TerraMosaic Daily Digest: Mar 22, 2026
Daily Summary
This March 22, 2026 digest distills 30 selected papers from 495 analyzed records. The strongest papers revolve around progressive failure under coupled forcing. The Jinpingcun landslide is reconstructed as a hydromechanical collapse in an anti-dip slope; a new corridor-scale deep network sharpens susceptibility mapping for the Sichuan-Tibet transport system; and structurally jointed Himalayan slopes are shown to move from critical stability to catastrophic failure under moderate infiltration or modest seismic loading.
A second cluster treats hazards as connected systems rather than isolated objects. River levee deformation is quantified probabilistically in heterogeneous soil, post-wildfire flood protection is evaluated as a low-cost intervention problem, cascade reservoir risk is propagated across dam networks, and landslide-generated surges are reinterpreted through water intrusion and porosity effects. Across slopes, levees, tunnels, and rivers, the most useful models are those that keep signals interpretable enough to remain linked to process.
Key Trends
Today's core contribution is not a new hazard type but a clearer way of representing how hazardous systems accumulate damage, exchange signals, and transmit failure.
- Progressive failure is overtaking static susceptibility as the main framing: anti-dip slopes, rock slopes, levees, and tunnel systems are analysed through staged deformation and connectivity rather than one-off trigger narratives.
- Indirect measurements are being physicalized: microseismicity, mineral composition, suspended sediment, and InSAR velocities are being translated into residual strength, failure time, collapse risk, or deformation state.
- Compound and chain hazards are increasingly explicit: landslide surges, post-fire floods, cascade reservoir failures, and sediment-starved bank collapse are all treated as propagated systems rather than local incidents.
- Hydrological extremes are being described structurally: sub-daily precipitation, flood-peak timing, moisture-source pathways, lake-regulated baseflow, and seasonal drought are being analysed through storage and routing logic.
- Machine learning survives when tied to mechanism: the strongest data-driven papers today succeed because they preserve interpretable links between input signals and physical controls on instability.
Selected Papers
This digest features 30 selected papers from 495 papers analyzed.
1. Progressive failure of an anti-dip slope: characteristics and mechanisms of the 2025 Jinpingcun landslide, Sichuan, China
Core Problem: Anti-dip slopes in seismically active, rainfall-prone terrain can fail progressively over long periods, but the coupling between structure, rainfall, and hydromechanical weakening is often poorly resolved after disaster.
Key Innovation: Field investigation and numerical analysis show how sustained rainfall, water-pressure buildup in weak layers, and cumulative freeze-thaw and seismic damage linked flexural toppling to large-scale progressive collapse in the Jinpingcun landslide.
2. A deep learning multi-branch network method for landslide susceptibility mapping in the Sichuan-Tibet transportation corridor
Core Problem: CNN-based susceptibility models often overemphasize merged feature channels, reducing their ability to capture the diverse environmental controls on corridor-scale landslides.
Key Innovation: A multi-branch deep network with adaptive weighting separates and fuses heterogeneous remote-sensing factors more effectively, improving corridor-scale landslide susceptibility mapping across geologically diverse subregions.
3. Static and dynamic analysis of joint-controlled rock slope failures in Udipur, Lamjung, Nepal
Core Problem: Open-jointed Himalayan rock slopes can be destabilized by moderate groundwater rise or modest seismic loading, but the interaction between discontinuity control and external forcing is hard to quantify.
Key Innovation: Field investigation, laboratory testing, kinematic analysis, and FEM/LEM modelling are combined to show how jointed rock mass, pore-pressure increase, and seismic acceleration jointly trigger progressive failure in the Udipur slope.
4. Probabilistic risk assessment of a flood-damaged river levee using SPH with Nyström-based heterogeneous soil fields
Core Problem: Deterministic levee analyses can miss localized deformation and overtopping risk because they neglect spatially variable, cross-correlated soil heterogeneity.
Key Innovation: A Nyström-generated random-field framework embedded in SPH enables large-deformation Monte Carlo assessment of a real flood-damaged levee, revealing how heterogeneity governs localized subsidence and failure likelihood.
5. Near-fault ground motions induced dynamic responses and failure mechanisms of tunnels considering seismic-karst coupling effects
Core Problem: Tunnel design in karst terrain remains vulnerable because near-fault pulse effects and cavity geometry can combine to shift both deformation mode and damage location.
Key Innovation: The study demonstrates how forward-directivity, fling-step, and non-pulse motions interact with lateral and basal karst voids to reorganize stress paths and migrate damage to distinct tunnel sectors.
6. Dynamic response of a gentle near-fault bedrock-overburden slope after rainfall based on shaking table tests
Core Problem: Post-rainfall seismic response of gentle bedrock-overburden slopes near faults is still insufficiently constrained, particularly before global sliding has fully developed.
Key Innovation: Sequential fault pre-displacement, rainfall, and shaking-table loading reveal staged cracking, elevation-dependent amplification, and pore-pressure redistribution that precede formation of a potential sliding weak zone.
7. Characteristics of overburden failure and evolution of the water-conducting fracture zone in downward mining of close distance coal seams: a case study of the Shaozhai Coal Mine
Core Problem: Downward extraction of close-distance coal seams increases collapse and water-inrush risk, but fracture-zone evolution is still often predicted with overly simple empirical rules.
Key Innovation: Theoretical, numerical, physical, and field evidence are integrated into a four-stage evolution model for the water-conducting fracture zone, clarifying how downward mining activates aquifer connectivity and inrush risk.
8. Parametric study on topographic amplification effect of soil slope under the coupling of soil nonlinearity and oblique incidence
Core Problem: Topographic amplification in soil slopes remains difficult to generalize because nonlinear soil behavior and oblique wave incidence modify both amplitude and resonance in coupled ways.
Key Innovation: A parametric two-dimensional seismic model shows how nonlinearity reduces amplification while oblique incidence intensifies and shifts it, providing more conservative spectral guidance for slope-hazard design.
9. Experimental study on face stability for shield tunnelling in sandy cobble strata at various rock contents
Core Problem: Mixed sandy-cobble ground poses highly variable face-stability conditions, yet the specific role of rock content in collapse development remains insufficiently quantified.
Key Innovation: Shield model tests show how rock content reorganizes cutterhead torque, collapse onset, and surface-subsurface failure geometry, clarifying stability differences between soil discharge and shield-retreat scenarios.
10. Ground reaction curve for shield tunnels under finite deformation conditions
Core Problem: Classical loosening-earth-pressure formulations ignore finite deformation effects, which can bias the load estimated on shield tunnels.
Key Innovation: A new ground-reaction-curve formulation accounts for overburden, gravity, and finite deformation, showing when Terzaghi-type loads systematically underestimate tunnel pressure.
11. On water intrusion and soil porosity in landslide-induced surge waves: A multi-layer SPH analysis of the Vajont disaster
Core Problem: Numerical surge-wave models usually treat the sliding mass as impermeable, overlooking how water intrusion and porosity affect energy transfer during catastrophic landslide impacts.
Key Innovation: A mixture-theory multi-layer SPH formulation shows that porous intrusion can amplify surges relative to impermeable assumptions and clarifies how porosity interacts with landslide mass in the Vajont setting.
12. An Elastoplastic Constitutive Model for Frozen Clay with DEM-Derived Micromechanical Damage
Core Problem: Degradation of interparticle cementation in frozen clay is difficult to quantify directly, limiting constitutive modelling under varying cooling and confinement conditions.
Key Innovation: DEM-derived micromechanical damage is embedded into a unified hardening framework, enabling more physically grounded prediction of frozen-clay stress-strain and volumetric behavior.
13. Spatial and morphometric analysis of a comprehensive dataset of loess sinkholes from a small basin in the Chinese Loess Plateau
Core Problem: Loess sinkholes remain far less well mapped and quantified than karst sinkholes, obscuring both their geomorphic significance and their role in subsurface soil erosion.
Key Innovation: Airborne and handheld laser scanning combined with photogrammetry create a 1,194-record basin-scale sinkhole dataset that links morphology, spatial patterns, and hidden soil-loss volumes in the Chinese Loess Plateau.
14. Low-stress creep behavior of multiple salt caverns under cyclic operations
Core Problem: Hydrogen-storage salt caverns undergo repeated pressure cycling, but low-stress creep remains underappreciated as a control on long-term closure and operational interaction between caverns.
Key Innovation: A three-dimensional creep model shows how low-stress pressure-solution creep accelerates cavern closure under cyclic operation, especially when cycle frequency is high and pressure swings are large.
15. Projected changes in sub-daily extreme precipitation: comparing temperature-scaling approaches and convection-permitting models across an Alpine gradient
Core Problem: Temperature-scaling methods are widely used for future short-duration rainfall extremes, but their reliability across complex mountain terrain remains uncertain.
Key Innovation: Comparison with convection-permitting model output shows where scaling performs well and where shifts in seasonality degrade it, refining how sub-daily precipitation change should be projected for Alpine basins.
16. A flood peak prediction in data-scarce mountain river basins considering the time distribution of rainfall
Core Problem: Data-scarce mountain basins need flood-peak models that respect rainfall timing and antecedent wetness without relying on heavily parameterized hydrological schemes.
Key Innovation: A runoff-generation and routing framework with explicit rainfall-time-distribution effects achieves markedly better flood-peak prediction than a benchmark conceptual model across multiple mountain catchments.
17. Associated influence of moisture sources and catchment drivers on the flood mechanism of a tropical basin
Core Problem: Flood mechanisms in tropical basins are often analyzed locally, without fully linking atmospheric moisture pathways to catchment-scale flood response.
Key Innovation: Event-scale analysis of monsoon flood cases shows how evolving oceanic and continental moisture sources, low-pressure systems, and cross-equatorial flow jointly shape flood generation in a tropical basin.
18. A lesson in preparedness: assessing the effectiveness of low-cost post-wildfire flood protection measures for the catastrophic flood in Kineta, Greece
Core Problem: Post-fire flood risk is escalating in Mediterranean basins, yet low-cost protection strategies are rarely evaluated within an integrated hazard-to-damage framework.
Key Innovation: Meteorological, hydrological, hydraulic, remote-sensing, and cost analyses are combined to show that post-wildfire protection treatments could have offset much of the Kineta flood impact at a fraction of damage cost.
19. Analysis of InSAR datasets from the EGMS and Veneto region using ground-based GNSS velocities for the monitoring of land subsidence in the Po river delta (Italy)
Core Problem: Open InSAR products are increasingly used for deformation monitoring, but their calibration quality and bias characteristics are not always known well enough for direct decision use.
Key Innovation: Dense GNSS control is used to validate and integrate multiple open InSAR products, producing a calibrated subsidence-monitoring workflow for the Po River delta and exposing where regional products require adjustment.
20. Future changes in seasonal drought in Australia
Core Problem: Australian drought projections remain uncertain because precipitation projections disagree strongly, obscuring whether drought intensification is robust across drought types and seasons.
Key Innovation: An ensemble of downscaled hydrological simulations shows robust future increases in meteorological, hydrological, and agricultural drought, especially in winter and spring, while also partitioning the major uncertainty sources.
21. Damming effects on bankfull channel geometry in the Lower Yellow River: adjustment patterns and a physics-informed machine learning framework
Core Problem: Channel adjustment after dam operation is often reduced to incision alone, even though width change can strongly reshape bankfull geometry in unstable downstream reaches.
Key Innovation: Twenty-two years of bathymetry and a physics-informed machine-learning model identify distinct width-elevation adjustment patterns and quantify how widening contributes to post-dam bankfull-area change in the Lower Yellow River.
22. A multi-source comprehensive risk quantification method for cascade reservoirs based on Monte Carlo simulation, HEC-RAS and Bayesian networks: Development and application in the Dadu river Basin
Core Problem: Traditional cascade-dam risk assessment often focuses on single flooding factors and empirical routing, missing the compound propagation of failure across reservoir systems.
Key Innovation: Monte Carlo simulation, HEC-RAS, and Bayesian networks are integrated into a multi-source cascade-reservoir risk framework that quantifies inter-dam transmission and compound failure likelihood more realistically.
23. Quantifying impacts of lakes on baseflow dynamics in headwaters of the Yellow River, Tibetan Plateau
Core Problem: The role of lakes in regulating baseflow under climate change in Tibetan headwaters remains ambiguous because their positive and negative influences can coexist.
Key Innovation: A quasi-paired catchment design and machine-learned no-lake scenario reveal how lakes both damp high-baseflow frequency and reweight the climatic controls on low-flow behavior.
24. A fully automated OPTRAM (aOPTRAM) for soil moisture retrieval: Evaluating multiple fitting functions, vegetation indices, land-cover types, and scales
Core Problem: Most OPTRAM implementations still rely on manual edge selection, limiting scalable soil-moisture retrieval across diverse land-cover conditions.
Key Innovation: An automated sliding-window edge-detection workflow removes manual calibration and systematically characterizes when different indices, fitting functions, and export scales yield the most reliable soil-moisture estimates.
25. Insights From Two Decades of Seismicity at Columbia Glacier
Core Problem: Long-duration seismic records from calving glaciers are underused as indicators of glacier-state evolution and changing terminus conditions.
Key Innovation: Two decades of glacier-quake records are tied to terminus position, bathymetry, velocity, and environment, showing how seismicity can track transitions in grounding, calving, and glacier evolution.
26. Micro-dynamic Disturbance Direction and Amplitude Effects on Shear Behavior of Karst Corrosion Discontinuities Under True Triaxial Stress
Core Problem: Karst-corroded discontinuities respond sensitively to disturbance direction and amplitude, but their true-triaxial shear behavior under micro-dynamic loading remains poorly understood.
Key Innovation: The study clarifies how disturbance orientation and amplitude reorganize shear response along karst corrosion planes under true triaxial stress, improving mechanical interpretation of discontinuity-controlled instability.
27. Three-Dimensional Reconstruction of Rock Discontinuities: A Logging-Constrained Optimization Approach Based on Simulated Annealing Algorithm
Core Problem: Reliable three-dimensional reconstruction of discontinuity networks remains difficult when borehole and surface observations must be integrated consistently.
Key Innovation: A simulated-annealing optimization constrained by logging information improves 3D discontinuity reconstruction, offering better structural input for stability analysis in fractured rock masses.
28. Study on the dual effects of In-situ stress recovery on unloading creep behavior and failure mechanism of deep red sandstone
Core Problem: The role of in-situ stress recovery in controlling unloading creep and delayed failure of deep sandstone remains ambiguous during excavation-related disturbance.
Key Innovation: The study clarifies how stress recovery can simultaneously stabilize and destabilize red sandstone during unloading creep, refining the mechanics of delayed deep-rock failure.
29. Frequency Spectrum and Energy Analysis of Stress Waves Vertically Incident on a Single Joint under Dynamic Loading
Core Problem: Stress-wave transmission across joints remains a key uncertainty in dynamic rock response and damage propagation under impact loading.
Key Innovation: Frequency-spectrum and energy analysis of waves incident on a single joint clarifies how discontinuities filter and dissipate dynamic loading in jointed rock masses.
30. Predicting Fault Gouge Shear Strength Under Small-Sample and Missing Feature Conditions: A Three-Stage Framework with Pretrained Tabular Inference
Core Problem: Fault-gouge shear strength is difficult to estimate reliably when laboratory datasets are small and feature coverage is incomplete, limiting fault-stability assessment.
Key Innovation: A staged tabular-inference framework is designed for small-sample and missing-feature conditions, improving predictive support for shear-strength estimation in fault-gouge materials.