The U.S. Geological Survey is studying the geologic, hydrologic, and geochemical properties of the glacial drift and underlying bedrock at a low-level radioactive-waste burial site in the Palos Forest Preserve, 22 kilometers southwest of Chicago. Data collected from the 33 test wells drilled into the drift plus data from 4 wells drilled into the underlying dolomite bedrock are presented. Data include maps showing the location of the test wells, a general description of the drift, well-construction information, and lithologic descriptions of cores from the wells finished in the drift. INTRODUCTION In early 1943, the U.S. Army Corps of Engineers leased land from the Cook County Forest Preserve District to build a radioisotope research facility. The facility formed part of the Metallurgical Laboratory operated by the University of Chicago for the Manhattan Engineer District (the Manhattan Project). The facility, known as Site A (fig. 1), housed the world's first nuclear reactor. Plot M was the burial site for the low-level radioactive waste resulting from operations at the Metallurgical Laboratory. The Argonne National Laboratory (ANL) has conducted a radiological monitoring program at Plot M since 1948. In 1954, at the request of ANL, the U.S. Geological Survey (USGS) studied the geologic, hydrologic, and chemical characteristics of the glacial drift at Plot M (W. J. Drescher, written commun., 1954) and advised ANL on its environmental monitoring program. In 1973, tritium was detected in water from a dolomite bedrock well at the Red Gate Woods picnic area. Consequently, the ANL monitoring program was expanded to include the Red Gate Woods well and other nearby forest preserve wells. This study of groundwater flow and tritium migration at Plot M began in 1978. The two main objectives of the study are to determine the geologic, hydrologic, and geochemical properties of the drift that control the migration of tritium from the burial site to the underlying bedrock and to determine the 80 160 FEET i 20 40 METERS *DH2 21 EXPLANATION Test wells in glacial drift Ar Test wells in dolomite bedrock. _ '204-Contour interval 1.5 meters National Geodetic Vertical batum of 1929 Figure 2. The Plot M burial site showing test-well locations, topography, and intermittent streams.

No abstract

Ground-water withdrawals by municipal wells in the Mattapoisett River valley are expected to triple in the next two decades. State and local concern about the long-term impacts of these increased withdrawals on ground-water levels and streamflow made it necessary to assess the ground-water resources of the valley and to develop a digital ground-waterflow model for management purposes.The model was calibrated under steady-state and transient conditions and simulates ground-water withdrawals by wells, leakage through streambeds, and leakage from the bordering till. Calculated results of the model are most sensitive to decreases in the values of model parameters, particularly streambed and aquifer hydraulic conductivity. Transient-responsetime tests of the model indicate that changes in long-term recharge rates would have to last at least 1 year for steady-state predictions to be realized.Ten pumping scenarios representing current and proposed withdrawals from the valley were simulated with conditions of reduced recharge. Under conditions simulating 1965 average annual recharge, predicted water levels in the aquifer are as much as 9 feet lower than average annual levels. At the highest withdrawal rates, the predicted drawdown in four wells exceeds the estimated available drawdown. For all pumping scenarios, at least 10 percent of the available ground water in the aquifer discharges to the Mattapoisett River. Under conditions representative of the 7-day 10-year low flow of the river, predicted water levels decline as much as 19 feet; moreover, at the highest withdrawal rates, available drawdown is exceeded in five wells. Simulated withdrawals in six scenarios use all of the available ground-water discharge. If this drought condition should occur and streamflow is not supplemented by surface water, the predictive results indicate that the downstream half of the river will stop flowing under most pumping plans.Test drilling and seismic refraction surveys conducted to aid model development indicate that the bedrock surface generally is flat except for a deep, narrow channel in the center of the valley. Continuous stream-stage data and baseflow data for the Mattapoisett River were used to increase previous estimates of flow duration, 7-day 2-year, and 7-day 10-year low flow. Water quality in both the aquifer and river may be characterized as slightly acidic and low in dissolved solids.

No abstract

The area around wells G and H, two former publicsupply wells for the city of Woburn, Massachusetts and currently designated as a U.S. Environmental Protection Agency "Superfund" site, was the focus of intensive hydrogeologic investigations from 1983 to 1988. The U.S. Geological Survey has provided assistance to the U.S. Environmental Protection Agency for the site since 1985. This report includes hydrogeologic information and describes a three-dimensional, digital groundwater flow model that was designed and calibrated by the U.S. Geological Survey for use by the U.S. Environmental Protection Agency to evaluate alternative pumping scenarios in developing an aquifer cleanup strategy. Wells G and Hand two nearby industrial-supply wells were constructed in a stratified-drift aquifer ranging in width from 0.5 to 1 mile and as much as 90 feet thick. The transmissivity of the aquifer in the vicinity of wells G and H ranges from ll,500to 14,000 feet squared per day, and the aquifer can sustain well yields of as much as 700 gallons per minute. Recharge to the aquifer is from precipitation. Under normal conditions, with only the industrial-supply wells pumping, groundwater discharges to the stream in most of the study area, and the river is a gaining stream throughout the year. When wells G and H and the industrial-supply wells are pumped simultaneously, infiltration of surface water significantly decreases streamflow in the area. A three-dimensional, digital groundwater flow model of the stratified-drift aquifer in the vicinity of wells G and H was designed and calibrated. The model represents a 0.8'-square-mile area and consists of nearly 5,000 active nodes in three model layers. Model gridspacing ranges from 20 x 20 feet to 200 x 200 feet. The model was calibrated to steady-state and transient conditions for December 1985 and January 1986. Throughout all model layers in the center of the model area, simulated hydraulic heads matched observed and estimated hydraulic heads to within 1 foot. Throughout the remainder of the model area, hydraulic heads matched within 5 feet except in some corners and sides of the active model area near tillbedrock boundaries. Under steady-state conditions, the simulated gain in streamflow in the model area was 0.27 cubic feet per second, which is within the range of observed gains in streamflow (0.10 to 0.62 cubic feet per second) measured during 1985 low-flow conditions. Under transient conditions, simulated streamflow losses in the Aberjona River were 1.25 cubic feet per second compared to a measured loss of 1.26 cubic feet per second at the end of a 30-day aquifer test during which withdrawals averaged 3.05 cubic feet per second. Sensitivity tests of the calibrated model were conducted to determine if the differences between simulated and observed/estimated data values could be attributed to the range of uncertainty in the values of input data and boundary conditions. Test results indicate that the model is least sensitive to variations in model boundary conditions, river stage, and rech...