The Prairie du Chien Group was injected with 2,754,000 gallons (368,200 cubic feet), or 10,430 cubic metres, of municipally treated water at about 100 gallons per minute (13.4 cubic feet per minute), or 6.3 litres per second, for 20 days. The injection-pipe system was designed to utilize pipe friction rather than a remote-controlled valve in the well to maintain positive pressure and eliminate air entrainment in the injection water and the escape cf dissolved gasses from the water. During the 20-day injection period the temperature of the injection water declined gradually from 15.0° to 11.2°C, and the flow rate decreased from 108 to 90 gallons per minute (6.8 to 5.7 litres per second).

Analyses of test data were, in some instances, based upon hydrologic judgement as well as observations. Results of aquifer tests before and after injections indicated that the transmissivity had decreased 18 percent during the intervening injection periods; however, the specific capacity remained the same, indicating no change in transmissivity during pumping. Analysis of water-level changes in observation wells during injection indicated a reduction in transmissivity of more than 50 percent; however, the specific capacity of the injection well decreased only about 5 percent during injection.

A comparison of water-level changes with the discharge or recharge rates of the three tests showed that the water-level changes in the two observation wells tapping the Prairie du Chien Group during the injection test were greater than those projected from the two aouifer pumping tests. The deviations in the water-level changes and in the analysis of aquifer-test data indicate that the methods used to analyze data from these wells may not be wholly applicable, inasmuch as anisotropic and nonhomogeneous conditions prevail in at least the Prairie du Chien part of the aquifer.

The native water and the injected water averaged 0.8 and 25 milligrams per litre chloride, respectively. The chloride, utilized as a tracer, showed that the injected water was detected only in the lower part of the nearest observation well, 99 feet (30.2 metres) from the injection well. The chemistry of the water and the rock formation showed little likelihood of plugging of the recharge well by chemical precipitation. Microbiological phenomena apparently did not become a significant factor in the recharge test.

The hydraulic gradient of the aquifer in October and December 1971 (before and after injection) was estimated to be N. 36° E., 0.0013, and N. 39° E., 0.0012, respectively, on the basis of measurements of water levels in the three wells in the Prairie du Chien Group. The single-well tracer-dilution method of calculation showed a hydraulic gradient of 0.0016. A longitudinal dispersivity of 280 feet (85 metres) was calculated. Such a value of dispersivity is typical of fractured reservoirs and shows that the Prairie du Chien Group is a heterogeneous aquifer.

The injection test demonstrated that it is hydrologically feasible to recharge the Prairie du Chien Group and the Jordan Sandstone artificially through wells completed in the Prairie du Chien Group. The fissures in the Prairie du Chien Group act as conduits through which water spreads. The water passes into the Jordan Sandstone from the Prairie du Chien over a larger area than it would if it were injected directly into the Jordan.