Residents of the southern New Jersey Coastal Plain are increasingly reliant on the unconfined Kirkwood-Cohansey aquifer system for public water supply as a result of increasing population and restrictions on withdrawals from the deeper, confined aquifers. Elevated nitrate concentrations above background levels have been found in wells in the surficial aquifer system in agricultural and urban parts of this area. A three-dimensional steady-state ground-water-flow model of a 400-square-mile study area near Glassboro, New Jersey, was used in conjunction with particle tracking to examine the effects of land use and travel time on the distribution of nitrate in ground and surface water in southern New Jersey.
Contributing areas and ground-water ages, or travel times, of water at ground-water discharge points (streams and wells) in the study area were simulated. Concentrations of nitrate were computed by linking land use and age-dependent nitrate concentrations in recharge to the discharge points. Median concentrations of nitrate in water samples collected during 1996 from shallow monitoring wells in different land-use areas were used to represent the concentration of nitrate in aquifer recharge since 1990. On the basis of upward trends in the use of nitrogen fertilizer, the concentrations of nitrate in aquifer recharge in agricultural and urban areas were assumed to have increased linearly from the background value in 1940 (0.07 mg/L as N) to the 1990 (2.5-14 mg/L as N) concentrations.
Model performance was evaluated by comparing the simulation results to measured nitrate concentrations and apparent ground-water ages. Apparent ground-water ages at 32 monitoring wells in the study area determined from tritium/helium-3 ratios and sulfur hexafluoride concentrations favorably matched simulated travel times to these wells. Simulated nitrate concentrations were comparable to concentrations measured in 27 water-supply wells in the study area. A time series (1987-98) of nitrate concentrations at base-flow conditions in three streams that drain basins of various sizes and with various land uses was compared to simulated concentrations in these streams. In all three of the streams, a reasonable fit to the measured concentrations was achieved by multiplying the simulated concentration by 0.6. Because nitrate appeared to move conservatively (not degraded or adsorbed) in ground water to wells, the apparent non-conservative behavior in streams indicates that about 40 percent of the nitrate in aquifer recharge is removed by denitrification in the aquifer near the streams and (or) by in-stream processes.
The model was used to evaluate the effects of various nitrate management options on the concentration of nitrate in streams and water-supply wells. Nitrate concentrations were simulated under the following management alternatives: an immediate ban on nitrate input, reduction of input at a constant rate, and fixed input at the current (2000) level. The time required for water to move through the aquifer results in a time lag between the reduction of nitrate input in recharge and the reduction of nitrate concentration in streams and wells. In the gradual-reduction alternative, nitrate concentrations in streams and wells continued to increase for several years after the reduction was enacted. In both the immediate-ban and gradual-reduction alternatives, nitrate concentrations remained elevated above background concentrations long after nitrate input ceased. In the fixed-use alternative, concentrations in streams and wells continued to increase for 30 to 40 years before reaching a constant level.
The spatial distributions of simulated nitrate concentrations in streams in 2000 and 2050 were examined with the assumption of no change in land use, nitrate concentration in recharge, or ground-water withdrawals. As expected, nitrate concentrations were highest in agricultural areas and lowest in largely undeveloped areas. Changes in concentration