GEOSTATISTICAL ANALYSIS OF GLACIAL HYDROGEOLOGY: EXTRAPOLATING FROM DATA-RICH SITES TO DATA-POOR SITES

Glacial depositional and erosional processes produce complex hydrogeological facies. Borehole data provide subsurface information that is vital to investigations ranging from local contaminated site assessments to regional water resources studies. However, data in either case are typically sparse, and hole-to-hole correlation of units may be difficult.

Evaluating the complexities and associated uncertainties in unit correlations may be performed through 3-D analysis of the data set's covariance and/or transition probabilities. Several data-rich Midwestern study areas show clear structure in 3-D variogram analyses of their aquifer/aquitard frameworks, as well as in their transition probabilities of multiple material types. One of these sites has 700+ boreholes per square km with sampling by continuous split spoons. This high-quality, high-density dataset overcomes the usual problem of studies relying on borehole data: too few data in the horizontal direction.

Geostatistical analysis of these data-rich sites provide a means of 1) determining the data density appropriate for understanding spatial hydrogeological structure, and 2) assessing uncertainty and variability in regional studies relying on large databases of relatively sparse drilling data.