Geohydrology and Simulation of Groundwater Flow in the Northern Atlantic Coastal Plain Aquifer System
By P. Patrick Leahy and Mary Martin
MODEL VERSION/TYPE: Modified version of the Trescott model, transient
AREA STUDIED: Northern Atlantic Coastal Plain (New York, New Jersey, Delaware, Maryland, Virginia, North Carolina)
AQUIFERS SIMULATED: 10 regional Coastal Plain aquifers
MOST RECENT WITHDRAWALS SIMULATED: 1980
MODEL SIZE: 11 layers, 85 rows, 32 columns
MINIMUM GRID SPACING:7 miles x 7 miles
MODEL ARCHIVE is available by email request at gs-nj-model-request@usgs.gov
REPORT ABSTRACT:
A groundwater flow model of the northern Atlantic Coastal Plain was designed and calibrated to increase understanding of the regional flow regime and changes in flow brought about by human activities. A multilayer flow model simulated groundwater flow in an aquifer system covering a 95,000-mi2 (square miles) area that included Long Island, N.Y., and the Coastal Plain of New Jersey, Delaware, Maryland, Virginia, North Carolina, and parts of the adjacent Continental Shelf. The aquifer system consists of a seaward-thickening sedimentary wedge of 10 regional aquifers (predominantly sand) separated by 9 confining units (silt and clay).
Prepumping and pumping (1900 to 1980) conditions were calibrated through trial-and-error adjustment of hydraulic characteristics until model-simulated heads and flows matched those measured or estimated. Hydrographs from 74 wells were used to calibrate the flow model through time. Model calibration also was done at a local scale using subregional models for North Carolina, Virginia, Maryland-Delaware, and New Jersey. The hydraulic characteristics for the regional and subregional models were compatible.
Areal distributions of aquifer transmissivity and confining-unit leakance were refined through calibration. Transmissivity was highest, more than 100,000 ft2/d (feet squared per day), in parts of the Castle Hayne aquifer in North Carolina. Transmissivity generally ranged from 500 to about 10,000 ft2/d for most aquifers in the system. Confining-unit leakance ranged from about 1.0 x 10-8 to 1.0 foot per day per foot. Lower values typically were found in the deeper units. Updip leakance values tended to be higher because of a decrease in thickness or an increase in vertical hydraulic conductivity of the confining unit. Sensitivity of the model response to hydraulic characteristics, confining-unit transient leakage, and the nature and location of the seaward model boundary was determined. Values of transmissivity, storage coefficient withdrawals, and confining-unit leakance were changed by varying amounts dependent on a subjective estimate of their uncertainty. Model heads were most sensitive to an order of magnitude change in confining-unit leakance and to confining-unit transient leakage. Additional data are needed to determine the importance of transient leakage in the hydraulics of the aquifer system. A variable-density flow model evaluated the sensitivity of the model to the seaward boundary. This model simulated flow in parts of the aquifer system that contained water having chloride concentrations greater than 10,000 mg/L (milligrams per liter). Along the 10,000-mg/L-chloride-concentration line, head differences between the calibrated constant-density flow model and the variable-density flow model generally were less than 10 ft (feet). Thus, probably only a small percentage of the water pumped from the system was derived from areas containing salty water.
Characteristics of the regional flow system described in the report are based on an analysis of the model-derived: (1) water budgets, (2) potentiometric surfaces, (3) vertical leakage between aquifers, and (4) lateral flow patterns and velocities within aquifers. Average areal groundwater recharge to the surficial aquifer of the Coastal Plain was estimated to be about 40,000 Mgal/d (million gallons per day), or 15.4 in/yr (inches per year). The majority of this recharge discharges to the nearest surface-water body. Model results indicate that under pre-pumping conditions, 592 Mgl/d, or about 0.5 in/yr, of the average areal recharge moved downward to the underlying confined system. This recharge occurred over approximately 25,000 mi2, or 26 percent of the total area. Discharge from the deeper aquifers occurred over the remaining 74 percent of the study area. In areas where the surficial aquifer is in direct contact with an underlying confined aquifer (no intervening confining unit), the maximum simulated prepumping rate of recharge from the surficial aquifer to the underlying confined system was about 16 in/yr, and the maximum rate of discharge from the underlying confined system to the surficial aquifer was about 20 in/yr.
Interpretation of simulated prepumping potentiometric surfaces indicates that (1) recharge to the confined aquifers occurred in areas of downward hydraulic gradient, generally along or near the Fall Line, and (2) discharge from the deeper confined aquifers occurred by upward leakage through confining units into the ocean or coastal estuaries and bays. The simulated prepumping potentiometric surfaces of the shallower aquifers show relatively local flow patterns. The influence of major rivers, estuaries, and embayments on the flow sytem is apparent. In contrast, the potentiometric surfaces of the deeper aquifers show a regional flow pattern. Although the streams and rivers affected flow in updip areas near outcrops, generally flow was not influenced by overlying surface-water bodies throughout the areal extent of the deeper aquifers. Notable exceptions included Raritan, Delaware, and Chesapeake Bays. These large surface-water bodies affected flow patterns in all aquifers and had significant influence on the location of the 10,000-mg/L chloride concentration. In the deeper aquifers, groundwater flow paths typically were several tens of miles long. Small lateral hydraulic head gradients and low values of hydraulic conductivity indicate that computed Darcy flow velocities were slow (less than 1 ft per year) along regional flow lines. By 1980, withdrawals had caused regional heads in several aquifers to decline to more than 100 ft below sea level in areas of North Carolina, Virginia, Delaware, and New Jersey, and to more than 50 ft below sea level in areas of Maryland. The size and shape of the pumping cones depended on the quantity of water being withdrawn, the location of the pumping center relative to the aquifer outcrop, and the hydraulic characteristics of the aquifers and confining units.
Withdrawals in 1980, primarily from the confined system, were estimated to be about 1,210 Mgal/d or about 3 percent of the estimated average annual groundwater recharge (40,000 Mgal/d) to the surficial aquifer. However, 1980 withdrawals were about twice the simulated recharge from the surficial aquifer to the confined-flow system prior to development (592 Mgal/d).
Pumpage of water resulted in (1) a reduction in aquifer storage, (2) an increase in recharge from the surficial aquifer to the confined system and (or) a decrease in discharge to or direct recharge from streams, and (3) a reduction in discharge from the confined system through overlying confining units to large surface-water bodies. Reduction in aquifer storage was negligible (less than 2 percent of the withdrawals). The system approached equilibrium in less than 5 years after each simulated change in withdrawals. An increase in recharge to the confined system derived from reduced discharge to streams was the principal source of pumped water. In 1980, the recharge to the confined system was 1,330 Mgal/d, and the area of recharge to the confined system from the surficial system was approximately 45 percent of the study area. The 1980 recharge represents an increase in recharge area of 19 percent over prepumping conditions. A smaller source of the pumped water was water that formerly discharged upward through overlying confining units in coastal areas.
Geohydrology and Simulation of Groundwater Flow in the Northern Atlantic Coastal Plain Aquifer System
By P. Patrick Leahy and Mary Martin
MODEL VERSION/TYPE: Modified version of the Trescott model, transient
AREA STUDIED: Northern Atlantic Coastal Plain (New York, New Jersey, Delaware, Maryland, Virginia, North Carolina)
AQUIFERS SIMULATED: 10 regional Coastal Plain aquifers
MOST RECENT WITHDRAWALS SIMULATED: 1980
MODEL SIZE: 11 layers, 85 rows, 32 columns
MINIMUM GRID SPACING:7 miles x 7 miles
MODEL ARCHIVE is available by email request at gs-nj-model-request@usgs.gov
REPORT ABSTRACT:
A groundwater flow model of the northern Atlantic Coastal Plain was designed and calibrated to increase understanding of the regional flow regime and changes in flow brought about by human activities. A multilayer flow model simulated groundwater flow in an aquifer system covering a 95,000-mi2 (square miles) area that included Long Island, N.Y., and the Coastal Plain of New Jersey, Delaware, Maryland, Virginia, North Carolina, and parts of the adjacent Continental Shelf. The aquifer system consists of a seaward-thickening sedimentary wedge of 10 regional aquifers (predominantly sand) separated by 9 confining units (silt and clay).
Prepumping and pumping (1900 to 1980) conditions were calibrated through trial-and-error adjustment of hydraulic characteristics until model-simulated heads and flows matched those measured or estimated. Hydrographs from 74 wells were used to calibrate the flow model through time. Model calibration also was done at a local scale using subregional models for North Carolina, Virginia, Maryland-Delaware, and New Jersey. The hydraulic characteristics for the regional and subregional models were compatible.
Areal distributions of aquifer transmissivity and confining-unit leakance were refined through calibration. Transmissivity was highest, more than 100,000 ft2/d (feet squared per day), in parts of the Castle Hayne aquifer in North Carolina. Transmissivity generally ranged from 500 to about 10,000 ft2/d for most aquifers in the system. Confining-unit leakance ranged from about 1.0 x 10-8 to 1.0 foot per day per foot. Lower values typically were found in the deeper units. Updip leakance values tended to be higher because of a decrease in thickness or an increase in vertical hydraulic conductivity of the confining unit. Sensitivity of the model response to hydraulic characteristics, confining-unit transient leakage, and the nature and location of the seaward model boundary was determined. Values of transmissivity, storage coefficient withdrawals, and confining-unit leakance were changed by varying amounts dependent on a subjective estimate of their uncertainty. Model heads were most sensitive to an order of magnitude change in confining-unit leakance and to confining-unit transient leakage. Additional data are needed to determine the importance of transient leakage in the hydraulics of the aquifer system. A variable-density flow model evaluated the sensitivity of the model to the seaward boundary. This model simulated flow in parts of the aquifer system that contained water having chloride concentrations greater than 10,000 mg/L (milligrams per liter). Along the 10,000-mg/L-chloride-concentration line, head differences between the calibrated constant-density flow model and the variable-density flow model generally were less than 10 ft (feet). Thus, probably only a small percentage of the water pumped from the system was derived from areas containing salty water.
Characteristics of the regional flow system described in the report are based on an analysis of the model-derived: (1) water budgets, (2) potentiometric surfaces, (3) vertical leakage between aquifers, and (4) lateral flow patterns and velocities within aquifers. Average areal groundwater recharge to the surficial aquifer of the Coastal Plain was estimated to be about 40,000 Mgal/d (million gallons per day), or 15.4 in/yr (inches per year). The majority of this recharge discharges to the nearest surface-water body. Model results indicate that under pre-pumping conditions, 592 Mgl/d, or about 0.5 in/yr, of the average areal recharge moved downward to the underlying confined system. This recharge occurred over approximately 25,000 mi2, or 26 percent of the total area. Discharge from the deeper aquifers occurred over the remaining 74 percent of the study area. In areas where the surficial aquifer is in direct contact with an underlying confined aquifer (no intervening confining unit), the maximum simulated prepumping rate of recharge from the surficial aquifer to the underlying confined system was about 16 in/yr, and the maximum rate of discharge from the underlying confined system to the surficial aquifer was about 20 in/yr.
Interpretation of simulated prepumping potentiometric surfaces indicates that (1) recharge to the confined aquifers occurred in areas of downward hydraulic gradient, generally along or near the Fall Line, and (2) discharge from the deeper confined aquifers occurred by upward leakage through confining units into the ocean or coastal estuaries and bays. The simulated prepumping potentiometric surfaces of the shallower aquifers show relatively local flow patterns. The influence of major rivers, estuaries, and embayments on the flow sytem is apparent. In contrast, the potentiometric surfaces of the deeper aquifers show a regional flow pattern. Although the streams and rivers affected flow in updip areas near outcrops, generally flow was not influenced by overlying surface-water bodies throughout the areal extent of the deeper aquifers. Notable exceptions included Raritan, Delaware, and Chesapeake Bays. These large surface-water bodies affected flow patterns in all aquifers and had significant influence on the location of the 10,000-mg/L chloride concentration. In the deeper aquifers, groundwater flow paths typically were several tens of miles long. Small lateral hydraulic head gradients and low values of hydraulic conductivity indicate that computed Darcy flow velocities were slow (less than 1 ft per year) along regional flow lines. By 1980, withdrawals had caused regional heads in several aquifers to decline to more than 100 ft below sea level in areas of North Carolina, Virginia, Delaware, and New Jersey, and to more than 50 ft below sea level in areas of Maryland. The size and shape of the pumping cones depended on the quantity of water being withdrawn, the location of the pumping center relative to the aquifer outcrop, and the hydraulic characteristics of the aquifers and confining units.
Withdrawals in 1980, primarily from the confined system, were estimated to be about 1,210 Mgal/d or about 3 percent of the estimated average annual groundwater recharge (40,000 Mgal/d) to the surficial aquifer. However, 1980 withdrawals were about twice the simulated recharge from the surficial aquifer to the confined-flow system prior to development (592 Mgal/d).
Pumpage of water resulted in (1) a reduction in aquifer storage, (2) an increase in recharge from the surficial aquifer to the confined system and (or) a decrease in discharge to or direct recharge from streams, and (3) a reduction in discharge from the confined system through overlying confining units to large surface-water bodies. Reduction in aquifer storage was negligible (less than 2 percent of the withdrawals). The system approached equilibrium in less than 5 years after each simulated change in withdrawals. An increase in recharge to the confined system derived from reduced discharge to streams was the principal source of pumped water. In 1980, the recharge to the confined system was 1,330 Mgal/d, and the area of recharge to the confined system from the surficial system was approximately 45 percent of the study area. The 1980 recharge represents an increase in recharge area of 19 percent over prepumping conditions. A smaller source of the pumped water was water that formerly discharged upward through overlying confining units in coastal areas.