TOWARDS A DIGITAL CRUST: EXPERIENCES FROM A LONG-TERM, 3-D GEOLOGIC MAPPING PROGRAM

Geologists at the Illinois State Geological Survey have been building county-sized, 3-D geologic maps of the upper 50-300 meters for over 20 years. The driving economic force for these studies has been characterization for groundwater flow modeling and prudent water resource management. Experiences from this program can provide insights relevant to continental and sub-continental scales of 3-D geologic modeling.

Early priorities need to be an understanding of data quality, and a plan for managing and integrating the range of data types to be used in the modeling. A relational database can be used to standardize descriptive and interpretive terms. This simplifies 3-D visualization and allows for an interactive evaluation of trends. Data structures evolved to include relational databases, standardized file structures, and open-standard data formats. Interactive interpretation was enabled using a dynamic suite of 2-D and 3-D software.

Development and description of the conceptual model for the geologic system being mapped is critical and can be documented using published reports, README files and formal metadata. A description of what is known about the depositional environments, and the tectonic, structural and diagenetic histories can be used to create a spatial description of the system being modeled. Application of 3-D maps to fluid flow simulations suggests that hierarchical models of basin and stratal architecture provide a valuable structure for nesting insight on scale-dependent variations in material properties. Other key items for documentation include what is known about modeled deposits, how it is known, and what is surmised. Descriptions of the potential for interpretation errors, where and why these errors are expected to occur, and the potential consequences of these errors on subsequent simulations are valuable to fluid flow modelers.

Other insights from these projects include insight on how interpolation can introduce predictable sources of error to models, an understanding that proprietary data structures are ephemeral and lead to loss of data, and insight that a robust approach for long-term management of 3-D geologic models is to convert individual mapped surfaces to networks of individual data points managed in a relational database system.