportunistic, multi-source sensing: Beyond satellites and gauges, studies fuse wide-swath altimetry (SWOT), microseismic classification, and even social-media imagery to infer water depths and detect hazard-relevant signals with near-real-time potential.
- Mitigation work couples engineered and nature-based interventions with hydro-mechanics: From polymer grouting for tunnel heave and flexible scour protection to bioengineered rhizosphere characterization, intervention studies increasingly quantify performance under realistic hydraulic and coupled loading.
Selected Papers
This digest features 81 selected papers from 597 papers analyzed (out of 2302 raw papers scanned; 597 new papers after deduplication) across multiple journals. Each paper has been evaluated for its relevance to landslide and broader geohazard research and includes links to the original publications.
1. Ensemble deep learning framework for landslide susceptibility mapping and road vulnerability index development in the Chittagong Hill Tracts, Bangladesh
Core Problem: Landslides pose a persistent and significant threat to life, livelihoods, and critical infrastructure, particularly roads, in regions like the Chittagong Hill Tracts, necessitating robust susceptibility mapping and vulnerability assessment.
Key Innovation: Developed an ensemble deep learning framework for highly accurate landslide susceptibility mapping and a comprehensive road vulnerability index, providing critical tools for disaster risk reduction in landslide-prone areas.
2. A novel perspective on fast landslide susceptibility mapping at regional scale: Probabilistic modelling and reliability updating
Core Problem: The necessity for fast and accurate computation of spatially distributed probability of landslide at a regional scale for effective landslide susceptibility assessment (LSA) and geological disaster risk mitigation, as traditional methods are computationally intensive.
Key Innovation: Development of an efficient computational framework for regional LSA based on probabilistic physical modelling and a geotechnical reliability updating approach (RUA). RUA uses low-discrepancy sequences and a weight index to rapidly update landslide probabilities for all mapping units, achieving comparable accuracy to Monte Carlo simulation (MCS) and first-order reliability method (FORM) with significantly higher computational efficiency (e.g., 3 min vs 55 h).
3. Debris flow susceptibility assessment with a small sample based on SMOTE-RF coupling method: a case study in eastern Inner Mongolia Mountains
Core Problem: Conducting accurate debris flow susceptibility (DFS) assessments when only a limited number of debris flow hazard points (small sample database) are available, which often leads to insufficient accuracy, poor generalization, and slow convergence in traditional methods.
Key Innovation: Proposing a novel framework integrating small-sample oversampling techniques (SMOTE) with the Random Forests (RF) model for DFS evaluation. This 'debris flow sample collection–SMOTE sample extension–Random Forests susceptibility assessment' workflow effectively addresses the challenges of limited samples, demonstrating superior performance and suitability for geological hazard evaluation with small-sample databases.
4. Seismic site characterization and site response analysis of the Shillong city, Northeast India
Core Problem: Shillong City, located in a highly seismically active region, lacks detailed seismic site characterization and site response analysis, which are crucial for accurate site-specific seismic hazard assessment and earthquake-resilient design.
Key Innovation: Detailed seismic site characterization of Shillong City using MASW to derive shear wave velocity profiles and classify sites (mainly NEHRP Class C), followed by 1D nonlinear site response analysis using simulated ground motions for the 1897 Assam earthquake, providing crucial insights into spatial variability of ground motion amplification for earthquake-resilient design.
5. Critical state and static liquefaction resistance of sand with low-plastic fines via equivalent void ratio concept under varied stress histories
Core Problem: Limited research explores how fines content and varied stress histories impact static liquefaction at critical state in sands, particularly using the equivalent void ratio concept and considering fabric techniques.
Key Innovation: Undrained triaxial tests demonstrating the effects of fines content and stress history on critical state behavior and static liquefaction resistance in clean and silty sands, unifying steady-state/critical-state data into an equivalent steady-state line using the equivalent granular void ratio concept, advancing understanding of flow potential.
6. MARS modelling for spatial analysis of coastal erosion susceptibility
Core Problem: Assessing and mapping the susceptibility of the entire Tuscan coastline to coastal erosion using an interpretable and robust modeling approach.
Key Innovation: Developed a Multivariate Adaptive Regression Splines (MARS)-based framework for coastal erosion susceptibility assessment, demonstrating high predictive performance, enhanced transparency for interpreting controlling factors and critical ranges, and producing reliable susceptibility maps that identify erosion hotspots.
7. The relationship between social vulnerability and wildfire structure loss across California
Core Problem: Lack of understanding of how social vulnerability intersects with wildfire exposure and impacts, and the assumption that risk is associated with affluence rather than disadvantage, hinders equitable resilience.
Key Innovation: Analysis of six domain-specific social vulnerability indices across California using a housing cluster framework, revealing that social vulnerability's link to exposure differs by region and vulnerability dimension, and that predictors of structure loss vary by vulnerability level, informing targeted mitigation and recovery strategies.
8. Seismic failure mechanisms of soil–structure-tunnel systems in liquefiable urban deposits: vector and scalar fragility assessment
Core Problem: Understanding and quantifying the seismic response and failure mechanisms of soil–structure–tunnel interaction (SSTI) systems in liquefiable urban deposits, and improving the reliability of seismic vulnerability assessment.
Key Innovation: Developed a rigorously validated finite element model to investigate liquefaction-induced SSTI behavior, revealing transient soil-arching and differential settlements. Introduced scalar- and vector-valued fragility functions using incremental dynamic analysis (IDA), demonstrating that vector intensity measures (IMs) significantly reduce epistemic uncertainty and provide more reliable predictions of deformation demands.
9. A comparative shaking table model study on the ground motion amplification effects of calcareous sand and quartz sand
Core Problem: Calcareous sand, prevalent in reef islands, exhibits distinct mechanical behavior under seismic loading compared to terrestrial quartz sand, and its dynamic characteristics and ground motion amplification effects are not fully understood, which is critical for seismic design in these regions.
Key Innovation: A comparative shaking table model study on dry calcareous and quartz sand sites, using an integrated analysis framework, revealing that calcareous sand sites show stronger attenuation of seismic energy (lower PGA) but dominant amplification in medium-to high-frequency ranges with more significant higher vibration modes, highlighting the importance of these effects for seismic design of structures on reef islands.
10. Effects of plant age on anisotropic saturated hydraulic conductivity of a synthesized rhizosphere using a cubic triaxial permeameter
Core Problem: Conventional permeameters cannot reliably measure anisotropic saturated hydraulic conductivity in intact vegetated soils, making it difficult to understand how plant age and root biomass influence soil hydraulic properties crucial for bioengineered geostructures and slope stabilization.
Key Innovation: Development and utilization of a cubic triaxial permeameter to reliably measure anisotropic saturated hydraulic conductivity in vegetated soils, demonstrating that increasing plant age and root biomass significantly reduce horizontal and vertical saturated hydraulic conductivity, providing insights for vegetation-based slope stabilization.
11. Development and laboratory validation of a field-scale piezovane shear test device for use in intermediate soils
Core Problem: Current field vane shear tests (FVST) do not directly measure porewater pressure (PWP), making it unclear whether measured torque reflects undrained, drained, or partially drained conditions, especially in intermediate soils like mine tailings, hindering accurate shear strength assessment for stability.
Key Innovation: Development and laboratory validation of a new "piezovane" device that directly measures PWP at the vane blade edges during shear testing, providing crucial data to resolve drainage conditions and improve the accuracy of undrained shear strength measurements in intermediate soils.
12. Investigation on scour evolution of bridge groups under tidal currents
Core Problem: Local scour around closely spaced bridge groups under tidal currents poses a significant threat to structural safety, and understanding its evolution mechanism and influencing parameters is crucial for effective protection measures.
Key Innovation: Combines flume experiments and numerical simulations to investigate scour evolution around bridge groups under tidal currents, quantitatively elucidating the influence of hydraulic and sedimentary parameters, and revealing wake-shielding effects and exacerbation of scour in closely spaced groups.
13. Experimental evaluation of flexible protection for local scour mitigation around multi-bucket jacket foundations
Core Problem: The exposed bucket tops of multi-bucket jacket foundations (MBJF) for offshore wind turbines exacerbate local scour, posing a significant threat to the foundation's stability, necessitating effective mitigation strategies.
Key Innovation: A self-deployable flexible protection method (FPM) for local scour mitigation around MBJFs, experimentally demonstrated to significantly modify the surrounding flow field and scour morphology, reducing maximum scour depth by up to 74.4% and providing a prediction formula for scour depth under FPM protection, offering valuable theoretical and technical support for offshore wind turbine foundation stability.
14. Experimental and numerical investigations on the hydrodynamics of multi-chambered porous breakwaters with a backwall
Core Problem: Coastal and harbor protection requires effective breakwaters to attenuate incoming waves, with a need to optimize porous designs for enhanced energy dissipation and resilience.
Key Innovation: Experimental and numerical investigations demonstrating that a triple-chambered porous breakwater significantly reduces wave reflection (minimum KR = 0.153) due to enhanced turbulence and vertical flow redirection, offering superior coastal protection, especially with optimized porosity (ε=0.3).
15. Understanding two-dimensional floodplain dynamics using wide-swath altimetry from the surface water and ocean topography (SWOT) mission: a case study of the 2024 Ganga River floods in North Bihar, India
Core Problem: Accurate and comprehensive assessment of two-dimensional floodplain dynamics during flood events is challenging, limiting effective flood management and response.
Key Innovation: Evaluates the potential of the newly launched Surface Water and Ocean Topography (SWOT) mission's wide-swath altimetry for assessing and understanding flood dynamics, demonstrated through a case study of the 2024 Ganga River floods.
16. Causal attention-based water depth estimation for complex flooding scenes in social media images
Core Problem: Existing deep learning methods for water depth estimation from social media flood imagery struggle to precisely focus attention on key submerged areas in complex scenes, leading to inaccuracies in disaster assessment.
Key Innovation: Developed a causal attention-based deep learning method that improves the precision of water depth estimation by effectively focusing on relevant submerged regions within complex flooding scenes captured in social media images.
17. Depth-dependent variations in shear strength of undisturbed root-soil composites of arbors: Insights from laboratory experiments on Malus halliana Koehne
Core Problem: The exact contribution of arbor roots to the shear strength of root-soil composites at different soil depths is not fully understood, despite its critical role in slope stability and landslide prevention.
Key Innovation: Investigating depth-dependent variations in shear strength and its parameters for undisturbed root-soil composites of *Malus halliana Koehne* through large-box direct shear tests. The study quantifies that roots increase soil shear strength by 0.3–20.7% and cohesion by 4.2–32.7%, identifying an optimal shear resistance depth (0.3–0.4 m). This provides important theoretical insights into root-soil interaction for bio-engineering solutions in slope stabilization.
18. Influence of non-plastic fines on cyclic behavior of Bushehr calcareous sand with presence of initial static shear stress
Core Problem: The cyclic behavior and liquefaction resistance of calcareous sands, particularly in coastal regions vulnerable to seismic loading, are critically influenced by the presence of non-plastic fines and initial static shear stress, but these effects are not fully understood.
Key Innovation: Performed undrained cyclic triaxial tests to systematically analyze the complex influence of varying non-plastic silt content and initial static shear stress on the liquefaction resistance of calcareous sand, establishing a soil-specific correlation for residual excess pore pressure ratio.
19. A Three-Dimensional Distinct Element Study of Fault Reactivation Mechanisms for CO2 Storage Reservoirs Using Fully Coupled Flow-Deformation Processes
Core Problem: Understanding and mitigating the risks of fault reactivation and induced seismicity during CO₂ injection in faulted reservoirs, which conventional continuum models may not realistically capture.
Key Innovation: Utilizes a 3D Distinct Element Code (3DEC) with fully coupled flow-deformation analysis to explicitly represent faults, revealing localized slip and block rotation, and demonstrating the critical role of fault geometry, stress conditions, and injection strategies in controlling fault stability for safe CO₂ storage.
20. Hydro-Mechanical Numerical Analysis of Fault Reactivation Considering Fault-Zone Architecture in Clastic Reservoirs
Core Problem: Many numerical studies overlook the critical influence of internal fault-zone heterogeneity (core and damage zone) on hydro-mechanical behavior and fault reactivation risks during production/injection scenarios in reservoirs.
Key Innovation: Developed a two-dimensional coupled hydro-mechanical framework explicitly representing fault-core and damage-zone domains with viscoelastoplastic behavior, demonstrating how internal fault-zone heterogeneity controls deformation, stress evolution, and transmissivity, with implications for fault reactivation and reservoir integrity in various geotechnical applications.
21. Neck cutoff morphodynamics of compound meanders with sandy banks capped by root-reinforced layers in the Northeastern Tibetan Plateau
Core Problem: Understanding the neck cutoff dynamics of compound meanders with sandy banks, particularly how tortuous geometries and flood extremes trigger breaching, and developing a criterion for anticipating such events.
Key Innovation: Identified first-order control of tortuous geometries on bend migration and cutoff timing, demonstrated that neck pressure-gradient can distinguish flood-triggered from stable bends, and provided a physically based criterion for anticipating flood-induced breaching in high-altitude rivers.
22. Experimental study on treating highway tunnel floor heave using polymer grouting material under complex Geological conditions
Core Problem: Highway tunnel floor heave, caused by complex geological conditions including water-bearing bedrock, water-sensitive soft rock, SO42- corrosion, gypsumization, and deviatoric stress, severely compromises traffic safety and requires effective treatment.
Key Innovation: Investigated a novel polymeric grout via full-scale field trials, demonstrating its superior performance over conventional materials in improving tunnel floor compactness, reducing mean deflection (42 µm vs. 19 µm), achieving faster stabilization, and offering high construction efficiency and durability against coupled hydrological, chemical, and mechanical loads.
23. Single- and multi-channel microseismic signal classification using dynamic-kernel attention-enhanced CNN-CapsNet
Core Problem: Substantial noise interference and inefficiency in manual signal identification hinder high recognition accuracy and practical application of microseismic (MS) monitoring for rock engineering safety.
Key Innovation: Proposal of DKECAlexCapsNet, a dynamic-kernel attention-enhanced CNN-CapsNet fusion framework, for high-accuracy single- and multi-channel MS signal classification, outperforming baselines and demonstrating robustness under noise.
24. Microstructure effects on the uniaxial compression mechanical properties of granite: A study using a novel Stochastic Grain-Based Model (SGBM)
Core Problem: Limitations of conventional Grain-Based Models (GBM) in accurately characterizing complex granite microstructure and its influence on mechanical properties.
Key Innovation: Develops a novel Stochastic Grain-Based Model (SGBM) that accurately reproduces complex grain interactions (e.g., interlocking, enveloping) in granite, revealing how coupled effects of average grain size and heterogeneity govern mechanical behavior and crack propagation modes, providing insights for deep rock mass engineering stability assessment.
25. Study on the influence of the soil–structure interaction effect on the seismic response of hybrid tower under different site conditions
Core Problem: Existing studies on the seismic performance of hybrid wind turbine towers often neglect the soil-structure interaction (SSI) effect, which can significantly influence the dynamic response under seismic action, especially for tall structures on various soil conditions.
Key Innovation: Establishment of a refined finite element model of a hybrid tower-foundation-soil system with viscoelastic boundaries to investigate the influence of SSI on the dynamic response of hybrid towers under seismic action for different site conditions (stiff and soft soil), revealing significant amplification of stresses and displacements due to SSI.
26. Seismic response of piles: An improved pseudo-static approach
Core Problem: Classical pseudo-static methods for estimating peak internal forces in piles under earthquake loading are often too simplistic, inadequately accounting for inertia-transmission effects from the superstructure and the complex coupling mechanisms of the pile-soil-structure interaction (PSSI) system.
Key Innovation: Development of an Improved Pseudo-Static Method (IPSM) for seismic pile response, incorporating a calibrated inertia-transmission reduction factor and effective superstructure properties derived within a rocking PSSI framework, validated against 3D FEA and field measurements, making it applicable to various pile-support stiffness conditions and layered soils.
27. Semi‐Volatile Organic Partitioning Improves Simulation of Biomass Burning Aerosol Mixing State Evolution
Core Problem: Current climate models misrepresent the three-phase evolution of black carbon coating thickness during biomass burning plume aging due to oversimplified treatment of organic aerosol volatility, leading to inaccurate climate impact predictions.
Key Innovation: Demonstrated that incorporating semi-volatile organic partitioning through the MATRIX-VBS model fundamentally improves simulation accuracy of biomass burning aerosol mixing state evolution, successfully capturing the observed three-phase pattern and providing essential improvements for climate model predictions in wildfire-affected regions.
28. Climate feedback of forest fires amplified by atmospheric chemistry
Core Problem: The full climate feedback of intensified forest wildfires is underestimated because the role of atmospheric chemistry in reducing oxidation capacity and amplifying methane accumulation is not fully accounted for.
Key Innovation: Integrated fire–atmospheric chemistry modelling revealed that wildfire emissions reduce global atmospheric oxidation capacity, amplifying methane accumulation and leading to a significant positive climate feedback that rivals other major warming factors.
29. Construction Disturbance Effects of Tunnelling in Completely Decomposed Granite and Reinforcement Using Interlocked Pipe Roofs
Core Problem: Tunnel excavation in completely decomposed granite causes significant disturbance to adjacent roadbeds and pile foundations, compromising structural stability and posing construction risks.
Key Innovation: Investigated stress-strain evolution of decomposed granite and numerically analyzed tunnel excavation's influence on adjacent structures, demonstrating that super-long interlocked pipe roofs effectively mitigate disturbances, preventing adjacent influence on roadbeds and pile foundations through arching and interlocking effects.
30. Experimental Investigation of Mechanical Anisotropy in Sandstone Containing Eliptical Cavities Under Combined Compression and Shear
Core Problem: Elliptical voids in sandstone induce pronounced anisotropy, predisposing the rock to shear or tensile failure and potentially triggering engineering hazards such as tunnel instability, requiring a better understanding of its mechanical behavior.
Key Innovation: Systematically investigated the effects of void inclination angle, aspect ratio, and normal stress on sandstone's mechanical properties and failure behavior under combined compression-shear loading, providing a theoretical foundation for stability assessment and hazard prevention in tunnel engineering through experimental and numerical analysis.
31. Experimental Study on Cyclic Shear Behavior of Saw-Tooth Filled Structural Planes
Core Problem: Accurately assessing the long-term stability in rock mass engineering requires a better understanding of the mechanical response and strength degradation mechanisms of saw-tooth filled structural planes under cyclic shear loading, considering the coupling effects of normal stress and filling degree.
Key Innovation: Conducted cyclic shear tests to elucidate the shear behavior and failure modes of saw-tooth filled structural planes, establishing a logarithmic strength degradation accumulation model that effectively quantifies the multi-factor coupling effect of normal stress and filling degree, providing theoretical support for long-term stability evaluation of rock mass engineering.
32. Assessment of thermal damage in intact rocks via linear and diffused ultrasonics coupled with full-field strains analyses
Core Problem: Limitations in quantitatively characterizing temperature-induced damage in rocks hinder objective assessment of thermal damage effects on ultrasonic signatures, and the applicability of diffused ultrasonic techniques (DUT) for thermally damaged rocks remains unexplored.
Key Innovation: A framework combining linear and diffused ultrasonic techniques with 2D digital image correlation to investigate ultrasonic signatures of thermal damage in rocks, providing real-time quantification of damage magnitude, intensity, and size, and revealing how crack characteristics influence ultrasonic wave behavior.
33. Experimental testing, characterisation methods, and mechanical behaviour of blast-induced rock damage from uniaxial strain to multiaxial stress states
Core Problem: A comprehensive investigation of blast-induced rock fragmentation and fracturing from near-field to far-field, considering in-situ stress conditions, is highly required for resource, energy, and space exploitation.
Key Innovation: Proposal of a workflow for blast-induced damage analysis from site-scale to lab-scale, systematically analyzing blasting and mechanical parameters, and updating advancements in dynamic loading and measurement techniques to improve understanding of rock behavior under extreme and multiaxial stress conditions.
34. Advances in understanding and modelling compacted expansive clays: a constitutive approach
Core Problem: Predicting the behavior of expansive clays under varying hydro-mechanical conditions is difficult due to strong coupling between hydraulic and mechanical processes, structural evolution, and water retention hysteresis, which are often overlooked in existing models.
Key Innovation: Proposal of ACMEG-Ex-S, an extended constitutive model for expansive clays, which independently describes capillary and adsorptive retention, incorporates hysteresis, and captures structural evolution within a single-structure mechanical framework, validated against experimental data.
35. Remote sensing-based monitoring of spatiotemporal waterlogging variations in groundwater-sensitive agroecosystems
Core Problem: Elevated groundwater-induced crop waterlogging stress (EGCWS) poses a significant threat to groundwater-sensitive agroecosystems, but its concealed nature makes accurate spatiotemporal monitoring challenging.
Key Innovation: Developed an innovative remote sensing-based method for effectively monitoring the spatiotemporal variations of waterlogging in groundwater-sensitive agroecosystems, addressing the challenge of its concealed nature.
36. Tail dependence of surge height and wind speed along the Dutch coast for storm clusters from large simulated datasets
Core Problem: Accurately representing the statistical dependence between extreme wind speeds and surge heights for flood risk assessment, particularly the upper tail dependence, as neglecting it can lead to underestimation of flood risk in probabilistic analyses.
Key Innovation: Exploring the statistical dependence between wind speed and surge height using a large synthetic dataset and storm clustering. The study found that while the BB8 copula generally provided the best fit, observed upper tail dependence did not decrease to zero for certain wind directions, indicating non-negligible dependence in joint extremes. This insight refines hydraulic load models for reliability assessments and design of flood defenses.
37. Reliable building inventory imputation for regional-scale risk assessment: An uncertainty-guided framework using spatially-enhanced Transformers
Core Problem: Pervasive missing data in building inventories undermines exposure models and downstream regional-scale risk analysis, and existing imputation methods lack explicit modeling of spatial dependencies and quantification of epistemic uncertainty.
Key Innovation: A novel uncertainty-guided framework leveraging a spatially-aware Feature Tokenization Transformer to impute missing building attributes, quantify imputation uncertainty using Monte Carlo Dropout, and iteratively refine estimates, significantly reducing imputation error and enhancing reliability for regional urban imputation tasks.
38. To be or not to be cooperative: Exploring the post-disaster recovery game among interdependent infrastructure networks
Core Problem: Existing analytical models for post-disaster recovery of interdependent infrastructures fail to capture real-world negotiation-based inter-sectoral cooperative behaviors, limiting the evaluation of benefits from collaborative recovery strategies.
Key Innovation: A government-guided cooperative framework integrating the Nash Bargaining (NB) model and a financial compensation mechanism into the recovery decision problem, solved using an outer approximation-based method, demonstrating that cooperative behavior significantly improves overall system recovery efficiency and promotes equitable burden distribution.
39. Measurement of 3-dimensional anisotropy of P-wave velocity and elastic modulus using cylindrical cores
Core Problem: Conventional methods for assessing elastic anisotropy in rocks are complex, time-consuming, and may underestimate anisotropic behavior, despite its critical influence on engineering applications and safety.
Key Innovation: Introduction of a technique for 3D characterization of elastic anisotropy (P-wave velocity and elastic modulus) from a single core/disc pair, enabling more accurate quantification of anisotropy and providing an equation to define these properties in any direction.
40. Interface behavior of concrete cut-off walls: a thermodynamics-based generalized elasto-plastic-damage model
Core Problem: Developing a unified model for stiff and flexible interface behaviors in concrete cut-off walls that accounts for state-dependent couplings among elasticity, plasticity, and damage, as well as interface roughness and contact fabric evolution.
Key Innovation: Proposes a thermodynamics-based generalized elasto-plastic-damage model incorporating interface contact fabric and its evolution, which successfully captures diverse shear hardening/softening and dilation/contraction behaviors in both stiff and flexible concrete cut-off wall interfaces, linking these to interface damage and contact fabric evolution.
41. Preferred Time Scales of Pacific Decadal Variability During the Last Millennium Related to Volcanic Activity
Core Problem: The preferred timescales of Pacific Decadal Variability (PDV) during the last millennium and the mechanisms behind them remain unresolved, with existing reconstructions showing limited consistency.
Key Innovation: Developed a new δ18O-based PDV index using a paleoclimate data assimilation framework and combined it with existing reconstructions and climate models to identify two dominant preferred timescales (multi-decadal and bi-decadal modes), establishing an important role for volcanic activity in shaping the temporal structure of PDV, with intensive volcanic periods favoring the bi-decadal mode.
42. PSO algorithm -improved particle breakage model for calcareous sands based on breakage energy
Core Problem: Existing breakage constitutive models for calcareous sand often rely on breakage indices from particle size distribution, which cannot capture the continuous evolution of particle breakage, hindering accurate prediction of soil behaviors crucial for marine geotechnical engineering.
Key Innovation: A novel PSO-improved breakage model for calcareous sand (PSO-MMC-PBE) that characterizes particle breakage using breakage energy (E_B) derived from a stress-dilatancy breakage model, incorporates it into an extended modified Mohr-Coulomb model, and uses a PSO algorithm to calibrate the initial friction angle, effectively simulating stress-strain behavior and micro-breakage processes.
43. A navigation risk assessment model for small-sample multimodal data with an improved Bayesian network
Core Problem: Traditional Bayesian network models struggle with precise risk assessments in marine environmental risk scenarios due to insufficient data availability and linguistic uncertainties.
Key Innovation: A small-sample multimodal model with an improved Bayesian network (SSM-IBN) that uses a multidimensional Bootstrap expansion for small datasets and a hesitant fuzzy agglomerative hierarchical clustering algorithm for subjective uncertainties, significantly reducing RMSE (18.2%–58.1%) in navigational risk assessment.
44. Adaptive neural optimal path following control for rotor-assisted vehicles via an improved event-triggered mechanism
Core Problem: Developing an efficient, robust, and less computationally intensive path following control scheme for rotor-assisted vehicles under external disturbances, while mitigating frequent signal updates and actuator wear.
Key Innovation: Proposing a critic-only adaptive dynamic programming (ADP) framework with robust neural damping and a dynamic integral-event-triggered (DIET) mechanism that automatically adjusts triggering thresholds and eliminates invalid triggering, enhancing robustness and reducing algorithm complexity.
45. Adaptive dense sampling planning and fixed-time trajectory tracking control for thermocline observation with an underwater glider
Core Problem: Enabling underwater gliders (UGs) to autonomously and efficiently conduct dense thermocline observations, requiring adaptive sampling planning and accurate trajectory tracking under constraints and disturbances.
Key Innovation: Developing a practical and easily deployable planning and control framework that includes an adaptive dense sampling strategy (dynamically adjusting gliding angle based on temperature gradient) and a fixed-time non-singular terminal sliding mode control scheme for robust trajectory execution, enhancing spatial sampling density and tracking accuracy.
46. Multi-Agent obstacle avoidance path planning for the arctic northeast passage
Core Problem: Improving trajectory smoothness and collision-avoidance responsiveness in multi-ship path planning within the dynamic and challenging Arctic sea-ice environment, especially with dynamic obstacles.
Key Innovation: Proposing a Bayesian-PPO algorithm that integrates Bayesian risk inference with control based on continuous actions, incorporating prospective risk prediction and multiple auxiliary rewards to enhance real-time avoidance of dynamic ice floes and static obstacles, leading to higher success rates and smoother trajectories.
47. Research on container ship trajectory prediction based on the EMD-PSO-GRU-RBFNN algorithm
Core Problem: Accurately predicting ship trajectories, which are characterized by strong time-varying, non-stationary, and highly nonlinear dynamics.
Key Innovation: Proposing a hybrid prediction framework (EMD-PSO-GRU-RBFNN) that first uses PSO-RBFNN for preliminary predictions, then applies EMD to decompose residual sequences into components for independent modeling by optimized GRU/RBFNN, and finally aggregates predictions, demonstrating strong stability and superior accuracy.
48. Underwater acoustic target classification using Fuzzy Inference System with Ensemble Learning
Core Problem: Accurately classifying underwater acoustic targets in complex marine environments, where signals are highly nonlinear, and existing methods lack sufficient accuracy and interpretability, especially with small sample datasets.
Key Innovation: Proposing the Underwater Acoustic Fuzzy Inference System with ensemble learning (UAFISEL), which uses a zero-order fuzzy inference system (ZOFIS) as a base learner to model nonlinear signals and boosts recognition through ensemble learning, offering strong generalization, effectiveness on small datasets, and enhanced interpretability.
49. Assessing Effects of Drought and Urbanization on Urban Trees Using Sentinel-2 Time Series
Core Problem: Identifying the effects of stress factors (drought, urbanization) on urban trees at a fine spatial resolution is challenging with coarse remote sensing data, making it difficult to differentiate species resilience and support urban tree management.
Key Innovation: Analysis of Sentinel-2 time series at fine spatial resolution (10m, 20m) using vegetation indices (ARVI, OSAVI) and traits (LAI, LCC) to derive phenological, productivity, and disturbance metrics, revealing species-specific responses to drought and urbanization along an urban-rural gradient, thus supporting urban tree management and policy decisions.
50. Machine learning-based paleobathymetric reconstructions using archaeal lipid biomarkers
Core Problem: Existing methods for accurate reconstruction of paleo-ocean depths, crucial for understanding tectonic evolution and global climate change, face substantial limitations.
Key Innovation: Developed machine learning models (random forest) trained on archaeal lipid biomarker (GDGT) distributions to achieve high-performance paleobathymetric reconstructions (R2 = 0.85, RMSE = 646 m), successfully reproducing tectonically driven bathymetric evolution in a 6-million-year record.
51. Freezing and thawing characteristics of saline farmland soils: analysis of hydrothermal parameters and driving mechanism
Core Problem: Dynamic responses of key hydrothermal parameters and the mechanisms governing water-salt evolution in saline farmland soils under natural freeze-thaw conditions remain insufficiently validated, despite their impact on soil physicochemical properties.
Key Innovation: Investigating freeze-thaw characteristics of saline farmland soils through hourly in-situ monitoring and laboratory experiments. The study systematically quantifies the spatiotemporal evolution of coupled water-heat-salt dynamics and parameter responses, elucidating the mechanisms underlying moisture recovery and salt redistribution. It identifies air temperature and snow cover as key drivers of vertical differentiation in heat and mass transfer, providing a reference for characterizing water-salt processes in comparable soils, which can influence soil stability.
52. Research on the testing method of rock elastic modulus based on the intelligent anchor detection system
Core Problem: Traditional laboratory methods for measuring rock elastic modulus do not accurately reflect the mechanical properties of rock under real environmental conditions, leading to less reliable data for underground engineering.
Key Innovation: Developed a new, rapid in-situ testing method for rock elastic modulus using an intelligent anchor detection system, establishing and validating drilling prediction models (e t - E eq , e r - E eq and e - E eq models) for equivalent elastic modulus.
53. Influence of fuel depth and sidewall restriction on combustion behavior of heptane pool fires in a short tunnel with natural ventilation
Core Problem: Existing models for tunnel fires lack the capability to simultaneously account for both fuel depth and proximity to the sidewall, hindering accurate prediction of combustion behaviors and fire dynamics in confined tunnel environments.
Key Innovation: This study experimentally investigates the combined effects of near-wall distance, fuel depth, and fuel size on flame morphology and mass loss rate of pool fires in a scaled tunnel model. It establishes dimensionless morphology prediction models, enhancing understanding of fire dynamics and offering insights for detection/suppression systems and hazard zone assessment.
54. Quantitative prediction on flame geometry of moving train fires in railway tunnels
Core Problem: Existing specifications for railway tunnel fire safety do not fully address the burning characteristics and flame morphology of a train fire while in motion, which is crucial for designing safe zones and emergency rescue strategies.
Key Innovation: This study experimentally investigates the flame morphology (height, tilt angle, width, length) of moving train fires in a scaled railway tunnel model under varying train velocities and heat release rates. It proposes quantitative prediction models for flame geometry parameters, offering critical references for fire safety design and emergency operating strategies in extra-long railway tunnels.
55. Study on fire damage and spalling characteristics of highway tunnel linings based on a thermal-hydro-mechanical coupling model
Core Problem: Highway tunnel linings, due to elevated moisture levels, are highly susceptible to complex and severe fire-induced damage and spalling, but the coupled effects of temperature and moisture on this degradation are not fully quantified.
Key Innovation: This study develops and validates a thermo-hydro-mechanical (THM) multiphysics coupling model to accurately capture coupled thermo-moisture migration and progressive damage evolution in tunnel linings during fires. It identifies initial water content as a key parameter governing the damage mechanism (thermally driven vs. thermo-hydraulic coupled), providing a quantitative understanding of spalling characteristics.
56. Deep Learning Identifies the Climate Warming Signal in Global Ocean Chlorophyll From Satellite Records
Core Problem: Detecting the influence of climate warming on global marine phytoplankton abundance from satellite chlorophyll-a records, which are often obscured by short record lengths and strong natural variability.
Key Innovation: Developed a deep-learning model, trained with Earth System Model simulations, to successfully identify the climate-warming signal in satellite-derived global chlorophyll-a fields, confirming the ongoing effect of climate warming on marine phytoplankton.
57. The Role of Land‐Atmosphere Feedbacks in Midlatitude Wintertime Surface Temperature Variability
Core Problem: Accurately representing synoptic near-surface temperature variability, crucial for predicting weather extremes, where existing models primarily attribute variability to tropospheric advection, potentially overlooking land-atmosphere interactions.
Key Innovation: Introduced a new scaling framework for temperature variance that incorporates local land-atmosphere feedbacks, demonstrating that longwave radiative damping dominates temperature variance responses over meridional temperature gradients when both land and circulation are perturbed.
58. Biogeophysical Impact of Land‐Use Scenarios on Holocene Surface Temperatures
Core Problem: Reconstructions and simulations disagree on Holocene temperature trends, and existing population-based estimates of anthropogenic land-use impact on regional climate vary widely.
Key Innovation: Examined transient Holocene climate model simulations forced with both population-based and a new pollen-based disturbed-land reconstruction, showing broad similarities in direct biophysical temperature effects but an earlier onset of disturbance in the pollen-based product, highlighting the cooling effect of land-use and remaining uncertainties in climate models.
59. Oceanic δ13C Fingerprints Caused by Laurentice Ice Sheet Discharges: Model‐Data Comparison During Heinrich Event 4
Core Problem: Investigating the sensitivity of oceanic circulation and δ13C of dissolved inorganic carbon to ice discharge events from the Laurentide ice sheet (Heinrich events) and comparing model simulations with observed anomalies.
Key Innovation: Used an isotope-enabled coupled climate-ice sheet model to simulate ice discharges triggered by reduced basal friction or increased oceanic melt rates, finding that reduced basal friction provided the best agreement with observed δ13C anomalies, though freshwater release needed amplification, suggesting additional processes are required to fully explain observed oceanic δ13C variations.
60. Shifts in the Annual Cycle and Associated Impacts on Northern Hemisphere Summer Onset Under Global Warming
Core Problem: The relative roles of long-term mean warming and changes in the annual cycle (AC) in altering the timing of seasonal onset under global warming remain unclear.
Key Innovation: Quantified how long-term warming and AC variations contribute to changes in Northern Hemisphere summer onset using observations and CMIP6 simulations, showing that while warming advances onset, AC changes create regional contrasts and can offset warming-induced advances, with greenhouse gases and aerosols affecting AC phase differently.
61. Streamflow Dynamics Across an Urbanizing Karst Watershed in the Ridge and Valley Province
Core Problem: Understanding how human activities (urbanization) and geological processes (karst) influence streamflow dynamics is crucial for effective freshwater management, especially given the unique challenges karst watersheds pose for predicting urbanization impacts.
Key Innovation: Investigated 24 years of streamflow data across an urbanizing karst watershed, demonstrating that urbanization increased stream flashiness, steepened recession rates, and decreased baseflow proportion. Showed that streams with high development were most sensitive to further urbanization, while those with substantial groundwater input were more stable, highlighting the need to account for hydrological complexity and groundwater recharge.
62. Fusing SWAN priors with tucker decomposition: A training-free framework for sparse wave field reconstruction
Core Problem: Reconstructing complete multi-parameter wave fields from sparse observations is challenging, especially in dynamic marine environments, and existing methods may lack robustness or require extensive training data.
Key Innovation: A training-free, physics-constrained framework that fuses SWAN-informed historical priors with Tucker low-rank decomposition, sparsity-based inversion, and temporal fusion to jointly reconstruct significant wave height, period, and direction from sparse sensor data, demonstrating high accuracy and stability even under typhoon conditions.
63. Prediction of FOWT mooring tension during non-stationary typhoon transit: a Transformer based framework with Bayesian optimization
Core Problem: Conventional steady-state simulations inadequately represent the transient coupling effects and non-stationary, extreme nonlinearity of rapidly transiting typhoons on Floating Offshore Wind Turbines (FOWT), making accurate prediction of mooring tension challenging.
Key Innovation: Development of a 12-h non-stationary typhoon model to capture the continuous "approach-peak-decay" lifecycle and a deep learning framework combining Transformer architecture with Gaussian process-based Bayesian optimization (GP-BO) to accurately predict FOWT mooring tension during typhoon transit, demonstrating exceptional generalization capability.
64. Multi-period modeling of floating production storage and offloading dock mooring line tensions using TimesNet-based deep learning framework
Core Problem: Existing research on FPSO dock mooring dynamics in complex marine environments faces challenges in multi-periodicity modeling, neglect of transient impact loads, and limited data fidelity, hindering accurate prediction of mooring line tensions.
Key Innovation: Proposal of a TimesNet-based deep learning model for multi-period modeling of FPSO dock mooring line tensions, trained on high-fidelity data from a calibrated ANSYS-AQWA model, demonstrating superior performance in ultra-short-term prediction and providing support for adaptive control and preventive maintenance.
65. Mooring lines structural health monitoring based on floating wind turbine response using an integrated ESAX-ResNet-50 model
Core Problem: Direct and visual monitoring of submerged mooring lines for floating offshore wind turbines is difficult, making early damage detection challenging despite their crucial role in stability and safe operation.
Key Innovation: Development of an integrated Enhanced Symbolic Aggregation Approximation (ESAX)-ResNet-50 model that indirectly monitors mooring lines by processing bitmap images of floating wind turbine dynamic responses to classify health status, enabling robust early damage detection.
66. Vulnerability analysis of offshore wind turbines considering the correlation of wind and waves
Core Problem: Existing vulnerability analyses for offshore wind turbines (OWTs) often neglect the crucial correlation between wind and waves and lack a comprehensive, multi-stage performance level division standard, leading to potentially inaccurate structural performance assessments.
Key Innovation: A vulnerability analysis methodology for OWTs that incorporates wind-wave correlations using a validated coherence function model, generates realistic load time histories, and proposes a unified three-stage damage evaluation indicator, demonstrating the necessity of considering correlations for accurate performance-based design.
67. AdaNGA: Adaptive neuronal genetic algorithms for unmanned surface vehicle dynamic path planning in dynamic environment
Core Problem: Dynamic path planning for unmanned surface vehicles (USVs) in dynamic water environments remains a significant challenge due to constantly changing obstacles and conditions.
Key Innovation: Introduces AdaNGA (Adaptive Neuronal Genetic Algorithm) which integrates neural networks with an adaptive genetic algorithm framework and a dynamic fitness function to enable real-time adaptation and efficient path optimization for USVs, demonstrating superior convergence speed and reduced travel time in complex dynamic environments.
68. A data-driven approach for determining navigation waters influenced by bridges – A framework and a case study
Core Problem: Quantitatively determining the specific navigation waters influenced by bridges is essential for improving maritime safety management and risk prevention.
Key Innovation: Proposes a data-driven framework using Automatic Identification System (AIS) data, integrating Gaussian Process Regression (GPR) for centerline extraction and Kernel Density Estimation (KDE), acceleration, and course difference analysis to describe vessel behavior, thereby quantitatively identifying and determining bridge area waters based on observed behavioral changes.
69. A Sample Spatiotemporal Expansion Method for Mapping Autumn-Harvested Crops in Cloudy and Rainy Areas
Core Problem: Crop mapping is challenged by limited sample availability and the inability of single-year field survey samples to support multiyear crop classification, especially in cloudy and rainy regions.
Key Innovation: A spatiotemporal sample expansion method that uses harmonic analysis of time series and time-weighted dynamic time warping to evaluate similarity, followed by sample thinning and random forest classification, achieving high-precision multiyear crop mapping with limited single-year field survey samples and improving overall accuracy by 2% compared to historical methods.
70. Evaluating Long-Term Effectiveness of Managed Aquifer Recharge for Groundwater Recovery and Nitrate Mitigation in an Overexploited Aquifer System
Core Problem: Over-extraction of groundwater leads to declining water levels and significant groundwater depressions, with a lack of long-term, regional quantification of MAR effectiveness for groundwater recovery and nitrate mitigation, especially concerning geological heterogeneity.
Key Innovation: Developed a coupled flow and multi-component reactive transport model to evaluate long-term MAR impacts, showing basin-wide groundwater level rise and nitrate reduction primarily by physical dilution, with geological heterogeneity governing spatial variability of water quality evolution.
71. Evolving synchronization of the Gulf Stream and Kuroshio-Oyashio Extension in a changing climate
Core Problem: The need for further investigation into the seasonality, predictability, and potential future changes of the recently reported decadal synchronization between the Gulf Stream and Kuroshio currents, especially in the context of a changing climate.
Key Innovation: Revelation of distinctive covariance between North Pacific and North Atlantic Western Boundary Current and Extension (WBCE) sea surface temperatures during boreal summer, linked to preceding Arctic sea-ice variability. Model simulations suggest that cold-season Arctic ice loss enhances atmospheric waves, contributing to summertime WBCE temperature anomalies, and that this covariability has increased historically but may weaken with future Arctic sea ice reductions.
72. Assessing the health risks of rice cadmium content standards in China
Core Problem: The adequacy of current cadmium (Cd) limit standards in rice for protecting public health in China requires comprehensive evaluation, given increasing understanding of Cd exposure levels and toxicity mechanisms.
Key Innovation: Found that a significant portion of rice Cd content in Jiangsu Province surpassed health thresholds, posing potential health hazards and lifetime carcinogenic risks, indicating that current Cd limit standards are insufficient and require more stringent safety thresholds.
73. A fractal–based approach to stream gradient analysis for active tectonic zones: Insights from the Morvarid–Sahneh–Mianrahan fault segments of the Main Recent Fault
Core Problem: Developing a robust quantitative metric to detect active deformation and tectonic influence across complex landscapes using stream gradient analysis, overcoming limitations of conventional methods.
Key Innovation: Introduced a novel fractal-based parameter (FD) for stream gradient analysis, demonstrating its effectiveness in capturing subtle morphotectonic variations, reduced sensitivity to outliers, and enhanced capability to distinguish structural discontinuities compared to traditional SL-based models.
74. Domain knowledge-enhanced dual-stream graph joint learning network for aeroengine remaining useful life prediction
Core Problem: Existing deep learning models for aeroengine remaining useful life (RUL) prediction insufficiently and incorrectly utilize multi-level domain knowledge (e.g., component-level physical topology and parameter-level essential dependencies), limiting accuracy and stability.
Key Innovation: A novel domain knowledge-enhanced dual-stream graph joint learning network (DK-DGJLN) that decouples and models component-level static physical topology and parameter-level dynamic dependencies using two distinct graphs, and then uses a DGJLN with a self-attention-based feature alignment module to extract and aggregate rich spatial-temporal degradation features for accurate RUL prediction.
75. A lightweight transformer with time dimension enhancement and global attention for health state monitoring in track circuits
Core Problem: Existing deep learning models face significant limitations in effectively capturing degradation rules and multi-scale temporal dependencies from multimodal track circuit data, hindering accurate identification and prediction of health states in safety-critical railway systems.
Key Innovation: A lightweight Transformer network, improved by Bayesian Optimization, that integrates a Time Dimension Enhancement Module (TDEM) to extract latent temporal features and multi-scale temporal dependencies, and a Global Attention Module (GAM) to refine multi-head self-attention, achieving high accuracy for track circuit health state monitoring.
76. A novel majority-oriented over-sampling technique and its application in imbalanced anomaly detection of aero-engines
Core Problem: Enhancing feature learning and generalization performance for anomaly detection in highly imbalanced industrial data, specifically aero-engines, where minority-class anomalies are hard to augment.
Key Innovation: Development of DSAFAN (deep self-attention feature augmentation network) incorporating MOOST (majority-oriented oversampling technique) to generate out-of-distribution synthetic data for minority classes and leverage multi-head self-attention for complex temporal dependency capture, validated on aero-engine data.
77. A proactive approach to assessing the resilience and safety of the future air traffic control system combining functional resonance analysis method and Bayesian belief networks
Core Problem: Proactively assessing the impact of increasing automation on the safety and resilience of future air traffic control systems.
Key Innovation: Combining the qualitative Functional Resonance Analysis Method (FRAM) with quantitative Bayesian Belief Networks (BBN) to provide a detailed, probabilistic assessment of system resilience and safety, identifying critical activities for proactive intervention in automated ATC systems.
78. New Bayesian optimisation framework for robust multi–objective design: decoupling performance and uncertainty
Core Problem: Existing robust optimization methods often neglect explicit control of performance stability, particularly the variance of objectives, under input uncertainty in multi-objective engineering design.
Key Innovation: Introducing a Variance–constrained Robust Bayesian Optimisation framework that decouples performance and uncertainty, using a robust Gaussian process and a variance–penalised scalarisation method to flexibly balance mean performance and variability, enabling context-dependent design selection while preserving data-efficiency.
79. Real-time graph-based traffic complexity evaluation for intelligent maritime supervision using multi-source data fusion
Core Problem: Challenges in accurate waterway risk identification due to missing ship trajectory data, complex ship interactions, and regional traffic patterns from AIS data.
Key Innovation: Developing a real-time graph-based evaluation framework for traffic complexity in intelligent maritime supervision, utilizing multi-source data fusion (AIS, video), LSTM for trajectory prediction, GNN for data fusion and interaction modeling, and DBSCAN for regionalization, enabling risk profiling and identification of high-risk areas.
80. Heterogeneous graph neural networks for building attribute prediction from hierarchical urban features and cross-view imagery
Core Problem: Extracting holistic representations to infer building attributes from multi-modal data across multiple spatial scales and vertical building characteristics remains a significant challenge, leading to scarce building property data essential for urban applications.
Key Innovation: A novel framework using heterogeneous graph neural networks (GNNs) is presented. It constructs a heterogeneous graph incorporating multi-dimensional urban elements and integrates cross-view visual information (satellite and street-level imagery) with feature fusion to predict building attributes with high accuracy and robustness, advancing scalable approaches for filling data gaps.
81. Real-time interaction method of the numerical-physical fusion method(NPFM) for mechanical behavior of tunnel structures
Core Problem: Investigating structural responses of large tunnel structures under complex working conditions (e.g., deep-buried, internal water pressure) is challenging with full-scale tests, and existing numerical-physical fusion methods lack sufficient interaction accuracy and stability.
Key Innovation: Proposed a real-time interaction method for the numerical-physical fusion method (NPFM) for tunnel structures, incorporating continuous data exchange, parameter updating, boundary load control, and a condition fitting strategy. Validated its effectiveness with full-scale test data, achieving response errors within 5% and confirming adaptability under varying loads.