CREATING QUANTITATIVE PATHWAYS TO CONNECT OUTCROP ANALOGS TO RESERVOIR PREDICTION THROUGH DIGITIZED STRATIGRAPHIC MEASURED SECTIONS

Stratigraphic measured sections capture observations from outcrops including bed thicknesses, observed grain size, sorting and shape, and sedimentary structures. These observations are the foundation for interpreting depositional processes, environment, and building predictive models of sediment distribution within sedimentary basins. These models are used for applications in petroleum resource exploration and extraction, carbon dioxide sequestration, and will be invaluable in developing methods for subsurface energy storage. Field data are transferred from a field notebook through hand drafting. Newer methods (e.g., the NSF funded StraboSpot project) are working toward the goal of collecting data digitally in the field. However, most of the existing stratigraphic measured sections have not yet been digitized, limiting their availability for use in newer modeling and data science methods.

Over a decade of work in the outcropping deepwater slope deposits of Cretaceous, Tres Pasos Formation of the Magallanes Basin has produced >10,000 meters data in > 300 measured sections. The outcrops reveal >100’s kms of slope to basin deepwater strata that are often >100’s meters thick at a single location. Consistent data collection across all study areas and the use of a common template for drafting ensured uniformity in translation of the data from field notebooks to paper archive. These stratigraphic measured sections are being digitized and inserted into a relational database developed specifically for this purpose. The database has multiple tables that capture grainsize data, bed thicknesses, and stratigraphic interpretations for each measured section. The measured section database is contextually linked to a document webserver to provide contextual information (i.e., manuscripts, maps, correlation panels, and images). The database produces pathways to translate field data, observations, and interpretation into a tangible, testable, and shareable form. It is ideal for development and testing of machine learning approaches to better predict sedimentologic architecture in the subsurface.