TerraMosaic Daily Digest: April 21, 2026
Daily Summary
This April 21, 2026 digest distills 32 selected papers from 1,778 analyzed records. This issue is anchored by papers that treat slope failure as a response to changing limits, not as a fixed landscape attribute. Qinling debris flows are organized into distinct initiation modes with explicit rainfall, moisture, and timing thresholds. Chin Coulee shows that reservoir-rim instability is inseparable from drawdown regime under climate pressure. The eastern Taiwan landslide paper goes further by linking monitoring data and discrete-element simulation to a failure acceleration threshold. Together, these studies make a consistent point: useful landslide analysis begins when triggering conditions are quantified as evolving bounds on system behaviour.
The same logic extends to material state and transport. The Jishishan loess and Sarangani studies show that liquefaction cannot be separated from sediment structure, groundwater setting, and field evidence. Particle orientation and air injection then demonstrate that cyclic resistance is a mutable property, not a single soil constant. Bailong, Batang Ombilin, and Central Rimo keep source renewal, hydraulic forcing, and downstream routing inside the hazard picture rather than treating them as secondary effects. The method papers are strongest when they reduce ambiguity in that same framework, whether through uncertainty-aware slope prediction, explainable subsidence screening, differentiable simulators, or automated constitutive calibration.
Key Trends
This set repeatedly replaces broad hazard labels with measurable control variables: trigger windows, source supply, material state, and uncertainty structure.
- Failure is treated as a threshold-crossing process: Qinling debris-flow initiation, Chin Coulee drawdown response, and the Taiwan earthquake-displacement study all express hazard through explicit trigger windows, response limits, or acceleration thresholds.
- Source supply and pathway evolution stay inside the model: Bailong weathering-controlled source supply, Batang Ombilin hydraulic forcing, and Central Rimo historical GLOFs all show that downstream consequence depends on how sediment and water are mobilized and routed after initiation.
- Liquefaction is being recast in terms of hidden state variables: The Jishishan loess study, Sarangani field evidence, particle-orientation tests, and air-injection mitigation paper all show that fabric, saturation state, and sediment structure materially reshape response.
- The strongest methods papers reduce ambiguity, not just error: Deep-kernel slope prediction, Bayesian liquefaction modelling, CycLiq auto-calibration, differentiable slope simulation, and explainable subsidence screening all make hazard assessment easier to inspect and defend.
Selected Papers
This digest features 32 selected papers from 1,778 papers analyzed, led by threshold-controlled slope response, source-and-routing mechanics in debris flows, and state-sensitive liquefaction modelling.
1. Study on initiation patterns and critical conditions of debris flows in typical small watersheds on the North piedmont of the Qinling Mountains
Core Problem: Small-watershed debris-flow initiation in the Qinling piedmont remains difficult to forecast because local triggering patterns and critical thresholds are poorly constrained.
Key Innovation: Field observations and flume experiments identify three distinct initiation modes and quantify regional thresholds for water content, cumulative rainfall, and initiation time in Qinling debris flows.
2. Assessing earthquake-induced displacements of large-scale landslide by in-situ monitoring and discrete element analysis
Core Problem: Large active landslides in seismic terrain still lack defensible displacement thresholds linking real-time monitoring to catastrophic failure scenarios.
Key Innovation: The study integrates GNSS, TDR, seismic measurements, and DEM simulation to derive a site-specific acceleration threshold for catastrophic sliding in an active landslide in eastern Taiwan.
3. Stability assessment of reservoir rim landslides under climate-induced drawdown scenarios: the Chin Coulee landslide
Core Problem: Climate-driven drawdown variability is becoming a first-order control on reservoir-rim instability, yet slope assessments rarely connect changing water-demand scenarios to reactivation risk.
Key Innovation: Using the Chin Coulee landslide, the paper builds a practical framework that links present and future drawdown scenarios to changing landslide stability under climate-sensitive reservoir operation.
4. Hydro-geotechnical integration for sustainable mitigation of river-induced landslides along the Batang Ombilin River, West Sumatra
Core Problem: River-induced landslides are often stabilized locally without resolving how hydraulic forcing at bends sustains the failure mechanism.
Key Innovation: This study integrates HEC-RAS hydraulics with geotechnical slope analysis to show that flow-energy reduction is the primary requirement for sustainable mitigation along the Batang Ombilin corridor.
5. A generic framework for investigating historical ice-dammed glacial lake outburst floods: A case study of Central Rimo Lake, Karakoram
Core Problem: Historical ice-dammed GLOFs remain under-documented, which limits risk assessment in glacierized mountain regions.
Key Innovation: The paper builds a remote-sensing framework to reconstruct multi-decadal ice-dammed GLOF histories and applies it to Central Rimo Lake, identifying 22 events and their controlling parameters.
6. Study on the liquefaction mechanism of the saturated loess in Zhongchuan County induced by the Jishishan MS 6.2 earthquake
Core Problem: The 2023 Jishishan disaster showed that saturated loess can fail as a liquefaction-driven flow-slip system, but the controlling mechanism is still insufficiently resolved.
Key Innovation: This paper analyzes the Zhongchuan loess failure to clarify how seismic loading, pore-pressure generation, and loess structure combined to produce a fatal liquefaction flow-slip event.
7. Soil Liquefaction in Sarangani Peninsula, Philippines Triggered by the 17 November 2023 Magnitude 6.8 Earthquake
Core Problem: Field-scale liquefaction events are most useful when mapped with enough geomorphic and geophysical detail to explain why some lowlands failed and others did not.
Key Innovation: The study combines geomorphic mapping, geophysical surveys, and grain-size evidence to explain the 2023 Sarangani liquefaction field and validate existing susceptibility maps.
8. Risk assessment of a typical debris flow gully in the Bailong River Basin based on the weathering characteristics of slope surface rocks
Core Problem: Debris-flow risk screening in Zhouqu is weakened when weathering-driven source supply is treated as static instead of as a dynamically degrading control.
Key Innovation: This paper uses time-series remote sensing, multi-elevation rock monitoring, and weighted risk modelling to connect slope-surface weathering rates directly to debris-flow risk evolution in the Bailong basin.
9. Landslide susceptibility patterns and their implications for erosional amphitheater formation on Bioko Island (Equatorial Guinea)
Core Problem: Remote volcanic landscapes often show clustered landslides whose geomorphic significance is easy to miss when susceptibility and landform evolution are analyzed separately.
Key Innovation: Using a new island-scale inventory and RF susceptibility modelling, the paper links focused landslide clustering on Bioko Island to precipitation, faults, and erosional amphitheater growth.
10. A Discontinuous deformation analysis method incorporating dynamic failure strength and its application to pulse-like ground motion-induced landslides
Core Problem: Pulse-like seismic waves are particularly damaging to slopes, but most coseismic landslide models still lack an explicit representation of dynamic failure strength under this loading style.
Key Innovation: The paper embeds dynamic cumulative damage and pulse-like ground-motion effects into discontinuous deformation analysis, improving simulation of pulse-driven earthquake landslides.
11. Particle orientation influences on liquefaction and reliquefaction resistance during shaking table tests: comparison with triaxial test results
Core Problem: Liquefaction resistance is usually parameterized with density and stress history, while particle fabric and its reorientation through liquefaction cycles remain underrepresented.
Key Innovation: Shaking-table tests show that particle orientation strongly controls both liquefaction and reliquefaction resistance, and that fabric reorientation helps explain why cyclic response changes after the first event.
12. Improvement of liquefaction resistance and cyclic response of sandy soil by air injection method
Core Problem: Liquefaction mitigation methods are often expensive or invasive, and the gains achievable through controlled partial saturation remain insufficiently quantified.
Key Innovation: The paper demonstrates that air injection can induce partial saturation, suppress excess pore-pressure buildup, and substantially improve cyclic liquefaction resistance in sandy soils.
13. Seismic behaviour of cantilever retaining walls on slopes with different inclinations in dynamic centrifuge model tests
Core Problem: Seismic design rules for retaining walls on slopes still rely heavily on simplified stability checks without enough large-scale experimental validation.
Key Innovation: Dynamic centrifuge tests show how increasing foundation inclination shifts retaining-wall failure from sliding toward overturning and global slope failure, while also benchmarking current design-code assumptions.
14. Slope stability prediction based on the combination of deep kernel learning and gaussian process
Core Problem: Machine-learning slope-stability prediction often captures nonlinear relations well but provides weak uncertainty quantification for safety-factor inference.
Key Innovation: This study combines deep kernel learning with Gaussian processes to improve slope-stability prediction while retaining probabilistic uncertainty estimates.
15. Monitoring and analysis of surface subsidence in mining areas of integrated SBAS-InSAR and probability integral method
Core Problem: Mining subsidence monitoring remains incomplete when InSAR deformation maps and theoretical subsidence models are used independently.
Key Innovation: The paper fuses SBAS-InSAR with a probability-integral model to generate more complete and accurate basin-scale mining-subsidence estimates.
16. Landslide susceptibility assessment in the Central Yunnan Plateau by assembling optimized statistical and machine learning models
Core Problem: Regional susceptibility models often include rainfall as a coarse background variable rather than optimizing the relevant event windows for the target terrain.
Key Innovation: This study optimizes rainfall, geology, and land-use predictors across multiple models and shows that 24-hour and 30-day rainfall windows are especially informative in the Central Yunnan Plateau.
17. Bayesian autocorrelation model selection and parameter identification using time-series monitoring data of pore water pressure of the slope under rainfall
Core Problem: Random-field slope analyses depend on autocorrelation structure, but it remains difficult to identify that structure directly from monitoring data.
Key Innovation: The paper develops a Bayesian workflow that infers autocorrelation models and scale-of-fluctuation parameters from pore-pressure time series in rainfall-affected slopes.
18. Experimental study on apparent viscosity and yield stress of clean sand in whole liquefaction process based on the Casson fluid model
Core Problem: Post-trigger liquefied soil is often treated with simplified rheology, limiting interpretation of the full transition from solid-like to fluid-like behaviour.
Key Innovation: Using the Casson fluid model, this study ties apparent viscosity and yield stress to liquefaction evolution and offers a compact rheological description of clean-sand behaviour.
19. Study on the Bayesian network probability model for soil seismic liquefaction with incomplete data
Core Problem: Liquefaction case databases are often incomplete, which weakens probabilistic models built on conventional fully observed training sets.
Key Innovation: This paper uses Bayesian-network modelling to estimate soil liquefaction probability under systematically incomplete field datasets.
20. From prediction to decision: An explainable AI framework for Urban ground subsidence risk management
Core Problem: Urban ground subsidence is difficult to manage because prediction models rarely expose which factors are driving risk strongly enough to support intervention decisions.
Key Innovation: The paper uses explainable AI to turn subsidence-risk prediction into a decision-oriented workflow that ranks the dominant drivers of urban ground settlement.
21. Liquefaction response of monopile offshore wind turbines under different loading conditions during operating and shutdown states
Core Problem: Offshore liquefaction analysis often treats foundation loading too generically, even though operating and shutdown states can generate different cyclic responses.
Key Innovation: This study evaluates seabed liquefaction around monopile wind turbines under operational and typhoon-related loading scenarios to clarify how loading state changes offshore liquefaction potential.
22. Energy-based approach to evaluate liquefaction potential of fiber-reinforced sand under isotropic and anisotropic consolidations
Core Problem: Fiber reinforcement is promising for liquefaction mitigation, but its influence under different consolidation paths is still not well described in energy terms.
Key Innovation: The paper applies an energy-based framework to quantify how fiber reinforcement alters liquefaction resistance under isotropic and anisotropic consolidation conditions.
23. Centrifuge modeling of tunnel uplift induced by liquefaction under successive seismic motions
Core Problem: Tunnel uplift in liquefiable ground is known to be severe, but its cumulative evolution under successive earthquakes remains poorly constrained by experiment.
Key Innovation: This study uses centrifuge modelling to resolve how repeated seismic loading amplifies uplift and deformation of buried tunnels in liquefiable soils.
24. Automatic calibration method for the CycLiq constitutive model with focus on liquefaction behaviour
Core Problem: Advanced liquefaction constitutive models remain hard to use operationally because calibration is laborious and strongly dependent on analyst experience.
Key Innovation: The paper develops an automated calibration strategy for the CycLiq model and shows that it can recover liquefaction behaviour reliably across benchmark tests.
25. Seismic resilience assessment of shield tunnel linings in sites susceptible to liquefaction
Core Problem: Seismic fragility studies often stop at damage probability and do not carry liquefaction effects through to recovery and resilience of buried infrastructure.
Key Innovation: This paper links liquefaction-induced tunnel deformation to recovery functions and a life-cycle resilience index for shield tunnels in layered sites.
26. Performance evaluation and optimal design of inclined alternating rigid-core high-pressure jet grouting piles slope retaining structure: A case study in Guiyang, China
Core Problem: Slope-retaining systems are often optimized locally, with limited case-based evidence on how new hybrid pile systems perform under realistic field conditions.
Key Innovation: The paper evaluates an inclined rigid-core jet-grouting pile retaining system and shows how configuration optimization improves deformation control and slope safety in Guiyang.
27. Deterministic and probabilistic stability analyses of single tunnel in a cohesive-frictional finite slope
Core Problem: Tunnel-slope interaction is usually assessed deterministically even though uncertainty in soil strength can materially change failure probability.
Key Innovation: This study combines finite-element limit analysis with FORM to evaluate both deterministic and probabilistic stability of a tunnel excavated in a cohesive-frictional finite slope.
28. Differentiable graph neural network simulator for forward and inverse modeling of multi-layered slope system with multiple material properties
Core Problem: Most fast slope simulators are built for simplified granular systems and do not reflect the layered material complexity of real engineered slopes.
Key Innovation: The paper develops a differentiable graph neural network simulator for multi-layered slope systems that supports both fast forward prediction and inverse estimation of material properties.
29. Downscaling Analysis of Remote Sensing Data Products Incorporating Physical Mechanisms Across Different Slope Positions in the New South Wales Catchment, Australia
Core Problem: Remote-sensing soil-moisture products often lose physical consistency across different slope positions, limiting their usefulness for slope-process interpretation.
Key Innovation: This study fuses remote sensing and hydrological modelling to generate physically constrained, high-resolution soil-moisture estimates across contrasting slope positions.
30. Rock outcrops and bedrock morphology impact on runoff and erosion processes in karst slopes
Core Problem: Karst-slope runoff and erosion are strongly shaped by surface rock cover and bedrock morphology, but their multi-layer hydrologic effects remain weakly resolved.
Key Innovation: Using controlled rainfall experiments, the paper separates how rock cover and undulating bedrock reorganize surface runoff, subsurface flow, and sediment yield on karst slopes.
31. Effect of net radiation on soil temperature field considering slope angle
Core Problem: Thermo-hydro-mechanical slope models usually approximate net radiation on horizontal surfaces and miss how slope angle changes thermal boundary conditions.
Key Innovation: The paper refines net-radiation estimation for inclined surfaces and shows that the improvement yields more realistic soil-temperature fields in slope models.
32. Model test study of polyurethane grouting repair for gas pipelines in mining subsidence areas
Core Problem: Ground subsidence in mining regions threatens buried pipelines, but trenchless repair methods are still weakly quantified under progressive deformation.
Key Innovation: The paper uses model testing to show how polyurethane grouting can restore support and reduce further damage to gas pipelines in subsiding mining areas.