2005 Salt Lake City Annual Meeting (October 16–19, 2005)

Paper No. 16
Presentation Time: 1:30 PM-5:30 PM

USING GEOSTATISTICAL METHODS TO MAP SAND AND TILL IN A SUCCESSION OF ILLINOIAN GLACIAL DEPOSITS, IN KANE COUNTY, ILLINOIS


KEEFER, Donald A. and RITTENHOUSE, Sarah C., Illinois State Geological Survey, 615 E. Peabody Drive, Champaign, IL 61820, keefer@isgs.uiuc.edu

Geologists at the Illinois State Geological Survey have developed a computerized 3-D geologic map of deposits for Kane County in northeastern Illinois. Seven lithostratigraphic units were defined within the unconsolidated Quaternary and Holocene deposits. The 3-D map was developed by mapping the top and bottom surfaces of each lithostratigraphic unit using computer interpolation software. One of the oldest Quaternary formations in the county is a complex succession of Illinoian age glacial tills and sand and gravels called the Glasford Formation. Wells penetrating the Glasford are too sparse to reliably map the tops and bottoms of the various sand and till members, so we chose to model the sediment distribution using a geostatistical approach that did not rely on surface maps. This approach first involved classifying the materials within the Glasford Formation as either sand or till. These sand-till data were then analyzed to identify patterns in sand thickness and extent at various locations within the county. This analysis showed that sand occurred more frequently and that individual sand deposits were thicker and more continuous horizontally over bedrock valleys than over bedrock upland areas. To reliably model these distinct sand distributions, the county was divided into 2 zones: areas over bedrock valleys, and areas over bedrock uplands. Indicator semivariograms were developed, using observed data characteristics and insight gained from the geologic conceptual model, to describe the variations in sand frequency, thickness and horizontal continuity within these two zones. The semivariograms were used in a geostatistical interpolation technique called sequential indicator simulation to create 50 different yet equally-probable 3-D maps of sand and till distribution. Each of the 50 maps were constructed so that they matched the observed data values and honored the three criteria that we used to define geologic reasonableness: the frequency, thickness and horizontal extent of the sand deposits as described by the semivariograms. The 50 maps were sorted based on the total amount of sand they contained, and one map was selected as the most representative with respect to the geologic conceptual model. This map was inserted into the 3-D map of the remaining deposits to represent the Glasford Formation.