Skip to content

rocknrene/ProboSed

Repository files navigation

ProboSed

DOI License: MIT

Probabilistic Sediment Stability Pipeline for IODP Multi-Expedition Analysis

ProboSed is a Python pipeline for extracting Fabric Preservation Index (FPI) scores and process flags from IODP Visual Core Description PDFs, and for running stochastic slope stability models calibrated from those scores. It supports cross-margin comparison across passive and convergent margin settings using data from IODP Expeditions 308, 386, and 405.


Scientific Background

Governing Equations

Slope displacement $q(t)$ and fault slip $s(t)$ are modeled as coupled Ornstein–Uhlenbeck stochastic differential equations:

$$ds = -k_f \cdot s \, dt + \sigma_s \, dW_s$$ $$dq = (-\gamma q + \alpha s) \, dt + \sigma_q \, dW_q$$

Slope failure is defined as a first-passage problem:

$$\tau = \inf\{ t > 0 : q(t) > \theta(x) \}$$

where the failure threshold $\theta(x)$ is derived from the FPI score:

$$\theta(x) = \frac{2x}{3}, \quad x \in \{0, 1, 2, 3\}, \quad \theta \in [0, 2]$$

FPI + Process Flag System

FPI Label Physical meaning Failure threshold $\theta$
3 Intact Primary fabric fully preserved 2.00
2 Partially preserved Minor disruption, fabric recognizable 1.33
1 Fabric destroyed Scaly/shear surfaces, original fabric lost 0.67
0 Structureless No fabric, homogeneous or fluidized 0.00

Process flag: T = Tectonic · G = Gravitational · D = Depositional · F = Fluid escape · U = Undetermined

MTD-context rule: fault terms appearing within mass-transport context are downgraded T -> G. Faults within MTDs are soft-sediment features, not tectonic structures.

Physical Parameter Constraints

Parameter Value Physical basis
$\sigma_q$ 0.6 Frontal prism $V_p$ = 1550–1750 m/s (Exp 405 LWD, Unit I)
$\gamma$ 1.0 Moderate restoring force; 1.5–2.0 for seismically strengthened sediment (Exp 386)
$\alpha$ 0.5 $P_c'$ = 17 MPa overconsolidation supports non-zero coupling (Valdez et al. 2015)
slip_mag 3.0 Scaled from ~50 m coseismic slip at C0019 décollement (Fulton et al. 2013)

Repository Structure

ProboSed/
├── notebooks/
│   ├── ProboSed_VCD_FPI_Catalog.ipynb
│   ├── ProboSed_Stochastic_Slope_Stability_Model.ipynb
│   ├── ProboSed_Parameter_Sensitivity_Analysis.ipynb
│   ├── ProboSed_Porewater_Geochemistry_Profiles.ipynb
│   ├── ProboSed_MTD_Physical_Property_Quantification.ipynb
│   ├── ProboSed_XRF_Geochemical_Lithology_Profile.ipynb
│   ├── ProboSed_LWD_Physical_Property_Analysis.ipynb
│   ├── ProboSed_Failure_Probability_vs_Depth.ipynb
│   ├── ProboSed_Core_Image_Patch_Extractor.ipynb
│   ├── ProboSed_Block_Size_Distribution_Imagery.ipynb
│   ├── ProboSed_Block_Size_Distribution.ipynb
│   ├── ProboSed_Core_Image_Texture_Analysis.ipynb
│   ├── ProboSed_VCD_Labeler.ipynb
│   ├── ProboSed_Stochastic_Forcing_Model.ipynb
│   ├── ProboSed_XCT_Density_Integration.ipynb
│   ├── ProboSed_Resistivity_Image_Panels.ipynb
│   ├── ProboSed_Structural_Mask_vs_Stress.ipynb
│   ├── ProboSed_Sediment_Stability_Pipeline.ipynb
│   └── ProboSed_Pipeline_Summary_Figures.ipynb
│
├── slope/
│   ├── __init__.py
│   ├── stability.py            Core physics -- OU model, Lyapunov, sensitivity
│   ├── toy_model.py            Standalone simulation script
│   └── sensitivity_run.py      Sensitivity analysis script
│
├── core_ml/
│   ├── __init__.py
│   └── labeler.py              JCORESMiner, StraterMiner, Exp386Miner, VCDLabeler
│                               FPI_LEXICON, PROCESS_FLAG_LEXICON
│
├── geochem/
│   ├── __init__.py
│   └── geochem_analysis.py     Porewater and headspace gas depth profiles
│
├── transport/
│   ├── __init__.py
│   └── agents.py               Agent-based sediment transport ensemble
│
├── utils/
│   ├── __init__.py
│   └── patcher.py              Core image patch extraction utility
│
├── requirements.txt
├── pyproject.toml
├── CITATION.cff
├── LICENSE
└── README.md

Notebooks

All notebooks are designed to run in Google Colab without local installation. Each notebook imports from the slope/, core_ml/, geochem/, transport/, and utils/ modules -- the .py files are the source of truth and the notebooks are interactive wrappers.

Run order for dissertation analysis: 01 → 02 → 03 → remaining notebooks in any order.

Notebook Purpose Key outputs
VCD_FPI_Catalog Parses VCD PDFs from Expeditions 308, 386, and 405 using OCR (tesseract) and three expedition-specific miners (StraterMiner, Exp386Miner, JCORESMiner). Assigns a two-axis classification: Fabric Preservation Index (FPI 0--3) and process flag (T/G/D/F/U) to each depth interval. Builds the multi-expedition deformation facies catalog. Run before all other notebooks. fpi_catalog_all_expeditions.csv, fpi_catalog_exp308.csv, fpi_catalog_exp386.csv, fpi_catalog_exp405.csv, fpi_summary_stats.csv, deformation_facies_catalog.csv, FPI_expedition_comparison.pdf, FPI_depth_profiles.pdf
Stochastic_Slope_Stability_Model Runs the coupled OU slope stability ensemble under slope-only and fault-coupled forcing. Cross-margin edition: derives sigma_q from public Vp data (PANGAEA MSCL for M0081; local CSV for Exp 308) and runs separate ensemble pairs for Expeditions 308 (passive), 386 (convergent), and 405 (convergent). Produces trajectory figures, cross-margin failure probability comparison, and Lyapunov exponent estimates. slope_model_trajectories_crossmargin.png, slope_model_pfail_crossmargin.png, slope_model_lyapunov_crossmargin.png, slope_model_crossmargin_summary.csv
Parameter_Sensitivity_Analysis One-at-a-time sensitivity analysis across physically motivated parameter ranges. Cross-margin edition: runs separately for each site's BASE parameters; uses FPI-anchored threshold values; per-site slope-only baselines replace hardcoded reference. Tests robustness of qualitative result across alpha, sigma_q, slip_mag, gamma, and theta for all three margins. sensitivity_results_crossmargin.csv, sensitivity_robustness_crossmargin.csv, per-site sensitivity figures
Porewater_Geochemistry_Profiles Five-panel porewater and headspace gas depth profiles for C0019J and C0019M. Overlays deformation intervals from the FPI catalog when the catalog CSV is available. Three geochemistry figures
ProboSed_MTD_Physical_Property_Quantification Statistical comparison (Mann-Whitney U) of resistivity, Vp, and porosity inside and outside FPI-derived deformation intervals. Validates that FPI captures real mechanical contrasts. C0019J_MTD_statistics.csv, C0019J_MTD_properties.csv, depth profile and box plot figures
XRF_Geochemical_Lithology_Profile Identifies the mafic geochemical discontinuity (SiO2 < 50 wt%) at 635.94 m CSF-A. Four-panel integrated physical and structural profile. C0019J_XRF_profile.png, C0019J_integrated_profile.png
LWD_Physical_Property_Analysis Loads the C0019H LWD LAS file and runs Mann-Whitney U comparisons of resistivity and Vp between deformed and intact intervals. Statistical comparison table
Failure_Probability_vs_Depth Monte Carlo failure probability versus depth using vane shear strength and MAD data. Quantile-based depth binning for sparse sampling. Failure probability depth profile
Core_Image_Patch_Extractor Sliding-window patch extraction from ruler-free core section photographs. Classifies patches using the FPI deformation catalog. Patch archive, patch_manifest.csv
Block_Size_Distribution_Imagery Block boundary detection from core photographs using brightness and lateral-variance signal. Power law fit to cumulative block size distribution. C0019J_block_sizes_imagery.csv, distribution figures
Block_Size_Distribution Block boundary detection from MSCL colorimetry L* profiles. Power law fit. Block size distribution, power law fit
Core_Image_Texture_Analysis GLCM texture metrics (contrast, correlation, energy, homogeneity), orientation statistics, and autocorrelation length from core image patches. Texture metric depth profiles
VCD_Labeler Interactive ipywidgets labeling interface for core image patches. Records fabric type, process flag, confidence, and notes. C0019J_VCD_labels.csv
Stochastic_Forcing_Model Two-component SDE model with optional SRCMOD Tohoku finite-fault slip-rate forcing. Compares response-only and seismically forced ensemble failure probabilities. Ensemble figures, s_forcing.csv
XCT_Density_Integration DICOM CT Hounsfield unit profiles per deformation interval from XCT scan data. Density profiles by interval
Resistivity_Image_Panels RAB resistivity image panel extraction for MTD-9 and MTD-10 from Hole C0019H. Resistivity image panels
Structural_Mask_vs_Stress Structural overprint flags versus Conin et al. JFAST/JTRACK borehole breakout stress data. Structural-stress comparison figure
Sediment_Stability_Pipeline Legacy -- retained for provenance only. Early VCD lexicon parser and geotechnical merge pipeline using the original single-axis stability score (0--3). Superseded by ProboSed_VCD_FPI_Catalog, which implements the two-axis FPI + process flag system. Do not use as part of the active analysis pipeline. Stability depth profile (legacy)
Pipeline_Summary_Figures Graphviz workflow diagram and two-panel mechanical regime depth profile. Workflow diagram, regime figure

Scientific Applications

ProboSed is designed for research in:

  • Submarine landslides and mass-transport deposits
  • Earthquake-triggered sediment transport at convergent margins
  • Probabilistic slope stability modeling
  • Marine sediment routing systems
  • IODP core data integration with geophysical observations

Development uses core and logging data from the Japan Trench margin, IODP Expeditions 386 and 405, Site C0019.


Contributing

Contributions are welcome. Please open an issue to discuss proposed changes before submitting a pull request.


License

MIT License — see LICENSE for details.


Acknowledgments

Development of ProboSed is part of doctoral research at The Ohio State University, School of Earth Sciences.

Advisors Dr. Brendan Crowell — Co-Advisor Dr. Jill Leonard-Pingel — Co-Advisor

Dissertation Committee Dr. Cole · Dr. Keep · Dr. Krissek

Field Science Dr. Christine Regalla — Co-Chief Scientist, IODP Expedition 405

Data This work uses data from the International Ocean Discovery Program (IODP), Expeditions 386 and 405, Japan Trench margin.


Selected References

Chester, F.M., et al. (2013). Structure and composition of the plate-boundary slip zone for the 2011 Tohoku-oki earthquake. Science, 342(6163), 1208–1211.

Fulton, P.M., et al. (2013). Low coseismic friction on the Tohoku-oki fault determined from temperature measurements. Science, 342(6163), 1214–1217.

Valdez, R.D., et al. (2015). Data report: permeability and consolidation behavior of sediments from the northern Japan Trench subduction zone, IODP Site C0019. Proc. IODP, 343/343T.

Expedition 405 Scientists (2025). Site C0019. In Proc. IODP, 405.

Alstott, J., Bullmore, E., & Plenz, D. (2014). powerlaw: A Python package for analysis of heavy-tailed distributions. PLOS ONE, 9(1), e85777.

Haralick, R.M., Shanmugam, K., & Dinstein, I. (1973). Textural features for image classification. IEEE Transactions on Systems, Man, and Cybernetics, 3(6), 610–621.

Citation

If you use ProboSed, please cite:

Castillo, R. et al. (in prep). When Slip Happens: From Megathrust Rupture to Nonlinear Shallow Prism Reorganization in the 2011 Tohoku-Oki Earthquake. PhD dissertation, Ohio State University.

See also CITATION.cff for machine-readable citation metadata.

About

ProboSed: Bridging legacy core logs and modern machine learning. An open-source suite for probabilistic sediment characterization and nonlinear failure analysis in subduction margins.

Topics

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors