Multi-Sourced Geologic Data Integration: A Time-Based Approach

Abstract

Complications in geological modeling are often made worse by the wide range of formats in which geologic data are stored. When crossing state lines, the names and interpretations of geologic formations may vary, individual formations can be grouped into one undifferentiated member, and the very existence of a formation may disappear when crossing a neighboring state border. The next generation of geologic mapping requires higher coordination in dealing with multiple datatypes. In order to ameliorate the complications associated with geologic data integration and modeling, I used a 4D time-based approach to integrate multiple different data sources. The format for data integration contained four columns, having an x and y coordinate in UTM, a z-dimension elevation value in meters and geologic contact age for each data point. The data were extracted from the Geological Map of North America (GMNA) for surface-bedrock records, while subsurface formation thicknesses were provided by drillers logs from at least 150 boreholes and wire-logs across several counties, companies and subsurface exploration projects in Michigan. The size, lack of major faulting and nearly ideal basin shape of the Michigan Basin allowed it to be a reasonable sample site to test this integration method and its ability to produce plausible and highly interactive 3D images of the subsurface. Post-integration, two methods of geostatistical kriging were performed in the SGeMS software: one in 3D with the dependent variable of interest being geologic time, the other in 2D with respect to the elevation. Further statistical analyses were performed in MATLAB to compare the plausibility of each method. This work will demonstrate approaches to large scale 3D data integration and mapping that can be used as a framework for the next generation of computer-based geologic modeling.