Analytical and numerical solutions of ground-water withdrawals in the unconfined part of the Kirkwood-Cohansey aquifer system of the Coastal Plain of New Jersey were evaluated for their usefulness in predicting the area of influence of a pumped well and in determining hydraulic characteristics of an aquifer. Additionally, simulations of ground-water withdrawal using a finite-difference model provided information on the ways in which prudent well-location strategies can disperse the local effects of withdrawal over a larger part of an aquifer system. The design of a monitoring network that is sensitive to the ground-water hydraulics of streams and wetlands of the Coastal Plain of New Jersey also was considered for its utility in providing hydrologic data necessary to establish the baseline hydrologic conditions near wetlands and streams and in signaling when ground-water levels are being adversely affected by withdrawals elsewhere in the system.

The application of methods based on the Theis analytical solution to ground-water flow in unconfined aquifers can lead to erroneous estimates of the size of the area of influence generated by ground-water withdrawals. Analysis oftime-drawdown data from an unconfined aquifer system are best evaluated by means of the Neuman solution, which accounts for the effects of gravity drainage; however, the pumped well must be far enough from streams so that ground water is not drawn from nearby streams. Time-drawdown data from a test well in Winslow County, N.J., were analyzed by means of the Neuman solution. Results indicate that the aquifer has a relatively high vertical to horizontal anisotropy of 1:198, and a specific yield of 0.028, an indication that the area of influence of a pumped well at the test site would be relatively large.

Results from a finite-difference ground-water-flow model of the northeastern part of the Mullica River Basin near Chesilhurst, N.J., show that the area influenced by a long-term withdrawal is best estimated from a steady state ground-water-flow analysis that includes the effects of average areal recharge. Withdrawal simulations indicate an order-of-magnitude difference between the size of the area of influence generated from a 3-day (72 hour) withdrawal and the size of the area produced under steady-state conditions. An aquifer characterized by a low specific yield will cause the area of influence to extend farther away from the pumped well.

The contributing area of flow to the pumped well includes areas on the water table that would, under natural conditions, be incorporated into the contributing areas of flow to streams. Ground water that is drawn to a pumped well is diverted from nearby streams; the withdrawal decreases the size of the contributing areas of flow to streams by an amount equal to the contributing area of flow to the well.

Withdrawals made from a well close to a stream divert ground water that would, under natural conditions, flow to the stream. The diverted ground water causes the area of influence of the well to be smaller than it would if the well were far from the stream. Water-table declines caused by withdrawals near streams are, to some degree, mitigated by ground-water diversion from streams. However, the withdrawals can significantly reduce ground-water seepage to nearby streams, especially along stream reaches and wetlands close to the well. Alternatively, these effects can be dispersed over a large part of the aquifer if wells are located on surface-water divides. 

Measurements of seasonal water-level fluctuations in the Mullica River Basin indicate that the greatest fluctuations in water levels are found in upland areas, where the average fluctuation is 3.4 feet. Fluctuations in hydraulic head in the wetland areas averages 1.3 feet. The bimodal average of ranges in water levels show that upland areas are more sensitive to recharge than lowland areas. The pattern of yearly mean water levels fluctuates irregularly about a long-term mean value. Abnormally low or high yearly average values that are brought on by periods of drought or excess recharge are short lived; over time, hydrologic conditions shift back to average levels under natural conditions.

Wetland areas in the New Jersey Coastal Plain are characterized by ground-water seepage into wide, shallow depressions. Periods of inundation are longest in the deepest part of the depression, whereas inundation of areas near the fringes of wetlands due to ground-water seepage is only seasonal. The seepage face in the fringe areas expand and contract in response to seasonal variation in water-table elevation and in response to precipitation.

Values of the aquifer storage coefficient and transmissivity can, in some cases, be determined by use of hydraulic head or streamflow recession analysis as an alternative to aquifer testing. The recession curves developed from hydro graphs of Middle Branch and McDonalds Branch in the New Jersey Coastal Plain indicate that the aquifer near McDonalds Branch has about 2.6 times the storage capacity of the aquifer adjacent to Middle Branch; this finding is consistent with the relatively small ranges of water-level changes measured in McDonalds Branch compared to those measured in Middle Branch.