|Abstract:||The Elm aquifer, which consists of sandy and gravelly glacial-outwash deposits, is present in several counties in northeastern South Dakota. An aquifer test was conducted northeast of Aberdeen during the fall of 1999 to determine the hydraulic properties of the Elm aquifer in that area. An improved understanding of the properties of the aquifer will be useful in the possible development of the aquifer as a water resource.
Historical water-level data indicate that the saturated thickness of the Elm aquifer can change considerably over time. From September 1977 through November 1985, water levels at three wells completed in the Elm aquifer near the aquifer test site varied by 5.1 ft, 9.50 ft, and 11.1 ft. From June 1982 through October 1999, water levels at five wells completed in the Elm aquifer near the aquifer test site varied by 8.7 ft, 11.4 ft, 13.2 ft, 13.8 ft, and 19.7 ft. The water levels during the fall of 1999 were among the highest on record, so the aquifer test was affected by portions of the aquifer being saturated that might not be saturated during drier times.
The aquifer test was conducted using five existing wells that had been installed prior to this study. Well A, the pumped well, has an operating irrigation pump and is centrally located among the wells. Wells B, C, D, and E are about 70 ft, 1,390 ft, 2,200 ft, and 3,100 ft, respectively, in different directions from Well A. Using vented pressure transducers and programmable data loggers, water-level data were collected at the five wells prior to, during, and after the pumping, which started on November 19, 1999, and continued a little over 72 hours.
Based on available drilling logs, the Elm aquifer near the test area was assumed to be unconfined. The Neuman (1974) method theoretical response curves that most closely match the observed water-level changes at Wells A and B were calculated using software (AQTESOLV for Windows Version 2.13-Professional) developed by Glenn M. Duffield of HydroSOLVE, Inc. These best fit theoretical response curves are based on a transmissivity of 24,000 ft2/d or a hydraulic conductivity of about 600 ft/d, a storage coefficient of 0.05, a specific yield of 0.42, and vertical hydraulic conductivity equal to horizontal hydraulic conductivity.
The theoretical type curves match the observed data fairly closely at Wells A and B until about 2,500 minutes and 1,000 minutes, respectively, after pumping began. The increasing rate of drawdown after these breaks is an indication that a no-flow boundary (an area with much lower hydraulic conductivity) likely was encountered and that Wells A and B may be completed in a part of the Elm aquifer with limited hydraulic connection to the rest of the aquifer.
Additional analysis indicates that if different assumptions regarding the screened interval for Well B and aquifer anisotropy are used, type curves can be calculated that fit the observed data using a lower specific yield that is within the commonly accepted range. When the screened interval for Well B was reduced to 5 ft near the top of the aquifer and horizontal hydraulic conductivity was set to 20 times vertical hydraulic conductivity, the type curves calculated using a specific yield of 0.1 and a transmissivity of 30,200 ft2/d also matched the observed data from Wells A and B fairly well.
A version of the Theim equilibrium equation was used to calculate the theoretical drawdown in an idealized unconfined aquifer when a perfectly efficient well is being pumped at a constant rate. These calculations were performed for a range of pumping rates, drawdowns at the wells, and distances between wells that might be found in a production well field in the Elm aquifer.
Although the aquifer test indicates that hydraulic conductivity near the well may be adequate to support a production well, the comparison of drawdown and recovery curves indicates the possibility that heterogeneities may limit the productive capacity of specific loca