TerraMosaic Daily Digest: July 8, 2026
Daily Summary
The July 8 literature advances landslide prediction by making model structure more accountable to process. An intrinsically interpretable Superposable Neural Network separates individual controls from pairwise interactions in coseismic susceptibility, while a regional rainfall-hazard framework couples statistical probability with physically based slope stability. Complementary studies show that deformation and rainfall must be represented at compatible time scales, and that normalized soil moisture can sharpen rainfall thresholds along transport corridors. The common advance is explicit representation of interactions, antecedent state, and mechanical constraints inside the prediction.
The mechanics papers identify where that state dependence originates. Preferential infiltration links rainfall and coal-mining disturbance in a loess-red clay landslide; mixing at heterogeneous soil interfaces promotes slip-zone formation; and field seismic signals, UAV imagery, and discrete-element simulation resolve fragmentation and entrainment in a heavily jointed landslide. A parametric analysis of vertical seismic forces further shows when the common practice of neglecting vertical acceleration can bias pseudo-static slope-stability estimates. Across these cases, external forcing alone is insufficient; internal structure and loading history govern where and how failure develops.
The observational layer is becoming more systematic. Multi-resolution SAR is used to validate regional geohazard identification, visual foundation models support cross-region few-shot landslide segmentation, and two complementary resources provide annual southeastern Tibetan Plateau glacial-lake inventories and a global multi-sensor benchmark for glacial-lake mapping. InSAR-machine-learning deformation forecasts, distributed acoustic sensing, and online geotechnical digital twins extend this emphasis from retrospective mapping to continuous monitoring. The wider issue connects these methods to compound flood-landslide hazards, GLOF exposure, subsidence, earthquakes, permafrost degradation, tsunami loading, and nonstationary flood regimes.
Key Trends
Five methodological directions connect the issue: interpretable model structure, coupled trigger histories, cross-region foundation-model transfer, repeatable sensing benchmarks, and compound nonstationary hazards.
- Interpretability is moving into the model architecture: Additive neural components, explicit interaction terms, physics-statistics coupling, and uncertainty-aware updating make factor effects inspectable rather than relying only on post-hoc explanation.
- Trigger history and internal structure are treated jointly: Rainfall, deformation, mining disturbance, interface mixing, fragmentation, entrainment, and vertical seismic loading are evaluated as coupled controls on failure initiation and mobility.
- Foundation models are being tested for geographic transfer: Few-shot adaptation, frequency decoupling, iterative refinement, Mamba mixtures, and spatially aware low-rank adaptation target cross-region generalization under sparse labels and domain shift.
- Longitudinal inventories and benchmarks formalize evaluation: Annual glacial-lake mapping, multi-sensor benchmark data, SAR deformation monitoring, DAS observations, and online digital twins provide repeatable tests for detection and warning systems.
- Hazard assessment is increasingly compound and nonstationary: CMIP6 flood-landslide scenarios, changing Arctic and typhoon-driven flood regimes, subsidence, permafrost thaw, earthquake sequences, and tsunami loading require models that retain interactions across hazards and time scales.
Selected Papers
The selected papers span interpretable and physics-coupled landslide susceptibility, rainfall-soil-moisture thresholds, mining-rainfall loess failure, interfacial slip-zone formation, fragmentation and entrainment, cross-region foundation-model segmentation, SAR geohazard monitoring, glacial-lake inventories and benchmarks, InSAR deformation forecasting, distributed acoustic sensing, digital twins, compound flood-landslide hazards, permafrost degradation, subsidence, earthquakes, tsunami loading, flood nonstationarity, and transferable uncertainty-aware AI and remote-sensing methods. This issue contains 124 selected papers from 2529 papers analyzed.
1. Glacial-Lake-Bench: A Global Multi-Sensor Benchmark Dataset for Evaluating Deep Learning Models for Glacial Lake Mapping
Core Problem: Global glacial-lake mapping lacks a common multi-sensor benchmark that tests models under clouds, shadows, frozen surfaces, and small-lake conditions.
Key Innovation: The dataset provides 19,115 Sentinel-1, Sentinel-2, and DEM image-label pairs plus a 1,105-scene challenge subset and leave-one-region-out baselines for geographic robustness.
2. Interpretable susceptibility assessment of coseismic landslides using a Superposable Neural Network
Core Problem: Coseismic susceptibility models must represent nonlinear interactions among terrain, lithology, rainfall, and shaking without losing physical interpretability.
Key Innovation: The study uses additive independent subnetworks to quantify individual and pairwise effects; interaction features raise AUC from 0.86 to 0.92 while exposing controls such as slope multiplied by PGA.
3. Compound multi-hazard assessment under CMIP6 climate scenarios: tracking seasonal flood-landslide and drought-fire interactions across Nepal
Core Problem: Single-hazard maps do not resolve how flood-landslide and drought-fire interactions shift seasonally as climate forcing changes across Nepal.
Key Innovation: The framework combines explainable machine learning with a 13-model bias-corrected CMIP6 ensemble to track compound hazard zones from a historical baseline through three future emissions pathways.
4. Response of coseismic landslide susceptibility to the temporal scales and interactions of deformation and rainfall
Core Problem: Coseismic landslide susceptibility depends on rainfall and deformation, but their relevant temporal scales and interaction effects are rarely tested directly.
Key Innovation: Using 5,007 Luding-earthquake landslides, the study shows that monthly rainfall and deformation outperform annual summaries and identifies where their interaction amplifies susceptibility.
5. Preferential infiltration-controlled failure of a loess-red clay landslide under coupled coal mining and rainfall: Insights from the Yaolvgou landslide
Core Problem: Loess slopes disturbed by underground mining can fail through coupled fissuring, preferential infiltration, weak-layer degradation, and groundwater uplift.
Key Innovation: Field investigation, laboratory tests, InSAR, and hydro-mechanical UDEC modeling reconstruct a multistage failure in which mining fractures route rainfall to a red-clay basal shear zone.
6. Dynamic fragmentation and entrainment characteristics of a heavily jointed landslide: DEM simulation, field seismic signal, and UAV image analysis
Core Problem: Fragmentation and entrainment govern landslide mobility, but their timing and internal evolution are difficult to infer from deposits or simulation alone.
Key Innovation: A calibrated DEM is cross-checked against UAV-derived particle geometry and field seismic signals, locating the strongest fragmentation and two major entrainment episodes during early motion.
7. Integrating physical and statistical models for regional rainfall-induced landslide hazard assessment
Core Problem: Regional event-based landslide models trade the predictive flexibility of statistical methods against the mechanical representation of physically based slope analysis.
Key Innovation: Logistic regression couples statistical indicators with mechanical stability outputs, improving ROC-AUC to 0.761 and retaining similar discrimination under a second rainfall event.
8. Promotion of landslide evolution by mixing at interfaces of heterogeneous soil layers in sliding zone
Core Problem: Rainfall-induced sliding zones often localize at heterogeneous soil contacts, but the role of interfacial mixing in hydraulic and strength contrasts is poorly constrained.
Key Innovation: Mixture tests, in situ measurements, and numerical analysis show that tuff-clay mixing weakens rapidly at high water content and creates a water-retaining barrier that concentrates pore pressure.
9. SETP_GLI: An annual 10-30 m glacial lake inventory for the southeastern Tibetan Plateau from 1990 to 2025
Core Problem: GLOF assessment in the southeastern Tibetan Plateau lacks annual, long-duration records that consistently capture small glacial lakes.
Key Innovation: SETP_GLI releases 36 annual vector layers from 1990 to 2025, adds uncertainty attributes and micro-lake detection, and documents an acceleration in regional lake-area growth after 2016.
10. Dynamic assessment of structures and infrastructures beyond traditional civil engineering applications: insights from shaking-table testing and advanced numerical simulations
Core Problem: This study presents a comprehensive overview of the dynamic performance of critical infrastructure systems using shaking-table-based experimental and hybrid methodologies and highlights their essential role in reproducing realistic earthquake loading conditions and advancing dynamic assessment beyond conventional civil engineering applications.
Key Innovation: It integrates experimental, numerical, and hybrid simulation findings across diverse infrastructure domains, including distributed power networks, modular and smart buildings, landslide-prone slopes, railway systems, and wind turbines, identifying both system-specific vulnerabilities and effective mitigation strategies.
11. Quantifying the Role of Vertical Seismic Forces in Pseudo-Static Slope Stability: A Parametric Study with Practical Analytical Indicators
Core Problem: Vertical acceleration is commonly omitted from pseudo-static slope analysis even where near-field vertical motion may materially change the factor of safety.
Key Innovation: The study introduces three analytical indicators and combines limit-equilibrium, finite-element, and case-history tests to quantify asymmetric factor-of-safety reductions under vertical loading.
12. Geohazard Risk Identification and Validation in Hunan Province Using Synergistic Multi-Resolution SAR Monitoring
Core Problem: As a natural event that poses a serious threat to human life, property, and the natural ecology, the effective identification, assessment, and early prevention of geological hazards are crucial.
Key Innovation: This paper proposes a multi-resolution SAR collaborative monitoring method using SBAS-InSAR technology for wide-area screening, followed by a joint PS/DS-InSAR processing framework to identify weak deformation signals at small spatial scales.
13. Cross-Regional Few-Shot Landslide Segmentation With Visual Foundation Model via Frequency Decoupling and Iterative Refinement
Core Problem: Few-shot landslide segmentation must generalize across regions despite sparse labels and large spectral and spatial domain shifts.
Key Innovation: The method adapts a visual foundation model through frequency decoupling and iterative refinement to reduce cross-region overfitting while preserving landslide boundaries.
14. Usage of normalized soil moisture for improving the performance of rainfall thresholds along transportation corridors
Core Problem: Rainfall-only thresholds along transportation corridors produce false alarms because they omit antecedent moisture conditions.
Key Innovation: The study combines highway landslide records with NOAA rainfall and SMAP soil moisture, finding that normalized moisture can identify many threshold exceedances that occur under below-average wetness.
15. VFM-MoME: A Remote Sensing Landslide Image Segmentation Network Guided by a Visual Foundation Model and a Mixture of Mamba Experts
Core Problem: Remote-sensing landslide segmentation must preserve irregular boundaries and transfer across heterogeneous terrain with limited labeled data.
Key Innovation: The proposed architecture combines a visual foundation model with a mixture of Mamba experts to specialize spatial-context modeling for landslide masks.
16. Vertical Structure of Air Temperature Excesses of the 2025 Mw 8.8 Kamchatka Earthquake
Core Problem: The Kamchatka Peninsula is one of the most seismically active regions in the world and experienced the Mw 8.8 earthquake in 2025.
Key Innovation: To examine whether tectonic processes were associated with enhanced atmospheric warming while minimizing external interference, this study analyzed multi-level air temperature fields from the NCEP (National Centers for Environmental Prediction) reanalysis data.
17. Geodetic Constraints on Fault Kinematics and Dynamics of 2021 Maduo Earthquake: Implications for Fault Friction
Core Problem: Postseismicaftersliphasbeeninferredtooccuronfaultbarrierssurroundingcoseismic asperities,whichcanbeexplainedbyvelocity-strengthening frictioninbarriersandvelocity-weakening frictioninasperities.
Key Innovation: However,afterslipalsoexhibitsspatialoverlapwithcoseismicslipforsomeevents.
18. Active learning for global optimization of highly non-stationary geotechnical engineering systems using Bayesian compressive sensing
Core Problem: Global optimization constitutes a crucial challenge in design optimization in geotechnical engineering, which aims to maximize the performance objective function of a geotechnical engineering system, thereby achieving the optimal output.
Key Innovation: For complex geotechnical engineering systems with computationally time-consuming models and highly non-stationary responses, the direct application of stochastic optimization algorithms usually requires numerous evaluations of the original model, resulting in significant computational expense.
19. Establishing erosion parameter relationships between portable scour testing device and erosion function apparatus for fine-grained soils under varying saturation levels
Core Problem: Scour due to erosion is the leading cause of bridge failures.
Key Innovation: This paper presents 36 controlled lab tests using the PSTD and EFA to evaluate cohesive soils with plasticity indices ranging from 0.5 to 40% under two saturation levels, 0-50% and 50-100%, to evaluate the combined effects of plasticity and saturation on erosion behavior.
20. Salinity effect on the hydro-mechanical performance of MgO-DG activated GGBS stabilized saline soil with high water content
Core Problem: Chloride-bearing high water content soils generated from reclamation, dredging, and site formation commonly exhibit poor mechanical stability and high seepage susceptibility, posing challenges for saline ground improvement.
Key Innovation: This study investigates the hydro-mechanical behavior of chloride-bearing soils stabilized with ground granulated blast-furnace slag (GGBS) activated by MgO and desulfurized gypsum (DG).
21. A framework for human-centered seismic resilience enhancement of water distribution networks
Core Problem: Seismic resilience analysis of water distribution systems has primarily centered on engineering metrics that evaluate hydraulic performance, while paying insufficient attention to human suffering caused by service interruptions.
Key Innovation: This paper develops a novel human-centered framework for explicitly quantifying and minimizing population suffering in seismic resilience enhancement.
22. A fully online Creep Digital Twin for real-time deformation monitoring, prediction, and early warning in geotechnical engineering
Core Problem: Time-dependent (creep) deformation is a key factor governing the long-term response of geotechnical structures, yet its reliable monitoring and prediction remain challenging due to the need to process continuously evolving field data.
Key Innovation: This paper presents a fully online Creep Digital Twin (CDT) framework for real-time creep deformation monitoring, prediction, and early warning.
23. City scale prediction of ground collapse dynamics and risks using hybrid simulation and deep learning methods
Core Problem: With the rapid development of urbanization in China, large-scale underground engineering construction has triggered an increasing number of urban ground collapse accidents.
Key Innovation: This study proposed GC-Net, a prediction framework trained on CFD-DEM simulation datasets to rapidly predict and visualize the spatiotemporal evolution of collapse.
24. Experimental study on dynamic behaviour of non-persistent granite specimens with coplanar rock bridges
Core Problem: Understanding fracture patterns of fractured rocks after dynamic impact is important for controlling the stability of rockslides and for effective prevention and mitigation.
Key Innovation: To investigate the effects of connectivity rate and joint angle on non-persistent granite specimens with coplanar rock bridges during dynamic impact, we adopted the central fractured rectangular granite specimens with different connectivity rates and different joint angles and performed the radial dynamic impact experiments on the specimens at the same.
25. Experimental study on hydro-jet-assisted vibratory probe compaction for collapsible soil-rock mixtures foundations
Core Problem: Soil-rock mixtures (SRMs) in colluvial and slope deposits are highly heterogeneous and, in their prevalent unsaturated state, exhibit strong water sensitivity and wetting-induced collapsibility that drive uneven settlement and foundation risk.
Key Innovation: This study proposes a hydro-jet-assisted vibratory probe compaction method (HVPC) for collapsible and heterogeneous SRMs and evaluates it in situ at a colluvial SRM site using time-frequency analysis, PCA, and field tests (MASW, DPT, and inundation loading).
26. Geometric investigation of dual-row retaining walls under tsunami-like solitary waves via smoothed particle hydrodynamics
Core Problem: Tsunami-induced overflow poses a major threat to coastal levees, yet the hydraulic performance of levees reinforced with dual-row retaining walls (DRRWs) remains insufficiently understood.
Key Innovation: This study leverages smoothed particle hydrodynamics (SPH) to evaluate how the geometry of DRRWs (i.e., tied double-sheet pile walls confining a soil core) modify overtopping and loading processes under tsunami-induced overflow conditions.
27. Instability of surrounding rock with different discontinuous joint occurrences in deep hard rock tunnels
Core Problem: To investigate the instability mechanisms of surrounding rock with different discontinuous joint occurrences in deep hard rock tunnels, modified true triaxial tests with one free face were performed.
Key Innovation: The results indicate that discontinuous joints substantially degrade the mechanical properties of the specimens, with the greatest degradation occurring when the DJS is parallel to the σ 2 and the DJDA is 30°.
28. Integrating seismic hazard simulation and multi-objective optimization for risk-resilient shale gas supply systems
Core Problem: Enhancing the reliability of shale gas supply systems under natural and induced seismicity remains a central challenge for risk-resilient energy planning.
Key Innovation: This study develops a hierarchical, resilience-oriented framework that integrates seismic hazard simulation, infrastructure vulnerability assessment, and multi-objective optimization for shale gas supply chains.
29. Managing scientific uncertainty in operational earthquake forecasting: A Delphi-mix approach to expert consensus and risk governance in Italy
Core Problem: Operational Earthquake Forecasting (OEF) represents the state-of-the-art in time-dependent seismic hazard assessment.
Key Innovation: We propose an OEF Action Agenda to move from a subjective panelist-judgment model to a structured risk governance framework.
30. Mesoscopic damage to rock under prestatic and cyclic disturbances: A grain-based model
Core Problem: In engineering practice, rock often experiences complex stress states.
Key Innovation: In this study, an improved grain-based discrete element model (GBM) was used to simulate the failure process of rock under prestatic and cyclic disturbances.
31. Study on dynamic response and damage mechanism of tunnel across reverse fault based on different earthquake action modes
Core Problem: Near-fault ground motions feature significant spatial variability and permanent displacement.
Key Innovation: Taking the Xianglu Mountain Tunnel as the prototype, this study constructs a non-consistent ground motion action mode, five equivalent earthquake action modes and three linear superposition earthquake action modes to analyze the seismic response and damage mechanism of cross-fault tunnels due to reverse fault considering permanent displacement.
32. Study on the influence of different cold shock methods on the mechanical properties and rockburst proneness of high-temperature monzogranite
Core Problem: Rockburst is a destructive geohazard commonly encountered in deep underground engineering.
Key Innovation: To improve rockburst prevention under complex geological conditions, this study proposes a novel control method based on high temperature-cold shock (rapid cooling of high-temperature rock).
33. Aerial Scene Recognition From Remote Sensing Images Based on Hybrid Multitemporal Information Fused Transformer-FRACTAL Network
Core Problem: Aerial scene classification, which aims to automatically label an aerial image with a specific semantic category, is a fundamental problem for understanding high-resolution remote sensing imagery.
Key Innovation: This study proposes a hybrid multitemporal information fusion architecture for aerial scene recognition from remote sensing images.
34. Anti-freezing superabsorbent polymer for soil stabilization: development and performance evaluation
Core Problem: Subgrades in seasonally frozen regions, particularly those subjected to complex freeze-thaw cycles and severe thermal gradients, are highly vulnerable to frost heaving and thaw settlement.
Key Innovation: This study presents a fundamental paradigm shift in cold region geotechnical engineering, offering a sustainable and thermo-mechanically stable solution for mitigating frost-induced degradation in critical infrastructure.
35. Development and application of an in-situ NMR triaxial testing system for exploring the multiphase fluid-solid coupling mechanism in unsaturated soils
Core Problem: The mechanical behavior of unsaturated soils is governed by solid-liquid-gas multiphase coupling interactions.
Key Innovation: In this study, an in-situ triaxial testing system tailored for unsaturated soils within LF-NMR environments was developed.
36. Ductile fracture behavior of buried X65 pipelines with defective girth welds under strike-slip fault movement
Core Problem: Buried steel pipelines crossing active faults are susceptible to permanent ground deformation (PGD) induced by fault movement.
Key Innovation: The results show that strike-slip-fault-induced PGD causes pronounced stress and strain localization within the fault-affected zone, and the defective girth weld on the tensile side is the most critical location.
37. Effect of heterogeneity on shear band formation in Malan loess: Insights from X-ray tomography
Core Problem: The influence of natural microstructural heterogeneity on strain localization in intact loess has not previously been demonstrated in three dimensions during deformation.
Key Innovation: Here, we provide such direct experimental evidence.
38. Energy-informed assessment of tunnel rock-support interaction using a triaxial stress-strain constitutive model with post-peak softening
Core Problem: This study addresses the limitations of conventional stress-displacement approaches in evaluating tunnel support performance, particularly under post-peak rock mass behaviour.
Key Innovation: Existing methods often fail to capture the progressive energy redistribution associated with damage evolution and support interaction, leading to an incomplete understanding of stability mechanisms.
39. Enhancing seismic robustness under soil parameter uncertainties: A novel supervisory adaptive nonlinear control algorithm
Core Problem: This study, for the first time, examines the impact of soil parameter uncertainties on the robustness of a novel supervisory nonlinear adaptive control algorithm designed for torsionally irregular structures equipped with magnetorheological (MR) dampers.
Key Innovation: By integrating the nonlinear stiffness matrix of structures founded on flexible bases, the proposed controller effectively captures coupled soil-structure nonlinearities within the Linear Quadratic Regulator (LQR) framework, thereby enhancing stability under severe ground motions and geotechnical uncertainties.
40. Explicit numerical simulation of quasi-static excavation in layered rocks: An FDEM approach based on the dynamic relaxation method
Core Problem: Due to the structurally control effect by bedding planes, the deformation and failure of layered rocks during tunnel excavation involve complex mechanical processes, including fracturing, bending and block spalling.
Key Innovation: To accurately reveal the progressive failure mechanisms under excavation, this study developed a numerical model for simulating quasi-static excavation process through Python and Fortran, which is suitable for the explicit cohesive zone model and finite discrete element method (CZM-FDEM).
41. Floods on trial: Transdisciplinary human accountability for improving disaster governance
Core Problem: The general public's main understanding is that flood disasters are caused solely by extreme hydrometeorological conditions.
Key Innovation: This study examines the role of psycho-behavioral and managerial failures exacerbating flood severity through extensive analysis of nine major flood events that occurred in Italy between 1994 and 2013.
42. Focused recharge in engineered landscapes: Contrasting pore and fissure groundwater responses to check dams on the Loess Plateau
Core Problem: Check dams are widely used across the Loess Plateau to control erosion and retain water, yet their effects on groundwater recharge remain poorly quantified.
Key Innovation: Using stable isotopes, residence-time analysis, and mixing models, we investigated groundwater recharge pathways in a check-dam catchment.
43. Identification of the main ground motion parameters influencing the significant duration on rock sites in Türkiye
Core Problem: The significant duration ( D 5 , 95 ) is an important parameter in geotechnical earthquake engineering, directly influencing the number of cycles in effective stress analyses and playing a pivotal role in evaluating soil liquefaction potential.
Key Innovation: In dynamic site response analyses, ground motions recorded on rock sites are essential because they serve as the excitation input, minimizing site-specific effects such as amplification or valley influences.
44. Improved analytical framework for influence zone of deep excavation considering additional stress
Core Problem: Deep excavations in densely built urban areas pose significant risks to surrounding structures, making accurate delineation of excavation influence zones critical for construction safety management.
Key Innovation: This study proposes an analytical framework that incorporates the excavation-induced additional stress field into the conventional stability analysis and establishes excavation influence zones based on safety factor distribution.
45. Lifecycle-oriented seismic resilience assessment of shield tunnels considering RC segment degradation
Core Problem: This study proposes a lifecycle-oriented seismic resilience assessment framework for shield tunnels considering degradation of reinforced concrete (RC) segments.
Key Innovation: Existing resilience evaluations generally assume intact initial conditions, neglecting the progressive degradation accumulated during long-term service.
46. MvReNowcast: Disentangling radar product features for improved nowcasting of raincloud evolution
Core Problem: Accurate radar echo extrapolation is fundamental for real-time monitoring and nowcasting severe rainfall events.
Key Innovation: To this end, this study proposes a physics-informed deep learning framework that explicitly integrates prior hydrometeorological information into the radar echo extrapolation process.
47. Probabilistic usability-based fragility curves for Italian reinforced concrete buildings using census and post-earthquake data
Core Problem: This study develops empirical fragility functions for reinforced concrete (RC) residential buildings expressed in terms of post-earthquake usability.
Key Innovation: The analysis uses more than 10,300 inspected buildings from the 2009 L’Aquila (Italy) earthquake, integrated with census data to represent uninspected constructions.
48. Rayleigh damping prediction for nonlinear seismic site response analysis using physics-informed neural networks with numerical and centrifuge modeling
Core Problem: Accurate nonlinear seismic site response analysis requires site-specific damping.
Key Innovation: This study proposes a physics-informed neural network (PINN) framework that inversely identifies Rayleigh damping from paired bedrock-surface acceleration records.
49. Seismic monitoring and site responses of the 2025 Ms 6.8 Dingri earthquake using distributed acoustic sensing
Core Problem: On January 7, 2025, an M s 6.8 earthquake occurred in Dingri County, Xizang Autonomous Region, China, with the epicenter located at (28.50°N, 87.45°E), as reported by the China Earthquake Networks Center.
Key Innovation: The results indicate that the earthquake sequence had a negligible influence on the monitoring site.
50. Separating typhoon and non-typhoon rainfall reveals structural hydrological regime shifts in coastal-to-inland basins
Core Problem: Long-term hydrological change in typhoon-prone coastal-to-inland river basins presents growing challenges for flood risk assessment under a changing climate.
Key Innovation: Here, we develop a generalizable basin-scale diagnostic framework that separates typhoon and non-typhoon rainfall, embeds them in a Budyko-based runoff analysis, and quantifies spatially heterogeneous hydrological sensitivities applicable to coastal-to-inland systems worldwide.
51. Video-Based Rainfall Opportunistic Sensing in Hydrology: A Lightweight Machine Learning Approach
Core Problem: Video-basedrainfallmeasurementisafrontiertopicinopportunisticsensing;however,rapid, accurate,androbustidentificationofrainfall-relatedrain-streakfeaturesfromdynamicvideosremainsakey challenge,especiallyformonitoringdeviceswithlimitedcomputationalresources.
Key Innovation: Toaddressthetrade-off betweenmeasurementaccuracyandefficiency,weproposealightweightmachine-learning approachtermedthe CoupledImageEnhancementandGradualAggregationNetwork(CIEGA-NET).
52. Widest-Path Reachability Fields for Connectivity-Preserving Slender Structure Segmentation
Core Problem: Segmenting slender curvilinear structures such as retinal vessels, cracks, and roads demands topological correctness, as even a single-pixel discontinuity can fragment a continuous network and invalidate downstream analysis.
Key Innovation: To address it, we propose Widest-Path Reachability Fields (WPRF), a differentiable Max-Min reachability objective that redirects gradient flow to connectivity bottlenecks.
53. A regional ground-motion model for H/V response spectral ratios in Southwestern China using Chinese site classes
Core Problem: This study develops a regional prediction model for H/V response spectral ratios (H/V ratios) of acceleration response spectra, compatible with the Chinese code-based site classification used in engineering practice.
Key Innovation: The dataset comprises 6099 records from 130 events (M w 4.5-8.0), based on strong-motion records from Southwest China (2007-2020).
54. A thermodynamic model of multiphysics in unsaturated frozen soil from the granular-scale
Core Problem: A granular-scale thermodynamic model is developed to describe the coupled thermo-hydro-mechanical (THM) behavior of unsaturated frozen soils subjected to freeze-thaw processes.
Key Innovation: Within a nonequilibrium thermodynamic framework, the model formulates conservation equations and thermodynamic and granular entropy production equations while coupling the dynamics of water-ice and water-vapor phase transitions.
55. A throat-based dynamic pore-network model for vibration-disturbed multiphase seepage in quasi-rigid porous geomaterials
Core Problem: Dynamic disturbances such as vibration and seismic excitation can alter phase distribution, residual trapping, and effective seepage capacity in geomaterials.
Key Innovation: To address this issue, this study develops a throat-based dynamic pore-network model for vibration-disturbed multiphase seepage in geomaterials.
56. A weakly coupled FVM-DEM framework for moisture diffusion and cyclic damage evolution in argillaceous sandstone
Core Problem: Moisture ingress coupled with cyclic loading plays a critical role in the strength degradation and progressive failure of argillaceous sandstone.
Key Innovation: To address these limitations, this study develops a weakly coupled finite volume method-discrete element method (FVM-DEM) framework for simulating transient moisture diffusion and cyclic damage evolution in argillaceous sandstone.
57. An ER-based multi-source information fusion method for adverse geological prediction during TBM tunnelling in fault zones: A case study
Core Problem: Accurately identifying adverse geological bodies before tunnel excavation is crucial for ensuring construction safety in complex geological environments.
Key Innovation: This study proposes a multi-source information fusion framework based on Evidential Reasoning (ER) theory to achieve accurate prediction of adverse geological ranges during TBM tunnel excavation.
58. ASFR-Net: Adversarial Alignment and Spatio-Frequency Refinement Network for Heterogeneous Remote Sensing Image Change Detection
Core Problem: The core challenge of heterogeneous change detection in remote sensing imagery lies in effectively decoupling genuine land-cover changes from significant modal disparities caused by distinct imaging mechanisms.
Key Innovation: To address this, we propose a novel, end-to-end adversarial spatio-frequency refinement network (ASFR-Net).
59. Assessing the impact of personal beliefs on flood risk perception: A case study of Tainan City, Taiwan
Core Problem: As climate change intensifies flood hazards in vulnerable coastal regions, understanding the social drivers of risk perception and adaptation has become increasingly urgent.
Key Innovation: This study examines how religious and fatalistic orientations relate to climate knowledge, flood risk perception, preparedness, and adaptation among residents of Tainan City, Taiwan.
60. Assessing the impact of pulse-like motions and column irregularity on the seismic fragility, loss, and resilience of bridges
Core Problem: The seismic performance of bridges with regular columns (BRC) is well documented.
Key Innovation: This study develops a unified framework to quantify the influence of column irregularity and motion type on seismic demands, fragility, losses, recovery time, and resilience of bridges.
61. Biocrust functions in mitigating soil water and wind erosion after freezing-thawing
Core Problem: Freezing-thawing regulates soil structural stability in cold winters, and further governs soil responses to water and wind erosion during subsequent warm seasons.
Key Innovation: As multifunctional communities widespread in cold-winter drylands, biocrusts have been found to influence soil erosion.
62. Digital twin modelling of shield tunneling undercrossing existing tunnels: A case study
Core Problem: Shield tunneling undercrossing existing tunnels presents significant challenges for deformation control due to complex ground-structure interactions and geotechnical uncertainties.
Key Innovation: The results demonstrate good agreement between predicted and measured deformations, confirming the model’s accuracy.
63. Field Monitoring and Numerical Evaluation of Water-Thermal-Mechancal Stability of Foam Concrete Subgrades in Permafrost Regions
Core Problem: To address the problems of frost heave and thaw settlement of subgrades caused by seasonal changes in permafrost regions, this study proposes a novel thermal insulation subgrade structure based on foamed concrete (FC).
Key Innovation: A one-year field test was conducted based on the G216 Xinjiang-Xizang Highway project.
64. Impact of floods on surface water quality: a systematic review and comprehensive assessment
Core Problem: Floods, as extreme flow events, are among the costliest and devastating natural hazards.
Key Innovation: This study aims to evaluate and assess multiple studies conducted globally to determine the impact of floods on WQ.
65. Seismic performance analysis of a high arch dam under mainshock-aftershock sequences based on the SBFEM-FEM coupling method
Core Problem: This research explores the structural response of high arch dams during mainshock-aftershock sequences since current seismic performance studies mostly concentrate on mainshock and ignore aftershocks.
Key Innovation: First, a self-written MATLAB program is used to couple the scaled boundary finite element method (SBFEM) and finite element method (FEM), developing a numerical analysis model based on the SBFEM-FEM approach.
66. Self-Supervised Pretraining Improves Cross-Site and Cross-Scale Robustness of Point Cloud Leaf-Wood Segmentation
Core Problem: The accuracy of existing leaf-wood segmentation methods for tree point clouds varies across forest types and sites.
Key Innovation: In this study, we pretrained Point-M2AE, a widely used SSL architecture for point clouds, on ShapeNet-55 augmented with 2,400 individual tree point clouds.
67. Shifting flood regimes in Arctic basins under climate change
Core Problem: Global warming intensifies the hydrological cycle, increasing the frequency and magnitude of floods worldwide, a trend particularly pronounced in the Arctic, where enhanced flooding exacerbates erosion and impacts ecosystems.
Key Innovation: However, the key drivers of Arctic flood variability remain unclear.
68. The antecedents of risk perception in the context of natural hazards: A multi-level meta-analysis
Core Problem: Risk perception is one of the major determinants of risk coping and, thus, of risk adaptation to natural hazards, whether resulting from climate change (e.g., sea level rise) or not (e.g., earthquakes).
Key Innovation: To understand how to increase an individual's risk perception of natural hazards, it is then crucial to identify the antecedents of risk perception.
69. Vision Language Action (VLA) Models for Unmanned Aerial Robotics and Bimanual Manipulation: A Review
Core Problem: Vision Language Action (VLA) models unify visual perception, natural-language understanding, and action generation within a single foundation model, allowing a robot to follow instructions such as fold the towel or fly to the red building directly from camera images.
Key Innovation: We show that the coordination strategies, training recipes, and action representations developed for bimanual VLAs transfer to unmanned aerial systems and identify fourteen research directions across both domains.
70. A novel approach to interval assessment of seismic reliability for metro stations considering geotechnical data sparsity: an epistemic uncertainty quantification perspective
Core Problem: The inherent issue of small sample sizes in geotechnical data and the resulting epistemic uncertainty are objective realities in geotechnical engineering.
Key Innovation: To effectively assess the seismic response and reliability of metro stations under sparse geotechnical data, the adaptive sparse polynomial chaos expansion (ASPCE) is firstly utilized in a given probability space of basic variables to obtain the shared physical analysis information of seismic response.
71. A physics-informed data-driven approach for long-term settlement of embankment widening
Core Problem: The accurate prediction of long-term ground settlement induced by embankment widening is critical for construction safety.
Key Innovation: In this study, a physics-informed data-driven approach is developed to recover the long-term settlement characteristics of widened embankments from measurement data.
72. A Regional Seismic Risk Assessment Framework Accounting for Urban Informality and Data Limitations: A Case Study of the Canton of San José, Costa Rica
Core Problem: Urban seismic risk assessment in the Global South is often constrained by informal construction and limited access to detailed building databases.
Key Innovation: Using the canton of San José, Costa Rica, as a case study, this study aims to (1) propose a regional seismic risk assessment framework that accounts for building informality and data limitations, (2) evaluate neighborhood-level variations in seismic risk within the canton, and (3) assess the influence of informality on risk estimates across neighborhoods.
73. Are traditional hazard-based weather warnings anticipating local impact?
Core Problem: Hazard-based weather warnings alert decision makers, responders, and the population about meteorological risks, yet often overlook data on vulnerability and exposure, relying instead on fixed or climatology-based thresholds.
Key Innovation: This raises questions about their effectiveness at higher resolutions, as weather impacts are ultimately determined by local characteristics.
74. CarbonCLIP: Enhance Carbon Prediction from Satellite Imagery via Integrated Street-View Semantics and Temporal Context Training
Core Problem: Accurately estimating urban carbon emissions is critical for sustainable urban planning, yet many existing approaches remain difficult to apply consistently across cities due to data-source heterogeneity and the lack of fine-grained semantic-temporal context in remote sensing data.
Key Innovation: We propose CarbonCLIP, a task-oriented multimodal distillation framework that improves satellite-based carbon emission prediction by transferring contextual knowledge into a unified satellite representation through dual-branch contrastive learning.
75. Characteristics of onshore and offshore ground motions from the 2024 Noto Peninsula Mw7.5 earthquake
Core Problem: This study analyzes the differences between far-field offshore and onshore ground motions during the 2024 M w7.5 Noto Peninsula earthquake.
Key Innovation: Using residual analysis relative to established onshore ground motion prediction equations under bounded V S30 assumptions, we examine whether offshore records show systematic deviations from onshore reference trends.
76. Deep learning approaches for streamflow flash drought prediction across the contiguous United States
Core Problem: Reliable drought prediction is crucial for early warning and decision-making regarding drought-related disasters.
Key Innovation: This study evaluates the ability of different deep learning architectures, especially the Temporal Fusion Transformer (TFT), in comparison to the widespread baseline Long Short-Term Memory (LSTM) model and a two-source LSTM incorporating static catchment attributes, to predict SFDs based on streamflow percentiles.
77. Deep learning-enhanced sequential updating of hydraulic parameters and state variables in groundwater models
Core Problem: As a fundamental component of the hydrological cycle, groundwater systems play a critical role in sustaining ecosystems, securing water resources, and supporting socio-economic development.
Key Innovation: In this study, we present an alternative strategy by introducing a deep learning (DL)-based update strategy into the EnKF framework, resulting in a hybrid ensemble filtering approach that integrates DL-based nonlinear updates with Kalman-type linear updates, termed EF(DL-K).
78. Enhanced time-domain pseudo-dynamic framework for evaluation of earthquake induced active earth thrust
Core Problem: Reliable evaluation of seismically induced active thrusts on retaining structures requires analytical formulations capable of representing the multi-frequency characteristics of earthquake ground motions.
Key Innovation: Existing modified pseudo-dynamic methods incorporate wave amplification and phase alteration effects; however, their applicability remains confined to single-frequency excitation, limiting their ability to represent realistic earthquake loading.
79. EventVGGT: Exploring Cross-Modal Distillation for Consistent Event-based Depth Estimation
Core Problem: Event cameras offer superior sensitivity to high-speed motion and extreme lighting, making event-based monocular depth estimation a promising approach for robust 3D perception in challenging conditions.
Key Innovation: To address this, we introduce EventVGGT, a novel framework that explicitly models the event stream as a coherent video sequence.
80. LoCA: Spatially-Aware Low-Rank Convolutional Adaptation of Vision Foundation Models
Core Problem: Pre-trained Vision Foundation Models (VFMs) provide strong visual representations for diverse downstream tasks.
Key Innovation: In this paper, we propose Low-Rank Convolutional Adaptation (LoCA), a convolution-aware PEFT framework that addresses spatial-channel entanglement by decoupling channel and spatial adaptation.
81. Monitoring and Prediction of Ground Deformation Using InSAR and Machine Learning Approaches in Tianjin City, China
Core Problem: Urban ground deformation must be monitored and forecast from spatially uneven displacement observations before subsidence affects infrastructure.
Key Innovation: The study combines InSAR deformation time series with machine-learning prediction to map and anticipate ground movement across Tianjin.
82. Multi-resolution downscaling of inundation extents using deep learning
Core Problem: The application of deep learning models for downscaling flood inundation extents has been brought into focus in recent years.
Key Innovation: Inspired by the hybrid modelling approach, this study uses a hydrodynamic model to generate a time-series of low-resolution flood inundation extents and depths.
83. Physically Grounded Monocular Depth via Nanophotonic Wavefront Encoding
Core Problem: Depth foundation models (DFMs) offer strong learned priors for 3D perception from single RGB images but lack physical depth cues, leading to ambiguities in metric scale.
Key Innovation: We introduce metalenses, an emerging class of ultrathin planar optical elements, as a solution to physically encode missing metric depth cues via nanophotonics.
84. Seismic fragility analysis of railway embankment structures considering shallow bedrock ground conditions
Core Problem: Railway embankment structures are a key component of linear infrastructure and may affect the operation of the railway network during earthquakes due to structural damage and safety-related operational restrictions.
Key Innovation: In this study, the seismic fragility of railway embankment structures was evaluated under ground conditions characterized by relatively shallow bedrock, with Korea considered as a representative example of such conditions.
85. Temperature Dependence of Fault Frictional Healing in Quartz Gouges at Hydrothermal Conditions
Core Problem: Faulthealingallowstheaccumulation ofstressalongfaultsduringinterseismicperiods.
Key Innovation: Inthis criticalstage,temperature-dependent fluid-rockinteractionswithinfaultzonesbecomecrucialatseismogenic depthincontrollingfaultstrengthrecovery.
86. Trees as hidden archives of mining-induced subsidence: A regional reconstruction from the former Ostrava-Karviná coalfield (Czechia)
Core Problem: Underground coal mining induces long-lasting ground deformation that may persist for decades after extraction ceases, posing challenges for environmental monitoring and land management.
Key Innovation: This study evaluates tree-ring growth disturbances as biological archives of mining-induced ground instability and presents a regional reconstruction for the former Ostrava-Karviná coalfield (Czechia).
87. VFM-Loc: Training-Free Cross-View Geo-Localization via Aligning Discriminative Visual Hierarchies
Core Problem: Cross-View Geo-Localization (CVGL) in remote sensing aims to locate a drone-view query by matching it to geo-tagged satellite images.
Key Innovation: To overcome these limitations, we present VFM-Loc, a training-free CVGL framework that leverages the generalizable visual representations from vision foundational models (VFMs).
88. WildCity: A Real-World City-Scale Testbed for Rendering, Simulation, and Spatial Intelligence
Core Problem: Humans can navigate an unfamiliar city and gradually form a coherent spatial mental map spanning tens of square kilometers.
Key Innovation: Can AI build spatial representations at a comparable scale?
89. A case study of spatial and temporal microseismic evolution and surface subsidence during room and pillar trial tests
Core Problem: Deep ‘room and pillar’ (R&P) mining without subsequent depillaring offers a low-impact alternative for extracting coal beneath sensitive surface infrastructure, yet the geomechanical behaviour of such operations at depths approaching 900 m remains poorly understood because seismicity and subsidence research have evolved largely independently.
Key Innovation: Key scientific and engineering contributions are represented by coupled analysis of induced microseismicity and surface deformation during a deep R&P trial in the Czech part of the Upper Silesian Coal Basin, where high-grade coal was extracted from a shaft protection pillar at depths of 700 to 900 m.
90. A comprehensive review of advanced laser scanning techniques for deformation monitoring and defect detection in shield tunnels
Core Problem: Ensuring the structural safety of shield tunnels is critical to the reliability and efficiency of modern urban transportation systems.
Key Innovation: Laser scanning (LS) has emerged as a transformative technology for tunnel deformation monitoring and defect detection because it enables rapid, non-contact acquisition of high-resolution three-dimensional point cloud data.
91. Coupled wind-rain-subsidence effects on soil erosion dynamics in an arid mining area: Insights from CFD-DEM and physical simulations
Core Problem: In arid and wind-sand-prone coal mining areas, surface subsidence caused by mining activities, combined with disturbances from wind and rainfall, results in a soil erosion process.
Key Innovation: This study focuses on a typical soil erosion area in the Shen Dong mine region, where a comprehensive research platform combining CFD-DEM coupled numerical simulation with physical modeling is developed.
92. Data-Driven Forecasting of three-Component Seismograms Using Transformer Architectures
Core Problem: Forecasting seismic waveforms beyond observed data remains challenging due to the nonlinear, dispersive, and multi-scale nature of seismic wave propagation.
Key Innovation: In this work, we introduce SeismoGPT, a transformer-based autoregressive model designed to forecast three-component seismic waveforms directly in the time domain.
93. Development of PXB-BVC Framework for Multivariate Flood-Risk Assessment Under Climate Change
Core Problem: Flood-risk assessment under climate change must integrate multiple interacting drivers rather than treating hazard, exposure, and vulnerability as fixed inputs.
Key Innovation: The study develops the PXB-BVC framework for multivariate flood-risk assessment across climate-change conditions.
94. Displacement of surrounding rock and stress analysis of lining structure in tunnel through fracture zone
Core Problem: As urban underground space development extends into areas with complex geological conditions, surrounding rock instability and threats to structural safety induced by fault fracture zones have emerged as critical challenges in tunnel construction.
Key Innovation: Based on the engineering background of a tunnel section in Qingdao Metro Line 8, this study selected typical working conditions involving the continuous crossing of two fracture zones.
95. Fabric anisotropy in cyclic liquefaction: DEM-CFM analysis of pre- and post-liquefaction response
Core Problem: This study investigates the interplay between inherent and load-induced fabric anisotropy in the pre- and post-liquefaction behavior of granular soils.
Key Innovation: To this end, the Discrete Element Method (DEM) is combined with the Coupled Fluid Method (CFM) to simulate the undrained response of granular assemblies under cyclic true-triaxial loading conditions.
96. Field comparative study of probe configurations in the vibratory probe compaction method for liquefiable silty sand ground
Core Problem: Field comparative tests were conducted at a liquefiable silty sand site along the Nanjing-Yancheng Expressway to evaluate the influence of probe configuration on the performance of the vibratory probe compaction method.
Key Innovation: The results indicate that all four probe configurations improved the ground condition, but the treatment effect was strongly dependent on the initial state of the soil.
97. Frequency-dependent seismic load-sharing mechanism of anchored sheet-pile walls and frame-beams combined structures in high-steep slope with bedrock
Core Problem: Although anchored sheet-pile and frame-beam (APFB) combined systems are effective for slope stabilization, their internal seismic load-sharing mechanisms remain insufficiently quantified.
Key Innovation: The results reveal notable frequency-dependent deformation and stress redistribution characteristics.
98. Land use pattern impacts on urban flood resilience in Wuxi: a multi-scenario analysis integrating PLUS simulation and explainable machine learning
Core Problem: Rapid urbanization challenges urban flood resilience, yet the mechanistic links between land-use patterns and long-term resilience dynamics remain under-explored.
Key Innovation: This study proposes an integrated simulation-evaluation-interpretation framework to examine these impacts in Wuxi, China.
99. Large-Scale Interseismic Crustal Deformation, Fault Slip Rate, Coupling and Earthquake Potential in the Upper Yellow River Basin
Core Problem: Earthquake-potential assessment in the upper Yellow River Basin requires spatially continuous estimates of interseismic deformation, fault slip, and locking.
Key Innovation: The study uses remote-sensing geodesy to jointly characterize crustal deformation, fault slip rates, coupling, and earthquake potential across the basin.
100. Linking microstructural changes induced by freeze-thaw and their consequences to unsaturated functions and volume-change behavior of treated expansive subgrades
Core Problem: This study focuses on multi-scale microstructure changes induced by freeze-thaw (FT) cycles of expansive subgrades with different mineralogical compositions, and their consequences in unsaturated functions and volume-change behavior.
Key Innovation: Microstructural changes were qualitatively analyzed using Scanning Electron Microscopy (SEM) and quantitatively measured through Mercury Intrusion Porosimetry (MIP).
101. Mechanisms and control of ground deformation during caisson sinking in soft ground under circular diaphragm wall confinement
Core Problem: Accurate prediction and control of ground deformation during deep caisson sinking are essential for safeguarding adjacent infrastructure in soft ground.
Key Innovation: This study investigates a wall-caisson composite structure with radial confinement by a circular diaphragm wall (CDW), first applied in a large water conveyance shaft project in Nantong, China.
102. Numerical testing proxies of seismic codes relative to topographic amplification effects: the case study of the Siena urban area (Central Italy)
Core Problem: The possible effects of surface topography on the seismic ground motion are addressed in some seismic codes through the use of suitable proxies.
Key Innovation: Anyway, this study also puts in evidence that any evaluation of effectiveness of the topographic amplification factors provided by the norm relies on a clearer definition of its aims in terms of expected accuracy and application domain.
103. ShieldGPT: A LoRA-Tuned large language model and its application in mechanised shield tunneling
Core Problem: Large language models (LLMs) are revolutionizing scientific discovery and engineering practices.
Key Innovation: This study develops ShieldGPT, a domain-specific LLM for shield tunneling via fine-tuning the Qwen2.5-7B-Instruct model with LoRA (Low-Rank Adaptation) method.
104. Trans-Himalayan ground motion characteristics of Mw 7.1 Dingri earthquake: A case study using observations recorded in India and physics-based simulations
Core Problem: This study investigates the trans-Himalayan ground motion characteristics of the January 7, 2025, M w 7.1 Dingri (Tibet) earthquake as recorded by Indian seismic stations.
Key Innovation: The observed ground motion exhibits distinct behaviour in amplitude, frequency, duration, and energy-related parameters.
105. Bifidelity Parameter Estimation Using Conditional Diffusion Models
Core Problem: We present a bifidelity method for uncertainty quantification of parameter estimates in complex systems, leveraging generative models trained to sample the target conditional distribution.
Key Innovation: In the Bayesian inference setting, traditional parameter estimation methods rely on repeated simulations of potentially expensive forward models to determine the posterior distribution of the parameter values, which may result in computationally intractable workflows.
106. Connecting Microseismicity to Lithology via a Model of Slip Avalanches
Core Problem: Fluid injection into the earth's crust can induce small and frequent earthquakes in the subsurface.
Key Innovation: Here we show that a simple micromechanical model of slip avalanches in slowly deforming solids predicts the slip statistics observed over drastically different spatial scales, namely meter-scale microseismic observations and nanometer- to micrometer-scale nanoindentation experiments can be described with this model.
107. Copula-driven adaptive surrogate modeling for regional post-earthquake resilience of high-speed railway bridge-overhead catenary system
Core Problem: Regional high-speed railway networks are lifeline corridors after earthquakes.
Key Innovation: This study presents a rapid resilience assessment framework for coupled bridge and catenary systems.
108. DYNA-PRUNER: Input-Adaptive Data-Model Co-Pruning for Efficient and Scalable Spatio-Temporal Media Prediction
Core Problem: Spatio-temporal prediction supports radar/satellite nowcasting and city-scale traffic monitoring, but modern models are often too expensive for real-time deployment.
Key Innovation: To enable automated, resource-aware architecture optimization in scalable media analysis, we propose Dyna-Pruner, an end-to-end framework for input-dependent co-pruning of data and model structure.
109. Interpretable machine learning models for permafrost hydrothermal response to climatic variables
Core Problem: The Qinghai-Tibet Plateau (QTP), the world's largest high-altitude permafrost region, is experiencing intensified environmental changes due to climate change.
Key Innovation: Based on long-term meteorological observations from the Beiluhe weather station, this study developed and optimized several machine learning models to simulate soil temperature and volumetric water content in permafrost regions.
110. Organo-mineral interactions in active layer and permafrost soils along aging Arctic landscapes
Core Problem: Rising temperatures are accelerating permafrost thaw, exposing large soil organic carbon (SOC) stocks to microbial decomposition with implications for global climate.
Key Innovation: Understanding how permafrost carbon is stored and protected through associations with minerals is critical for predicting its vulnerability to decomposition upon thaw.
111. Physics-guided spatiotemporal neural models for fuel density prediction
Core Problem: This paper presents a physics-guided machine learning (PGML) framework for fuel density prediction, integrating physics constraints and domain knowledge into deep learning models to enhance model accuracy and stability.
Key Innovation: We explore three deep learning architectures -- ConvLSTM, Adaptive Fourier Neural Operator (AFNONet), and Video Vision Transformer (ViViT) -- to model the spatiotemporal evolution of fuel density.
112. Quantitative evaluation of ground disturbance induced by shield tunneling considering soil spatial variability: insights from field monitoring and random finite element analysis
Core Problem: Ground deformation induced by shield tunneling poses critical challenges to the safety and serviceability of urban underground infrastructures.
Key Innovation: This study presents a reliability-based framework for quantifying the influence of soil spatial variability on tunneling-induced ground disturbance by integrating random field theory, the random finite element method, and Monte Carlo simulations.
113. Representation of meteotsunamis in km-scale regional simulations coupled at high frequency
Core Problem: Meteorological tsunamis, or meteotsunamis, are anomalous waves triggered by atmospheric disturbances such as thunderstorms, gravity waves, squalls, or cyclones.
Key Innovation: This study evaluates the capability of the Met Office's atmosphere-ocean-wave regional coupled system (UKC4) and Météo-France’s atmosphere-ocean regional coupled system (AROBASE) to capture and predict meteotsunamis.
114. Synergistic SAR and Wide-Swath Interferometric Altimetry Observations for Estimating Flood Dynamics and Water Storage Variations in East Dongting Lake
Core Problem: Floodplain dynamics and water-storage changes are difficult to estimate when inundation extent and water-surface elevation are observed separately.
Key Innovation: The study combines SAR inundation mapping with wide-swath interferometric altimetry to estimate flood evolution and storage variation in East Dongting Lake.
115. Temporary Floodplain Ponds Shape Vegetation Mosaic in a Natural River Valley: Evidence from SAR and Optical Remote Sensing
Core Problem: Short-lived floodplain ponds are difficult to observe consistently, obscuring their influence on vegetation patterning in natural river valleys.
Key Innovation: The study integrates SAR and optical imagery to relate temporary pond dynamics to the spatial mosaic of floodplain vegetation.
116. Vertical uplift capacity of a horizontal strip anchor embedded in the slope face under seismic load
Core Problem: This study estimates the seismic vertical uplift capacity (Q uE ) of a horizontal strip anchor embedded in the face of a stable, cohesionless soil slope using lower bound, upper bound, and 15-node mixed Gauss elements in finite element limit analysis (FELA).
Key Innovation: The effects of soil friction angle (ϕ), slope inclination (β), embedment ratio (D/B), and horizontal and vertical seismic acceleration coefficients (α h and α v ) have been explicitly considered for Q uE estimation.
117. A novel model for predicting the stress-strain behavior of freeze-thaw-impacted soils
Core Problem: Repeated freeze-thaw cycles alter soil strength and deformation, but constitutive predictions must represent progressive degradation across cycles.
Key Innovation: The study introduces a hyperbolic decay model for freeze-thaw-induced strength deterioration and stress-strain response.
118. CoFINN: Conservation Flux Informed Neural Networks for Physics Problems Governed by Conservation Laws
Core Problem: We present CoFINN (Conservation Flux Informed Neural Networks), a physics-informed deep learning framework for predicting compressible flow fields governed by conservation laws.
Key Innovation: Unlike conventional data-driven convolutional neural networks (CNNs), which optimize only pixel-wise similarity metrics, CoFINN embeds finite-volume conservation physics directly into the training process.
119. Efficient Bayesian Deep Ensembles via Analytic Predictive Inference
Core Problem: We introduce an efficient Bayesian deep ensemble method for predictive regression designed to enhance interpretability while maintaining competitive predictive performance and computational efficiency.
Key Innovation: Our method combines the statistical rigor of Bayesian inference with the scalability of deep ensembles, providing calibrated uncertainty estimates that enable its use not only for standalone prediction but also as a component within broader learning systems.
120. Hydrodynamic characteristics of tsunami wave and its comparison with solitary wave over an uneven coral reef with lagoon
Core Problem: Real-world coral reefs feature complex and diverse topographies, yet most existing studies rely on oversimplified step-like models that ignore these structural realities.
Key Innovation: To address these questions, this study employs a combination of physical experiments and numerical simulations to systematically compare the hydrodynamic differences between tsunami waves and conventional solitary waves over an uneven coral reef with lagoon.
121. Institutional priority lock-in and the limits of disaster response financing: Evidence from South Africa's 2022 floods
Core Problem: Funding disaster response is generally considered crucial to facilitate timely relief and recovery, but in most developing countries, delays in spending and delivery remain common.
Key Innovation: This study explores the financing of disaster response in South Africa, drawing on the 2022 flood response in KwaZulu-Natal and the Eastern Cape provinces.
122. Mechanical Performance and Parameter Sensitivity of a Self-Balanced Composite Retaining Structure for High Fill Slopes
Core Problem: High fill slopes (H > 15.0 m) often require retaining systems that sustain large loads, yet remain difficult to deformationally control and economically prohibitive.
Key Innovation: This paper proposes a self-balanced composite retaining structure (SBCRS) integrating a gravity retaining wall, inclined piles, and vertical anchor cables.
123. Optimal Conformal Prediction under Epistemic Uncertainty
Core Problem: Conformal prediction (CP) is a widely used frequentist framework to quantify uncertainty by constructing prediction sets with user-specified marginal coverage guarantees.
Key Innovation: In this paper, we consider the question of how to optimally employ CP on top of a more expressive formalism, namely credal sets, which can express both aleatoric and epistemic uncertainty.
124. Propagation of fluid-driven aseismic slip fronts under upper crustal conditions
Core Problem: Fracture mechanics models suggest that aseismic slip may either lag behind or outpace the fluid pressure front, depending on injection conditions and the fault’s initial stress state.
Key Innovation: We show that aseismic slip lags the pressure front at low injection rates and low initial stress but rapidly outpaces it under high-rate injection or near-failure initial stress conditions.