TerraMosaic Daily Digest: April 24, 2026
Daily Summary
The April 24 set is anchored by process-resolving hazard science rather than by inventories alone. Post-seismic landslide reactivation is treated as an uncertainty-calibrated, physically guided multimodal problem; post-fire debris-flow mitigation is tested by carrying plot-scale runoff physics into catchment-scale initiation probabilities; and volcanic quiescence is reframed as a period during which trapped, water-rich magma can silently enlarge hazardous reservoirs. Across slopes, frozen ground, rockbursts, caldera collapse, and mixed-bed sediment transport, the stronger papers ask what physical variable changes first: basal or surface mobility, frozen-core geometry, pore pressure, structural-plane resistance, or reservoir pressure.
A second cluster is operational but not merely applied. Urban flooding papers interrogate whether infrastructure variables, social-media evidence, tide-gauge autocorrelation, and human forecasters change warning skill; seismic papers move from catalog description to anisotropy, creep, precursor statistics, site amplification, and cascading rupture; and engineering papers expose tunnel, dam, soil-pile, and cavern stability as boundary-condition problems. The result is a day in which AI is useful when constrained by physics, monitoring is useful when it isolates mechanism, and risk assessment is useful when it keeps the chain from initiating event to societal consequence intact.
Key Trends
The methodological center is constraint: physics constrains AI, field plots constrain mitigation models, tide gauges constrain surge nowcasts, and geophysical observables constrain hidden failure processes.
- Process models are being tied to actionable probabilities: PIMSEL, post-fire debris-flow treatment modeling, rainfall-infiltration slope stability, frozen-ground geophysics, and sediment-mobility simulations connect physical controls to risk tiers, exceedance probabilities, or identifiable thresholds.
- Operational hydrology is testing the limits of machine learning: Urban flood susceptibility, storm-surge nowcasting, flood severity from social media and remote sensing, satellite precipitation, thunderstorm prediction, and forecaster-in-the-loop benchmarking all ask which information actually improves warning skill.
- Hidden seismic and volcanic processes are becoming observable: Shear-wave anisotropy, magnetotelluric precursors, aseismic creep, random-fault cascades, rate-and-state friction, fluid-saturated granite failure, caldera collapse, and long-lived magma-reservoir growth all expose sources before or beneath the event surface.
- Engineering hazards are increasingly boundary-condition problems: Tunnel-face collapse, working-face rockburst, tunnel drainage, soft-rock squeezing, slate anisotropy, jointed cavern stability, soil-pile response, and bridge aftershock fragility are analyzed through stress, water, geometry, and material state.
- Risk work is preserving trigger-to-impact chains: NaTech reviews, coastal boulder archives, river-island flood settlement, flood relocation policy, hurricane resilience, wildfire spread, volcanic disturbance, and disaster social-media labeling keep the physical event connected to exposure and response.
Selected Papers
This digest features 66 selected papers from 1,401 papers analyzed. The sequence opens with physics-constrained landslide reactivation, post-fire debris-flow mitigation, and volcanic reservoir growth, then moves through frozen-ground slope stability, sediment mobility, urban flood warning, seismic-source diagnostics, tunnel and dam boundary conditions, and broader risk chains linking triggers to exposure and response.
1. PIMSEL: A Physically Guided Multi-Modal Semi-Supervised Learning Framework for Earthquake-Induced Landslide Reactivation Risk Assessment
Core Problem: Post-seismic landslide inventories are large but unevenly labeled, making long-lived reactivation risk difficult to triage after major earthquakes.
Key Innovation: PIMSEL combines physically guided multimodal fusion, semi-supervised graph clustering, prototype mutation, and calibrated uncertainty to classify reactivation risk across 7482 co-seismic landslides from three Sichuan earthquake sequences.
2. Shredding post-fire debris flow likelihood: A field-constrained catchment-scale model of wood-shred surface treatment effectiveness
Core Problem: Surface mulches are widely used after wildfire, but their effect on catchment-scale debris-flow initiation probability has remained difficult to quantify.
Key Innovation: Field runoff plots with controlled wood-shred cover are linked to a process-based debris-flow model, reducing maximum annual exceedance probability from 72% to 32% or 4% depending on treatment cover.
3. A volcano reawakens after more than 100,000 years of “silent” magma reservoir growth
Core Problem: Long volcanic quiescence can be mistaken for extinction even while magma reservoirs continue to grow below the surface.
Key Innovation: Geochemistry, isotopes, and more than 1250 zircon ages from Methana show how volatile-rich magmas can be trapped for over 100,000 years, building reservoirs that may transition toward caldera-forming hazard.
4. Integrated geomorphological–geophysical characterization of periglacial processes and frozen ground distribution in the easternmost edge of the Darwin range, Tierra del Fuego (54° S)
Core Problem: The southernmost permafrost edge in the Americas is poorly constrained, leaving uncertainty in how frozen cores and debris inheritance affect slope and sediment routing.
Key Innovation: Geomorphological mapping and ERT reveal compartmentalized frozen ground, rock glaciers, solifluction lobes, and ice-bearing deposits that can structurally control postglacial slope stability.
5. Hump‐Shaped Enhancement of Coarse‐Grain Transport in Sediment Mixtures Induced by Fine Grains of Different Sizes
Core Problem: Mixed gravel-sand transport controls channel response after extreme sediment pulses, yet fine-grain effects on coarse mobility remain poorly parameterized.
Key Innovation: Discrete-element simulations identify fine-grain burial depth as a unifying control on hump-shaped enhancement of coarse-grain transport.
6. Large Deformation, Pressure‐Driven Mechanistic Modeling of the 2018 Caldera Collapse at Kīlauea Volcano, HI
Core Problem: Collapse mechanics at large calderas require models that can represent pressure-driven deformation, slip planes, and distributed plastic strain during rapid subsidence.
Key Innovation: A smoothed-particle hydrodynamics model of Kilauea 2018 uses observed chamber-pressure histories to bound critical collapse pressure and resolve stress arching and localized shear evolution.
7. True-triaxial laboratory tests with relevance to working-face rockburst in deep level bored tunnels
Core Problem: Working-face rockburst is difficult to reproduce in laboratory settings, limiting mechanism identification and warning design.
Key Innovation: A true-triaxial test system with internal support, high-speed micro-imaging, 3D pit reconstruction, and reverse modeling reproduces pure working-face rockburst and resolves its five-stage failure sequence.
8. Assessing urban flood susceptibility in Seoul, South Korea using machine learning models: effects of urban infrastructure and sampling variability
Core Problem: Urban flood susceptibility models often ignore infrastructure variables and sampling variability, weakening transfer to dense cities.
Key Innovation: A Seoul flood inventory is modeled with RF, XGB, MLP, LSTM, and TabPFN; adding road, sewer, and detention-system density improves performance and SHAP identifies road density as the leading factor.
9. The Critical Role of Surge Autocorrelation in Data‐Driven Nowcasting of Typhoon‐Induced Storm Surge
Core Problem: Short-lead storm-surge nowcasts must be accurate with sparse observations, yet machine-learning designs often overcomplicate inputs.
Key Innovation: Comparative modeling shows that local surge autocorrelation alone can support efficient sub-3-hour nowcasts and can outperform more complex recurrent, transformer, and XGBoost setups.
10. The Value of Forecasters‐in‐the‐Loop in Real‐Time Flood Forecasting in the Age of Machine Learning
Core Problem: Claims that machine learning can replace operational flood forecasting are rarely tested against real forecaster-in-the-loop systems.
Key Innovation: A direct comparison with the California-Nevada River Forecast Center shows that human-guided operational forecasts outperform ML models for streamflow and flood alerts up to 96 hours.
11. Shear-wave anisotropy in the 2022 Mw 5.6 Cianjur earthquake area, Indonesia
Core Problem: The 2022 Cianjur earthquake damaged a region where the causative fault was absent from hazard maps.
Key Innovation: Shear-wave splitting from 442 aftershocks maps stress-induced and structure-controlled anisotropy, indicating a conjugate fault system and off-fault deformation zone.
12. Estimating urban flood severity using social media text and remote sensing data: a case study of Henan, China
Core Problem: Rapid flood-impact assessment in cities must connect physical inundation with social exposure and reported consequences.
Key Innovation: The study fuses social-media text and remote-sensing evidence to estimate urban flood severity in Henan, providing a faster impact-oriented layer for emergency decisions.
13. Revealing agricultural irrigation-driven deformation in drylands using InSAR
Core Problem: Dryland irrigation can create deformation patterns that are difficult to separate from broader subsidence or groundwater signals.
Key Innovation: InSAR time series reveal agricultural irrigation-driven deformation and its spatiotemporal evolution, linking water management to surface-motion diagnostics.
14. Automated TBM tunnel-face collapse detection system based on spatiotemporal fusion and attention mechanism with muck characteristics and operational data
Core Problem: Tunnel-face collapse detection needs real-time use of muck characteristics and operational data rather than delayed manual inspection.
Key Innovation: A spatiotemporal fusion and attention-based system combines TBM muck and machine data to detect tunnel-face collapse conditions automatically.
15. Slope stability analysis with surrogate modeling: incorporating rainfall infiltration and spatial variability of soil parameters
Core Problem: Rainfall infiltration and spatially variable soil parameters make slope-stability analysis computationally expensive and uncertain.
Key Innovation: A surrogate-modeling framework incorporates infiltration and random soil variability to accelerate probabilistic slope-stability assessment.
16. Correlation between electrical resistivity variation and shear failure behaviors of rock joint
Core Problem: Rock-joint failure monitoring needs measurable precursors that respond to shear damage rather than only final displacement.
Key Innovation: Electrical-resistivity changes are correlated with shear-failure behavior of rock joints, supporting resistivity as a damage-sensitive monitoring signal.
17. State-of-the-art review of NaTech risk assessment of industrial facilities exposed to hydro-meteorological hazards
Core Problem: Industrial facilities exposed to floods, storm surges, and tsunamis lack an integrated NaTech risk-assessment view across hazard, vulnerability, and exposure.
Key Innovation: The review synthesizes fragmented methods for hydro-meteorological NaTech risk and frames hazardous-material release consequences under multi-hazard industrial exposure.
18. A Comprehensive Review of Polygenetic Signatures, Methodological Advances, and Implications for Coastal Boulder Deposits (CBDs) Assessment
Core Problem: Coastal boulder deposits preserve extreme-wave histories but are often ambiguous between storm and tsunami origins.
Key Innovation: A systematic review of 77 studies integrates UAV, TLS, LiDAR, geochronology, and hydrodynamic modeling to treat boulder deposits as polygenetic coastal-hazard archives.
19. Spatial Analysis of Earthquake Risk in Şanlıurfa City Center
Core Problem: Earthquake losses in expanding cities depend on planning history, groundwater, and proximity to faults, not only regional shaking.
Key Innovation: GIS-AHP mapping of Sanliurfa identifies short-distance risk contrasts after the 2023 earthquakes and ties high-risk zones to fault proximity and groundwater conditions.
20. Tectonic Deformation Analysis with ALOS-Based Digital Elevation Models in the Longshou Shan Mountains (NW China)
Core Problem: Active deformation around overlapping Tibetan Plateau fault systems must be resolved to evaluate local seismic hazard.
Key Innovation: ALOS DEMs and geomorphic indices quantify mountain-front, basin-shape, hypsometric, and channel-steepness signals in the Longshou Shan.
21. Spatial and Magnitude Distribution of Seismic Events in Santorini Island, January–February 2025: Tectonic or Volcanic Earthquakes?
Core Problem: The 2025 Santorini seismic sequence required distinguishing tectonic earthquakes from possible volcanic reactivation signals.
Key Innovation: Spatial and magnitude analysis of local seismic-network data separates event distributions and clarifies process interpretation during the January-February 2025 swarm.
22. Analysis on Multi-Factor Synergistic Hazards Mechanism of Wet Micro-Downburst: a Multi-Source Data Fusion Analysis Based on Passenger Vessel Capsizing Accident in Qianxi, Guizhou on 4 May 2025
Core Problem: Rare high-impact downburst events are difficult to reconstruct because radar, surface, lightning, video, and damage evidence are often analyzed separately.
Key Innovation: Minute-scale multi-source fusion resolves the dynamic, thermal, and microphysical chain behind the 2025 Qianxi passenger-vessel capsizing accident.
23. Magnetotelluric Monitoring of Earthquake Precursors
Core Problem: Electromagnetic earthquake precursor monitoring remains difficult to site and interpret.
Key Innovation: The paper reviews magnetotelluric monitoring windows, fault-sensitive apparent-resistivity signals, and ULF-ELF anomalies observed before recent Chinese earthquakes.
24. Exploratory Statistical Analysis of Precursors to Moderate Earthquakes in Japan
Core Problem: Moderate earthquakes are common but precursor statistics remain noisy and uncertain.
Key Innovation: Exploratory statistical analysis tests candidate precursor patterns for Japanese moderate events, offering a cautious data-driven baseline for short-term seismic monitoring.
25. 80 years of research on tsunamigenic earthquakes in the Makran subduction zone (1945–2025): a review- part B: tsunamis in the MSZ
Core Problem: The Makran subduction zone has sparse instrumental constraints but high regional tsunami consequences.
Key Innovation: An 80-year review organizes earthquake and tsunami evidence in the Makran subduction zone to support regional source characterization.
26. Assessment of seismic disaster risk in the Dingri earthquake based on fault-induced ground motion effects
Core Problem: Earthquake disaster risk after the Dingri event depends on fault-induced ground-motion effects that are not captured by exposure alone.
Key Innovation: The study integrates ground-motion effects into seismic disaster-risk assessment for the Dingri earthquake.
27. Identifying Deep Seismogenic Sources in Southern Piedmont (North-Western Italy) via the New Tool TESLA for Microseismicity Analysis
Core Problem: Microseismicity can reveal deep seismogenic structures, but source identification requires tools that resolve weak catalogs.
Key Innovation: The TESLA workflow is used to identify deep seismogenic sources in southern Piedmont from microseismic analysis.
28. Fault Structure Characterization in the Gulf of Evia (Central Greece): Insights from an Enhanced, Relocated Seismic Catalog (2018–2023)
Core Problem: Fault architecture in the Gulf of Evia remains uncertain despite frequent seismicity.
Key Innovation: An enhanced relocated seismic catalog from 2018-2023 is used to characterize fault structure and seismic-source geometry.
29. Seasonal and Episodic Variation of Aseismic Creep Displacement Along the West Valley Fault, Philippines
Core Problem: Fault-creep displacement can vary seasonally and episodically, complicating seismic-hazard interpretation.
Key Innovation: Monitoring along the West Valley Fault quantifies aseismic creep displacement changes and their episodic behavior.
30. Partial Ruptures, Cascading Multi‐Fault Ruptures, and Aftershocks in 2D Random Fault Network
Core Problem: Partial ruptures, cascading fault interactions, and aftershocks remain difficult to represent in simplified earthquake-source models.
Key Innovation: A 2D random fault-network model explores how partial ruptures cascade across multiple faults and generate aftershock patterns.
31. On the Application of 1 Versus 2 State Variable Rate‐and‐State Friction Laws: An Example From the Nankai Trough Megasplay Fault Zone
Core Problem: Rate-and-state friction laws with different state-variable structures can imply different fault-slip evolution in subduction settings.
Key Innovation: The Nankai megasplay example compares one- and two-state-variable friction formulations to clarify model choice for fault-zone behavior.
32. Dilatant Strengthening in Fluid‐Saturated Thermally Cracked Westerly Granite: Competing Effects of Strain Rate and Fluid Diffusion
Core Problem: Crustal rock strength under fluid pressure depends on the competition between deformation rate, permeability, and pore-pressure diffusion.
Key Innovation: Experiments on thermally cracked Westerly granite measure internal pore-pressure drops and show how dilatant strengthening bridges drained and undrained strength limits.
33. Site amplification estimated from strong motion records at Türkiye strong-motion stations using spectral separation technique
Core Problem: Regional earthquake loss estimation needs site amplification constraints from recorded strong motions.
Key Innovation: Spectral separation of Turkiye strong-motion records estimates station-specific amplification for seismic-response assessment.
34. Seismic response and fragility assessment of high-speed railway track bridge system under the action of mainshock-aftershock sequence
Core Problem: High-speed railway bridges can be affected by mainshock-aftershock sequences rather than single design events.
Key Innovation: Fragility assessment of track-bridge systems evaluates seismic response under sequential shaking.
35. A physics-informed transformer framework for seismic response modeling and fragility evaluation of soil–pile systems
Core Problem: Soil-pile fragility modeling needs efficient representations of nonlinear dynamic response.
Key Innovation: A physics-informed transformer predicts seismic response and supports fragility evaluation of soil-pile systems.
36. ST-SRNet: A deep learning framework for seismic response prediction of subsea tunnels
Core Problem: Subsea tunnels require rapid seismic-response prediction under complex marine and ground-motion conditions.
Key Innovation: ST-SRNet combines finite-element simulations with deep learning to estimate seismic responses at tunnel monitoring points.
37. Study on the in-situ stress characteristics and causes of large deformation failure in soft rock tunnels in Southwest China
Core Problem: Soft-rock tunnels in high-stress terrain can experience large deformation and failure, but stress causes remain difficult to isolate.
Key Innovation: In-situ stress characterization and mechanism analysis explain large deformation failure in southwest China soft-rock tunnels.
38. XAI-Aided Method for Accurately Assessing the Stability of Unlined Double Tunnels in Rock Masses
Core Problem: Unlined double tunnels in rock masses require stability assessment that remains interpretable under complex geotechnical controls.
Key Innovation: An XAI-aided method evaluates tunnel stability and exposes the drivers behind model predictions.
39. Monitoring of Surrounding Rock Deformation in Underground Roadways Using 3D Laser Scanning and Analysis of Environmental Influencing Factors
Core Problem: Underground roadway stability requires dense 3D deformation monitoring under changing environmental conditions.
Key Innovation: 3D laser scanning is used to monitor surrounding-rock deformation and analyze environmental controls in underground roadways.
40. A 3D Ubiquitous Multiple Joint Model for stability analysis in large underground caverns within jointed rock masses
Core Problem: Large underground caverns in jointed rock masses need stability models that represent multiple joint families in 3D.
Key Innovation: A ubiquitous multiple-joint model is proposed for stability analysis of large underground caverns.
41. Hydraulic performance of circular tunnels with interior drainage boundaries subjected to water level fluctuation
Core Problem: Tunnel drainage boundaries respond to water-level fluctuations, changing pore pressure and long-term stability.
Key Innovation: An analytical model estimates hydraulic performance of circular tunnels with interior drainage under fluctuating surface or groundwater levels.
42. Anisotropic failure characteristics of slate tunnels under large in situ stress
Core Problem: Slate tunnels under high in-situ stress can fail anisotropically, but orientation-dependent mechanisms remain underconstrained.
Key Innovation: The study resolves anisotropic failure characteristics of slate tunnels under large in-situ stress.
43. A two-phase dual-point SPH model with a new porosity updating method for water-soil interaction
Core Problem: Rapid water-soil interaction problems need numerical methods that handle porosity change and two-phase motion.
Key Innovation: A two-phase dual-point SPH model with porosity updating improves simulation of water-soil interaction relevant to collapse and landslide-like failures.
44. Damage evolution and hydro–mechanical coupled creep model of tuffaceous siltstone under wet–dry cycles
Core Problem: Tuffaceous siltstone weakens under repeated wet-dry cycling, affecting slopes and tunnel rocks.
Key Innovation: A hydro-mechanical coupled creep model links damage evolution to time-dependent deformation under wet-dry cycles.
45. Thermal cycling-induced degradation of carbonaceous mudstone soft interlayers: Energy dissipation-driven shear strength prediction
Core Problem: Mining slopes with carbonaceous mudstone interlayers degrade under thermal cycling, but strength prediction remains weak.
Key Innovation: Energy-dissipation analysis connects thermal-cycle damage to shear-strength degradation for climate-sensitive slope stability assessment.
46. Ground-breathing effect of the Loess Plateau: Insights from the Chinese 72 pentads
Core Problem: Seasonal ground motion in loess terrain can reflect coupled thermal, hydrological, and mechanical processes.
Key Innovation: The ground-breathing effect is characterized using the Chinese 72 pentads, linking short-period deformation to Loess Plateau process dynamics.
47. Dynamics of kinetic energy in rill headcut erosion and sediment yield characteristics within granitic red soil regions
Core Problem: Rill headcuts drive sediment yield and soil loss, but kinetic-energy partitioning across profile configurations is not well resolved.
Key Innovation: Laboratory experiments quantify energy dissipation, particle sorting, and sediment pulses in granitic red soil headcut erosion.
48. Regional Soil Erosion Assessment Using Remote Sensing and Field Validation: Enhancing the Erosion Potential Model
Core Problem: Soil-erosion models require field validation to support regional management rather than only index-based mapping.
Key Innovation: Remote sensing and field validation enhance the Erosion Potential Model for regional soil-erosion assessment.
49. Event‐driven shifts in river and reservoir sediment sources: Cottonwood River and John Redmond Reservoir, USA
Core Problem: Reservoir sedimentation mitigation requires separating chronic channel sources from floodplain contributions during extreme flows.
Key Innovation: Sediment fingerprinting shows a hydrogeomorphic threshold in which routine flows are bank-dominated but large floods mobilize cropland floodplain sediment.
50. Characterizing settlement phases, social structure and processes in the context of recurring floods, riverbank erosion and evolution of a tropical riverine island of the Ganga River, India
Core Problem: Riverine islands experience settlement, social, and geomorphic change under recurring floods and bank erosion.
Key Innovation: The study links settlement phases, social structure, and river evolution on a Ganga River island to disaster-response conditions.
51. Biotic-geomorphic coupling by livestock grazing enhances hillslope hydrological function and stability
Core Problem: Vegetation and grazing can reorganize hillslope hydrological function, but their geomorphic stability effects are often simplified.
Key Innovation: Long-term field evidence shows biotic-geomorphic coupling by livestock grazing can maintain connected trails, soil moisture, and functional hillslope resilience.
52. The DLOD&MCCA Framework for Accurate Mapping of Reservoir Dams in Arid Regions from Remote Sensing Imagery: A Multimodal Fusion and Constraint Approach
Core Problem: Dam inventories in arid regions remain incomplete where remote-sensing signatures are sparse or confused with natural terrain.
Key Innovation: A multimodal remote-sensing framework maps reservoir dams with fusion and constraint mechanisms.
53. Data-Driven Deep Learning Model for Detecting Ionospheric Electric Field Perturbations and Seismic Correlation
Core Problem: Ionospheric electric-field perturbations may correlate with seismic activity, but detection requires noise-resistant learning methods.
Key Innovation: A data-driven deep-learning model is developed for detecting ionospheric perturbations and exploring seismic correlation.
54. Public Access Dimensions of Landscape Changes in Parks and Reserves: Case Studies of Erosion Impacts and Responses in a Changing Climate
Core Problem: Protected landscapes face climate-driven erosion and flooding that disrupt public access as well as conservation functions.
Key Innovation: Case studies from New Zealand parks combine VGI and DEM analysis to separate physical impacts, human responses, and access-management tradeoffs.
55. Anticipatory Justice: Relocation policy impacts in flood-affected communities of Southern Québec, Canada
Core Problem: Retreat policies after floods can redistribute vulnerability and loss beyond direct physical damage.
Key Innovation: A mixed-method study of Southern Quebec evaluates relocation policy impacts, home-equity loss, place attachment, and justice implications.
56. Risk perception and adaptation of riparian farmers to extreme weather events in the Maule River basin (Chile)
Core Problem: Riparian farmers face repeated extreme weather but adaptation depends on risk perception and local constraints.
Key Innovation: A Maule River basin case study evaluates how farmers perceive and adapt to climate-related river hazards.
57. Integrated decision-support framework for enhancing coastal community resilience against hurricane-induced hazards
Core Problem: Coastal resilience planning requires integrated decision support for hurricane-induced hazards, including surge exposure.
Key Innovation: A decision-support framework is developed to enhance coastal community resilience under hurricane hazards.
58. Huayu: Advanced real-time satellite-based precipitation estimation
Core Problem: Real-time precipitation estimation limits flash-flood and hydrological warning in data-sparse regions.
Key Innovation: Huayu provides an advanced real-time satellite-based precipitation estimation system with relevance for rapid hydrometeorological monitoring.
59. Predicting thunderstorm risk probability at very short time range using deep learning
Core Problem: Very-short-range thunderstorm risk prediction requires rapid identification of convective hazard probability.
Key Innovation: A deep-learning model estimates thunderstorm risk probability at very short lead times.
60. A Bibliographic Analysis of Multi-Risk Assessment Methodologies for Natural Disaster Prevention
Core Problem: Natural-disaster prevention increasingly requires methods that combine interacting hazards rather than evaluate them separately.
Key Innovation: A bibliographic analysis reviews multi-risk assessment methodologies and identifies methodological gaps for disaster prevention.
61. Utilizing LLMs and ML Algorithms in Disaster-Related Social Media Content
Core Problem: Disaster response needs structured social-media data, but crisis posts are usually unlabeled and semantically heterogeneous.
Key Innovation: LLMs and clustering are used to structure, label, and extract emergent themes from disaster-related tweet datasets.
62. Metaheuristic optimizer combined with advanced machine learning algorithm for accurate wildfire susceptibility in Western Canada: A novel approach
Core Problem: Wildfire susceptibility maps require stable pseudo-absence strategies and interpretable machine-learning outputs.
Key Innovation: Metaheuristic-optimized XGBoost and SHAP produce a province-wide wildfire susceptibility framework for Alberta.
63. FireCast-fusion: Physics-Guided fusion of UAV RGB–thermal imagery and environmental data for near-term wildfire spread prediction
Core Problem: Wildfire spread forecasting needs fast fusion of UAV observations with environmental controls.
Key Innovation: FireCast-fusion combines UAV RGB-thermal imagery, physics guidance, and environmental variables for near-term spread prediction.
64. Volcanic Eruptions and Moss Heath Wildfires on Iceland’s Reykjanes Peninsula: Satellite and Field Perspectives on Disturbance and Recovery
Core Problem: Volcanic eruptions can trigger secondary wildfire disturbance whose ecological recovery is rarely quantified.
Key Innovation: Satellite and field observations map moss-heath wildfire disturbance and recovery after Reykjanes eruptions.
65. Tracing cascading economic impacts of volcanic disruption through regional value chains: Implications for disaster risk reduction
Core Problem: Volcanic disruption can propagate through regional value chains, producing cascading losses beyond the eruption footprint.
Key Innovation: A regional value-chain approach traces economic disruption pathways and their implications for disaster-risk reduction.
66. A review of similar physical simulation in geotechnical engineering: Advances in the past 25 years
Core Problem: Physical modeling remains central to slope, tunnel, dam, and pile-foundation problems, but methods have diversified rapidly.
Key Innovation: A 25-year review synthesizes advances in similitude theory, materials, loading systems, and multiphysics measurement for geotechnical hazards.