1

Hydrologic forcing of submarine groundwater discharge: insight from a

seasonal study of radium isotopes in a groundwater dominated salt marsh

estuary

Matthew A. Charette

Department of Marine Chemistry and Geochemistry

Woods Hole Oceanographic Institution

Woods Hole, Massachusetts 02543 USA

Submitted to:

Limnology and Oceanography

May 16, 2006

Running head: Salt marsh groundwater discharge

*e-mail, mcharette@whoi.edu; tel, 508-289-3205; fax, 508-457-2193

2

Acknowledgements

I thank Craig Herbold, Matt Allen, and Adam Rago for assistance in the field and

laboratory and Paul Henderson and Jack Cook for figure preparation. Henrieta Dulaiova

provided useful comments that improved the manuscript. This work is a result of research

sponsored by NSF (OCE-0346933) and the NOAA Sea Grant College Program Office,

Department of Commerce, under Grant No. NA86RG0075, Woods Hole Oceanographic

Institution, Sea Grant Project No. 22850063. The views expressed herein are those of the author

and do not necessarily reflect the views of NOAA or any of its subagencies. Matching funds

were provided by the Cove Point Foundation.

3

Abstract

A seasonal study of radium-derived submarine groundwater discharge (SGD) and

associated nitrogen fluxes was carried out in a salt marsh estuary between 2001-2003 (Pamet

River Estuary, MA). Twelve-hour time-series at the estuary inlet of salinity and radium were

used to determine the relative importance of fresh versus saline SGD, respectively. In addition,

the distinct radium (228Ra/226Ra) isotopic signature of marsh peat pore water and aquifer-derived

brackish groundwater were used to further partition the Ra-derived SGD estimate. Of these three

groundwater sources, only the marsh-derived groundwater was constant over time. The ratio of

brackish to fresh SGD was inversely correlated with water table elevation in the aquifer,

suggesting that Ra-derived SGD was enhanced during dry periods. The various SGD fluxes were

responsible for an average annual dissolved inorganic nitrogen (DIN) input of between 1.7 and

7.1 mol m-2 y-1 and a soluble reactive phosphate (SRP) flux of 0.13 to 0.54 mol m-2 y-1. When

taking into account the dissolved nutrient load being exported from the marsh inlet, it was

determined that approximately 30% of the SGD-derived DIN and SRP flux is exported to coastal

waters (Cape Cod Bay) while 70% is retained by the salt marsh ecosystem.

4

Introduction

Submarine groundwater discharge (SGD) has been increasingly recognized as an

important source of nutrients for a wide variety of coastal marine environments (Johannes, 1980;

Valiela et al., 1990; Krest et al., 2000). In some settings, subsurface input of nutrients has been

shown to be more important than surface inputs including both rivers and atmospheric deposition

(Valiela et al., 1992). One major limitation in quantifying SGD-derived fluxes has been reliable

techniques for determining the flux. Chemical tracers have become a popular tool for the study

of SGD (e.g. Moore, 1996; Cable et al., 1996). A primary desirable characteristic of chemical

tracers of SGD is that they have a unique or dominant source in groundwater and are not present

in any significant quantity in other vector (e.g. river water, rainfall) for water flux to the coastal

ocean.

Radium isotopes have proven to be useful tracers of total SGD in many environments on

both small and large scales from salt marshes (Rama and Moore, 1996; Krest et al., 2000;

Charette et al., 2003) to estuaries (Charette et al., 2001; Kelly and Moran, 2002; Yang et al.,

2002) and to the continental shelf (Moore, 1996). The chemical behavior of radium is such that

its Kd decreases significantly in saline environments, mainly due to cation exchange processes

(Li and Chan, 1979). Thus, radium is usually only enriched (relative to surface water) in brackish

to saline groundwater; fresh SGD that does not interact with saline groundwater in “subterranean

estuary” will not acquire a radium signal and hence may not be quantified (Mulligan and

Charette, 2006). This characteristic of the Ra-derived SGD tracer has led to some controversy

and confusion in the literature, usually due to the observation that Darcy’s Law and other

traditional hydrogeologic-based groundwater flux estimates are often significantly lower than

5

Ra-based SGD (Burnett et al., 2001; Smith and Zawadzki, 2003). Such differences can often be

attributed to the tendency for SGD to include a substantial component of seawater that has been

recirculated through coastal marine sediments (Burnett et al., 2006).

The existence of four naturally occurring radium isotopes (224Ra, t1/2 =3.66 days; 223Ra,

t1/2 =11.4 days; 228Ra, t1/2 =5.75 years; 226Ra, t1/2 =1600 years) make Ra particularly useful for

quantifying multiple sources of SGD, such as fluid originating from confined versus surficial

aquifers (Crotwell and Moore, 2003; Moore, 2003; Charette and Buesseler, 2004). This approach

is made possible through two primary mechanisms. First, aquifers with different principal

mineral or sediment types can have differing degrees of uranium (238U -> 226Ra; 235U -> 223Ra)

and thorium (232Th -> 228Ra -> 224Ra) series isotopes. Secondly, the frequency of seawater

circulation through an aquifer can leave sediments, which are the ultimate source of Ra isotopes

in groundwater, enriched in the shorter lived isotopes and depleted in the longer lived isotopes

due to the relative differences in ingrowth rates from their thorium parents.

Here, I examine radium-derived submarine groundwater discharge and associated

nitrogen fluxes in a salt marsh estuary (Pamet River Estuary, MA). This site was chosen because

groundwater is the only source of freshwater to the system. By performing both salinity and

radium mass balances during five time periods over the course of ~20 months, I determined the

relative importance of fresh versus saline SGD, respectively. Also, I used the distinct radium

(228Ra/226Ra) isotopic signature of marsh peat pore water and aquifer-derived saline groundwater

to estimate SGD fluxes for each of these sources using a three endmember mixing model. The

results presented here suggest that radium isotopes will be useful for determining the dominant

6

form of SGD-derived nitrogen in salt marsh estuaries: i.e. ‘new’ vs. ‘recycled’ nitrogen, and also

salt marsh primary productivity.

7

Study Area

The Pamet River estuary is located on outer Cape Cod in Truro, MA (Fig. 1). It has an

area of 930,000 m2 and a fairly large tidal range of >3 m, which means that it is very well flushed

during each tidal cycle. Salinity in the surface waters range from 0 to 32. The freshwater

endmember is almost exclusively derived from groundwater; the sediments of Cape Cod are

highly permeable, therefore surface water runoff is almost non-existent except in the case of

large rainfall events. The high salinity endmember is that of Cape Cod Bay, the body of water

that exchanges with the Pamet during each tidal cycle. There are two groundwater lenses that

supply the Pamet with freshwater, the Pamet Lens to the north and the Chequesset Lens to the

south (Eichner et al., 1997). The hydraulic conductivity of the sediments is high and the

hydraulic gradient of the water table elevations adjacent to the marsh are steep. These factors

lead to a significant groundwater influence for this marsh system.

8

Methods

To investigate the dynamics of submarine groundwater discharge in salt marsh systems, a

two-year seasonal study of Ra isotopes in the Pamet River Estuary, Truro, MA was undertaken.

The study design involved a two-pronged approach. First, during each period (July 2001, March

2002, July 2002, November 2002, and March 2003), a 12-hour time series sampling of Ra

isotopes and nutrients was conducted at the inlet to the estuary (Fig. 1). For radium isotopic

analysis, each hour 100 L of water was filtered into a polyethylene barrel, which was then slowly

pumped (~1 L min-1) through MnO2 coated acrylic fiber to extract the Ra (Moore and Reid,

1976). Subsamples for nutrients and salinity were also collected at this time. Tidal height,

temperature, and salinity were monitored every 15 min. using a YSI 600XLM CTD. By using

this sampling approach, a whole-marsh estimate of Ra export and hence SGD was obtained

simply by knowing net Ra flux (ebb tide minus flood tide inventories) and the tidal prism

(630,000 m3). In order to maintain consistency between sampling periods, sample collection was

always conducted on the highest spring tide of the month.

We also collected more than 40 groundwater and pore water samples from various

locations along the marsh edge and from within various sediment types (marsh peat, quartz sand)

of the system. We used a drive-point piezometer system called Retract-A-Tip (AMS, Inc.) to

accomplish this task (Charette and Allen, 2006). Briefly, the stainless steel piezometer was

driven to the depth of interest. Samples were pumped through Teflon tubing using a peristaltic

pump. For Ra analysis, groundwater was pumped directly through the fiber and the filtrate (10-

20 L) was collected to determine the sample volume. Samples for nutrients were collected into

30 mL, acid cleaned scintillation vials using a plastic syringe and Millipore Sterivex filter. Basic

9

water properties including salinity, pH, dissolved oxygen, and redox potential were recorded

using a YSI 600R and 650MDS handheld computer. The majority of the groundwater samples

were collected during July 2002, though a number of samples were collected during the winter

(March) and fall (November) time periods as well.

For the sediment desorption experiment, aquifer sands were collected from two locations:

(1) from the inland portion of the aquifer such that the sediments had not been subject to salt

water intrusion, and (2) from the marine portion of the aquifer, where they had been exposed to

fully saline groundwater (~28-30) for an extended period of time (~decades). Groundwater was

then collected from each location and passed through a column of MnO2 fiber to remove the

radium. The Ra-free water was filtered and mixed such that we had five treatments for each

sediment type (salinity = 0, 7.5, 15, 22.5, 30). Each treatment was run in triplicate and consisted

of a ~400 g aliquot of sediment added to 20 L of Ra-free groundwater. After 48 hours, the water

was filtered and the desorbed Ra was concentrated onto MnO2 fiber for analysis by the

techniques described below.

Back in the laboratory, the MnO2-fiber was rinsed and partially dried. Activities of 223Ra

(t1/2 = 11.4 days) and 224Ra (t1/2 = 3.66 days) were measured on a delayed coincidence counter as

described by Moore and Arnold (1996). The fiber was then ashed in a muffle furnace (820°C for

16 hours), ground, and homogenized before being packed in a counting vial and sealed with

epoxy to prevent 222Rn loss (Charette et al., 2001). Once 222Rn had reached secular equilibrium

with its parent, activities of 226Ra (t1/2 = 1600 years) and 228Ra (t1/2 = 5.75 years) were

determined by γ-counting in a well detector (Canberra, model GCW4023) by the ingrowth of

10

214Pb (352 keV) and 228Ac (911 keV). Calibration of the well detector was achieved by counting

four ashed MnO2 fiber standards having the same activity range and geometry of the samples.

Nutrient analyses (nitrate, phosphate, ammonium, silicate) were performed using standard

methods on a Lachat QuickChem 8000 Flow Injection Analyzer.

11

Results

Radium isotope activities from the five time series samplings are presented in Table 1.

For all time periods, there was an inverse correlation between tidal stage and radium activity

(Fig. 2). In general, the flood tide (derived from Cape Cod Bay) radium values were consistently

low–within the range of samples collected in other shelf regions (e.g. Moore, 2000). However,

there was significant variability in Ra activity between time periods for the ebb tide samples. For

example, 226Ra in March 2002 peaked at 62 dpm 100 L-1 compared with only 20 dpm 100 L-1

during March 2003.

For each individual time period, radium in surface water was generally constant above a

salinity of ~25, then generally increased with decreasing salinity (Fig. 3). However, interestingly,

radium isotopes did not display a consistent trend with salinity when compared between time

periods. The 226Ra maximum for March 2002 occurred at a salinity of ~27, while the highest

226Ra for other periods was found at lower salinities (e.g. ~11 for March 2003). These differences

were not correlated with the time of year, e.g. winter versus summer, as had been observed for

several prior studies (e.g. Bollinger and Moore, 1993; Kelly and Moran, 2002).

The same temporal pattern in Ra isotopes was observed during the time-series sampling

regardless of the time of year (Fig. 4). During flood tide, salinity was high and Ra was low, both

values typical for the source of this water: Cape Cod Bay. During early ebb tide, values were

very similar to the Cape Cod Bay water. Then, approximately 3-4 hours after peak high tide,

salinity began to drop while Ra increased. Looking in closer detail, during peak ebb tide, salinity

began to increase while Ra dropped only slightly. The salinity minimum is derived from

12

freshwater that has accumulated in the main channel of the estuary during the previous 12 hours.

In the meantime, the low permeability salt marsh sediments continue to drain the high salinity

water that occupied their pore spaces during the previous flood tide. This intermediate salinity,

high Ra water is derived from this tidal pumping/inundation process.

The temporal pattern and trend with salinity in the time-series nutrients was very similar

to that of radium, indicating that they both have a common source in groundwater (Table 1, Fig.

5). During March 2002, dissolved inorganic nitrogen (DIN) tracked the increase in radium

during the outgoing tide (Fig. 4). Nutrient concentrations in the water exiting the marsh were

significantly higher than the flood tide water, a result of significant nutrient input combined with

rapid estuarine flushing.

There was no discernable trend between groundwater salinity and radium or nutrients

(Table 2). This is likely a function of the fact that numerous kinds of pore water samples were

collected (groundwater seeps, marsh sediments, coarse-grained aquifer sediments) and that they

were collected over wide spatial scales. For example, 226Ra ranged from 2 (below average bay

water) to 181 dpm 100 L-1 (~ 10x average estuarine surface water). On average, though, radium

isotopes and nutrients were 2-3x enriched in groundwater relative to surface water.

13

Discussion

On an aerial basis, salt marshes are among the most productive ecosystems on earth.

They play host to a wide range of plant and wildlife species, and are hatchery for many

commercially important finfish. Many salt marshes were created as a result of glacial processes

during the ice ages, and they are characterized as an accumulation of fine-grained sediment over

the top of coastal plain or glacial outwash sediments. This creates a unique hydrogeologic

environment, which for an ideal case is illustrated in Figure 6.

In a coastal salt marsh, groundwater typically enters surface water via two pathways (Fig.

6). The shallow flow path is groundwater that passes from the unconfined aquifer through the

marsh peat. This source also includes tidally pumped water that inundates these marsh sediments

during each flood tide; as a result, the salinity of this groundwater is often indistinguishable from

surface water, and can even be hypersaline due to utilization of the freshwater by marsh plants.

The intermediate flow path involves groundwater flowing beneath the marsh peat, where a

deposit of silt and clay acts as a confining layer. This confining layer is often breeched by the

permeable sediments of tidal creeks, which allows the groundwater to escape into the surface

water system. Such a breech is present in the main channel of Pamet River system; this

groundwater is usually brackish.

Analysis of groundwater samples collected along the estuary edge revealed that these two

groundwater sources had distinct Ra isotopic (228Ra/226Ra) signatures (Fig. 7a). The marsh

groundwater endmember, had high 228Ra relative to 226Ra (activity ratio of ~10-15) while the

aquifer-derived groundwater had relatively low activity ratios (1-2). The explanation for this is

14

related to two factors: (1) the frequency with which seawater circulates through the sediment and

(2) the regeneration rate of the particular radium isotope from its parent thorium isotope.

Therefore, because the marsh sediments are repeatedly flooded on tidal time-scales, and because

228Ra has a significantly shorter half-life than 226Ra, marsh sediment-derived groundwater has

unusually high 228Ra/226Ra activity ratios (Rama and Moore, 1996; Krest et al., 2000; Charette

and Buesseler, 2004).

These two distinct groundwater activity ratios allow us to construct a three endmember

mixing model for data obtained during our time series sampling at the inlet. In this model, the

three sources of Ra to the estuary are marsh groundwater, brackish aquifer-derived groundwater,

and coastal ocean seawater introduced to the Pamet from Cape Cod Bay during tidal mixing (Fig.

7b). The following three equations:

fco + fmarsh + fbrackish = 1 (1)

228Raco × fco + 228Ramarsh × fmarsh + 228Rabrackish × fbrackish = 228Rasurf (2)

226Raco × fco + 226Ramarsh × fmarsh + 226Rabrackish × fbrackish = 226Rasurf (3)

containing three unknowns (fco,fmarsh,fbrackish) are linear and can therefore be solved by

substitution. The terms fco,fmarsh, and fbrackish are the fractions of water derived from the coastal

ocean, marsh groundwater, and surficial aquifer, respectively, in a given sample. The

endmember Ra activities are indicated by the subscripts, and 228Rasurf and 226Rasurf are the

activities in the surface water (surf) sample of interest. The endmember activities we used were

as follows (all units are dpm 100 L-1): 228Ramarsh=516 and 226Ramarsh=36, 228Rabrackish=180 and

15

226Rabrackish=180 (Fig. 5). The coastal ocean endmember was based on the minimum value

observed during flood tide for a given period (typically: 228Raco=15 and 226Raco=8). The brackish

surficial aquifer Ra endmember was chosen as the sample with both the highest 226Ra activity

and lowest 228Ra/226Ra AR (1.0).

The model was run for each data point collected during each of the 12-hour inlet time-

series sampling periods (Table 3). As an example, in Figure 8, we show the results for March

2002 only. During flood tide and the first few hours of the following ebb tide, nearly 100% of the

Ra can be accounted for having originated from Cape Cod Bay. During the latter portion of the

ebb tide, both the brackish and marsh contributions begin to increase. The marsh groundwater

contribution to the Ra inventory peaks at 10% while the brackish groundwater contribution tops

out at 30%. Except for variation in the magnitude of the Ra contributions, this pattern was

similar for all time periods.

For each time period, we estimated the net fraction of each endmember discharging from

the marsh by averaging the ebb tide ‘f’ values (Table 4). We also estimated the % fresh

groundwater flux from the Pamet by conducting a simple salt balance for the ebb minus flood

salinity values (which assumes all of the freshwater originated in the subsurface). Interestingly,

while the brackish groundwater contribution varied by a factor of three for the five time periods,

the marsh groundwater contribution was constant regardless of the season. This result is not

entirely unexpected and actually provides somewhat of an independent check on our mixing

model: if dominated by tidal pumping of surface water, the amount of fluid that flows through

the marsh pore waters should not vary with the season.

16

If the marsh groundwater contribution is relatively constant over time, then what factors

explain the variability in the brackish, aquifer-derived groundwater contribution? To answer this

question, we first converted the fbrackish in Table 1 to a water flux estimate by multiplying by the

tidal prism volume and 1.9 tides per day. It is important to note that the Ra-derived brackish

SGD does include some (non-constant) fraction of the salt-balance derived fresh SGD. The

results are shown in Figure 9, which also shows the water table elevation from a USGS

monitoring well (USGS J561 MA-TSW) in the Pamet groundwater lens (north side of the

marsh). While fresh SGD correlates with water table elevation, the brackish groundwater flux

showed the opposite trend: highest in March 2002, during the lowest aquifer water level of the

two-year period, with the ratio of brackish plus marsh SGD to fresh SGD showing an inverse

correlation (Fig. 10). I hypothesize that this relationship can be accounted for by enhanced

groundwater-seawater interaction in the shallow aquifer during this time.

My hypothesis can be explained as follows: water table height, a function of net recharge

to the aquifer, will determine the location of the groundwater-seawater interface. During periods

of net aquifer recharge, this interface will remain closer to the shoreline. Conversely, drought

conditions (such as in March 2002) will result in a lowering of the water table and landward

movement of the boundary. When the level of the water table drops, salt water intrudes deeper

into sediments that have not been exposed to saline groundwater for a period of time and desorbs

quantitatively more radium (and possibly nutrients like ammonium), which is subsequently

delivered to the marsh via SGD. Periods of higher groundwater elevation were characterized by

(relatively) lower brackish groundwater flux, a result of the brackish groundwater interacting

17

with aquifer sediments that had already been exposed to seawater and therefore had been

weathered of the majority of their surface bound radium.

This hypothesis is supported by simple desorption experiment, whereby sediment from

the freshwater portion of the Cape Cod aquifer and sediment from the marine (saline

groundwater) end were exposed to Ra-free seawater of varying salinity (Fig. 11). For all

treatments (excluding freshwater), the aquifer sands that had never been exposed to salt-water

intrusion yielded ~25x more 226Ra than the sediments that had been continually exposed to saline

groundwater. In addition, the 228Ra/226Ra activity ratio in the marine sands was significantly

higher (3.2) than the fresh water sands (1.3, close to the parent isotope AR of ~1), which

demonstrates the effect of faster 228Ra regeneration relative to that of 226Ra (Fig. 12) in

previously desorbed sediments.

The fresh SGD fluxes as estimated from the salinity balance for the Pamet River time

series data clearly do not correlate with the brackish or marsh SGD fluxes as predicted from the

three endmember mixing model. If the source of the freshwater in the Pamet is indeed

groundwater, then it must be freshwater that bypasses the fresh-saline groundwater mixing zone

where the radium signal is typically imparted (Charette et al., 2003; Mulligan and Charette,

2006). Radon, an inert noble gas, is enriched in groundwater of any salinity and therefore should

be an ideal tracer of total SGD (fresh plus brackish plus marsh). Indeed, Allen and Charette

(2004), in a multi-isotope tracer study of the Pamet River, reported that the radon-derived SGD

value (1.5 m3 m-2 d-1) was balanced by the sum of the radium-derived SGD (0.77 m3 m-2 d-1) and

the salinity balance (0.70 m3 m-2 d-1).

18

Past studies have stressed that SGD was a major source of nutrients to salt marshes

(Valiela et al., 1978; Johannes, 1980; Krest et al. 2000) and the Pamet River is no exception.

During all time periods, nutrients displayed a significant inverse correlation with salinity,

indicating a brackish or fresh groundwater endmember (Figs. 4 & 5). In addition, nitrate,

ammonia and phosphate concentrations were low near high tide and high near low tide (Table 1),

suggesting input by SGD. Using the average groundwater nutrient concentrations compiled in

Table 2 and the groundwater fluxes from Table 4, we can estimate nutrient fluxes to the marsh

via SGD for each time period. The results for dissolved inorganic nitrogen (DIN = nitrate plus

ammonium) are summarized in Table 5. Given that the Ra approach alone may not capture all of

the groundwater input to the system, two methods for SGD-derived DIN input were used: (1)

using the SGDfresh estimate (based on the salt-balance) and (2) using the sum of the SGDmarsh and

SGDbrackish (Ra-derived) estimates. In both cases, a single groundwater DIN average (41 µM)

was used. This approach should provide upper and lower limit estimates, respectively. For the

salt-balance method, DIN inputs ranged from 8.1-41 x 103 mol d-1 compared with 3.0-7.0 x 103

mol d-1 for the Ra-derived SGD estimate. Using a marsh area of 9.3 x 105 m2, these fluxes

correspond to an average annual DIN input of between 1.7 and 7.1 mol m-2 y-1. Given the

average groundwater soluble reactive phosphate (SRP) of 3.1 µM, the corresponding SRP input

to the Pamet River ranges from 0.13 to 0.54 mol m-2 y-1. These values are in the upper range of

SGD-derived DIN fluxes to coastal embayments as compiled by Hwang et al. (2005).

Are these inputs balanced by nutrient export from the marsh? For all time periods,

nutrient concentrations peak during ebb tide, indicating a net source from the marsh to coastal

19

waters (e.g. DIN in Fig. 4). However, despite the relatively high surface water nutrient

concentrations in this estuary, except in the case of silicate, extrapolating the channel time series

data to a salinity of zero (Fig. 5) yields values less than the groundwater averages (Table 2),

which suggests nutrient uptake by the marsh ecosystem.

To calculate nutrient export from the marsh, we used the average ebb tide minus the

average flood tide concentration and multiplied the excess by the tidal prism volume (6.3 x 105

m3). For DIN, these output fluxes, which range from 1.7-11 x 103 mol d-1, are reported in Table

5. In all cases, except during times of extreme freshwater input (July-01 and Mar-03), DIN

output exceeds input from SGD. Assuming that nutrient uptake by the marsh is occurring during

all time periods (check reference on this), the two exceptions suggest that the Ra-approach does

not capture total groundwater flow and associated nutrients during these extreme events.

Using the average annual outputs minus inputs for this system (SGDfresh estimate), we

estimate a net salt marsh DIN uptake of 5.1 mol m-2 y-1 (0.38 mol m-2 y-1 for SRP). This DIN

balance implies that approximately 30% of the SGD-derived DIN flux is exported to coastal

waters (Cape Cod Bay) while 70% is retained by the salt marsh ecosystem (SRP balanced in

exactly the same way-30/70%). For the Nauset Marsh system, which very similar to the Pamet

and is located only 20 km to the South, Roman et al. estimated that 72% of the total productivity

was attributed to the marsh grass Spartina. Using a C:N ratio of 25 for this species (Valiela and

Teal, 1974 as cited in Gallagher, 1975), we estimate a net salt marsh primary productivity of 130

mol C m-2 y-1. This compares quite favorably with the estimate of Roman et al. (1990) for the

nearby Nauset Marsh (160 mol C m-2 y-1). It is important to note that the roles of dissolved and

20

particulate organic nitrogen were not considered in these budget calculations. In the case of the

productivity estimate, we also did not account for denitrification as a potential alternative DIN

sink. In either situation, however, the marsh was clearly a net source of DIN to coastal waters.

21

Conclusions

The implications of these results are several-fold. First, the Ra-derived SGD approach

may not capture all of the fresh groundwater input to the coastal zone. Despite this fact,

numerous studies have shown that they are excellent predictors of groundwater nutrient flux to

the coastal zone (Krest et al., 2000; Charette et al., 2001; Moore et al., 2002). This is due to the

fact that Ra accounts for the chemical transformations (e.g. denitrification, ammonium

mobilization) that occur when fresh and salty groundwater interact prior to discharge into surface

waters. Secondly, water table elevation likely plays a major role in the flux of groundwater-

derived elements to coastal waters, and perhaps in opposition to conventional wisdom (Michael

et al., 2005; e.g. the wet season may not be associated with the peak flux in groundwater-derived

chemicals). Lastly, radium isotopes are valuable tools for quantifying different groundwater

sources in coastal regions. For example, this technique could be used to determine what fraction

of nutrient inputs to an estuary is a recycled (e.g. marsh sediment pore water) versus a new

source (e.g. aquifer-derived groundwater) to the system.

Numerous studies within the past several decades have reached the same conclusion:

SGD plays a major role in the productivity of coastal ecosystems (e.g. Johannes, 1980; Capone

and Bautista, 1985; Valiela et al., 1990). However, our understanding of SGD to coastal waters

and its potential for carrying a significant nutrient burden is still far from complete (Slomp and

van Cappellen, 2004). Further studies shall concentrate on defining better the variability in space,

time and composition of SGD, and refinement of geochemical tracers and other techniques for

quantifying the flux.

22

References

Allen, M. C., and M. A. Charette. 2004. The utility of a continuous radon monitor for tracing water

masses and submarine groundwater discharge in the coastal ocean. ASLO/TOS Ocean Research

Conference.

Bollinger, M. S., and W.S. Moore. 1993. Evaluation of salt marsh hydrology using radium as a tracer.

Geochim. Cosmochim. Acta 57: 2203-2212.

Burnett, W. C. and others 2006. Quantifying submarine groundwater discharge in the coastal zone via

multiple methods. Science of the Total Environment In Press, Corrected Proof.

Burnett, W. C., G. Kim, and D. Lane-Smith. 2001. A continuous monitor for assessment of Rn-222 in

the coastal ocean. Journal of Radioanalytical and Nuclear Chemistry 249: 167-172.

Cable, J. E., W. C. Burnett, J. P. Chanton, and G. L. Weatherly. 1996. Estimating groundwater discharge

into the northeastern Gulf of Mexico using radon-222. Earth and Planetary Science Letters 144:

591-604.

Capone, D. G., and M. F. Bautista. 1985. A groundwater source of nitrate in nearshore marine

sediments. Nature 313: 214-216.

Charette, M. A., and M. C. Allen. 2006. Precision groundwater sampling in coastal aquifers using a

direct push shielded-screen well-point system. Ground Water Monitoring & Remediation 26: 87-

93.

Charette, M. A., and K. O. Buesseler. 2004. Submarine groundwater discharge of nutrients and copper

to an urban subestuary of Chesapeake bay (Elizabeth River). Limnology and Oceanography 49:

376-385.

Charette, M. A., K. O. Buesseler, and J. E. Andrews. 2001. Utility of radium isotopes for evaluating the

input and transport of groundwater-derived nitrogen to a Cape Cod estuary. Limnology and

23

Oceanography 46: 465-470.

Charette, M. A., R. Splivallo, C. Herbold, M. S. Bollinger, and W. S. Moore. 2003. Salt marsh

submarine groundwater discharge as traced by radium isotopes. Marine Chemistry 84: 113-121.

Crotwell, A. M., and W. S. Moore. 2003. Nutrient and radium fluxes from submarine groundwater

discharge to Port Royal Sound, South Carolina. Aquatic Geochemistry 9: 191-208.

Eichner, E. M., T. C. Cambareri, K. Livingston, B. Sobczak, and B. Smith. 1997. Pamet River

Investigation: Groundwater Assessment Study, p. 39. Cape Cod Commission Water Resources

Office.

Gallagher, J. L. 1975. Effect of an Ammonium Nitrate Pulse on the Growth and Elemental Composition

of Natural Stands of Spartina alterniflora and Juncus roemerianus. American Journal of Botany

62: 644-648.

Hwang, D. W., G. B. Kim, Y. W. Lee, and H. S. Yang. 2005. Estimating submarine inputs of

groundwater and nutrients to a coastal bay using radium isotopes. Marine Chemistry 96: 61-71.

Johannes, R. E. 1980. The ecological significance of submarine discharge of groundwater. Mar. Ecol.

Progr. Ser. 3: 365-373.

Kelly, R. P., and S. B. Moran. 2002. Seasonal changes in groundwater input to a well-mixed estuary

estimated using radium isotopes and implications for coastal nutrient budgets. Limnology and

Oceanography 47: 1796-1807.

Krest, J. M., W. S. Moore, L. R. Gardner, and J. T. Morris. 2000. Marsh nutrient export supplied by

groundwater discharge: Evidence from radium measurements. Global Biogeochemical Cycles

14: 167-176.

Li, Y.-H., and L-H Chan. 1979. Desorption of Ba and 226Ra from river-borne sediments in the Hudson

Estuary. Earth Planet. Sci. Lett. 43: 343-350.

24

Michael, H. A., A. E. Mulligan, and C. F. Harvey. 2005. Seasonal oscillations in water exchange

between aquifers and the coastal ocean. Nature 436: 1145-1148.

Moore, W. S. 1996. Large groundwater inputs to coastal waters revealed by Ra-226 enrichments. Nature

380: 612-614.

Moore, W. S., and R. Arnold. 1996. Measurement of 223Ra and 224Ra in coastal waters using a delayed

coincidence counter. J. Geophys. Res. 101: 1321-1329.

Moore, W. S. 2000. Ages of continental shelf waters determined from 223Ra and 224Ra. J. Geophys. Res.

105: 22,117-123,894.

---. 2003. Sources and fluxes of submarine groundwater discharge delineated by radium isotopes.

Biogeochemistry 66: 75-93.

Moore, W. S., J. Krest, G. Taylor, E. Roggenstein, S. Joye, and R. Lee. 2002. Thermal evidence of water

exchange through a coastal aquifer: Implications for nutrient fluxes. Geophysical Research

Letters 29.

Mulligan, A. E., and M. A. Charette. 2006. Intercomparison of submarine groundwater discharge

estimates from a sandy unconfined aquifer. Journal of Hydrology In Press, Corrected Proof.

Rama, and W. S. Moore. 1996. Using the radium quartet for evaluating groundwater input and water

exchange in salt marshes. Geochimica Et Cosmochimica Acta 60: 4645-4652.

Roman, C. T., K. W. Able, M. A. Lazzari, and K. L. Heck. 1990. Primary productivity of angiosperm

and macroalgae dominated habitats in a New England Salt Marsh: a Comparative analysis.

Estuarine, Coastal and Shelf Science 30: 35-45.

Slomp, C. P., and P. Van Cappellen. 2004. Nutrient inputs to the coastal ocean through submarine

groundwater discharge: controls and potential impact. Journal of Hydrology 295: 64-86.

Smith, L., and W. Zawadzki. 2003. A hydrogeologic model of submarine groundwater discharge:

25

Florida intercomparison experiment. Biogeochemistry 66: 95-110.

Valiela, I., J. Costa, K. Foreman, J. M. Teal, B. Howes, and D. Aubrey. 1990. Transport of

Groundwater-Borne Nutrients from Watersheds and Their Effects on Coastal Waters.

Biogeochemistry 10: 177-197.

Valiela, I. and others 1992. Couplings of Watersheds and Coastal Waters - Sources and Consequences of

Nutrient Enrichment in Waquoit Bay, Massachusetts. Estuaries 15: 443-457.

Valiela, I., J. M. Teal, S. Volkmann, D. Shafer, and E. J. Carpenter. 1978. Nutrient and particulate fluxes

in a salt marsh ecosystem: tidal exchanges and inputs by precipitation and groundwater.

Limnology and Oceanography 23: 798-812.

Yang, H. S., D. W. Hwang, and G. B. Kim. 2002. Factors controlling excess radium in the Nakdong

River estuary, Korea: submarine groundwater discharge versus desorption from riverine

particles. Marine Chemistry 78: 1-8.

26

Table 1. Nutrient concentrations and radium activities from the Pamet River estuary inlet time series.

Tidal Height Salinity NO3-/NO2

- PO43- NH4

+ SiO4- 224Ra 223Ra 226Ra 228Ra 228Ra/226Ra

(m) AR

July 2001

0.849 31.20 0.8 0.5 2.4 14.2 69.0 2.3 13.8 32.2 2.3

0.224 28.33 2.2 0.7 4.1 73.2 116.4 5.4 17.1 43.4 2.5

0.021 23.35 4.8 0.8 4.2 95.4 162.9 9.6 20.4 51.8 2.5

0.005 22.09 4.0 0.7 5.4 116.6 177.4 9.0 27.2 69.3 2.6

0.030 17.64 7.8 0.6 5.0 122.6 193.9 8.2 16.6 56.4 3.4

0.503 25.76 3.1 0.8 6.1 67.7 265.3 18.5 17.3 60.9 3.5

1.205 30.21 0.9 0.3 0.2 4.6 120.5 6.2 11.6 33.7 2.9

1.982 31.15 0.0 0.0 0.0 0.0 64.3 2.3 9.3 16.3 1.7

2.628 31.18 0.0 0.0 0.0 0.0 34.1 1.1 6.9 14.9 2.2

2.882 31.24 0.0 0.0 0.0 0.0 31.9 1.6 9.3 12.8 1.4

2.665 30.95 0.0 0.0 0.0 0.0 28.4 1.7 8.6 14.7 1.7

2.202 30.69 0.1 0.4 0.9 5.8 45.2 1.9 9.5 23.1 2.4

1.583 30.31 0.3 0.5 1.7 10.8 73.9 2.5 11.7 15.8 1.3

average 28.0 1.8 0.4 2.3 39.3 106.4 5.4 13.8 34.3 2.3

± 4.4 2.4 0.3 2.3 48.3 74.3 5.0 5.7 20.1 0.7

March 2002

0.696 31.64 0.3 0.5 0.3 0.1 82.7 5.4 12.1 37.8 3.1

1.530 31.73 0.2 0.4 0.3 0.0 39.5 2.3 11.4 26.2 2.3

2.322 31.78 0.1 0.5 0.1 0.0 21.4 1.3 10.7 22.5 2.1

2.911 31.77 0.1 0.5 0.2 0.0 18.0 1.2 9.8 22.4 2.3

2.993 31.77 0.1 0.5 0.1 0.0 25.3 1.4 10.5 25.8 2.4

2.602 31.76 0.1 0.6 0.2 0.0 21.6 1.7 10.4 26.4 2.5

1.953 31.66 0.1 0.5 0.2 0.1 35.0 1.6 10.7 27.1 2.5

1.110 30.75 0.3 0.5 0.3 0.3 45.2 2.5 11.2 30.1 2.7

0.196 29.09 1.1 0.7 0.3 0.9 92.4 6.1 13.4 36.4 2.7

0.063 25.82 2.5 0.6 0.7 1.5 149.4 8.7 23.0 49.3 2.1

0.016 26.92 5.2 0.6 0.7 1.6 321.6 13.3 61.5 103.6 1.7

0.027 29.07 12.3 0.8 0.3 1.4 337.0 11.1 58.1 101.2 1.7

average 30.3 1.9 0.6 0.3 0.5 99.1 4.7 20.2 42.4 2.4

± 2.1 3.6 0.1 0.2 0.6 114.3 4.2 18.8 29.1 0.4

July 2002

2.484 31.22 0.1 0.1 0.5 1.2 28.1 1.8 9.6 18.2 1.9

2.457 31.20 0.0 0.1 1.1 1.2 29.6 1.8 9.2 18.9 2.1

2.135 31.20 0.1 0.1 0.6 1.3 30.7 1.5 8.9 18.8 2.1

1.645 31.11 0.2 0.2 0.2 2.4 42.5 1.9 10.5 28.1 2.7

0.982 30.96 0.3 0.1 0.0 3.0 62.1 3.1 11.2 23.3 2.1

0.328 29.59 0.8 0.5 1.7 10.2 101.1 4.4 17.0 43.7 2.6

0.012 27.53 1.5 0.4 0.9 17.8 145.2 10.8 18.2 56.0 3.1

0.002 25.37 2.0 0.4 1.0 25.3 185.0 7.9 24.7 68.8 2.8

0.325 25.07 3.3 0.4 0.5 25.0 219.6 9.5 25.9 78.6 3.0

0.864 29.72 2.3 0.5 0.4 8.8 157.4 7.5 13.0 48.4 3.7

1.445 31.20 0.2 0.2 0.2 1.6 55.5 2.2 11.3 27.7 2.4

2.054 31.19 0.2 0.2 0.5 1.3 47.2 1.9 9.9 17.0 1.7

2.677 31.27 0.1 0.1 0.1 1.0 44.1 1.4 12.3 22.1 1.8

average 29.7 0.8 0.3 0.6 7.7 88.3 4.3 14.0 36.1 2.5

± 2.3 1.1 0.2 0.5 9.2 66.2 3.4 5.8 20.9 0.6

November 2002

0.405 28.30 3.4 0.3 3.6 28.6 310.0 25.7 19.6 94.8 4.8

1.047 31.37 0.4 0.1 0.1 1.1 99.6 6.3 9.1 29.4 3.2

1.956 31.04 0.0 0.0 0.1 0.6 17.0 1.4 12.5 36.1 2.9

2.764 31.35 0.0 0.0 0.0 1.2 41.1 3.2 8.0 20.5 2.6

3.325 31.21 0.0 0.1 0.2 0.0 10.0 0.5 7.7 15.1 2.0

3.280 30.85 0.0 0.1 0.0 0.1 6.6 0.4 7.4 17.1 2.3

2.844 32.05 0.0 0.1 0.0 0.4 29.4 2.3 7.1 17.0 2.4

2.172 31.93 0.1 0.0 0.1 0.5 50.3 2.3 7.7 19.8 2.6

1.319 31.47 0.3 0.0 0.0 5.1 62.5 4.1 8.3 23.1 2.8

0.386 29.97 0.5 0.1 0.4 12.9 86.8 7.3 10.2 28.1 2.8

0.189 28.50 0.9 0.0 0.0 28.8 174.6 9.8 30.4 59.4 2.0

0.120 28.11 2.0 0.0 0.0 37.5 306.8 18.7 37.2 73.1 2.0

0.096 27.41 3.8 0.1 0.8 47.1 395.7 20.7 34.4 90.7 2.6

average 30.3 0.9 0.1 0.4 12.6 122.4 7.9 15.3 40.3 2.7

± 1.6 1.3 0.1 1.0 16.8 132.2 8.5 11.2 28.9 0.8

March 2003

0.024 14.39 16.6 1.0 4.2 96.1 135.0 12.2 18.2 53.3 2.9

0.023 11.34 17.3 0.4 4.9 112.3 156.4 11.4 19.4 57.7 3.0

0.857 31.62 0.8 0.3 0.4 7.3 61.9 5.0 12.9 34.4 2.7

1.666 32.26 0.0 0.6 0.2 2.1 19.1 1.6 12.9 24.1 1.9

2.485 32.04 0.0 0.3 0.2 2.3 14.2 0.5

2.953 32.36 0.0 0.2 0.3 1.7 21.6 1.4 9.4 18.5 2.0

2.845 32.36 0.0 0.4 0.2 1.8 11.4 0.9 8.2 17.5 2.1

2.338 32.30 0.0 0.3 0.2 3.3 15.4 1.4 9.0 18.7 2.1

1.637 31.76 0.0 0.3 0.2 4.7 23.2 2.4 7.1 15.0 2.1

0.820 27.83 3.9 0.2 3.4 27.6 38.3 3.1 8.6 23.0 2.7

0.183 22.65 4.6 0.3 0.7 53.9 105.7 8.1 15.7 43.8 2.8

0.079 21.44 7.7 0.3 1.5 63.0 135.6 13.1 18.3 52.5 2.9

0.048 17.35 14.8 0.2 3.0 82.9 148.8 12.5 19.7 53.6 2.7

average 26.1 5.1 0.4 1.5 35.3 68.2 5.7 13.3 34.3 2.5

± 7.8 6.8 0.2 1.8 41.1 58.6 5.0 4.8 16.7 0.4

__________(dpm 100 L-1)____________________(!mol L-1)__________

27

Table 2. Dissolved nutrient concentrations and radium activities in groundwater from the Pamet River estuary watershed.

Sampling Salinity NO3-/NO2

- PO43- NH4

+ SiO4- 224Ra 223Ra 226Ra 228Ra 228Ra/226Ra

Period AR

July-2001 N/A 3.8 1.1 0.3 16.3 166 25 13 48 3.7

July-2001 N/A 58.6 0.6 0.4 346.4 180 26 5 41 8.5

July-2001 N/A 59.7 0.1 0.8 352.5 122 8 6 35 5.6

July-2001 N/A 58.8 0.4 0.5 298.2 13 1 ND ND ND

July-2001 N/A 0.2 1.2 11.7 616.8 185 23 186 269 1.4

July-2001 N/A 0.0 2.3 21.3 877.8 548 54 298 848 2.8

July-2001 N/A 0.0 0.0 0.0 0.0 6 1 4 8 2.0

July-2002 20.76 3.3 7.1 0.6 28.4 703 55 20 155 7.7

July-2002 8.41 4.7 9.6 0.7 28.4 435 28 23 80 3.5

July-2002 20.90 2.7 0.5 0.1 13.6 486 28 21 86 4.2

July-2002 12.89 40.4 18.7 0.5 58.3 242 14 10 50 5.2

July-2002 10.50 66.8 1.0 0.7 44.6 210 9 12 49 4.0

July-2002 10.23 74.2 1.2 4.8 44.6 250 18 14 64 4.6

July-2002 14.19 161.9 10.6 0.6 30.7 187 15 19 76 3.9

July-2002 18.07 92.3 1.1 2.2 45.7 118 6 17 57 3.3

July-2002 30.12 27.4 1.2 0.0 17.3 393 17 22 158 7.2

July-2002 14.28 50.9 5.3 0.3 18.1 233 15 13 91 7.0

July-2002 20.82 18.0 2.0 0.7 61.6 263 21 16 68 4.1

July-2002 31.18 12.1 2.3 0.2 16.2 641 52 17 109 6.3

July-2002 29.74 0.1 12.4 162.3 239.9 923 44 54 315 5.8

July-2002 24.24 39.4 0.6 0.7 30.0 217 18 13 76 5.8

July-2002 7.97 94.1 1.0 0.5 63.8 44 5 ND ND ND

July-2002 28.45 11.3 0.4 0.6 19.8 395 44 17 87 5.1

July-2002 28.30 0.3 5.4 8.0 28.1 224 17 35 63 1.8

July-2002 29.97 40.9 1.3 0.6 16.2 361 30 36 126 3.5

July-2002 1.32 0.8 6.6 116.8 114.4 14 1 ND ND ND

July-2002 0.86 0.5 2.9 63.3 102.2 28 1 3 23 8.5

July-2002 13.59 0.4 13.6 161.9 323.0 275 6 181 186 1.0

July-2002 32.85 3.0 1.1 1.5 36.8 1264 80 33 490 15.0

July-2002 34.21 10.0 1.3 3.3 73.1 986 30 36 516 14.2

July-2002 0.00 0.0 0.0 0.0 0.0 534 15 75 78 1.0

July-2002 0.00 0.0 0.0 0.0 0.0 574 20 86 134 1.6

July-2002 0.00 0.0 0.0 0.0 0.0 480 18 73 128 1.8

Nov-2002 28.30 0.6 0.6 3.3 119.3 120 20 ND ND ND

Nov-2002 28.30 0.6 0.5 4.1 123.4 181 27 4 43 10.6

Nov-2002 28.30 0.6 0.5 5.5 125.1 211 21 10 50 4.8

March 2003 28.96 0.1 6.0 50.4 172.1 766 60 38 251 6.6

March 2003 28.75 0.0 2.3 24.0 123.7 1030 118 24 268 11.1

March 2003 14.05 44.8 4.5 1.0 87.0 350 40 19 141 7.4

March 2003 0.00 94.5 0.1 0.6 55.9 164 12 5 66 13.8

March 2003 10.17 0.4 3.4 1.6 29.1 127 22 2 35 14.4

March 2003 24.40 23.9 2.5 0.6 32.8 460 33 15 188 12.3

March 2003 23.79 14.8 0.2 3.0 82.9 730 46 ND ND ND

average 18.3 26.0 3.1 15.4 114.3 368.4 26.5 38.8 146.2 6.1

± 11.1 36.4 4.2 39.0 172.4 299.8 22.7 59.7 164.7 4.0

__________(!mol L-1)__________ ________________________(dpm 100 L-1)________________________

28

Table 3. Three end member mixing model for the Pamet River estuary time series experiments.

Tidal Height Salinity fco fmarsh fbrackish ffresh

(m)

July 2001

0.85 31.20 0.93 0.03 0.04 0.00

0.22 28.33 0.90 0.04 0.06 0.09

0.02 23.35 0.87 0.06 0.08 0.25

0.01 22.09 0.82 0.08 0.11 0.29

0.03 17.64 0.88 0.07 0.05 0.43

0.50 25.76 0.87 0.08 0.05 0.17

1.21 30.21 0.94 0.03 0.03 0.03

1.98 31.15 0.98 0.00 0.02 0.00

2.63 31.18 0.99 0.01 0.01 0.00

2.88 31.24 0.98 0.00 0.02 0.00

2.67 30.95 0.98 0.00 0.01 0.01

2.20 30.69 0.97 0.02 0.02 0.02

1.58 30.31 0.97 0.00 0.03 0.03

March 2002

0.70 31.64 0.95 0.04 0.01 0.00

1.53 31.73 0.97 0.02 0.01 0.00

2.32 31.78 0.98 0.01 0.01 0.00

2.91 31.77 0.98 0.01 0.00 0.00

2.99 31.77 0.97 0.02 0.01 0.00

2.60 31.76 0.97 0.02 0.01 0.00

1.95 31.66 0.97 0.02 0.01 0.00

1.11 30.75 0.96 0.03 0.01 0.03

0.20 29.09 0.94 0.04 0.02 0.08

0.06 25.82 0.88 0.04 0.07 0.18

0.02 26.92 0.63 0.08 0.29 0.15

0.03 29.07 0.64 0.08 0.27 0.08

July 2002

2.48 31.22 0.99 0.01 0.00 0.00

2.46 31.20 0.99 0.01 0.00 0.00

2.13 31.20 0.99 0.01 0.00 0.00

1.65 31.11 0.97 0.02 0.01 0.01

0.98 30.96 0.98 0.01 0.01 0.01

0.33 29.59 0.92 0.04 0.04 0.05

0.01 27.53 0.89 0.07 0.04 0.12

0.00 25.37 0.84 0.08 0.08 0.19

0.33 25.07 0.82 0.10 0.08 0.20

0.86 29.72 0.92 0.06 0.01 0.05

1.45 31.20 0.97 0.02 0.01 0.00

2.05 31.19 0.99 0.00 0.01 0.00

2.68 31.27 0.97 0.01 0.02 0.00

November 2002

0.41 28.30 0.81 0.14 0.05 0.12

1.05 31.37 0.97 0.03 0.01 0.02

1.96 31.04 0.94 0.03 0.03 0.03

2.76 31.35 0.99 0.01 0.00 0.02

3.33 31.21 1.00 0.00 0.00 0.03

3.28 30.85 1.00 0.00 0.00 0.04

2.84 32.05 1.00 0.00 0.00 0.00

2.17 31.93 0.99 0.01 0.00 0.00

1.32 31.47 0.98 0.01 0.00 0.02

0.39 29.97 0.96 0.02 0.01 0.06

0.19 28.50 0.83 0.05 0.13 0.11

0.12 28.11 0.77 0.06 0.16 0.12

0.10 27.41 0.75 0.10 0.14 0.14

March 2003

0.02 14.39 0.89 0.06 0.05 0.56

0.02 11.34 0.87 0.07 0.06 0.65

0.86 31.62 0.94 0.03 0.03 0.02

1.67 32.26 0.96 0.01 0.03 0.00

2.49 32.04 0.00 0.00 0.00 0.01

2.95 32.36 0.98 0.00 0.01 0.00

2.85 32.36 0.99 0.00 0.01 0.00

2.34 32.30 0.99 0.00 0.01 0.00

1.64 31.76 1.00 0.00 0.00 0.02

0.82 27.83 0.98 0.01 0.01 0.14

0.18 22.65 0.91 0.04 0.04 0.30

0.08 21.44 0.89 0.06 0.06 0.34

0.05 17.35 0.88 0.06 0.06 0.46

29

Table 4. Radium-derived submarine groundwater discharge rates for the Pamet River estuary.

Sampling Tidal Prism SGDmarsh SGDbrackish SGDfresh

Period (106 m3) fmarsh fbrackish ffresh

[a] [b] [c] [d] [e] = [(b) x (a)] [f] = [(c) x (a)] [g] = [(d) x (a)]

July 2001 0.63 0.04 0.05 0.14 43 58 168

March 2002 0.63 0.04 0.10 0.07 53 117 89

July 2002 0.63 0.04 0.03 0.05 42 30 65

November 2002 0.63 0.03 0.06 0.06 40 68 75

March 2003 0.63 0.03 0.03 0.27 40 40 328

ebb tide average

(103 m3 d-1)

30

Table 5. Dissolved inorganic nitrogen fluxes in the Pamet River estuary.

Sampling DIN output

Period (SGDfresh) (SGDmarsh+SGDbrackish) (SGDfresh) (SGDmarsh+SGDbrackish)

July 2001 21 4.2 6.5 15 -2.3

March 2002 12 7.0 4.2 7.5 2.8

July 2002 8.1 3.0 1.8 6.3 1.2

November 2002 8.9 4.6 1.7 7.1 2.9

March 2003 41 3.3 11 30 -7.9

average 18 4.4 5.1 13 -0.7

avg. (mol m-2 y-1)*7.1 1.7 2.0 5.1

*assuming marsh surface area = 9.3 x 105 m2

Output-Input (method)

-------------------------------- (103 mol d-1) --------------------------------

DIN input (method)

31

Figure Captions

Figure 1. Map of the Pamet River estuary with the location of the time-series sampling indicated

by a star.

Figure 2. Radium isotopes versus tidal height for the time series data.

Figure 3. Radium isotopes versus salinity for the time series data.

Figure 4. Time series of 226Ra and salinity at the Pamet River inlet during March 2002. The same

pattern was typically observed during all time periods, with only the order of magnitude of the

226Ra activities changing.

Figure 5. Nutrients versus salinity for all time-series data. The curve fit includes all data.

Figure 6. Hydrogeology of a salt marsh ecosystem (adapted from Howes et al., 1996) with

typical 228Ra/226Ra activity ratios for the two groundwater endmembers.

Figure 7. 228Ra versus 226Ra for (a) all groundwater samples, with activity ratios indicated by the

dashed line, and (b) groundwater and surface water with three endmembers used in the mixing

model.

Figure 8. Results from the three endmember mixing model for the inlet time series samples

collected during March 2002.

32

Figure 9. Marsh, brackish, and fresh groundwater flux (cubic meters per day) to the Pamet River

estuary for five time periods between July 2001 and March 2003. Also shown is the groundwater

elevation (meters above sea level) from a nearby USGS monitoring well (J561 MA-TSW).

Figure 10. Ratio of the Ra-derived SGD (brackish + marsh) to the salt-balance derived SGD

(fresh) versus water table elevation in a nearby USGS monitoring well (J561 MA-TSW).

Figure 11. 226Ra activity released per milligram of sediment as a function of salinity during the

desorption experiment. The error bars represent the standard deviation of the triplicate

measurements for each treatment (error bars are smaller than the symbol for the marine sands).

Figure 12. Average desorbed 228Ra and 226Ra for the two sediment types. Also shown is the

average 228Ra/226Ra activity ratio for each treatment.

33

Figure 1.

34

Figure 2.

35

Figure 3.

36

Figure 4.

37

Figure 5.

38

Figure 6.

39

Figure 7.

40

Figure 8.

41

Figure 9.

42

0

0.5

1

1.5

2

1.2 1.4 1.6 1.8 2

Bra

ckis

h+

Mars

h S

GD

/Fre

sh S

GD

Groundwater Elevation (m above m.s.l.)

Mar-02

Nov-02

Jul-02

Jul-01

Mar-03

Figure 10.

43

0

40

80

120

0 5 10 15 20 25 30 35

Freshwater SandsMarine Sands

22

6R

a (

dpm

mg

-1)

Salinity

Figure 11.

44

Figure 12.