USGS1999
The vertical hydraulic conductivity per unit thickness (streambed leakance) of unconsolidated sediment immediately beneath the channel of the Rockaway River near a municipal well field at Dover, N.J., is between 0.2 and 0.6 feet per day per foot and is probably near the low end of this range. This estimate is based on evaluation of three lines of evidence: (1) Streamflow measurements, which indicated that induced infiltration of river water near the well field averaged 0.67 cubic feet per second; (2) measurements of the rate of downward propagation of diurnal fluctuations in dissolved oxygen and water temperature at three piezometers, which indicated vertical Darcian flow velocities of 0.6 and 1.5 feet per day, respectively; and (3) chemical mixing models based on stable isotopes of oxygen and hydrogen, which indicated that 30 percent of the water reaching a well near the center of the well field was derived from the river. The estimated streambed-leakance values are compatible with other aquifer properties and with hydraulic stresses observed over a 2-year period, as demonstrated by a set of six alternative groundwater flow models of the Rockaway River valley. Simulated water levels rose 0.5 to 1.7 feet near the well field when simulated streambed leakance was changed from 0.2 to 0.6 feet per day per foot, or when a former reach of the Rockaway River valley that is now blocked by glacial drift was simulated as containing a continuous sand aquifer (rather than impermeable till). Model recalibration to observed water levels could accommodate either of these changes, however, by plausible adjustments in hydraulic conductivity of 35 percent or less. The ground-water flow models incorporate a new procedure for simulating areal recharge, in which water available for recharge in any time interval is accepted as recharge only where the water level in the uppermost model layer is below land surface. Water rejected as recharge on upland hillsides is allowed to recharge aquifers at the base of the hillsides. Inclusion of uplands in models of valley-fill aquifers and use of the new procedure increases model complexity and data requirements, but automates the simulation of recharge to those aquifers from the uplands, even in transient-state simulations with multiple periods of varied stresses, and facilitates delineation of upland areas that contribute water to well fields. The area from which ground water flowed toward the Dover well field decreased with an increase in simulated streambed leakance or an increase in simulated hydraulic conductivity of upland till. Concentrations of solutes in ground water near the Dover well field reflect the mixing of native ground water with water infiltrated from the Rockaway River. Chemical reactions in the aquifer, chiefly the weathering of carbonate minerals by dissolved carbon dioxide, affect the pH and the concentrations of both solutes and dissolved gases. Concentrations of sodium, chloride, and sulfate appear to be related to man's activities, such as road deicing, or to decay of organic matter in the aquifer.