A statistical framework for automated stratigraphic interpretation from drillhole lithology data with uncertainty quantification.
Many legacy drillhole datasets contain only lithological descriptions without stratigraphic information, limiting their utility for regional 3D geological modeling. litho2strat addresses this challenge by:
- Systematically exploring all geologically plausible stratigraphic orderings for individual drillholes
- Correlating solutions across multiple drillholes to identify geologically consistent interpretations
- Combining constraints from lithological descriptions with stratigraphic relationships from regional maps
- Quantifying uncertainty by generating multiple plausible stratigraphic interpretations
This approach enables the transformation of lithology-only drillhole logs into stratigraphic sequences with quantified uncertainty, providing valuable constraints for subsurface modeling, resource estimation, and data acquisition planning.
If you use this code, please cite:
Ogarko, V. and Jessell, M.: Automated stratigraphic interpretation from drillhole lithological descriptions with uncertainty quantification: litho2strat 1.0, Geosci. Model Dev., 19, 1007–1025, https://doi.org/10.5194/gmd-19-1007-2026, 2026.
This paper provides a detailed explanation of the methodology implemented in this code.
numpy, matplotlib, networkx
pip install .python3 litho2strat.py -p <Parfile>python3 litho2strat.py -p ./parfiles/Parfile_SA.txtThe graph is obtained by running: pydeps litho2strat.py --max-bacon 3 --exclude matplotlib numpy networkx