Case Study/ Origin and Extent of Fresh Paleowaters … · Case Study/ Origin and Extent of Fresh Paleowaters on the Atlantic Continental Shelf, USA ... Orange and yellow shaded patterns

Case Study/

Origin and Extent of Fresh Paleowaterson the Atlantic Continental Shelf, USAby Denis Cohen1, Mark Person2, Peng Wang3, Carl W. Gable4, Deborah Hutchinson5, Andee Marksamer6,Brandon Dugan7, Henk Kooi8, Koos Groen8, Daniel Lizarralde9, Robert L. Evans9, Frederick D. Day-Lewis10, andJohn W. Lane Jr.10

AbstractWhile the existence of relatively fresh groundwater sequestered within permeable, porous sediments beneath

the Atlantic continental shelf of North and South America has been known for some time, these waters have neverbeen assessed as a potential resource. This fresh water was likely emplaced during Pleistocene sea-level low standswhen the shelf was exposed to meteoric recharge and by elevated recharge in areas overrun by the Laurentide icesheet at high latitudes. To test this hypothesis, we present results from a high-resolution paleohydrologic modelof groundwater flow, heat and solute transport, ice sheet loading, and sea level fluctuations for the continentalshelf from New Jersey to Maine over the last 2 million years. Our analysis suggests that the presence of freshto brackish water within shallow Miocene sands more than 100 km offshore of New Jersey was facilitated bydischarge of submarine springs along Baltimore and Hudson Canyons where these shallow aquifers crop out.Recharge rates four times modern levels were computed for portions of New England’s continental shelf thatwere overrun by the Laurentide ice sheet during the last glacial maximum. We estimate the volume of emplacedPleistocene continental shelf fresh water (less than 1 ppt) to be 1300 km3 in New England. We also presentestimates of continental shelf fresh water resources for the U.S. Atlantic eastern seaboard (104 km3) and passivemargins globally (3 × 105 km3). The simulation results support the hypothesis that offshore fresh water is apotentially valuable, albeit nonrenewable resource for coastal megacities faced with growing water shortages.

1Department of Geological & Atmospheric Sciences, Iowa StateUniversity, Ames, IA 50011

2Corresponding author: Hydrology Program, New MexicoTech, NM 87801; (575) 835-6506; fax (575) 835-6436;[email protected]

3University Information Technology Services, Indiana University,Bloomington, IN 87801

4Los Alamos National Laboratory, Los Alamos, NM 875455U.S. Geological Survey, Woods Hole, MA 025436Indiana University, Department of Geological Sciences,

Bloomington, IN 878017Department of Earth Science, Rice University, Houston, TX

770058VU University Amsterdam, Amsterdam9Department of Geology and Geophysics, Woods Hole

Oceanographic Institution, Woods Hole, MA 0254310U.S. Geological Survey, Office of Groundwater, Branch of

Geophysics, Storrs, CT 06269Correction added after online publication September 15, 2009:

on page 7, in Equation 9, the value in the square root should haveread: ‘‘1 − (ξ/L)2’’. We apologize for this error.

Received April 2009, accepted August 2009.Copyright © 2009 The Author(s)Journal compilation ©2009NationalGroundWaterAssociation.doi: 10.1111/j.1745-6584.2009.00627.x

IntroductionHydrologists have long regarded the marine environ-

ment as the domain from which salt water invades coastalaquifers in response to groundwater pumping (e.g., Henry1959; Gill 1962; Schaefer and Walker 1981). However,data from well bores indicate that relatively fresh ground-water exists beneath the Atlantic continental shelf faroffshore North and South America (Figure 1). One of themost striking examples of fresh groundwater offshore isfound on the Atlantic continental shelf of New England,where a series of well bores (Hathaway et al. 1979; Schleeand Fritsch 1982; Mountain et al. 1994; Austin et al.1998; Malone et al. 2002) reveal nearly fresh water (lessthan 5 parts per thousand total dissolved solids, or ppt tds)200 m beneath the seafloor more than 100 km offshoreof New Jersey and Long Island (Figure 1, section 3).Emplacement mechanisms that have been proposed toaccount for this fresh water include (1) ocean-directed,lateral inflow of fresh water during Pleistocene sea-levellow stands from confined aquifers whose recharge areaswere, and often still are, above sea level today (right-handside of Figure 2A; Meisler et al. 1984; Pope and Gordon

NGWA.org GROUND WATER 1

Figure 1. Salinity distribution (contours in ppt) for four cross-sectional transects (1 to 4) across the Atlantic continental shelfof North and South America (from Hathaway et al. 1979; Johnston 1983; Kooi and Groen 2001; Marksamer et al. 2007). Thelocation of cross sections 1 to 4 is shown in the inset (numbered red section lines). The fresh water beneath these shelves waslikely emplaced via several mechanisms, as discussed in the text. Orange and yellow shaded patterns represent crystallineand continental shelf sedimentary rocks, respectively.

1999); (2) localized, vertical paleorecharge on subaeriallyexposed continental shelf during Pleistocene sea-level lowstands (central part of Figure 2A; Kooi and Groen 2001;Kooi et al. 2000) facilitated by gaps in confining unitsdue to either nondeposition or erosion and local topo-graphic variations in the water table; or (3) sub-ice-sheetrecharge along portions of the continental shelf overrunby the Laurentide ice sheet (Siegel and Mandle 1984;Person et al. 2003, 2007a) between 21 and 14 KyBP(Figure 2B; Uchupi et al. 2001). Proglacial lakes thatform in front of ice sheets could also contribute recharge(Marksamer et al. 2007; Lemieux et al. 2008a, 2008b,2008c; Figure 2B). To date, no study has estimated thevolume of emplaced fresh water on the continental shelfin New England or for passive margins globally. What isclear from inspection of Figure 1 is that the distributionof present-day fresh and salt water cannot be explainedby assuming hydrodynamic equilibrium conditions withmodern sea level (Figure 2C).

The presence of large volumes of offshore freshwater along the southeastern United States (Figure 1,section 2) can be explained, in part, by present-day mete-oric recharge to confined aquifers that outcrop on land andthe presence of high-permeability subseafloor karstified

units. Along the New England coast, however, onshorehydraulic heads are too low to drive meteoric water faroffshore (Kooi and Groen 2001). In addition, confinedaquifers north of Martha’s Vineyard and Nantucket out-crop or subcrop below sea level (not shown). Therefore,onshore recharge cannot be happening today. The freshand brackish subseafloor pore waters along the north-eastern U.S. Atlantic coast thus can be considered primeexamples of paleogroundwater emplaced during Pleis-tocene sea-level low stands that escaped salinization dur-ing Holocene sea level rise (e.g., Hathaway et al. 1979;Meisler et al. 1984; Kooi and Groen 2001; Person et al.2003). These conditions have also been reported along theAtlantic continental margin in Europe (Edmunds 2001).Because pore water data are extremely limited, there is adearth of information regarding the distribution and ori-gin of these offshore paleowaters. Paleohydrologic modelsrepresent a powerful tool to assess offshore groundwaterresources.

To estimate the amount of fresh water present oncontinental shelf environments, we analyzed four salinitycross-sectional profiles from Suriname (Kooi and Groen2001), Florida (Johnston 1983), New Jersey (Hathawayet al. 1979), and Massachusetts (Marksamer et al. 2007)

2 D. Cohen et al. GROUND WATER NGWA.org

Figure 2. Conceptual models for fresh water emplacement.(A) Meteoric recharge under modern sea level conditions.Sea water–fresh water interface is assumed to be in equi-librium with modern sea level conditions; (B) fresh waterdriven offshore by sub-ice-sheet recharge from the Lauren-tide ice sheet during the last glacial maximum, 21,000 yearsbefore present; and (C) vertical infiltration of meteoric waterthrough confining units induced by local flow cells on thecontinental shelf during sea-level lowstands associated withglaciation. Salt water is shaded yellow, fresh water is white.Gray and orange shaded patterns represent confining unitsand bedrock, respectively.

(presented in Figure 1 and Table 1). The lithologic com-position of these sections varies from siliclastic (Suri-name, New Jersey, Nantucket) to carbonate dominated(Florida). The limit of the 1 ppt isochlor varies consid-erably from about 10 km off New Jersey to 120 km offFlorida. Inspection of Table 1 suggests that the volumeof continental shelf fresh water appears to be inverselycorrelated with mean continental shelf water depth. Theaverage depth (and hence slope) of the continental shelf

would control the degree of subaerial exposure of theshelf during sea water low stands. However, mean waterdepth is only one factor controlling fresh water occur-rence. Other factors, such as the presence/absence ofglacial loading, the elevation and areal extent of onshorerecharge areas for confined aquifers, and recharge rates,are also important. The average volume of fresh water(salinity less than 1 ppt) per kilometer width for thesefour cross sections is estimated to be about 3.8 km3. Thisestimate assumes an effective sediment porosity of 0.2and includes fresh water in both coarse and fine-grainedsediments. Assuming that the length of the U.S. Atlanticcontinental shelf is 3310 km and assuming an average of3.8 km3 of available fresh water per kilometer of coast-line (i.e., using all four section lines), then the volumeof available fresh water along the eastern seaboard of theUnited States is about 12,000 km3. Globally, continentalshelf width varies considerably from almost zero near sub-duction zones off of South America to more than 900 kmoff of Siberia. If we restrict our analysis to passive mar-gins, then the total global length of the continental shelf isabout 80,000 km. Together with a fresh water volume of3.8 km3/km, this yields a global estimate of sequesteredfresh water along passive margins of about 300,000 km3.The order-of-magnitude calculation suggests that passivemargins may represent an important unconventional freshwater resource. However, there is considerable uncertaintyin these estimates.

The purpose of this study is to evaluate the mech-anisms of emplacement, distribution, and volume of thisunconventional, nonrenewable water resource using high-resolution numerical modeling. Obviously, this work rep-resents only a first step toward estimating the extent ofthis potential resource, and additional hydrologic and geo-physical exploration is required to verify the simulationresults. We present order-of-magnitude estimates of pas-sive margin fresh water volumes globally, but focus ouranalysis on New England, where unusually fresh porefluids are found in boreholes located up to 100 km offthe coast (wells 6001, 6009, 6011, 6020; Figure 3A).These wells have salinity below 3 ppt within Pleistocene,Pliocene, and Miocene sand units from 100 to 500 mbelow seafloor (Kohout et al. 1977; Folger et al. 1978).Because the response time for solute transport on con-tinental shelf aquifer systems is long (at least 106 years

Table 1Atlantic Continental Shelf Fresh Water Estimate

Cross SectionDistance Fresh Water Lens

Extends Offshore (km)Vertical Fresh Water Lens

Thickness (km)Mean ContinentalShelf Depth (m)

Volume FreshWater1 (km3/km)

Suriname 40 0.4 45 3.2Florida 120 0.5 32 12New Jersey 5 0.2 65 0.2Nantucket 50 0.3 37 3

1Assuming a porosity of 0.2.

NGWA.org D. Cohen et al. GROUND WATER 3

Figure 3. (A) Areal footprint of three-dimensional paleohy-drologic model (dashed purple line) along Atlantic conti-nental shelf in New England and location of cross sectionsand wells presented in Figures 8 and 9. (B) Cross-sectionaltransects depicting hydrostratigraphic units represented inthree-dimensional model: silt/clay (blue), fine sand (red), andmedium to coarse sand (green) units. Note that the shal-low sands crop out in submarine canyons. Imposed surfacehydraulic heads (C) and topography/bathymetry of conti-nental shelf (D) during LGM, 21,000 years before present.Areas of high imposed heads (greater than 250 m) delin-eate the approximate location of the Laurentide ice sheet inFigure 3C.

assuming a solute diffusion coefficient of 3 × 10−10 m2/sand a confining unit thickness of 100 m), we must recon-struct hydrogeologic conditions during the Pleistocene inorder to arrive at a representative account of present-daysalinity conditions.

Study AreaSediments of the Atlantic continental margin of New

England have been studied for their petroleum and min-eral resources (Perlmutter and Geraghty 1963; Hathawayet al. 1979; Mattick and Hennessy 1980; Poag 1982; Mal-one et al. 2002) and probed to assess water resourcesand chemical fluxes (Kohout et al. 1977; Meisler et al.1984; Pope and Gordon 1999; Buxton and Modica 1992;Person et al. 1998, 2003). A series of boreholes werecompleted along the Atlantic continental shelf during the1970s and 1980s as part of the Atlantic Margin Cor-ing (AMCOR) project (Hathaway et al. 1979), as con-tinental offshore stratigraphic test (COST) wells (Schleeand Fritsch 1982), and most recently through the OceanDrilling Program (ODP) Legs 150 and 174A (Mountainet al. 1994; Austin et al. 1998; Malone et al. 2002). Sedi-ments that accumulated on the Atlantic continental marginfollowing the Mesozoic break-up of Africa and NorthAmerica locally exceed 10 km in thickness. Tertiary andCretaceous unconsolidated sands extend tens of kilome-ters offshore before grading into finer-grained facies. TheTertiary and Cretaceous shelf deposits of New Englandare capped by a roughly 100- to 200-m-thick layer ofmarine sediments and Pleistocene glacial tills and outwashsands. Multiple glacial/interglacial stages are documentedon Nantucket and Martha’s Vineyard (Kaye 1964; Oldaleet al., 1982).

The permeability of medium to fine-grained sedi-mentary units along the continental shelf is estimatedto range between about 10−10 to 10−18 m2 (Table 2).These have been measured through aquifer tests or esti-mated through model calibration. Most of the directmeasurements were made in the near-shore environmentwhere sediment burial depths are minimal; thus, off-shore values may be overestimated. Hydrogeologic modelcalibration (e.g., Pope and Gordon 1999; Person et al.2003) has been used to assess hydrogeologic propertiesof sediments far offshore. Sediment compressibility forthe Pliocene-Oligocene sands has been estimated to be4.4 × 10−7 Pa−1, which would yield a specific storagecoefficient (Ss) of about 10−2.3 m−1 (Dugan and Flem-ings 2002). Pope and Gordon (1999) estimated the spe-cific storage coefficient for deeper New Jersey confinedaquifer units to be about 10−6 m−1 based on calibrationof their sharp interface model. Measured porosity of thePleistocene and Miocene sands ranged between 0.30 and0.48 (Person et al. 1998). Off New Jersey, Miocene sandsappear to be significantly overpressured and porosity hasbeen estimated to be as high as 0.7 (Dugan and Flemings2002). Scholle (1977, 1980) reported porosity variationsbetween 0.6 and 0.15 within COST wells B-2 and B-3for the mid-Atlantic region. Gill (1962) reported porosityvariations within the Cohansey Sand off Cape May, NewJersey, to range from 0.41 to 0.28.

Estimates of modern recharge rates (∼0.5 m/a) forNew England using environmental isotopes (Knott andOlimpio 1986), potential evapotranspiration estimates(Steenhuis et al. 1985), or regional groundwater flowmodels (Guswa and LeBlanc 1981; Person et al. 1998;


Table 2Aquifer and Confining Unit Properties from In Situ and Core Tests

Unit Permeability Range (m2) Hydraulic Conductivity Range (m/d)

Pleistocene glacial outwash sands 10−11.5 –10−10.1 22–70Pliocene/Oligocene silt/clayey silt 10−18.4 –10−14.8 0.000001–0.0012Cretaceous sandstones (Magothy and Raritian)1 10−12.7 –10−11.1 0.15–22

1Sources: Perlmutter and Geraghty 1963; Guswa and Le Blanc 1981; Buxton and Modica 1992; Person et al. 1998; Pope and Gordon 1999.

Pope and Gordon 1999) indicate that about 40% to 50%of the annual precipitation goes to recharge. On NantucketIsland, Cape Cod, and Martha’s Vineyard, Pleistoceneunconsolidated outwash sands are the only source of freshwater (Person et al. 1998). On Long Island and New Jer-sey, deeper Cretaceous aquifers are also exploited forfresh water, because shallow aquifers have become con-taminated by point and nonpoint source pollutants (Katzet al. 1980) and shallow salt water intrusion (Schaeferand Walker 1981). Water table elevations vary from 2 mabove sea level on Nantucket Island (Person et al. 1998)to 24 m on Long Island (Buxton and Modica 1992) andmore than 38 m in New Jersey (Pope and Gordon 1999).As indicated earlier, confined offshore aquifers in Maineand Massachusetts receive no modern meteoric rechargebecause they outcrop or subcrop below sea level.

Radiocarbon ages for glacial tills of the continentalshelf off New England indicate that the Laurentide icesheet reached its southern most extent across Long Islandand Nantucket about 21 ka before retreating onto themainland by approximately 14 ka (Uchupi et al. 2001)(Figure 4). Ice sheet thickness was 1000 m near CapeCod and 500 m on Long Island (Denton and Hughes1981). Several proglacial lakes formed following the lastglacial maximum (LGM) as the Laurentide ice sheetretreated (Uchupi et al. 2001; Mulligan and Uchupi 2003)(Figure 4). Seepage from proglacial lakes in NantucketSound, Cape Cod Bay, Block Island Sound, and LongIsland Sound could have provided extensive rechargeto confined aquifers of the New England continentalshelf (Marksamer et al. 2007). When these lakes burst,rapid sedimentation on the continental shelf would havechanged the land surface morphology, induced rapidsediment loading, and perhaps induced infiltration of freshwater.

A series of prior hydrologic models constructedfor the continental shelves off New Jersey (Meisleret al. 1984; Pope and Gordon 1999), Long Island, andMassachusetts (Person et al. 2003) concluded that bothonshore and sub-ice-sheet recharge played important rolesin explaining the occurrence of sequestered fresh water.Because of their restrictive simplifying assumptions (e.g.,sharp interface theory, cross-sectional geometry), thesemodels lacked the ability to provide an accounting ofthe total amount of sequestered fresh water on theNew England’s continental shelf. Our analysis overcomesthis limitation by combining a detailed representation

of three-dimensional transport processes, an extensiveregional scale (125,000 km2), and a relatively high spatialdiscretization (more than 5 million tetrahedral elements)of the sediments that control the flow patterns in thesubsurface.

Mathematical ModelWe developed a high-resolution, three-dimensional,

variable-density groundwater flow model for this study(GW_ICE ) to reconstruct the extent of fresh water onthe Atlantic continental shelf in New England during thePleistocene. GW_ICE is a modified (parallelized) ver-sion of the serial finite-element based paleohydrogeo-logic model GEOFE (Person et al. 2007b). Documen-tation describing the governing transport equations forgroundwater flow, heat, solute transport, numerical imple-mentation, and partial model validation exercises for theserial version of this code (GEOFE ) are available athttp:\\ww.ees.nmt.edu/person.

A continental shelf paleohydrologic model needs tore-create hydrologic conditions through time by incor-porating, as realistically as possible, the temporal evo-lution of sea level, temperature, subsea topography and,where appropriate, glaciations. This requires representa-tion of changes in land surface temperature, sea levelvariations, glaciations, and salinity conditions on geologictime scales.

Groundwater FlowThe groundwater flow equation solved in this study

represents the effects of variable-density flow, sea levelchange, and ice sheet loading through time following theapproach of Provost et al. (1998):

∇x · [Kμr∇x(h + ρrz)] = Ss

[∂h

∂t− ρi

ρo

∂η

∂t

](1)

where ∇x is the gradient operator, K is the hydraulicconductivity tensor, h is hydraulic head, η is ice sheetthickness, z is elevation, ρo is the density of waterat the standard state (10◦C, salinity of 0.0 mg/L, andatmospheric pressure), ρr is the relative density (definedbelow), ρi is the ice density, μr is the relative viscosity(defined below), and Ss is the specific storage. Equation 1is similar to other variable-density flow models except thatit includes an ice sheet loading term (second term on right-hand side of Equation 1). Equation 1 assumes a loading


Figure 4. Distribution of proglacial lakes (black) and position of Laurentide ice sheet (thick red lines) during Late Quaternaryon the Atlantic continental shelf (after Uchupi et al. 2001). The following abbreviations are included: L.A. is Lake Amsterdam;L.At is Lake Attlebury; L.F. is Lake Fishkill; L.B. is Lake Bascom; L.BIS is Lake Block Island Sound; L.C. is LakeConnecticut; CH.S. is Champlain Sea; L.H. is Lake Hudson; L.Hi is Lake Hitchcock; L.K. is Lake Kiskaton; L.M. is LakeMarrimack; L.Mi is Lake Middletown; L.N. is Lake Narragansett; L.P. is Lake Passaic; L.S. is Lake Schohaire; L.Sc. isLake Sacandaga; T. is Lake Tomhannock; L.T. is Lake Tilson; L.Ta. is Lake Tauton; L.W. is Lake Warrensberg; L.Wa. isLake Warwarsing; and L.V. is Lake Vermont.

efficiency of 1.0 (i.e., all of the load of the ice sheetis transferred to the fluid; Lemieux et al. 2008a). Thismay overestimate induced pore pressure increases by thismechanism at shallow depths (less than 300 m). However,at greater depths, a loading efficiency of 1 is probablyreasonable. Support for this assumption at shallow depthscomes from preconsolidation stress data collected acrossNorth America and Europe, which indicate that tillsoverrun by ice sheets are less consolidated than wouldbe expected based on the reconstructed ice sheet loads.Estimated pore water pressures within glacial till basedon consolidation data are between 78% and 98% ofthe ice-overburden pressure for maximum and minimumice thickness estimates (Piotrowski and Kraus 1997;Hooyer and Iverson 2002). In other words, most of theice sheet weight was supported by elevated pore fluidpressures. Our analysis neglects the effects of Pleistocenesedimentation on simulated fluid pressures (see Duganand Flemings 2000; Marksamer et al. 2007). Althoughthis prevents us from accurately reconstructing anomalousheads in fine-grained sediments, it probably has littleeffect on the salinity distribution in relatively coarse-grained, near-shore sands. Our analysis also neglectssubgrid scale density instabilities induced by buoyancyeffects (Simmons et al. 2001; Post and Kooi 2003) and

permeability heterogeneity within individual lithologicunits (Gelhar 1993).

We solved a variable-density form of Darcy’s law(Garven and Freeze 1984) for the specific dischargevector:

�q = −Kμr∇x[h + ρrz] (2)

where the relative density (ρr) and relative viscosity (μr)terms are given by:

ρr = ρf − ρo

ρo(3)

μr = μo

μf(4)

where ρf is the density of groundwater, μo is the viscosityof water at standard state, and μf is the viscosity ofthe fluid at elevated temperature, pressure, and salinityconditions. Because Atlantic shelf sediments have agentle slope, we neglected off-diagonal terms of thehydraulic conductivity tensor (i.e., the principal directionsof anisotropy are aligned with vertical and horizontaldirections).

Solute TransportWe used a conventional advective/dispersive equation

to represent time-dependent transport of a conservative


solute:

φ∂C

∂t= ∇x[φD∇xC] − �q∇xC (5)

where D is the standard three-dimensional hydrodynamicdispersion-diffusion tensor (Konikow and Grove 1977)and C is solute concentration (solute mass fraction; kgsolute/kg solution). D is a function of solute diffusivityas well as longitudinal and transverse dispersivity. Ourapproach neglects the potential effects of salt exclusion onsalinity changes due to the effects of freezing (Starinskyand Katz 2003).

Heat TransportThe thickness of continental shelf sediments (up to

3 km in this study) requires that we represent the tempera-ture effects on fluid density. GW_ICE solves a conductiveand convective-dispersive heat transfer equation:

[cfρfφ + csρs(1 − φ)]∂T

∂t= ∇x[λ∇xT ] − �qρfcf∇xT

(6)

where λ is the thermal dispersion-conduction tensor; T

is temperature; cs and cf are the specific heat capacitiesof the solid and liquid phases, respectively; and ρsis the density of the solid phase. Permafrost duringPleistocene glaciations can cause important reductions inpermeability near the land surface (Kleinberg and Griffin2005; Bense and Person 2008). In the model, the thermaldispersion-conduction tensor is a function of the solidthermal conductivity, the fluid thermal conductivity, theporosity, the longitudinal and transverse dispersivities, thefluid density, the fluid heat capacity, and the magnitudeand direction of the Darcy flux (de Marsily 1986). Forall time steps where the atmospheric temperature wasbelow −2◦C, surface elements in our model beyond theice sheet toe were assigned a permeability 1000 timeslower than their Holocene values (Kleinberg and Griffin2005) to represent near-surface freezing. Permeabilitywas reassigned to normal levels for these elements iftemperatures rose above freezing in subsequent time steps.Because the surface mesh thickness (�z) varied between0.6 and 15 m, our permafrost algorithm would not allowfor the development of a thick (greater than 100 m)permafrost zone.

Equation of StateThermodynamic equations of state are used to com-

pute the density and viscosity of groundwater at elevatedtemperature, pressure, and salinity conditions. GW_ICEuses the polynomial expressions of Kestin et al. (1981):

1

ρf= a(T ) + b(T )P + c(T )P 2 + Cd(T ) + C2e(T )

−PCf (T ) − C2Pg(T ) − h(T )

2P 2 (7)

μf = μo[1 + B(T , C)P ] (8)

where a(T ), b(T ), . . . , h(T ), and B(T , C) are third- andfourth-order temperature- and concentration-dependent

polynomials and P is pressure (P = [h − z]ρfg). Thesepolynomial expressions are valid for temperatures between10◦C and 150◦C and salinities between 0 and 6 molalNaCl. Fluid density is less sensitive to changes in fluidpressure than it is to changes in temperature and salinityfor the range of conditions encountered in sedimentarybasins. Between −3◦C and 10◦C, fluid density remainsfixed at standard state conditions although permeabilityreduction prevents fluid flow from occurring below −2◦C.

Boundary ConditionsPrior studies have used different strategies to rep-

resent the hydraulic boundary condition beneath theice sheet, including specified flux (Provost et al. 1998;Breemer et al. 2002), specified head (Bense and Person2008), or a mixed boundary (Lemieux et al. 2008a). Forthis study, we specified a hydraulic head equivalent to90% of the ice sheet height beneath the glacier (Boul-ton et al. 1995). Subglacial heads of this magnitude havebeen observed beneath ice streams (Engelhard and Kamb1997). We assume that recharge flux in excess of localmelt rates is supplied along a network of sub-ice-sheetchannels, which were supplied by melt water from thenortheast along an esker network (Shreve 1985). Directinflux of glacial surface melt water can also be a sourceof sub-ice-sheet groundwater recharge. In Greenland, sur-face melt water lakes have been observed to drain rapidlythrough fracture networks in areas where the ice sheetthickness is less than 1000 m (van der Veen 2007; Zwallyet al. 2002). The thickness of the ice sheet (η) was esti-mated using a polynomial expression presented by vander Veen (1998):

η = H√

1 − (ξ/L)2 (9)

where H is the maximum ice sheet thickness at aparticular time step, L is the ice sheet length at that timestep, ξ is the distance from the margin of the domainmargin in an upgradient direction, and η is ice sheetthickness at distance ξ from the margin of the basin.Both H and L were scaled using the Greenland Ice CoreProject (GRIP) sea level data (Figure 5) such that duringthe LGM, L, and H were at their maximum values. Themaximum thickness of the ice sheet during the LGM isshown in Figure 3C. The weight of the ice can causelithosphere deflection (Peltier 1998). This resulted in areduction of the hydraulic heads assigned to nodes beneaththe ice sheet on the continental shelf surface by as muchas 100 m and diminished the hydraulic gradients beneaththe ice sheet. While the full isostatic compensation ofice sheet loading is 28% (Marshall and Clarke 2002;Tarasov and Peltier 2004, 2007), the ice sheet was out onthe continental shelf for only a few thousand years andprobably did not attain isostatic equilibrium (Redfield andRubin 1962; Barnhardt et al. 1995). The deflection of theland surface due to ice sheet loading at the LGM is shownin Figure 3D. We used the ice sheet thickness to deflectthe lithosphere isostatically by 10%. Sea level in NewEngland was varied by 0 to −120 m relative to modern


Figure 5. Isotopic reconstruction of Late Pleistocene sealevel (bold black line) and atmospheric temperature (redline) vs. time in years before present (data source: Imbrieet al. 1984 and Johnsen et al. 1995).

conditions (Figure 5). Subaerial nodes not overrun by theice sheet were assigned a hydraulic head equal to the landsurface elevation. Nodes below sea level were assigned ahead consistent with sea level elevation. A temperature of4◦C was specified for all nodes below sea level. This wasbased on modern records of deep (greater than 2000 m)ocean temperatures and is likely lower than modernthermal conditions on the continental shelf. Land surfacetemperatures for unglaciated regions were prescribedusing the continental temperature reconstruction estimatedfrom the SPECMAP and GRIP ice core record (Figure 5).Beneath the ice sheet, temperatures were fixed at 0◦C. Abasal heat flux typical of the North American craton (Sasset al. 1989) of 0.06 W/m2 was imposed along the base ofthe model. Insulated boundary conditions were prescribedfor all sides for heat transfer. Boundary conditions forsolute transport equations were a constant concentrationof 0 or 35 ppt on the sea floor, depending on whetherthe local land surface elevation was above or below sealevel. For submarine canyons, a spring or advective type(zero lateral concentration gradient) boundary conditionwas imposed.

Initial ConditionsWe assigned hydrostatic initial conditions for ground-

water flow using the local elevation of the water table.Initial temperatures varied linearly with depth using agradient of 30◦C/km. Assigning initial salinity conditionswas problematic because of the long response time (∼106

years) associated with solute diffusion. Salt water was ini-tially assigned to be present within all sediments. We thenran the model for 2 million years holding sea level con-stant at Late Pliocene levels (−53 m), thereby adjustingto the imposed pre-Pleistocene sea level conditions. Atthe start of the Pleistocene, we varied sea level by repeat-edly recycling the GRIP sea level record (Figure 5). Weassume that the Laurentide ice sheet extended out ontothe continental shelf in New England three times (Kaye1964). We used the Wisconsinan ice sheet history on theshelf as our guide to represent the Nebraskan and Illinoianglacial cycles.

Numerical ImplementationWe solved variable-density groundwater flow, heat,

and solute transport equations sequentially using the

finite element method. Advective solute transport wassolved using a Lagrangian-based modified method ofcharacteristics (MMOC; Zheng and Bennett 1994). Atime step size of 10 years was used. We used theParMETIS parallel mesh partitioning library (Karypis andKumar 1996) and Aztec (Tuminaro et al. 1999) toolkitsto iteratively solve the matrices formed in GW_ICE. Toverify the performance of our parallel code, we performeda scalability analysis using two test grids comprised of104,758 and 1,277,068 nodes, respectively. The matricessolved in the serial version used the SPLIB iterativesolver (Bramley and Wang 1995). We enforced consistentparameters and solution schemes in both the serial andparallel versions. For the relatively coarse mesh, thespeedup plateaus around 128 processors (Figure 6A). Forthe more refined mesh (Figure 6B), we found that linearspeedup prevailed, even using 256 processors. For theNew England continental shelf models, we ran GW_ICEon 256 processors using the NSF Teragrid (BigRed atIndiana University) and Encanto at the New MexicoSuper Computer Application Center. The New Englandstudy described in the following paragraph required about5 days of central processing unit (CPU) time.

The convergence may have been faster and thetotal simulation time reduced had we solved all threeequations simultaneously as it is done in some codes(e.g., Hayba and Ingebritsen 2001). However, this wouldhave required replacing the MMOC method used inGW_ICE, which typically uses a smaller Courant-limitedtime step size, with an upwinding scheme (e.g., Noorishad

Figure 6. Comparison of serial and parallel model perfor-mance. Parallel version of GW_ICE used Aztec solver pack-age. The serial version (GEOFE ) used SPLIB (Bramley andWang 1995). Both solver packages used the GMRES itera-tive scheme for this test. Model performance using a coarse(A, 104,758 nodes) and more refined (B, 1,277,068 nodes)grid is presented. Simulations were performed on NCSA I64cluster. The dashed line indicates optimal speedup.


et al. 1992) to minimize numerical instabilities associatedwith advection-dominated solute transport. Upwindingschemes tend to smear solute concentration fronts whichwe wanted to avoid.

New England GridOur finite element model represents paleohydrologic

conditions on the continental shelf (Figure 3A) fromNew Jersey to Maine in both glaciated and unglaciatedareas. We divided the Quaternary, Tertiary, and Creta-ceous sediments into three lithologic units comprised ofmedium-coarse sands, fine sands, and silt/clays (Table 3;Figure 3B). Our hydrogeologic framework model wasbased on existing onshore and offshore lithologic welllogs and aerial structure maps of Pleistocene-Upper Cre-taceous strata from Klitgord and Schneider (1994) andKlitgord et al. (1994). The model domain extends to adepth of about 3 km near the continental slope. The modeldomain extends about 250 km offshore and about 750 kmparallel to the coastline. Many of the deeper fine- andmedium-grained sand units are encased in finer-grainedsilt and clay layers (e.g., cross section A-A’ in Figure 3B).The notable exceptions are in submarine canyons whereMiocene sands are exposed (Hampson and Robb 1984).The Tertiary and Cretaceous shelf deposits are cappedby about 100- to 200-m-thick units of marine and Pleis-tocene glacial tills, outwash sands, and silt/clays. Thestratigraphy used in our model was based on coastalborings (Kohout et al. 1977), AMCOR wells (Hathawayet al. 1979), and COST wells (Mountain et al. 1994).We include details of the continental shelf morphologyand stratigraphy at the 1- to 100-m scale. Using grid-generation algorithms developed at Los Alamos NationalLabs (LaGriT©; Gable et al. 1996), we constructed ahigh-resolution tetrahedral finite element mesh of theAtlantic continental shelf using 5.8 million elements and987,995 nodes. The grid spacing in the x–y directionswas about 4 km per element. In the z-direction, grid spac-ing varied between 3 and 45 m. We chose this relativelyrefined vertical discretization to ensure that there was ahigh enough grid resolution near sand–silt interfaces toavoid numerical dispersion. We chose a minimum elementwidth of 3 m for representing strata pinch outs in orderto avoid high element aspect ratios.

Table 3Hydrogeologic Properties Assigned in

Paleohydrologic Model

Unit kmax (m2)1 φ2 Ss (m−1)3

Medium sand 10−12 0.2 10−3

Fine sand 10−14 0.2 10−4

Silt/clay 10−17 0.3 10−4

1Permeability.2Porosity.3Specific storage coefficient.

ResultsA full model calibration (Hill 1998) to estimate all

parameters is impractical given the model’s complexity,long simulation period, and regional scale, as well asthe inaccessibility of most of the area represented bythe model domain to field measurements. Addressing theproblem as a hypothesis test, we conducted a limitedmodel calibration exercise in which permeability condi-tions were varied for the three lithologic units over theirknown range of uncertainty (±102 m2). We did not varyspecific storage in our sensitivity study. We found thatonly a narrow range of permeability was able to reproduceobserved salinity profiles. Confining unit permeability wasfound to be one of the most important factors controllingoffshore salinity distribution. If confining unit permeabil-ity was too low, fresh water was restricted to the near-shore environment. Parameters resulting in the best fits tosalinity profiles are listed in Table 3. Our best-fit param-eters fall within the range of values reported from priorstudies (Table 2) for aquifers and confining units. Param-eters held constant for all three units include permeabilityanisotropy (ratio of vertical to horizontal permeability;100), longitudinal dispersivity (100 m), transverse disper-sivity (10 m), and solute diffusivity (3 × 10−10 m2/s).

Once the model was calibrated, we ran one additionalsimulation in which the glacial loading effects wereremoved. This was done to assess whether or not glacialloading influenced the distribution of fresh water on thecontinental shelf in the vicinity of the ice sheet. This wasaccomplished by modifying the specified head boundarycondition, removing the effects of the ice sheet, andsetting the glacial loading term (∂η/∂t) in Equation 1 tozero.

Computed heads for the model that included theeffects of glacial loading varied considerably laterally andwith depth (Figure 7). Figure 7 presents computed headpatterns across the top, southern, and seaward edges ofmodel domain. Prior to the LGM (37 ka in Figure 7), sealevel was about 80 m lower than today and heads werehighest (30 m above modern sea level) in New Jersey andin sand/silt units far offshore. The highest head conditionsin the offshore environment are found at depth in isolatedsand units, which do not outcrop and which are withinoverlying relatively thick, silt units. This is because ofboth density effects and remnant heads associated withhigher sea level boundary conditions and prior (Illinoian)glacial loading. This same pattern is seen today (0 ka inFigure 7). During the LGM (21 ka in Figure 7), glacialloading resulted in highest heads beneath the ice sheet(up to 700 m). High heads (up to about 30 m abovesea level) also occurred in both aquifers and confiningunits alike as the effects of glacial loading propagatedout across the continental shelf. This was facilitated bylow permeability conditions associated with permafrostformation (temperatures below −2◦C) beyond the toe ofthe ice sheet in periglacial areas. At 37 ka, sea levelwas 80 m lower than present, exposing a large portion ofthe continental shelf to meteoric recharge and enhancingthe shore-normal hydraulic gradient. The topographic


Figure 7. Changes in computed heads (in meters) duringLate Pleistocene and Holocene along the Atlantic continentalshelf in New England. The locations of New Jersey (NJ),Long Island (LI), and Nantucket Island (NT) are listed toprovide a geographical frame of reference. The dashed whiteline represents the position of the coastline through time.

lowlands to the north and east of Cape Cod drovegroundwater northeastward from Massachusetts towardMaine. Flow directions were reversed as the Laurentideice sheet overran the continental shelf at 21 ka. As sealevel continued to rise and the ice sheet retreated, meteoricrecharge became restricted to New Jersey and LongIsland. Heads in excess of modern sea level (+30 m)are indicated at depth today (Figure 7), due in part todensity effects and to fossil heads associated with ice sheetloading (Marksamer et al. 2007).

Fresh water is sequestered in sediments from thepresent-day coastline to about 50 km offshore. Withinthis region, fresh water occurs within both aquifers andconfining units (Figure 8), matching observations fromMartha’s Vineyard (well ENW-50 in Figure 9). Betweenabout 50 to 100 km offshore, a complex transition zoneexists, characterized by brackish to fresh water withinsand units and brackish to near sea water in confiningunits. This transition zone is observed beneath NantucketIsland (Figure 9, well 6001) wells. Our model predictsthese patterns (section D-D’, Figure 8) in Massachusetts,but places them farther offshore. The position of thistransition zone is variable from south to north along theshelf. Our model predicts fresh water conditions beneathwestern Long Island (section B-B’, Figure 8). Some

Figure 8. Simulated concentrations (in parts per thousand)along cross sections extracted from the three-dimensionalfinite element model of the Atlantic continental shelf. Cylin-ders depict concentration of offshore AMCOR, COST, andODP wells (well radii not to scale). The wells were raised500 m above the sea floor in this image, so that the crosssections would not obscure them. The locations of the crosssections and wells are presented in Figure 3A.

brackish water is predicted to be present in the basal sandunit across eastern Long Island (section C-C’, Figure 8).Stumm (2001) documents that salt water intrusion hasoccurred in response to pumping during the 1970s withinthe Lloyd aquifer (basal Cretaceous sand, section B-B’)along Long Island’s northwestern shore on the Great NeckPeninsula. The more significant discrepancy between themodel and field observations occurs along the southernshore of western Long Island, where Perlmutter et al.(1959) reported the presence of sea water within theupper Pleistocene outwash sands, brackish water withinthe upper Cretaceous Magothy aquifer, and fresh waterwithin the basal Cretaceous sands (Lloyd aquifer) alongsouthwestern Long Island. The area where salt water isobserved in the Magothy-Jameco aquifer has complexhydrogeology with local erosion features that may playa role in the presence of salt water (Buxton and Shernoff1999). It appears that our model overpredicts the extent offresh water beneath southwestern Long Island extendingunder the ocean, although predevelopment salinity dataare not available. On the other hand, predevelopmentsalinity data of Pope and Gordon (1999) suggest thatour model appears to underpredict the amount of freshwater within the Upper Cretaceous sands beneath NewJersey. Within the shallow Miocene sand unit off NewJersey (section A-A’, Figure 8), fresh to brackish wateris transported to 100 km offshore. This is because theseshallow sands outcrop in Baltimore and Hudson Canyons,creating submarine springs which discharged brackishwater to the ocean during sea-level low stands.

Models for which the effect of glaciation wasremoved predicted higher salinity in regions proximal tothe ice sheet toe (e.g., blue dashed lines, Figure 9). Forseveral wells (6014, 6011, 6001), the computed salini-ties in the upper 200 m of the sea floor were about 5 ppt


Figure 9. Comparison of simulated (lines) and observed (red squares) concentrations (in parts per thousand) for offshoreAMCOR, COST, and ODP wells on Atlantic continental shelf. Simulations including (ice sheet, solid black line) and removing(sea level, blue dashed line) the effects of glacial loading are presented. The locations of the wells are presented in Figure 3A.

higher than for the model run, where the effects of glacia-tions were removed. In one well (6016), the computedsalinities were 15 ppt saltier when glaciations effects wereremoved. This suggests that glaciations had a pronouncedeffect on continental shelf salinities proximal to the icesheet toe. However, some wells far from the ice sheet(wells 6009, 6010, 6015, 6018) were not impacted byglaciation. We calculated the volume of sequestered freshwater in New England to be about 1300 km3 by inte-grating over the sand-dominated elements in our modelcontaining fresh water of less than 1 ppt.

During the modeled LGM, predicted recharge/discharge rates in the vicinity of the ice sheet reachedabout 2 m/year, about four times higher than modern esti-mates for Nantucket Island (Knott and Olimpio 1986).High hydraulic heads predicted near the toe of the icesheet are consistent with glaciotectonic features observedwithin tills on Nantucket Island and Martha’s Vineyard(Oldale 1985; Oldale and O’Hara 1984; Boulton et al.1995). There is a consistent pattern of recharge beneaththe ice sheet and focused discharge just beyond the icesheet toe (Figure 10A). Permeable sand units helped prop-agate the effects of high imposed heads beneath theice sheet outward over a wide region beyond the icesheet toe, thereby increasing flow rates. In unglaciatedregions during the LGM, higher than modern recharge

rates occur due to enhanced shore-normal hydraulic gra-dients associated with sea-level low stands (Figure 10B).Along the continental slope, paleo submarine springsdischarge groundwater (about 1 cm/year) along Hudsonand Baltimore Canyons (Figure 10C), where Miocenesands outcrop (Hampson and Robb 1984). Flow direc-tions fluctuate through geologic time within these canyons(Figure 10C–E). During periods of sea-level low standsand glaciation, brackish (16 ppt) groundwater dischargeoccurs (Figure 10C). Today, sea water invades these sandsalong the canyon (Figure 10D) and in the near-shore envi-ronment (red areas of along the sea floor in Figure 10D),where fresh water sands outcrop at the sediment–waterinterface, density-driven convection cells (Simmons et al.2001; Post and Kooi 2003) are seen in the model (areasbelow sea level with alternating fresh and sea water pat-terns in Figure 10D). Calculated temporal (Figure 10E)variations in submarine canyon discharge during sea-level low stands of the LGM help explain why freshwater extends farther off New Jersey than in areas far-ther to the north (Johnston 1983). Computed dischargerates (1 cm/year) are consistent with groundwater sappingfeatures observed during submersible dives along Balti-more Canyon off New Jersey (Robb 1984). It appearsthat submarine canyons influence the fresh water distri-bution only in relatively shallow, permeable deposits thatoutcrop (section A-A’, Figure 3).


Figure 10. Computed magnitude of surface groundwaterflux (in log m/year) during the last glacial maximum (LGM;A) and present day (B). Computed groundwater salinity(total dissolved solids in parts per thousand) concentrationduring the LGM (C) and present day (D). Computed tem-poral variations in vertical groundwater flux along HudsonCanyon (E) during the last glacial cycle (positive for dis-charge, negative for salt water reflux). The location of thedischarge point is shown in Figure 10D (black circle). Thelocations of New Jersey (NJ), Long Island (LI), and Nan-tucket Island (NT) are listed in Figure 10D to provide ageographical frame of reference.

Discussion and ConclusionsWhile our paleohydrologic model represents one of

the largest (spatially) and highest resolution regionalgroundwater models ever developed, it is idealized;nonetheless, it facilitates useful insight on the emplace-ment mechanisms, distribution, and volume of offshorefresh water. We did not account for surface water drainagepatterns (Lemieux et al. 2008a) or the effects of Plio-Pleistocene sedimentation on anomalous pore pressures(Dugan and Flemings 2000). We did not account forlocal sea level fluctuations associated with the effectsof a proglacial forebulge (Barnhardt et al. 1995). It islikely that during periods of glaciation, high hydraulicheads resulted in enhanced permeabilities due to loweffective stress (Boulton et al. 1995), as evidencedby glaciotectonic features observed on Nantucket andMartha’s Vineyard (Oldale and O’Hara 1984).

Overall, our model did a reasonable job of predict-ing offshore well salinity profiles (Figure 9). As notedearlier, our model may have overpredicted the extent offresh water beneath Long Island’s western shorelines.

Stumm (2001) reports salt water intrusion occurred fol-lowing about two decades of heavy (more than 3 billiongallons/year) pumping within the Magothy and Lloydaquifers on the Great Neck Peninsula, Long Island. Itshould be noted, however, that the Lloyd aquifer termi-nates (and crops out) along Long Island Sound proximal(2 to 3 km) to these well fields. The more problematic areais along the southern shore of western Long Island, wheresaline water has been observed in the Magothy-Jamecoaquifer since measurements have been made (Buxton andShernoff 1999). Possible erosional breaks in the confiningunit overlying the Magothy aquifer (not represented in ourregional scale model) could be the cause of this observedsalt water intrusion. However, it is important to note thatour hypothesis testing model is regional in scale and itis not possible to perfectly match local salinity variationseverywhere without excess parameter tuning, which weelected not to pursue.

Our model results indicated that glacial loadingreduced the salinity of groundwater in wells situated nearthe ice sheet margin between about 5 to 15 ppt. Wellsfar from the ice sheet margin in New Jersey were notsignificantly impacted by glaciations. Some of these NewJersey wells (e.g., well 6009, Figure 9) had remarkablylow salinity within shallow Miocene sands and werelocated far (greater than 100 km) offshore. Our modelresults indicate that this was caused by enhanced offshoreflow in these shallow sands which cropped out in HudsonCanyon. Our finite element model results suggest that bothice sheet loading and sea-level low stands have played asignificant role in the emplacement of the offshore freshgroundwater on New England’s continental shelf. Webelieve that the occurrence of distal offshore fresh water isprimarily controlled by sea level fluctuations and proximalfresh water is most strongly ice influenced by ice sheetsat high latitudes. This suggests that data from differentparts of the continental shelf are important to disentangleand constrain fresh water emplacement mechanisms.

Our back-of-the-envelope and model-based estimatesprovide only a first-cut assessment of the potential foroffshore fresh water. These calculations require field ver-ification. Porosity on the Atlantic continental shelf variesbetween about 0.6 and 0.1. If we vary the porositybetween 0.1 and 0.4, the resulting estimated fresh watervolume ranges from 500 to 2000 km3 for New England.Uncertainty regarding the thickness of fresh water satu-rated sediments and the seaward extent of the fresh waterlens could also change our estimates of fresh water vol-ume, probably by a factor of ±2. In addition, these esti-mates do not consider how much water can be extracted.As in petroleum extraction, practical recovery is likely tobe substantially less than 100%. Finally, there is consid-erable variability in offshore fresh water estimates amongthese four sections (12 to 0.2 km3/km). Hydrologic and(or) geophysical field data collected in the future could beused to refine the model, resolve more heterogeneity inhydraulic properties, and reduce prediction uncertainty.

With demands for fresh water increasing in coastalareas, sequestered continental shelf fresh water represents


a large, valuable, untapped, albeit nonrenewable resource.Many coastal megacities (e.g., Karachi, Pakistan; Rahmanet al. 1997) are facing acute water shortages, and localgroundwater resources are becoming contaminated. Theseproblems will become worse under conditions of climatechange (Nicholls 2004). To put these fresh water estimatesinto perspective, during the 20th century the cumulativevolume of fresh water withdrawn from the High PlainsAquifer in the United States for agricultural and otherpurposes is estimated to be about 270 km3 (Konikow2002). The volume of useable offshore water could beactually much greater than stated earlier if reverse osmosisfacilities (Schiermeier 2008; Shanon et al. 2008) wereused to process brackish (less than 10 ppt) offshoregroundwater. It is important to note, however, that inmany coastal settings, this fresh water is not beingreplenished today. Pumping of offshore aquifers willdeplete the resources and will eventually lead to salt waterintrusion.

Development of these resources would be expensiveand would require offshore drilling and construction ofpipelines to transfer water inland, and/or the use ofmodern land-based petroleum drill rigs capable of drillinglarge-scale (greater than 10 km) horizontal wells to accessthe offshore resource. Marine seismic surveys wouldneed to be conducted to ensure that drilling hazardsare avoided, such as natural gas deposits hosted instructural traps. In addition, extensive hydrologic andgeophysical characterization would be necessary to mapthe distribution of groundwater salinity offshore andidentify viable drilling targets. Electromagnetic methods(e.g., transient and frequency-domain controlled sourcetechniques) are well suited to the first task (e.g., Hoefeland Evans 2001), whereas marine multichannel seismicmethods are well suited to the second.

Our study may have implications for estimating sub-marine groundwater discharge to oceans because cur-rent levels reported in the literature (e.g., Li et al. 1999;Taniguchi 2002; Michael et al. 2005) may not be represen-tative of long-term (i.e., Pleistocene) average conditions(Figure 10).

AcknowledgmentsThis work was supported by the U.S. National

Science Foundation and the Malcolm & Sylvia BoyceEndowment at Indiana University. We gratefully acknowl-edge computational support from the NSF TeragridProgram at Indiana University and the New MexicoNMCAC supercomputer center, and support from theUSGS Groundwater Resources Program. We thank Jean-Michel Lemieux and Steve Ingebritsen for their thoroughreviews of this paper.

ReferencesAustin, J.A., N. Christie-Blick, M. Malone, S. Berne, M.K.

Borre, G. Claypool, J. Damuth, H. Delius, G. Dickens,P. Flemings, C. Fulthorpe, S. Hesselbo, K. Hoyanagi,

M. Katz, H. Krawinkel, C. Major, F. McCarthy, C. Mc-Hugh, G. Mountain, H. Oda, H. Olson, C. Pirmez,C. Savrda, C. Smart, L. Sohl, P. Vanderaveroet, W. Wei,and B. Whiting. 1998. Continuing the New Jersey Mid-Atlantic sea-level transect; covering Leg 174A of thecruises of the drilling vessel JOIDES Resolution, Halifax,Nova Scotia, to New York, New York. Proceedings of theOcean Drilling Program, Part A: Initial Reports 174A:1071–1073. Reston, Virginia: USGS.

Barnhardt, W.A., W.R. Gehrels, D.F. Belknap, and J.T. Kelley.1995. Late Quaternary relative sea-level change in thewestern Gulf of Maine: Evidence for a migrating glacialforebulge. Geology 23, no. 4: 317–320.

Bense, V.F., and M.A. Person. 2008. Transient hydrodynam-ics within intercratonic sedimentary basins during glacialcycles. Journal of Geophysical Research 113, 1–17.

Boulton, G.S., P.E. Caban, and K. van Gijssel. 1995. Ground-water flow beneath ice sheets. Part 1, large scale patterns.Quaternary Science Reviews 14, no. 6: 545–562.

Bramley, R., and X. Wang. 1995. A User’s Guide for theCADYF-SPLIB Project. Bloomington: Department of Com-puter Science, Indiana University. ftp://ftp.cs.indiana.edu/pub/techreports/TR453.ps.Z.

Breemer, C.W., P.U. Clark, and R. Haggerty. 2002. Modelingthe subglacial hydrology of the late Pleistocene LakeMichigan Loge, Laurentide Ice Sheet. Geological Societyof America Bulletin 114, no. 6: 665–674.

Buxton, H.T., and P.K. Shernoff. 1999. Ground-water resourcesof Kings and Queens Counties, Long Island, New York.U.S. Geological Survey Water-Supply Paper 2498. Reston,Virginia: USGS.

Buxton, H., and E. Modica. 1992. Patterns and rates of ground-water flow on Long Island, New York. GroundWater 30,no. 6: 857–866.

Denton, G.H., and T.J. Hughes. 1981. The Last Great Ice Sheets.New York: John Wiley & Sons.

de Marsily, G. 1986. Quantitative Hydrogeology, GroundwaterHydrology for Engineers. New York: Academic Press.

Dugan, B., and P.B. Flemings. 2002. Fluid flow and stabilityof the US continental slope offshore New Jersey from thePleistocene to the present. Geofluids 2, no. 2: 137–146.

Dugan, B., and P.B. Flemings. 2000. Overpressure and fluidflow in the New Jersey continental slope: implications forslope failure and cold seeps. Science 289, 288–291.

Edmunds, W.M. 2001. Paleowaters in European coastalaquifers: The goals and main conclusions of thePALAEAUX Project. Geological Society Special Publica-tions 189, 1–16.

Engelhard, H., and B. Kamb. 1997. Basal hydraulic system ofa west Antarctic ice stream: Constraints from boreholeobservations. Journal of Glaciology 43, 207–230.

Folger, D.W., R. Hathaway, R.A. Christopher, P. Valentine, andC. Poag. 1978. Stratigraphic test well. Nantucket Island,Massachusetts. U.S. Geological Survey Circular C 773.Reston, Virginia: USGS.

Gable, C.W., H.E. Trease, and T.A. Cherry. 1996. Geologicalapplications of automatic grid generation tools for finiteelements applied to porous flow modeling. Numerical GridGeneration in Computational Fluid Dynamics and RelatedFields: 1007-1015.

Garven, G., and R.A. Freeze. 1984. Theoretical analysis of therole of groundwater flow in the genesis of stratabound oredeposits. 1. Mathematical and numerical model. AmericanJournal of Science 284, 1085–1124.


Gelhar, L.W. 1993. Subsurface Stochastic Hydrology. Engle-wood Cliffs, New Jersey: Prentice Hall.

Gill, H.E. 1962. Ground-water resources of Cape May County,New Jersey: Salt-water invasion of principal aquifers.New Jersey Department of Conservation and EconomicDevelopment Special Report 18.

Guswa, J.H., and D.R. Leblanc. 1981. Digital models of ground-water flow in the Cape Cod aquifer system, Massachusetts.Water Resources Investigation Open-File Report: 80-67.

Hampson, J., and J. Robb. 1984. A geologic map of the conti-nental slope off New Jersey: Lindenkohl Canyon to TomsCanyon. U.S. Geological Survey Miscellaneous Investiga-tions Map I-1608. Reston, Virginia: USGS.

Hathaway, J.C., C.W. Poag, P.C. Valentine, R.E. Millerk, D.M.Schultz, F.T. Manheim, F.A. Kohout, M.H. Bothner, andD.A. Sangrey. 1979. U.S. Geological Survey core drillingon the Atlantic Shelf, 1979. Science 206, no. 4418:515–527.

Hayba, D.O., and S.E. Ingebritsen. 2001. The computer modelHYDROTHERM, a three-dimensional finite-differencemodel to simulate ground-water flow and heat transport inthe temperature range of 0 to 1,200 degrees C. U.S. Geo-logical Survey Water-Resources Investigations 94-4045.Reston, Virginia: USGS.

Henry, H. 1959. Saltwater intrusion into fresh-water aquifer.Journal of Geophysical Research 64, no. 11: 1911–1919.

Hill, M.C. 1998. Methods and guidelines for effective model cal-ibration. U.S Geological Survey Water- Resources. Inves-tigations Report 98-4005. Reston, Virginia: USGS.

Hoefel, F, and R. Evans. 2001. Impact of low-salinity porewateron seafloor electromagnetic data: A means of detectingsubmarine groundwater discharge? Estuarine Coastal andShelf Science 52, 179–189.

Hooyer, T.S., and N.R. Iverson. 2002. Flow mechanism of theDes Moines lobe of the Laurentide ice sheet. Journal ofGlaciology 48, 575–587.

Imbrie, J., J. Hays, D. Martinson, A. McIntyre, A.C. Mix, J.Morley, N.G. Pisias, W.L. Prell, and N.J. Shackleton.1984. The orbital theory of Pleistocene climate: Supportfrom a revised chronology of the marine δ18O record.In Milankovitch and Climate, 269–305. Boston, Mas-sachusetts: D. Reidel Publishing Company.

Johnsen, S.J., D. Dahl-Jensen, W. Dansgaard, and N. Gunde-strup. 1995. Greenland paleotemperatures derived fromGRIP borehole temperature and ice core isotope profiles.Tellus 47B, 624–629.

Johnston, R. 1983. The saltwater-freshwater interface in the Ter-tiary limestone aquifer, Southeast Atlantic outer-continentalshelf of the U.S.A. Journal of Hydrology 61, no. 1–3:239–249.

Karypis, G., and V. Kumar. 1996. A parallel algorithm for mul-tilevel mesh partitioning and sparse matrix ordering. 10th

International Parallel Processing Symposium, Honolulu,Hawaii, 314–319.

Katz, B.G., J.B. Lindner, and S.E. Ragone. 1980. An evaluationof sources of nitrogen as causes of ground-water contam-ination in Nassau County, Long Island. GroundWater 18,607–616.

Kaye, C.A. 1964. Outline of pleistocene geology of Martha’sVineyard, Massachusetts. U.S. Geological Survey Pro-fessional Paper P0501-C. C134–C139. Reston, Virginia:USGS.

Kestin, J., H.E. Khalifa, and R. Corriea. 1981. Tables of thedynamics and kinematic viscosity of aqueous NaCl solu-tions in the temperature range of 20–150◦C and the pres-sure range of 0.1–35 MPa. Journal of Physical ChemistryReference Data 10, 71–87.

Kleinberg, R., and D. Griffin. 2005. NMR measurements ofpermafrost: Unfrozen water assay, pore-scale distribution ofice, and hydraulic permeability of sediments. Cold RegionsScience and Technology 42, 63–77.

Klitgord, K.D., and C.M. Schneider. 1994. Geophysicaldatabase of the East Coast of the United States, northernAtlantic margin: Velocity analyses. U.S. Geological SurveyOpen-File Report: 94-192. Reston, Virginia: USGS.

Klitgord, K.D., C.W. Poag, C.M. Schneider, and L. North.1994. Geophysical database of the East Coast of the UnitedStates, northern Atlantic margin: Cross sections and griddeddatabase (Georges Bank basin, Long Island platform, andBaltimore Canyon trough). U.S. Geological Survey Open-File Report: 94-637. Reston, Virginia: USGS.

Knott, J.F., and J.C. Olimpio. 1986. Estimation of recharge ratesto sand and gravel aquifer using environmental tritium.U.S. Geological Survey Water-Supply Paper 2297. Reston,Virginia: USGS.

Kohout, F.A., J.C. Hathaway, D.W. Folger, M.H. Bothner, E.H.Walker, D.F. Delaney, M.H. Frimpter, E.G.A. Weed, andE.V.C. Rhodehamel. 1977. Fresh groundwater stored inaquifers under the continental shelf, Implications froma deep test, Nantucket Island, Massachusetts. WaterResources Bulletin 13, no. 2: 373–386.

Konikow, L.F. 2002. Ground-water depletion and overexploita-tion; a global problem. Geological Society of AmericaAbstracts with Programs 34, 229.

Konikow, L.F., and D.B. Grove. 1977. Derivation of equationsdescribing solute transport in groundwater. U.S. Geologi-cal Survey Water-Resources Investigations 77-19. Reston,Virginia: USGS.

Kooi, H., and K. Groen. 2001. Offshore continuation of coastalgroundwater systems; predictions using sharp-interfaceapproximations and variable-density flow modelling. Jour-nal of Hydrology 246, 19–35.

Kooi, H J., K. Groen, and A. Leijnse. 2000. Modes of seawaterintrusion during transgressions. Water Resources Research36, no. 12: 3581–3589.

Lemieux, J.-M., E. Sudicky, W. Peltier, and L. Tarasov. 2008a.Simulating the impact of glaciations on continental ground-water flow systems. I. Relevant processes and model for-mulation. Journal of Geophysical Research 113, F03017.

Lemieux, J.-M., E. Sudicky, W. Peltier, and L. Tarasov. 2008b.Simulating the impact of glaciations on continental ground-water flow systems. II. Model application to the Wiscon-sinian glaciation over the Canadian landscape. Journal ofGeophysical Research 113, F03018.

Lemieux, J.M., E.A. Sudicky, W.R. Peltier, L. Tarasov. 2008c.Dynamics of groundwater recharge and seepage over theCanadian landscape during the Wisconsinian glaciation.Journal of Geophysical Research 113, F01011.

Li, L., D.A. Barry, F. Stagnitti, and J.Y. Parlange. 1999. Subma-rine groundwater discharge and associated chemical inputto a coastal sea. Water Resources Research 35, no. 11:3253–3259.

Malone, M., M. Claypool, G. Jonathan, and G.R. Dickens.2002. Variable methane fluxes in shallow marine systemsover geologic time: The composition and origin of pore


waters and authigenic carbonates on the New Jersey shelf.Marine Geology 189, no. 3–4: 175–196.

Marksamer, A.J., M. Person, F. Day-Lewis, J.W. Lane, D.Cohen, B. Dugan, H. Kooi, and M. Willett. 2007. Integrat-ing geophysical, hydrochemical, and hydrologic data tounderstand the freshwater resources on Nantucket Island,Massachusetts. In: Subsurface Hydrology Data Integrationin Subsurface Hydrology, 1–54. Washington, DC: AGUWater Resources Monograph.

Marshall, S.J., and P. Clark. 2002. Basal temperature evolutionof North American ice sheets and implications for the100-kyr cycle. Geophysical Research Letters 29, 769–793.

Mattick, R.E., and J.L. Hennessy. 1980. Structural framework,stratigraphy, and petroleum geology of the area of oil andgas lease sale No. 49 on the U.S. Atlantic continental shelfand slope. U.S. Geological Survey Circular C812. Reston,Virginia: USGS.

Meisler, H., P.P. Leahy, and L.L. Knobel. 1984. Effect ofEustatic sea-level changes on saltwater-freshwater in NorthAtlantic Coastal Plain. U.S. Geological Survey Water-Supply Paper 2255. Reston, Virginia: USGS.

Michael, H.A., A.E. Mulligan, and C.F. Harvey. 2005. Seasonaloscillations in water exchange between aquifers and thecoastal ocean. Nature 436, no. 7054: 1145–1148.

Mountain, G.S., K.G. Miller, P. Blum, P.G. Alm, M.P.Aubry, L.H. Burckle, B.A. Christensen, J. Compton, J.E.Damuth, J.F. Deconinck, L. de Verteuil, C.S. Fulthorpe, S.Gartner, G. Guerin, S.P. Hesselbo, B. Hoppie, M.E.Katz, N. Kotake, J.M. Lorenzzo, S. McCracken, C.M.McHugh, W.C. Quayle, Y. Saito, S.W. Snyder, W.G. tenKate, M. Urbat, M.C. Van Vossen, and A. Vecsei. 1994.New Jersey continental slope and rise; covering Leg 150of the cruises of the drilling vessel JOIDES Resolution,Lisbon Harbor, Portugal, to St. John’s Newfoundland. InProceedings of the Ocean Drilling Program, Part A: InitialReports, 902-906. College Station, Texas, Ocean DrillingProgram.

Mulligan, A., and E. Uchupi. 2003. New interpretation of glacialhistory of Cape Code may have important implications forgroundwater contaminant transport. EOS Transactions 84,no. 19: 177–183.

Nicholls, R. 2004. Coastal megacities and climate change.GeoJournal 37, 369–379.

Noorishad, J., C.F. Tsang, P. Perrochet, and A. Musy. 1992.A perspective on the numerical solution of convection-dominated transport problems: A price to pay for the easyway out. Water Resources Research 28, no. 2: 551–561.

Oldale, R.N. 1985. Geologic map of Nantucket and nearbyislands, Massachusetts, Miscellaneous InvestigationsSeries. U.S. Geological Survey Report I-1580. Reston,Virginia: USGS.

Oldale, R.N., and C.J. O’Hara. 1984. Glaciotectonic origin ofthe Massachusetts coastal end moraines and a fluctuatinglate Wisconsinan ice margin. Geological Society of AmericaBulletin 95, 61–74.

Oldale, R.N., P.C. Valentine, T.M. Cronin, E.C. Spiker, B.W.Blackwelder, D.F. Belnap, J.F. Wehmiller, and B.J. Szabo.1982. Stratigraphy, structure, absolute age, and pale-ontology of the upper Pleistocene deposits at SankatyHead, Nantucket Island, Massachusetts. Geology 10, no. 5:246–252.

Peltier, W.R. 1998. Postglacial variations in the level of thesea: Implications for climate dynamics and solid-earthgeophysics. Reviews of Geophysics 36, 603–689.

Perlmutter, N.M., and J.J. Geraghty. 1963. Geology and ground-water conditions in southern Nassau and southeasternQueens Counties, Long Island, New York. U.S. GeologicalSurvey Water Supply Paper 1613-A. Reston, Virginia:USGS.

Perlmutter, N.M., J.J. Geraghty, and J.E. Upson. 1959. Therelation between fresh and salty groundwater in southernNassau and southeastern Queens Counties, Long Island,New York. Economic Geology 54, 416–435.

Person M., J. McIntosh, V. Bense, V.H. Remenda. 2007a. Pleis-tocene hydrology of North America: The role of ice sheetsin reorganizing groundwater flow systems. Reviews of Geo-physics 45, 1–28.

Person, M., P. Roy, H. Wright, E. Ito, T. Winter, D. Rosen-berry, W. Gutowski, D. Cohen, and P. Morin. 2007b.Hydrologic response of the Crow Wing watershed, Min-nesota to mid-Holocene climate change. Geological Societyof America Bulletin 119, no. 3–4: 363–376.

Person, M., B. Dugan, J. Swenson, L. Urbano, C. Stott, J. Tay-lor, and M. Willett. 2003. Pleistocene hydrogeology of theAtlantic continental shelf, New England. Geological Soci-ety of America Bulletin 115, 1324–1343.

Person, M., J. Taylor, and S.L. Dingman. 1998. Sharp interfacemodels of salt water intrusion and well head delineationon Nantucket Island, Massachusetts. GroundWater 36,731–742.

Piotrowski, J.A., and A.M. Kraus. 1997. Response of sedimentsto ice-sheet loading in northwestern Germany: Effectivestress and glacier-bed stability. Journal of Glaciology 145,495–502.

Poag, C.W. 1982. Stratigraphic reference section for GeorgesBank Basin-depositional model for New England PassiveMargin. American Association of Petroleum GeologistsBulletin 66, no. 8: 1021–1041.

Pope D.A., and A.D. Gordon. 1999. Simulation of ground-water flow and movement of the freshwater-saltwaterinterface in the New Jersey Plain. U.S. Geological SurveyInvestigations Report 98-4216. Reston, Virginia: USGS.

Post, V.E.A., and H. Kooi. 2003. On rates of salinization by freeconvection in high-permeability sediments; insights fromnumerical modeling and application to the Dutch coastalarea. Hydrogeology Journal 11, 549–559.

Provost, A.M., C.I. Voss, and C.E. Neuzil. 1998. SKI SITE-94 Glaciation and regional ground-water flow in theFennoscandian Shield. SKI Report 96. Stockholm: SwedishNuclear Power Inspectorate.

Rahman, A., H. Lee, and M. Khan. 1997. Domestic watercontamination in rapidly growing megacities of Asia: Caseof Karachi, Pakistan. GeoJournal 44, 339–360.

Redfield, A.C., and M. Rubin. 1962. The age of salt marshpeat and its relation to recent changes in sea level inBarnstable, Massachusetts. Proceedings of the NationalAcademy of Sciences of the United States of America 48,no. 10: 1728–1735.

Robb, J. 1984. Spring sapping on the lower continental slope,offshore New Jersey. Geology 12, no. 5: 278–282.

Sass, J.H., D.D. Blackwell, D.S. Chapman, J.K. Costain, E.R.Decker, L.A. Lawver, C.A. Swanberg, D. Blackstone, C.Brott, H. Heasler, A. Lachenbruch, B.V. Marshall, P. Mor-gan, R. Munroe, and J. Steele. 1989. Heat flow from thecrust of the United States. Physical Properties of Rocksand Minerals 2, no. 2: 503–548.

Schiermeier, Q. 2008. Purification with a pinch of salt. Nature452, no. 7185: 260–261.


Schaefer, F.L., and R.L. Walker. 1981. Saltwater intrusion intothe Old Bridge Aquifer in the Keyport-Union Beach areaof Monmouth County, New Jersey. U.S. Geological Sur-vey Water-Supply Paper Report W2184. Reston, Virginia:USGS.

Schlee, J.S., and J. Fritsch. 1982. Seismic stratigraphy of theGeorges Bank Basin complex offshore New England.American Association of Petroleum Geologists Memoir 43,223–251.

Scholle, P.A. 1980. Geologic studies on the COST No. B-3Well, U.S. Mid-Atlantic Outer Continental Shelf Area. U.S.Geological Survey Circular 833. Reston, Virginia: USGS.

Scholle, P.A. 1977. Geologic studies on the COST No. B-2Well, U.S. Mid-Atlantic Outer Continental Shelf Area. U.S.Geological Survey Circular 750. Reston, Virginia: USGS.

Shanon, M.A., P.W. Bohn, M. Elimelech, J.G. Georgiadis, B.J.Marinas, and A.M. Mayes. 2008. Science and technologyfor water purification in coming decades. Nature 452,301–310.

Shreve, R.L. 1985. Late Wisconsin ice-surface profile calculatedfrom esker paths and types, Katahdin esker system, Maine.Quaternary Research 23, 27–37.

Siegel, D.I., and R.J. Mandle. 1984. Isotopic evidence forglacial melt water recharge to the Cambrian-OrdovicianAquifer, north-central United States. Quaternary Research22, 328–335.

Simmons, C.T., T.R. Fenstemaker, and J.M. Sharp. 2001.Variable-density groundwater flow and solute transport inheterogeneous porous media: Approaches, resolutions andfuture challenges. Journal of Contaminant Hydrology 52,245–275.

Starinsky A., and A. Katz. 2003. The formation of naturalcryogenic brines. Geochimica et Cosmochimica Acta 67,1475–1484.

Steenhuis, T., C. Jackson, S. Kung, and W. Brutsaert. 1985.Measurement of groundwater recharge on eastern Long

Island, New York, U.S.A. Journal of Hydrology 79,no. 1–2: 145–169.

Stumm, F. 2001. Hydrogeology and extent of saltwater intrusionof the Great Neck Peninsula, Great Neck, Long Island, NewYork. U.S. Geological Survey Water-Resources Investiga-tions Report 99-4280. Reston, Virginia: USGS.

Tarasov, L., and W.R. Peltier. 2007. Coevolution of continentalice cover and permafrost extent over the last glacial-interglacial cycle in North America. Journal of GeophysicalResearch 112, F02S08.

Tarasov, L., and W.R. Peltier. 2004. A geophysically con-strained large ensemble analysis of the deglacial history ofthe North American ice-sheet complex. Quaternary ScienceReviews 23, 359–388.

Taniguchi, M. 2002. Tidal effects of submarine groundwaterdischarge into the ocean. Geophysical Research Letters 29,no. 12: 2.1–2.3.

Tuminaro, R.S., M. Heroux, S.A. Hutchinson, and J.N. Shadid.1999. Official Aztec User’s Guide: Version 2.1. SandiaNational Laboratories Technical Report, SAND99-8801J.http://www.cs.sandia.gov/CRF/aztec1.html.

Uchupi, E., N. Driscoll, R.D. Ballard, and S.T. Bolmer. 2001.Drainage of late Wisconsin glacial lakes and the mor-phology and late quaternary stratigraphy of the New Jer-sey–southern New England continental shelf and slope.Marine Geology 172, 1170–145.

van der Veen, C.J. 2007. Fracture propagation as means ofrapidly transferring surface meltwater to the base ofglaciers. Geophysical Research Letters 34, L01501.

van der Veen, C.J. 1998. Fundamentals of Glacier Dynamics.Rotterdam, Netherlands: A.A. Balkema.

Zheng, C., and G.D. Bennett. 1994. Applied Contaminant Trans-port Modeling. New York: Van Rostrand Reinhold.

Zwally, H.J., W. Abdalati, T. Herring, K. Larson, J. Saba, andK. Steffen. 2002. Surface melt induced acceleration ofGreenland ice-sheet flow. Science 297, 218–222.


Case Study/ Origin and Extent of Fresh Paleowaters … · Case Study/ Origin and Extent of Fresh Paleowaters on the Atlantic Continental Shelf, USA ... Orange and yellow shaded patterns

Documents