Northeastern Section - 37th Annual Meeting (March 25-27, 2002)

Paper No. 0
Presentation Time: 2:45 PM

CHARACTERIZATION OF BEDROCK AQUIFERS IN CONNECTICUT (2) HYDROLOGIC SIMULATION


STARN, J. Jeffrey1, STONE, Janet R.2 and MONDAZZI, Remo A.2, (1)U.S. Geol Survey, 101 Pitkin Street, East Hartford, CT 06108, (2)U.S. Geological Survey, 101 Pitkin Street, East Hartford, CT 06108, jjstarn@usgs.gov

An aquifer test in Old Lyme, Conn. indicates that the fractured-rock aquifer is strongly heterogeneous; however, watershed-scale modeling indicates the aquifer can be treated as homogeneous and anisotropic. This result may be useful for regional ground-water models. The dominant feature producing anisotropy seems to be fractures along foliation-parallel partings, indicating that characterizing rock fabric can be useful in ground-water models.

Borehole geophysical data indicate several sets of fractures, including subhorizontal unroofing joints, steeply dipping foliation-parallel partings (striking N), and cross-cutting fractures (striking E-W). Outcrop data indicate that the foliation-parallel fractures are continuous but that E-W fractures are not. Drawdown measured in 15 bedrock wells and 9 piezometers during the aquifer test, and geophysical data collected in 4 wells, were used to construct a conceptual model of ground-water flow. One well seems to be directly connected to the pumped well through a highly permeable fracture. Drawdown in 4 wells is consistent with radial flow in a well-connected, subhorizontal fracture zone that receives water from an overlying source, such as the glacial deposits. Drawdown in the remaining wells is consistent with linear flow in less well-connected fractures.

A nonlinear regression ground-water-flow model was calibrated in a 2.4-mi2 area defined by watershed boundaries. The calibration data set consisted of water levels, measured by drillers, in 822 wells. Three alternative models were posed-a homogeneous and isotropic aquifer (the basic model), the basic model with horizontal anisotropy, and the basic model with a heterogeneous hydraulic conductivity distribution. Heterogeneity was created by defining model parameters for hydraulic conductivity in areas above and below 200 ft in altitude. Parameter estimates for each model are realistic, and the residuals of the regression appear unbiased. The anisotropic model seems to be slightly better than other models.

Evidence suggests greater hydraulic conductivity in the N-S direction than in other directions. The watershed-scale regression model estimates a 50:1 relation of the hydraulic conductivity N-S to E-W. A radial plot of aquifer diffusivity also corroborates the N-S direction.