TerraMosaic Daily Digest: October 1, 2026
Daily Summary
The October 1 issue is led by studies that preserve continuity across hazard stages. A rainfall-chain framework converts shallow-slope instability into explicit sediment supply and routes one evolving solid-water mixture, diagnosing the 2023 Beijing event as short, high-concentration pulses followed by rapid dilution rather than a sustained debris flow. After the 2020 Lake Fire, four years of observations and simulations show that seasonal hydraulic conductivity and grain size matter more for post-fire debris-flow likelihood than vegetation cover alone. A mine-slope study extends the same process view through five stages of damage under repeated excavation and wet-dry cycling.
Underground and cold-region studies turn hidden state changes into measurable thresholds. Six years of monitoring show how failure of a confined sandstone aquifer redistributes load above longwall panels; a karst-tunnel model and physical tests identify a water-pressure threshold near 1.4 MPa for the studied configuration; and true-triaxial tests isolate acoustic-emission precursors whose timing changes under saturation. Across subarctic riverbanks, water content dominates sediment strength while both spring thaw and autumn freezing contribute to erosion, indicating that shorter frozen seasons can widen the annual window of vulnerability.
Multi-hazard and flood papers focus on operational evidence. A Puerto Rico framework represents concurrent amplification, residual vulnerability and recovery between Hurricane Maria and later earthquakes. RHITA reconstructs European heat, cold, rainfall and wind events in space and time using open forecasts and a 1950-2024 reference archive. A forensic analysis of the 4 July 2025 Central Texas flood evaluates the full NOAA forecasting chain against high-water marks, while Hainan radar experiments show that the best precipitation-estimation architecture changes with rainfall intensity.
The methods collection is strongest where it exposes limits rather than only reporting accuracy. Differentiable hybrid hydrology constrains the process backbone so an LSTM cannot silently bypass it; block cross-validation quantifies how much spatial information models exploit; and the Spatio-Directional Variance Reduction Method handles anisotropic geotechnical variability from irregular samples. Foundation-model-assisted RGB-D reconstruction, multisource remote-sensing fusion and road segmentation broaden the observation toolkit, while title-level entries on freeze-thaw ground damage, tunnel displacement and lining delamination, earthquake loss and hurricane debris are retained without inferring unavailable results.
Key Trends
The day's research links evolving material state to hazard transitions, operational evidence and model accountability.
- Hazard chains are being modelled as continuous transitions: The Beijing framework carries rainfall, slope release, sediment concentration and flood routing through one evolving mixture, while the mine-slope study follows progressive cracking to abrupt translational failure.
- Seasonality and moisture state control failure windows: Post-fire debris-flow simulations distinguish wet- and dry-season conductivity, cold-region riverbanks fail during both thaw and freezing, and saturated sandstone changes strainburst intensity and acoustic precursors.
- Thresholds are being tied to physical failure mechanisms: Karst-tunnel water pressure, longwall-aquifer breakage and rockburst warning indicators are interpreted through coupled deformation, seepage, load transfer and fracture evolution rather than empirical alerts alone.
- Multi-hazard impacts retain memory of earlier events: Puerto Rico reconstruction shows that residual hurricane damage changes later earthquake loss, and RHITA plus the Guadalupe analysis frame operational hazard information through event history, lead time and impact evidence.
- Model skill is being checked against hidden shortcuts: Constrained hybrid hydrology tests whether the physical module remains meaningful, spatial block validation measures location-driven skill, and wildfire projections disclose calibration and recursive-input limitations.
Selected Papers
The October 1 selection centers on rainfall-induced landslide-to-flood chains, seasonal post-fire debris-flow likelihood and time-dependent mine-slope failure. It also covers riverbank erosion, mine-aquifer failure, karst-tunnel water inrush, acoustic precursors of strainburst, multi-hazard recovery, flash-flood forecast verification, extreme-weather tracking, fault-zone structure, tunnel deformation, earthquake loss and intensity-dependent precipitation estimation. Transferable methods are included when they address physical consistency, spatial validation, constrained remote-sensing inference or infrastructure monitoring; entries lacking reliable abstracts are explicitly marked.
1. An integrated modeling method for rainfall-induced geohazard chains: coupling landslide mobilization and flow-state transition
Core Problem: Mountain-hazard models often split slope failure, debris-flow transition and downstream flooding into separate calculations that do not conserve event timing and sediment supply.
Key Innovation: The framework converts hydrologically activated slope instability into time-, geometry- and volume-explicit sediment input and routes one evolving solid-water mixture; for the 2023 Beijing event it diagnoses short high-concentration pulses followed by rapid dilution, with about 3.29 × 10⁴ m³ released and only about 7% added by scour.
2. Seasonal influence on post-fire debris flow likelihood after the 2020 Lake Fire
Core Problem: Post-fire debris-flow likelihood declines with time after burning, but seasonal soil-state changes can alter risk independently of vegetation recovery.
Key Innovation: Four years of field and satellite observations after the 2020 Lake Fire constrain simulations showing that dry-season unsaturated hydraulic conductivity is about an order of magnitude higher than in the wet season and has more influence on likelihood than vegetation cover.
3. Time dependent failure evolution of counter tilted rock slopes subjected to mining disturbance and wet dry cycles
Core Problem: Counter-inclined rock slopes can accumulate damage under repeated mining and wet-dry cycles before failing abruptly through interacting toppling, seepage and locked-segment rupture.
Key Innovation: Physical modelling and discrete-element simulation resolve a five-stage evolution from collapse and tensile cracking to creep and translational slip, with mining stress redistribution and water-driven fracture deterioration forming a reinforcing failure pathway.
4. Riverbank Erosion Susceptibility Under Varying Sedimentological and Freeze-Thaw Conditions in Cold Climate Regions
Core Problem: Climate-driven changes in freezing and soil moisture alter when fine-grained high-latitude riverbanks become mechanically vulnerable.
Key Innovation: Three northern rivers are compared through geotechnical testing, in-situ thermal-moisture monitoring, aerial change detection and climate scenarios; water content dominates shear strength and erosion occurs during both spring thaw and autumn freezing.
5. Confined water migration and failure of the extra-thick sandstone aquifer above multi-longwall faces-a case study
Core Problem: Failure of an extra-thick confined aquifer above longwall panels changes both water migration and load transfer in ways that can amplify mine hazards.
Key Innovation: Six years of groundwater monitoring reveal phased water-level response and sharp changes when goaf geometry reaches square stages; drainage shifts more roof load to surrounding rock, increasing stress concentration and dynamic-disaster risk.
6. Study on the water inrush failure characteristics and disaster-inducing threshold of tunnels overlying karst caves under stress-strain-seepage coupling
Core Problem: Tunnels above water-filled karst caves require a defensible pressure threshold for the transition from seepage and deformation to catastrophic inrush.
Key Innovation: A stress-strain-seepage model calibrated from triaxial tests and checked against physical experiments identifies cave pressure as the primary control; the studied Shuangbei configuration becomes unstable at about 1.4 MPa.
7. Identification of acoustic emission precursors for strainbursts in dry and saturated sandstone under true triaxial transient unloading
Core Problem: Groundwater changes both failure intensity and the acoustic precursors used for rockburst warning.
Key Innovation: True-triaxial unloading tests combine high-speed imaging and acoustic emission; saturated sandstone shows reduced burst intensity and a distinct quiet period, while variance and Ib-value indicators provide the clearest preliminary precursors.
8. A quantitative methodology for analysing physical damage and recovery dynamics from concurrent and consecutive hazards: forensic insights from Puerto Rico
Core Problem: Damage from overlapping hazards depends on event order, residual vulnerability and recovery, which single-event loss models usually omit.
Key Innovation: A transparent Python framework represents concurrent amplification and path-dependent cumulative damage, then reconstructs Hurricane Maria and the 2019-2020 Puerto Rico earthquakes to show that ignoring unrepaired hurricane damage understates later losses.
9. RHITA: a web tool for real-time detection of extreme weather events
Core Problem: Operational users need rapid hazard characterization that is locally calibrated and comparable with a long historical record.
Key Innovation: RHITA clusters local-quantile exceedances into evolving space-time events, reports intensity, extent, duration and rarity, and operates with ECMWF open forecasts plus ERA5 records from 1950-2024.
10. Brief communication: Rise of the Guadalupe River - a multifaceted post event analysis of 4 July 2025, flood in Central Texas
Core Problem: The fatal 4 July 2025 Central Texas flood requires evaluation of how forecast skill and impact estimates changed with lead time, including at ungauged locations.
Key Innovation: The study evaluates NOAA precipitation, National Water Model streamflow and inundation products using 306 forecast maps and a USGS high-water-mark benchmark, keeping skill and impact performance separate.
11. Decadal Wildfire Risk Projection in Guangxi, China, Using a Two-Stage Deep Learning Framework with Recursive Environmental Factor Forecasting
Core Problem: Long-horizon wildfire mapping must forecast its environmental drivers without presenting recursively generated inputs as observed conditions.
Key Innovation: A two-stage FactorLSTM and LSTM-Transformer produces monthly 1 km conditional projections for Guangxi while explicitly reporting rare-event calibration problems, residual overestimation and the limits of a fixed land-cover assumption.
12. Estimation of Error Margin in the Quantification of Large Wood Volumes in River From In Situ and UAV Measurements
Core Problem: Large wood (LW) affects worldwide rivers by adding complexity to their flow structure, their sediment transport dynamics and by contributing to the spatial and temporal occurrences of fluvial morphologies.
Key Innovation: Thus, we were able to compare the volume measurements on 40 identical LW jams (LWJs) twice with only one month of interval. Our results shed light on the uncertainty of LWJ volume measurement and allow us to propose an equation to estimate the uncertainty of LW budget at various spatial and temporal scales.
13. Stress Field and Mantle Flow Patterns in the Ryukyu Subduction Zone From a New Moment Tensor Catalog
Core Problem: We present a new regional Ryukyu Area Centroid Moment Tensor (RACMT) catalog that documents detailed transitions in deformation pattern and geodynamic processes across the Ryukyu subduction system.
Key Innovation: We propose that efficient toroidal mantle flow around the western slab edge makes a major contribution to the westward increase in back-arc opening rate, fore-arc stretching, and widespread lateral extension within the deep PSP. The resulting focal mechanisms are broadly consistent with GCMT, while centroid depths for shallow earthquakes are significantly improved.
14. An RGB-D Approach Assisted by Foundation Models for Reconstructing Bedforms With Artificial Vegetation
Core Problem: Detecting sand beds in laboratory sediment experiments is challenging, especially when artificial vegetation or other structures are present.
Key Innovation: To address these challenges, we propose a hybrid workflow that transfers two-dimensional (2-D) masks from the Segment Anything Model to registered RGB-D point clouds, enabling automatic three-dimensional (3-D) separation of sand-bed and vegetation points. Across hundreds of monitoring points, we find low errors in scour depth and bed elevation, indicating that spatial variability is captured accurately.
15. In situ stress inversion method in reservoirs: an efficient framework based on coupled numerical simulation and reinforcement learning
Core Problem: Canadian Geotechnical Journal, Volume 63, Issue, Page 1-24, January 2026.
Key Innovation: To improve efficiency and accuracy, this study proposes an in situ stress inversion method integrating numerical manifold models, efficient surrogate models, and deep reinforcement learning. The proposed method provides an effective tool for mitigating geological hazards and optimizing exploitation, demonstrating strong potential for practical engineering applications.
16. Influence of plant species and root growth period on the mechanical behavior of loess soil in drained triaxial tests
Core Problem: Canadian Geotechnical Journal, Volume 63, Issue, Page 1-23, January 2026.
Key Innovation: Vegetation is increasingly used for shallow loess-slope stabilization, yet its growth-dependent reinforcement mechanisms remain insufficiently quantified. These findings provide a basis for incorporating vegetation growth periods and species selection into the engineering assessments of loess slopes, and for guiding the development of constitutive models.
17. Improving large-scale river routing models with ESA long-term CCI discharge data assimilation: advancing accuracy for climate studies
Core Problem: Large-scale hydrological models like CTRIP and MGB are essential for simulating river dynamics and supporting large-scale climate studies.
Key Innovation: To evaluate the added-value of these produtcs across contrasting modelling and hydrological contexts, we assimilate altimetry-derived discharge, multispectral-imagery-derived discharge, and WSE anomalies into two existing ensemble Kalman Filter frameworks: HyDAS in CTRIP, a global-scale physically based and uncalibrated river-routing model, and HYFAA in MGB, a calibrated semi-distributed regional hydrological model. Future work includes merging altimetry and multispectral discharge data, improving discharge retrieval algorithms using SWOT data, and refining data assimilation techniques to support climate studies and river system modeling in complex, climate-impacted basins.
18. Imaging the shear deformation zone along the Kefalonia transform fault zone
Core Problem: Fault traces alone do not reveal the three-dimensional width, segmentation and secondary activation of a seismically active deformation zone.
Key Innovation: Relocated aftershocks from the 2014 Kefalonia and 2015 Lefkada earthquakes yield reproducible inner and outer shear-zone widths and map coherent principal planes plus oblique relay structures along the eastern Mediterranean's most hazardous strike-slip system.
19. CM-S6: Cross-Modal Selective Scan via Parameter-Level Modulation for Multisource Remote Sensing
Core Problem: Multimodal remote sensing classification benefits from complementary spectral, structural, and elevation information.
Key Innovation: Although Mamba enables linear-complexity sequence modeling, multimodal Mamba methods usually fuse modalities before selective scanning, limiting the direct use of auxiliary information in parameter generation. Paired tests detected no significant OA difference from MCAMamba on Houston 2013 or from HLMamba and PICNet on Augsburg, while CM-S6 shows significant OA gains over MSFMamba on Augsburg and MUUFL Gulfport and over DSFormer on MUUFL Gulfport.
20. Few-Shot Remote Sensing Scene Classification Leveraging Collaboration of Foundation Models
Core Problem: Foundation models have shown remarkable generalization ability for few-shot learning (FSL).
Key Innovation: To address these issues, we propose a framework for few-shot remote sensing scene classification leveraging collaboration of foundation models. Recent multimodal large language models evaluated in the zero-shot setting are additionally reported as references, while the 2-shot results of our method illustrate the performance when limited labeled data are available.
21. Point-Based and Spatial Deep Learning for Radar Quantitative Precipitation Estimation Across Precipitation Intensity Ranges over Hainan Island
Core Problem: Radar QPE models can rank differently at low, moderate and high rain rates, making aggregate scores inadequate for severe-weather use.
Key Innovation: Point and spatial deep-learning models are compared by observed intensity; a point model leads overall, while the spatial CNN gives a modest but statistically supported RMSE improvement for high-intensity rainfall.
22. The Spatio-Directional Variance Reduction Method (SD-VRM): a robust framework for anisotropic scale of fluctuation from irregularly distributed geotechnical data
Core Problem: Characterizing horizontal soil spatial variability from irregular site investigations remains a challenge because conventional random field methods require linear, equidistant data.
Key Innovation: This study develops a generalized Spatio-Directional Variance Reduction Method (SD-VRM) framework combining global polynomial detrending and equivalent horizontal bin widths (ε h) to decouple non-uniform spacing constraints from authentic directional anisotropy. Ultimately, the proposed SD-VRM serves as a robust quantitative engineering tool to bridge irregular site characterizations with downstream stochastic mechanics, preventing the non-conservative overestimation of safety margins along preferential failure paths.
23. Spatial information explained by prediction models under block cross-validation
Core Problem: Spatial information in geographic data arising from the dependence between variable values and their locations fundamentally shapes the performance and evaluation of spatial prediction models.
Key Innovation: This study develops the spatial information explained (SIE) metric to assess the extent to which spatial prediction models explain spatial information during modelling, through quantifying the relative reduction in spatial information from a response variable to its prediction residuals. Specifically, KNN, the neural network, and XGBoost achieve the highest multi-scale SIE of approximately 0.50 together with high prediction accuracy, SIE rankings correlate weakly with accuracy rankings across the ten models, and random cross-validation overstates SIE relative to block cross-validation for all ten models, by 2.1% to 17.0% with a mean of 9.9%.
24. Satellite-based analysis of freeze-thaw driven ground damage using subsurface temperature and soil moisture
Core Problem: Title-level focus: Satellite-based analysis of freeze-thaw driven ground damage using subsurface temperature and soil moisture.
Key Innovation: The verified title, authors and DOI establish the study scope; methods, results and quantitative validation cannot be assessed because a reliable abstract was unavailable.
25. Forecasting adjacent-construction-induced radial displacement of MRT shield tunnels using dynamic deep learning
Core Problem: Adjacent construction can impose time-varying radial deformation on operating metro shield tunnels, requiring forecasts that respond to evolving construction conditions.
Key Innovation: The verified title, authors, PII and Crossref DOI establish a dynamic deep-learning forecast of construction-induced MRT tunnel displacement; model inputs, validation design and performance cannot be assessed because a reliable abstract was unavailable.
26. Non-contact acoustic-vibration identification and spatial delineation of multiple shallow subsurface delaminations in tunnel linings
Core Problem: Title-level focus: Non-contact acoustic-vibration identification and spatial delineation of multiple shallow subsurface delaminations in tunnel linings.
Key Innovation: The verified title, authors and DOI establish the study scope; methods, results and quantitative validation cannot be assessed because a reliable abstract was unavailable.
27. Evapotranspiration intensification will drive terrestrial water storage decline in the Turpan-Hami Basin
Core Problem: In the hyper-arid Turpan-Hami Basin (THB) of northwestern China, where water availability critically constrains socio-ecological sustainability, reliable hydrological characterization is essential yet insufficient.
Key Innovation: To overcome these dual challenges, this study developed a bias-corrected, fully coupled CWatM-MODFLOW model for the Turpan-Hami Basin and introduced a novel multi-objective calibration scheme. (5) Attribution and sensitivity analyses demonstrate that the annual responses of dS to PRE and ET ( bP more than 0, bET < 0) are approximately symmetric and opposite in sign, with ∣ b E T ∣ · Δ E T dominating dS under stronger forcing; monthly diagnostics indicate a negative lag from ET to TWSA.
28. Constrained differentiable hybrid hydrologic modeling across multiple process backbones: accuracy gains and risks to physical consistency
Core Problem: Hybrid hydrologic models can achieve strong streamflow scores while the learned postprocessor bypasses a poorly calibrated physical backbone.
Key Innovation: Across eight conceptual models, an explicit loss on the process module improves its realism without sacrificing final skill; the best constrained hybrid reaches median NSE 0.742 versus 0.724 for the LSTM and shows larger gains in arid basins.
29. Earthquake-induced loss assessment of steel frame buildings with buckling-restrained braced frames subjected to spectrally equivalent long- and short-duration ground motions
Core Problem: Spectrally equivalent ground motions with different durations may produce different damage and loss outcomes in buckling-restrained braced steel frames.
Key Innovation: The verified title, authors, PII and Crossref DOI establish a comparative loss assessment under long- and short-duration motions; structural models, loss components and quantitative findings cannot be assessed because a reliable abstract was unavailable.
30. Debris-Cast: A Two-Stage Machine Learning Framework for Hurricane Debris Prediction Using Multi-Source Geospatial Data
Core Problem: Title-level focus: Debris-Cast: A Two-Stage Machine Learning Framework for Hurricane Debris Prediction Using Multi-Source Geospatial Data.
Key Innovation: The verified title, authors and DOI establish the study scope; methods, results and quantitative validation cannot be assessed because a reliable abstract was unavailable.
31. Crustal and lithospheric mantle thinning variations beneath the East African Rift System
Core Problem: Knowledge gaps remain regarding strain localization mechanisms that may initiate the transition to late-phase rifting and break-up along active rifts hosted in thick continental lithosphere.
Key Innovation: Combined with estimated crustal thicknesses, we estimate the thinning factors of the crust and lithospheric mantle, as well as the deformation decoupling parameter, defined as the ratio of the lithospheric mantle to the crustal thinning factors. Our results reveal intriguing magnitudes of lithospheric mantle necking beneath both the eastern (magma-rich) and western (magma-poor) branches.
32. Undrained adhesion of consolidated mudstone muck under varying water contents and stress histories for shield tunnelling
Core Problem: Canadian Geotechnical Journal, Volume 63, Issue, Page 1-18, January 2026.
Key Innovation: To investigate the effects of effective consolidation stress, water content, and stress history on adhesion strength, consolidated undrained rotary shear tests were conducted on saturated moderately weathered mudstone muck using a modified rotary shear apparatus. For overconsolidated muck, unloading resulted in progressively faster decreases in adhesion strength as effective consolidation stress decreased, while reloading caused progressively faster increases as effective consolidation stress rose.
33. A wave-based model for predicting K0 values
Core Problem: Canadian Geotechnical Journal, Volume 63, Issue, Page 1-8, January 2026.
Key Innovation: A more reliable alternative is to use shear-wave velocity ([math]), which is measured at near-zero strain, to determine K0. The model showed reasonably good performance across a range of soil types and geological histories, including different ages and isotropic over-consolidation ratios (R0).
34. Nonlinear dynamic response analysis for hub ship lock structures under seismic action
Core Problem: The ship lock structures are responsible for enabling vessels to navigate between different water levels on coastal rivers and canal waterways.
Key Innovation: With the development of marine water transportation, the risk of earthquakes on ship lock structures has gradually increased. The findings are expected to provide awareness of considering corrosion effects in the life-cycle design and seismic analysis of ship lock structures.
35. A Review of Remote Sensing Image Translation Based on Generative Adversarial Networks
Core Problem: Generative adversarial networks (GANs) have become an important approach for translating remote sensing images across sensing modalities, acquisition domains, and degraded observation conditions.
Key Innovation: The review also examines the use of translated imagery in change detection, object detection, semantic segmentation, and data augmentation. The literature shows a strong concentration on SAR-to-optical translation, accompanied by growing use of multiscale feature extraction, attention mechanisms, contrastive learning, cross-modal alignment, and pretrained representations.
36. From Prior to Frequency: A Progressive Structure Recovery Network for Remote Sensing Road Segmentation
Core Problem: Roads in remote sensing imagery often contain narrow branches, complex boundaries, and long-range connections.
Key Innovation: To address these challenges, we propose a Progressive Structure Recovery Network (PSRNet) for remote sensing road segmentation. Experiments on the Massachusetts Roads, DeepGlobe, and CHN6-CUG datasets demonstrate the effectiveness of PSRNet for road segmentation.
37. Three-dimensional seismic isolation analysis of LNG storage tanks incorporating vertical isolation devices and friction pendulum bearings
Core Problem: Liquid-filled tanks are effective storage infrastructure for water, oil, and liquefied natural gas (LNG).
Key Innovation: In this work, the focus is given on sliding-based systems, namely, single friction pendulum bearings (SFPBs), triple friction pendulum bearings (TFPBs), and mainly on the recently developed quintuple friction pendulum bearings (QFPBs). In addition, a surrogate model has been developed for simulating the dynamic response of the superstructure (tank and liquid content) to achieve an optimal balance between computational efficiency and accuracy.