Science of the Total Environment 438 (2012) 86–93

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Nutrient fluxes via submarine groundwater discharge to the Bay of Puck, southernBaltic Sea

Beata Szymczycha a, Susanna Vogler b, Janusz Pempkowiak a,c,⁎a Institute of Oceanology, Polish Academy of Sciences, ul. Powstancow Warszawy 55, 81712 Sopot, Polandb Leibniz-Institut fur Ostseeforschung, Warnemünde, Seestraße 15, 18119 Rostock, Germanyc Faculty of Construction and Environmental Engineering, Koszalin University of Technology, ul.Śniadeckich 2, 75453 Koszalin, Poland

H I G H L I G H T S

► SGD fluxes were measured, and seepage water collected, by means of seepage meters.► Both groundwater and recirculated seawater end-members were quantified.► Nitrogen(N) and Phosphorus(P) loads were scaled up to the Puck Bay and Baltic Sea.► N and P loads to the Puck Bay were significant at the background of other sources.► SGD contributed significant P proportion of Baltic Sea nutrient budget.

⁎ Corresponding author at: Institute of Oceanology, PPowstancow Warszawy 55, 81712 Sopot, Poland. Tel.: +

E-mail address: pempa@iopan.gda.pl (J. Pempkowia

0048-9697/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.scitotenv.2012.08.058

a b s t r a c t

a r t i c l e i n f o

Article history:Received 21 June 2012Received in revised form 14 August 2012Accepted 14 August 2012Available online 11 September 2012

Keywords:Gulf of GdanskLoadsNitrogenPhosphorousSeepage water

Submarine groundwater discharge (SGD) has been recognized as an important exchange pathway betweenhydrologic reservoirs due to its impact on biogeochemical cycles of the coastal ocean.This study reports nutrient concentrations and loads delivered by SGD into the Bay of Puck, the southern BalticSea. Measurements were carried out between September, 2009 and October, 2010 at groundwater seepagesites identified by low salinity of pore water. Groundwater fluxes, measured using seepage meters, ranged from3 to 22 L m−2 day−1. Average concentrations of nutrients in groundwater samples collected were as follows:0.4 μmol L−1 nitrate (NO3), 0.8 μmol L−1 nitrite (NO2), 18.2 μmol L−1 ammonium (NH4) and 60.6 μmol L−1 or-thophosphate (PO4). Levels of NH4 and PO4 were significantly higher in samples from SGD sites than in seawater.Seawater and SGD samples showed similar NO2 concentrations but SGD samples exhibited lower NO3 levels thanthose observed in seawater samples. Measured seepage water fluxes and nutrient concentrations were used tocalculate nutrient loads discharged into the study area while the literature groundwater flux and the measurednutrient concentrations were used to estimate nutrient loads discharged into the Bay of Puck. The estimates sug-gest that SGD delivers a dissolved inorganic nitrogen (DIN) load of 49.9±18.0 t yr−1 and a PO4

‐ load of 56.3±5.5 t yr−1 into the Bay of Puck. The projected estimates are significant in comparison with loads delivered tothe bay from other, well-recognized sources (705 t yr−1 and 105 t yr−1 respectively for DIN and PO4). Nutrientdischarge input loads were projected to the entire Baltic Sea The extrapolated values indicate SGD contributesa significant proportion of phosphate load but only an insignificant proportion of DIN load. Further studies arenecessary to better understand SGD contributions to the nutrient budget in the Baltic Sea.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Submarine groundwater discharge (SGD) is a one of the waterpathway of exchange between marine and continental water reser-voirs that strongly influences the hydrogeochemistry of coastalzones (Bokuniewicz, 1992; Burnett et al., 2006; McCoy and Corbett,2009). In comparison with easily quantified and typically largepoint sources of surface water inputs (e.g., rivers and streams)

olish Academy of Sciences, ul.48 58 5517281.

k).

rights reserved.

which are easily quantified, groundwater inputs are harder to evaluatedue to difficulties in gauging fluxes from these sources and accuracy ofdischarges estimates is limited by the spatial and temporal variability ingroundwater discharges. (Burnett et al., 2003; Mulligan and Charette,2009). Groundwater from many areas is enriched with chemical con-stituents and thus may serve as a substantial source of nutrients, tracemetals and organic compounds entering the marine reservoir andinfluencingmarine biota. SGD is thus likely to contribute tomarine geo-chemical cycles and may cause environmental perturbations in coastalzones. Numerous studies have documented the ecological impact ofgroundwater flow into coastal areas. The scale of SGD nutrient loadscan also affect coastal eutrophication (Valiela et al., 2002; Slomp and

87B. Szymczycha et al. / Science of the Total Environment 438 (2012) 86–93

Van Cappellen, 2004; Andersen et al., 2007). As such, SGD poses acritical threat to biodiversity around the world (Carlton, 2006; Eamuset al., 2006). Quantifying groundwater transport of nutrients andtheir fate in coastal environments can enhance understanding of thehydrogeochemical processes in these areas.

The Baltic Sea receives a relatively high loads of material originat-ing from anthropogenic sources. SGD represents a source of inputsinto the Baltic Sea that has not been rigorously quantified. Peltonen(2002) estimated the total volume of SGD entering the Baltic Seato be 4.4 km3 yr−1, or about 1% of total river run-off. SGD studiesconducted along the Baltic Sea coastline over the years have mostlyfocused on groundwater fluxes in the Gulf of Finland, the EckernfőrdeBay, the Gulf of Gdansk and the Bay of Puck (Zekster, 1973; Piekarek-Jankowska et al., 1994; Piekarek-Jankowska, 1996; Bussmann andSuess, 1998; Falkowska and Piekarek-Jankowska, 1999; Kaleris et al.,2002; Peltonen, 2002; Viventsova and Voronov, 2003; Korzeniowski,2003; Schlüter et al., 2004; Kryza and Kryza, 2006; Pempkowiak et al.,2010). Groundwater discharge into the Gulf of Finland (surface areaequal to 29,500 km2) was estimated to be 0.6 km3 yr−1 (Viventsovaand Voronov, 2003), whereas the discharge rate into the EckernfőrdeBay (171 km2) ranges from 0.04 to 0.4 km3 yr−1 (Schlüter et al., 2004).Piekarek-Jankowska et al. (1994) estimated groundwater discharge of0.03 km3 yr−1 into the Bay of Puck (359.2 km2). Uścinowicz (2011)concluded that SGD into the Bay of Puck and Gdansk Bay significantlyexceeded SGD levels in other regions of the Baltic Sea.

Despite greater recognition of the role of SGD in coastal ecosystemsand land-locked seas, uncertainties remain concerning the chemicalcomposition of groundwater seeping into the Baltic Sea. Anomalousnutrient concentrations for the Gulf of Gdansk water column have beenattributed to SGD (Falkowska andPiekarek-Jankowska, 1999). No studiesup to this point have attempted to quantify nutrient concentrations andloads delivered by known SGD sites or input into the Baltic Sea as awhole. This work reports nutrient concentrations and load estimates aswell as overall groundwater volume discharged into the study area andthe Bay of Puck. The results allowed us to project nutrient loads to theentire Baltic Sea in order to assess the potential scale of nutrients deliv-ered by SGD relative to other sources. Thus the results of this study con-tribute to understanding of the role of SGD in Baltic Sea nutrient budgets.

2. Description of the study area

The Bay of Puck (Fig. 1) is the eastern part of the Gulf of Gdansk,along the southern Baltic Sea coast. A narrow spit known as the HelPeninsula separates the bay from the open waters of the Baltic. Thebay has a total area of 359.2 km2 and consists of an outer southeasternregion with an average depth of 20.5 m, and an inner northwestern

Fig. 1. Location of the study area.

region with an average depth of 3.1 m. The basin slopes downward ina northeasterly direction reaching a depth of 54 m near the tip of theHel Peninsula (Nowacki, 2003). Salinity in the Bay of Puck is around 7,similar to that of the open Baltic Sea. Relatively uniform salinity ofseawater throughout the bay reflect limited river run-off as well asexchange between its inner and outer regions (Korzeniowski, 2003).

The Hel Peninsula developed during the Pleistocene and Holoceneepochs. Geomorphic landforms surrounding the Bay of Puck consistof wave-dominated sedimentary plains and dune deposits formingin micro tidal zones. Coastal erosion is the dominant source ofsediments within the study area. Waves, storm surges, currents andwinds drive erosion, transport, accumulation and redeposition ofsediments in the coastal zone. The average river run-off into theBay of Puck equals to 0.25 km3 yr−1 and the average precipitationamounts to 0.20 km3 yr−1 (Cyberski, 1992).

Seismic–acoustic investigations of the study area have imaged per-meable layers of Holocene to Pleistocene sands and silts and underlyingTertiary silt layers (Piekarek-Jankowska et al., 1994). Groundwaterappears within Quaternary and Tertiary units that include interbeddedsands andglacial clayswith peripheral fluvial (sand and gravel) deposits.Beneath the weakly permeable Cretaceous layers of the aquifer, occa-sional SGD seepage sites are confined to near-shore, coarse-grainedaquifers (Falkowska and Piekarek-Jankowska, 1999).

Pempkowiak et al. (2010) identified a submarine groundwater dis-charge site located off the Hel Peninsula covering about 9200 m2. At thestudy area medium ormoderately well sorted, mostly symmetrical sandoccurs. The site is located in shallow, sandy, wave-dominated areas, andhosts irregular groundwater seepage. The sediments of the study areaare influenced by threemajor processes: seepage of groundwater, perco-lation of seawater into the surface layer of sediments and sedimentaryprocesses including mineralization of deposited organic matter in thesediments. The last process is limited to the uppermost layer of the sed-iments, while seawater can be forced into the sediments to the depth offew decimeters. Therefore the study on groundwater chemical composi-tion needs to concentrate on pore-water residing on the depthwhere nosalinity increase caused by groundwater–seawater mixing actuallyoccurs. This depth depends on several factors: seepage intensity, thegranulometric properties of the sediments, water depth, sea bottomrelief and wave action. Thus there are several water types to be dealtwith: seawater (occurring above the seafloor) and pore water (waterin sediments–interstitial water). Pore water can be unaffected ground-water (salinity equal to zero), or mixed groundwater and seawater(seepage water; salinity less than that of seawater). The extent of thewater types is specific to both particular section within the study areaand hydrodynamic conditions at the time of sampling.

This report continues earlier investigations by Pempkowiak et al.(2010) and describes the chemical composition of discharge fromthe previously identified SGD site.

3. Materials and methods

Four sampling campaigns were carried out near the shoreline of theHel Peninsula during the following periods: 31.08 to 3.09.2009, 2 to6.11.2009, 28.02 to 1.03.2010 and 5 to 7.05.2010. Seepage water sam-pling points were selected based on a salinity survey of the study areacarried out in August, 2009 (31.08.2009).

3.1. Salinity survey

On 31st August, 2009 salinity profiles were recorded along paralleltransects that extended seaward from the beach. Along each of the 20transects, samples were collected from 15 sampling points spacedabout 15 m apart (Fig. 2). Moreover in the periods: 2 to 6.11.2009,28.02 to 1.03.2010 and 5 to 7.05.2010 less extensive salinity measure-ments (4–6 sampling stations) in the study area were performedusing seepage meters and groundwater lances.

Fig. 2. Salinity distribution in sediment pore water samples collected at (a) 5 cm and (b) 25 cm depth, on 30.08.2009. Points G and G′ correspond to positions of groundwater lancesand points S1, S2, and S3 — correspond to seepage meter placement positions.

88 B. Szymczycha et al. / Science of the Total Environment 438 (2012) 86–93

Seawater depth along the transects ranged from 0.5 to 2 m inaccordance with distance from the shore.

Push-point lances (see Beck et al., 2007b) equipped with 50 mlpolyethylene syringes were used to sample pore water at depths of5 cm and 25 cm within the sediment. Salinity and temperature mea-surements of sediment pore water and seawater were performedwith a salinometer (WTWMulti 3400i Multi-Parameter Field Meters)having 0.02 and 0.1 °C accuracy.

3.2. Pore water sampling for nutrient analysis

Groundwater lances were used to collect pore water for salinityand nutrient analysis. After a 24 h equilibration period followinginsertion, the devices were used to withdraw 35 ml of pore water atdepths of 0, 4, 8, 12, 16, 24 and 30 cm below sediment–water inter-face. Groundwater measuring less than 0.5 often occurred at depthsgreater than 15 cm below the sediment–water interface. Pore waterat this depth was thus interpreted as groundwater seepage into thesurroundings.

3.3. Seepage water fluxes to the study area

Seepage meters (Pempkowiak et al., 2010) were used to measureseepage water fluxes to the study area. This method is acceptable atsites where the outflow is controlled by small scale geological heteroge-neity (Andersen et al., 2007). However, in geologically heterogeneousareas SGD resultsmay vary as they depend upon the exact seabed selec-tion for deployment of the seepagemeter (Burnett et al., 2006). Seepagemeters consisted of a polyethylene (PTE) chamber (surface: 0.785 m2)with an inlet at one end and a sample port with a PTE collector at theother end. Groundwater seeping through the sediment was trapped inthe PTE collector. The volume of collected over ameasured time intervalprovided the seepage flux rate.

The PTE chambers were deployed one day before actual samplingof seepage water. Than the PTE collector was installed and seepagewater rate was collected over the following time periods: on 2.09.09and 4.11.09 the PTE collectors were left for 45 min, on 28.02.10 —

145 min, and on 5.04.2010 — 60 min.Salinity of the collected samples varied from2 to 4. The groundwater

fraction for each sample was calculated using the end-member method(Burnett et al., 2006; Szczepańska et al., 2012). Groundwater flux wascalculated as the ratio of groundwater volume plus device surface area

divided by the observation time interval. Briefly, the end-membermethod is based on the following mass balance relationships:

VS ¼ VG þ VSW

SSVS ¼ SGVG þ SSWVSW

where S andV are salinity and volume. Subscripts S, G and SWrepresentthe sample, groundwater and seawater fractions, respectively. The twounknowns (VG, VSW)were calculated using the above equations and themeasured values for SS, SG, SSW and VS.

3.4. Nutrient analysis

Water samples (sediment pore water and seawater; 35 ml vol-ume) were passed through 0.45 μm syringe-driven membrane filtersand transferred into polyethylene bottles. The sample bottles wereinitially washed in 2 M nitric acid and then rinsed with MilliQ wateruntil the measured background nutrients were below detection limits(5 times×30 ml). The bottles were also rinsed with a small volume ofthe collected sample water prior to the final sample collection. Thecollected samples were then transported to the laboratory and ana-lyzed for nutrient concentrations within 48 h. The samples were re-frigerated at 4 °C between transport and analysis stages.

Nutrients were analyzed using colorimetric methods described byStrickland and Parsons (1967) and Salley et al. (1986). Repeated anal-yses of both sediment pore water and seawater samples (n=7 foreach) using internal and external (spike) standards established theaccuracy and precision of nutrient analyses. The standards included0.5, 1.5, 2.0 mg L−1 for NO3, NH4, PO4 and 0.005, 0.05, 0.5 mg L−1

for NO2. The average relative standard deviations (precisions) were0.4% for NO3, 1.4% for NO2, 0.3% for NH4 and 1.4% for PO4. Analyses in-dicated recoveries of 96.2% for NO3, 101.4% for NO2, 98.3% for NH4 and99.1% for PO4.

4. Results

4.1. Salinity concentrations in pore water samples

Pore water salinity was used to locate SGD sites within the studyarea. Previous studies have used this method to identify SGDlocations and measure seepage rates (Zohdy and Jackson, 1969;

Fig. 3. Salinity profiles from groundwater impacted area (solid line profiles collectedfrom location G) obtained on 3.09.2009, 5.11.2009, 28.02.2010 and 5.05.2010, and a sa-linity profile from an area free of groundwater impact (dashed profiles collected fromlocation G′) obtained on 3.09.2009.

89B. Szymczycha et al. / Science of the Total Environment 438 (2012) 86–93

Millham and Howes, 1994; Rapaglia, 2007). Fig. 2 shows the porewater salinity distribution within the study area in August, 2009with clearly distinguished areas of higher and lower salinity. Thegroundwater salinity distribution is rather variable. This can be attrib-uted to the granulometric properties of the sediments, sea bottomrelief, wave action and hydrodynamic conditions at the time of sam-pling. A groundwater impacted area (site G) and an area without ap-parent groundwater impact (site G′) were selected from the salinitydistribution map (Fig. 2). Fig. 3 shows the salinity profiles for bothG and G′ sampling points. Pore water salinity of the G profile (mea-sured on 3.09.09, 5.11.09, 28.02.10, 5.05.10) decreases with depthfrom about 7 to about 0.2. In November, 2009 the lowest observed sa-linity (2.1) occurred at 30 cm depth. The G salinity profiles show twodistinct segments. The first segment extends from the surface toabout 15 cm below the sediment–water interface and represents amixing zone between seawater and groundwater. This zone is affect-ed by wave-induced seawater intrusion into the sediment (Li et al.,1999; Ullman et al., 2003; Beck et al., 2007a, 2010; Pempkowiak etal., 2010). In addition to heterogeneities within the aquifer, coastalseepage sites host a recirculated seawater component. Measuredfluxes from Great South Bay, New York for example reached valuesof 150 L m−2d−1. Discharge measurements of 50 L m−2d−1 fornear-shore environments decreased to 30 L m−2d−1 at sites located100 m offshore. Secondary convection due to instabilities in thewater density structure at the sediment–water interface can thus ob-scure the seepage distribution. Marine water can penetrate seepagesites to depths of tens of centimeters due to thermohaline density ef-fects (Bokuniewicz, 1992). The end member approach provides aseepage flux interpretation that recognizes potential contributionfrom recirculated seawater.

The groundwater-seawater mixing zone gradually grades into freshgroundwaterwithout a sharp boundary between zones. Previous studiesof SGD sites have reported salinity profiles similar to those describedhere (Ullman et al., 2003; Beck et al., 2007a, 2010; Pempkowiak et al.,2010). In contrast to SGD sites, pore water salinity profiles of the G′ siteshowed nearly constant salinity of around 7. At reference point G′ alsoa seepagemeterwas deployed. Themeasured salinity at the PTE collectorwas 7.2 and the nutrient concentrations were similar to those character-istic of seawater: nitrate (NO3) 0.76–1.34 μmol L−1, nitrite (NO2) 0.13–0.98 μmol L−1, ammonium (NH4) 0.56–1.98 μmol L−1, orthophosphate(PO4) 1.91–4.21 μmol L−1, thus samples collected at the G site can beregarded as containing traces of groundwater at the most.

4.2. Nutrient concentrations from SGD sites in the study area

Fig. 4 shows characteristic porewater nutrient profiles (samples col-lected on 3.09.2009, 5.11.2009, 28.02.2010 and 5.05.2010). In

September 2009, NO3 decreased with depth from 1.03 to0.36 μmol L−1 while and NO2 decreased from 0.26 to 0.1 μmol L−1.Measured NH4 and PO4 concentrations followed opposite trends withNH4 increasing from 5.6 to 367.5 μmol L−1 and PO4 increasing from0.09 to 56.5 μmol L−1. In November 2009, the NO2 concentration pro-file showed decreases with depth from 0.5 to 0.3 μmol L−1, and NO3

concentrations decreased with depth from 0.9 to 0.2 μmol L−1. BothNH4 and PO4 concentrations were higher in pore water collected fromdeeper sediment layers (43.1 and 57.9 μmol L−1, respectively) thanfrom surface layers (0.9 and 1.05 μmol L−1, respectively). Profiles mea-sured in February, 2010 resembled thosemeasured in November, 2009.Profiles collected inMay, 2010 resembled those collected in September,2009. The NH4 concentration maxima varied significantly with time.The lowest NH4 concentration was measured from samples collectedin February, 2010 and the highest NH4 concentrations (b300 μmol L−1)occurred in samples collected in September, 2009. The other nutrientsanalyzed showed less variation in their concentrations as a function oftime.

Fig. 5 shows inverse relationships between salinity and both NH4

and PO4 concentrations. The PO4 profile exhibits a nonlinear decreasewith salinity. The apparently non-conservative behavior of PO4 sug-gests diagenetic processes taking place on ground- and sea-watermixing. Relative to phosphates, NH4 concentrations exhibit more con-servative behavior.

The patterns observed in NH4 and PO4 data resemble those reportedin Ullman et al. (2003) wherein nutrient concentrations in sedimentpore waters from Cape Henlopen, Delaware (a sandy beach face)showed signs of dispersive mixing between saline and groundwaterend-members, as well as a diagenetic contribution (or removal).

In many coastal environments, groundwater discharge representsonly a small proportion of freshwater flux into the ocean and thusmay not be a significant input pathway of dissolved chemical sub-stances. If nutrient concentrations are sufficiently high however, SGDcan deliver substantial input loads. Table 1 shows average concentra-tions of nutrients in groundwater and seawater samples collectedfrom the study area. Nutrients concentrations in pore water collectedfrom the depths of 15 cm, or deeper, below sediment–water interface(salinity smaller than 0.1) were used as characteristic of groundwater.Derivation of these values assumed that groundwater salinity rangedfrom 0 to 1 (except samples collected in November, 2009 when salinityranged from 1 to 2.1) and seawater salinity ranged from of 6.9 to 7.2.Pore water samples having salinities ranging from 1 to 6.9 (or 2.1 to6.9 for November, 2009) were assumed to represent mixed groundwa-ter–seawater samples.We used groundwaterfluxmeasures to calculateweighted average values for concentrations of each species. Table 1shows the yearly average concentration values calculated as arithmeticmeans of the average seasonal concentrations.

Variations in nutrient concentrations over the course of the yearindicate pronounced seasonal changes in groundwater nutrient con-tent. Ammonia reached its highest average value in samples collectedduring the summer months (239 μmol L−1) and its lowest value(55 μmol L−1) in samples collected during the spring. The highest av-erage PO4 concentration was observed in samples collected duringthe autumn (76.6 μmol L−1) and the lowest PO4 concentrations oc-curred in samples collected during the spring (49 μmol L−1). Concen-trations of NO3 and NO2 were less temporally variable. Average NH4

and PO4 concentrations measured from seawater were lower thanthose measured from groundwater. Concentrations of NO3 werehigher than those observed in seawater whereas NO2 concentrationswere equal to those observed in seawater.

4.3. Nutrient fluxes into the study area

Seepage water fluxes and nutrient concentrations measured ingroundwaterwas used to calculate nutrient load input via SGD. Seepagewater flux was measured using a seepage meter. Seepage water

Fig. 4.Nutrient profiles in sediment pore water collected from the groundwater impacted area (location G) on 3.09.2009, 5.11.2009, 28.02.2010 and 5.05.2010, and from an area freeof groundwater impact (location G′).

Fig. 5. The relationships between NH4 and PO4 concentrations and salinity. Theresults include data collected on 31.08–3.09.2009; 2–6.11.2009; 28.02–1.03.2010; and5–7.05.2010.

90 B. Szymczycha et al. / Science of the Total Environment 438 (2012) 86–93

percolating into the device from the sediment displaces water trappedin the chamber, forcing it up through a port into a PTE measurementbag. The change in the volume of water within the bag over ameasuredtime interval equals the seepage flux. The contribution of groundwaterto flux estimates was established using the end-member method. Thusthis method uses sensitive and precise measurements of end-membercompositions to receive groundwater and seawater components ofseepage fluxes.

Salinity is often used as a tracer for SGD identification (Rapaglia,2007). Table 2 shows salinity of seepage water samples, seepage waterfluxes and groundwater fluxes. The highest flux (21.3 Ld−1 m−2) wasobserved on 2.09.09 and the lowest flux (3.0 Ld−1 m−2) on 28.02.10.Within the fluxes results we can observe lower fluxes in February, 2010and May, 2010 whereas higher fluxes were measured in September,2009 and November, 2009. The SGDs are well correlated with averagemonthly precipitation characteristic of the area (Korzeniowski, 2003).

The calculated average groundwater fluxes and the measurednutrient concentrations were used to derive nutrient fluxes at samplingpoints (Table 3).

5. Discussion

5.1. Nutrient fluxes to the Bay of Puck

Nutrient loads from SGD are usually calculated as the products ofseepage water flux and concentration of chemical species measured ininland groundwater sources (Oberdorfer et al., 1999). This methodassumes that nutrient concentrations from coastal and inland ground-water sources do not change. Due to the questionable nature of thisassumption, the actual nutrient concentrations were measured ingroundwater samples collected throughout the study by means of

Table 1Average nutrient concentrations in seawater and groundwater samples collected from the study area.

Nutrients Concentrations [μmol L−1]Averagea

Range: minimum–maximum

September, 2009 November, 2009 February, 2010 May, 2010 Average±σ

Seawater Groundwater Seawater Groundwater Seawater Groundwater Seawater Groundwater Seawater Groundwater

PO4 0.360.32–0.4

55.254.8–55.6

0.0090.007–0.011

76.676.3–76.9

1.71.01–2.4

61.239.4–83.01

0.370.32–0.42

49.030.9–67.1

0.61±0.370.41–1.01

60.6±5.942.5–78.7

NH4 5.85.6–6.01

239.0108.8–369.2

0.90.84–0.94

117.089.1–144.9

2.31.4–3.2

60.538.7–82.3

1.10.1–2.1

55.050.8–59.2

2.5±1.12.0–3.01

118.2±42.717.3–219.1

NO3 0.570.17–0.97

0.370.27–0.47

0.980.96–1.01

0.250.243–0.257

2.061.26–2.86

0.580.57–0.59

1.150.9–1.4

0.330.15–0.51

1.2±0.310.8–1.6

0.4±0.070.3–0.5

NO2 0.250.18–0.32

0.280.27–0.29

0.430.30–0.56

0.30.15–0.45

0.220.02–0.42

0.230.21–0.25

0.350.25–0.45

0.170.07–0.27

0.3±0.050.2–0.4

0.2±0.030.14–0.26

a Arithmetical means.

91B. Szymczycha et al. / Science of the Total Environment 438 (2012) 86–93

groundwater lances and seepage meters. We assume a continuous sup-ply of nutrients transported by groundwatermigration through the sed-iment driven by the hydraulic gradient. Nienchesky et al. (2007) used asimilar approach, measuring average nutrient concentrations in watersamples collected from permanent land-based wells, beach groundwa-ter, and samples from the surf zone, inner shelf and open ocean. Howev-er still, there is limitation to the implemented method, since in themixing layer biogeochemical transformations of the nutrients may sub-stantially alter the true discharge of this nutrient species.

The sum of the measured NO3, NO2 and NH4 concentrations formedthe basis of the dissolved inorganic nitrogen (DIN) concentration.The PO4 concentration served as a measure of dissolved phosphates.A groundwater flux estimate of 0.03 km3 yr−1 was adopted fromPiekarek-Jankowska et al. (1994) and Korzeniowski (2003). The seep-age discharged rates estimated by Piekarek-Jankowska et al., 1994have been made for the entire Puck Bay. The calculation was basedon hydrogeological method and allowed to estimate the role of SGDin the water balance of the entire Puck Bay. The seepage dischargedrates estimated during this study were characteristic of the activeSGD at the study area and thus exceed those calculated by Piekarek-Jankowska et al. (1994).

Table 4 gives the loads calculated as products of fluxes and aver-age yearly concentrations of nutrients. The NH4 concentration was47 times higher in groundwater than in seawater and the PO4 concen-tration was almost 99 times higher. In their study along a coastal la-goon barrier in southern Brazil, Nienchesky et al. (2007) alsoshowed that groundwater is enriched in NO3 and NH4, but containedonly slightly elevated levels of PO4. The ranges of NH4 and PO4 levelsin groundwater samples from their study however were lower than

Table 2Seepage and groundwater fluxes measured from the study area.

Date Seepagemeter

Salinity measuredfrom PTE bags

Seepage waterflux [L d−1 m−2]

Groundwater flux,calculated from theend-membermethod[L d−1 m−2]

Measuredvalues

Average

2.09.09 S1 4.1 64.4 28.5 21.3S2 5.6 23.2 5.3S3 5.2 89.7 25.6

4.11.09 S1 6.3 99.9 12.8 18.4S2 6.5 187.1 18.7S3 6.1 150.4 23.6

28.02.10 S1 5.8 7.1 1.4 3.0S2 5.6 10.2 2.3S3 4.7 15.1 5.4

5.05.10 S1 5.9 19.3 3.6 3.6S2 3.7 7.3 3.7S3 6.5 36.7 3.7

those reported here (Table 4). A study by Lee et al. (2009) of theMasan Bay, a site recognized as the most eutrophic embayment inKorea, reported similar DIN concentrations to those reported here.The PO4 concentrations in groundwater samples from the Lee et al.(2009) study however were approximately ten times lower thanthose reported here (see Table 4). The PO4 concentrations reportedhere also exceed those measured at seepage locations in other studies(Beck et al., 2007b; Leote et al., 2008; Lee et al., 2009). Hosono et al.(2012) reported SGD amounting to 6000 m3 day−1 over a 20,000 m2

area as well as DIN and DP fluxes of 920 mol day−1 and 56 mol day−1

respectively. Street et al. (2008) described sites along the Hawaii coastwith SGD fluxes ranging 0.02 to 0.65 m3 m−2d−1, delivering nitrogenloads of 0.04 to 40 mmol m−2d−1 and phosphate loads of 0.01 to1.6 mmol m−2d−1. Given their scale relative to previous examples,the SGD site described here is interpreted as a major source of nutrientinput into the surrounding coastal ecosystem.

5.2. Nutrient fluxes to the Baltic Sea

River transport fromdrainage basin networks deliversmajor loads ofnutrients to marine environments (Bokuniewicz, 1992; Turner et al.,1999; Burnett et al., 2003). Analysis of inputs from major rivers allowsrelatively precise estimates of freshwater and nutrient inputs into theBaltic Sea (HELCOM, 2005). Non-point pathways in the subsurface areincreasingly recognized as a major input mechanism for certain coastalareas (Beck et al., 2007b; Lee et al., 2009; Pempkowiak et al., 2010).Burnett et al. (2003) define “submarine groundwater discharge” asany and all water flow from the seabed to the ocean along coastal mar-gins, regardless of fluid composition or driving force. SGD typically ex-hibits low specific flow rates that make detection and quantificationdifficult. These fluids transfer biogeochemically important inputs intothe coastal marine environment. Slomp and Van Cappellen (2004) sug-gest that groundwater seeps may contribute to P limitation in the vicin-ity of SGD sources due to specific N/P ratios. The Bay of Puck SGD siteexhibits small N/P ratios and thus would not contribute to P limitation

Table 3Nutrient input into the study area.

Samplingcampaign

Average nutrientconcentrations[μmol L−1]

SGD flux[L d−1 m−2]

Nutrient loads[μmol d−1 m−2]

PO4 NH4 NO3 NO2 PO4 NH4 NO3 NO2

September,2009

55.2 239.0 0.4 0.3 21.3 1175.8 5090.7 7.9 6.0

November,2009

76.6 117.0 0.3 0.3 18.4 1409.4 2152.8 4.6 5.5

February,2010

61.2 60.5 0.6 0.2 3.0 183.6 181.5 1.7 0.7

May, 2010 49.0 55.0 0.3 0.2 3.6 176.4 198.0 1.2 0.6

92 B. Szymczycha et al. / Science of the Total Environment 438 (2012) 86–93

in surrounding coastal environments. Gradual seepage of groundwaterthrough sediments can occur anywhere given an aquifer with sufficienthydrologic pressure relative to sea level and a permeable coastal inter-face. The Baltic Sea sediments are mostly impermeable (Peltonen,2002) but some locations along the southern Baltic coastline includesufficiently porous bottom layers that permit freshwater seepage(Peltonen, 2002; Uścinowicz, 2011).

Although a number of studies have addressed SGD along coastalareas of the Baltic Sea (Zekster, 1973; Piekarek-Jankowska et al., 1994;Piekarek-Jankowska, 1996; Bussmann and Suess, 1998; Falkowskaand Piekarek-Jankowska, 1999; Kaleris et al., 2002; Peltonen, 2002;Viventsova and Voronov, 2003; Korzeniowski, 2003; Schlüter et al.,2004; Kryza and Kryza, 2006; Uścinowicz, 2011), these earlier worksdid not report nutrient concentrations. Given the absence of previousSGD nutrient load estimates, we project the nutrient inputs observedhere to the Baltic Sea nutrient budgets. Nutrient projections assumedthat SGD along the Baltic Sea coast contains DIN and PO4 concentrationssimilar to those observed in seepage water from the Bay of Puck site(Table 4) and combined these estimates with groundwater water flowestimates from literature sources (Peltonen, 2002). The error envelopesof estimates were calculated from standard deviations of the averageyearly concentrations observed at the study site. The projectionsindicated that nutrient input into the Baltic Sea amounts to a meanvalue of 7320 t yr−1 for DIN (1080 to 13,560 t yr−1) and a meanvalue of 8260 t yr−1 for DP (6430 to 10,080 t yr−1).

The limitations of the data notwithstanding, the projected nutri-ent load calculations constrain SGD inputs to the Baltic Sea withinan order of magnitude, rather than approximating actual loads. Sever-al factors may influence the nutrient load projection. These includeuncertainties in the representativeness of concentrations for a givensampling campaign, representativeness of concentrations over theentire year and whether the study area provides an accurate ground-water flux estimate for the entire Baltic Sea. The following paragraphsdiscuses shortly each of these uncertainties.

The campaigns for collecting sediment pore water samples andmeasuring seepage water flux in this study lasted several days. Bothfluxes and nutrient concentrations were rather constant over thetime span of each of the four campaigns (Table 1). On the otherhand, both fluxes and nutrient concentrations varied in samples col-lected during subsequent seasons. Studies of groundwater collectedwith piezometers within two to three-month long periods from in-land locations in the vicinity of the study area indicate that concentra-tions of major ions are relatively constant along the Bay of Puckcoastline (Przewtocka, 2007). The sampling frequency on which nu-trient load projections were based may thus be close to the limit nec-essary to provide reliable mean estimates.

Pomerania, the area surrounding the study site, resembles otheragricultural areas along the southern Baltic Sea coast (Peltonen,2002). Few studies have addressed the location or flux intensity ofgroundwater discharges along the Baltic Sea coastline but the sedi-mentary and geomorphic structure of the area indicate that seepsoccur primarily along the densely populated southern and easterncoastal areas where land use patterns are dominated by agricultural

Table 4PO4 and DIN fluxes from external sources into the Bay of Puck.

Source Fluxes into the Bay of Puck

[t yr−1] [kmol yr−1]

PO4 DIN PO4 DIN

Atmospherea 18 485 581 34,626Rivers and point sourcesa 70 220 2260 15,706Resuspensiona 97 825 3132 58,899SGDb 56 50 1808 3570

a Korzeniowski (2003).b This study.

activity. Uścinowicz (2011) suggested that the Bay of Gdansk is thedominant source of groundwater discharge into the Baltic Sea. Locat-ed within the Bay of Gdansk, the present study area can serve as ancharacteristic example of regional SGD. The sampling site selectedfor the study is a typical agricultural area with average fertilizerusage of 150 kg km−2annum−1 nitrogen and 10 kg km−2annum−1

phosphate. Nutrient concentrations measured in groundwater fromthe study area can thus be considered typical of agricultural areas prev-alent along the Baltic Sea coast. Given these factors, we interpret theloads listed in Table 4 as representative of the upper limit of nutrientfluxes from individual SGD sites along the southern Baltic Sea coast.

The projected estimates of nutrient input into the Baltic Sea areprimarily intended to draw attention to the significance of SGD in hy-drologic nutrient cycles. The projections demonstrate that SGD sitesmay transport several thousand tons of DIN and DP into the BalticSea. Nutrient budgets for marine reservoirs however do not typicallyinclude SGD sources. The Baltic Sea water column is relatively eutro-phic and groundwater seepage within the coastal zone is likely to en-hance eutrophication. Because primary production in coastal marineenvironments is often limited by nitrate, groundwater derived nitratemay contribute significantly to eutrophication and harmful algalblooms (Andersen et al., 2007).

Our results and the nutrient projections based on them indicate thatSGD contributes amajor proportion of nutrients to theBaltic Sea. Overallnutrient estimates for the Baltic Sea typically include inputs from directatmospheric deposition to surface water, river inputs and point sourcesdischarging directly into the sea. Table 5 lists DIN and PO4 input to theBaltic Sea from these sources. Waterborne inputs contributed thegreatest proportion of nitrogen (75%) and phosphorous (nearly 100%)to the Baltic Sea in 2000. These estimates do not include nutrient loadsfrom SGD but they provide an important point of comparison for theprojected estimates reported here. Compared to other sources, theprojected SGDDIN load constitutes only a small fraction of total nitrogendelivered to the Baltic Sea. The projected SGD PO4 load however issignificant relative to loads delivered from other sources. Comparisonamong the sources indicates that SGD contributes around 20% of thePO4 input but only around 1% of the DIN input relative to the totalrespective dissolved nutrient fluxes into the Baltic Sea (HELCOM,2005). Given that coastal areas with permeable sediment layers arehighly susceptible to groundwater impact, these values indicate thatSGDdischarge plays a significant role in thehydrogeochemistry of coast-al ecosystems in the Baltic Sea.

6. Conclusions

This study provides initial estimates of SGD nutrient concentra-tions and fluxes into the Bay of Puck. The PO4 and NH3 concentrationsfrom SGD sources entering the Bay of Puck are elevated relative toseawater concentrations whereas NO2 and NO3 concentrations arecomparable to those observed in seawater. The SGD phosphate loadis significant relative to loads entering the bay from the atmosphereand fluvial sources, whereas DIN load is less significant relative tothese sources.

Nutrient loads observed at the SGD site studied here wereprojected to the entire Baltic Sea in order to establish overall SGD nu-trient contributions within an order of magnitude. The projections

Table 5PO4 and DIN fluxes from external sources into the Baltic Sea.

Fluxes into the Baltic Sea PO4 [t yr−1] DIN [t yr−1]

Atmospherea b1730 196,000Rivers and point sourcesa 30,200 641,000SGDb 8668 7176

a Fluxes in 2000 (Helcom, 2005).b This study.

93B. Szymczycha et al. / Science of the Total Environment 438 (2012) 86–93

indicated significant SGD contribution to overall phosphate input intothe Baltic Sea. Further studies are critical to fully understand theapparently significant role of SGD nutrient inputs into the Baltic Seamarine ecosystem.

Acknowledgments

We are grateful to the anonymous reviewers for providing com-ments and suggestions that were used to improve the manuscript.The study reports results obtained within the framework of the fol-lowing projects: AMBER‐a BONUS_B EU FP6 Project, the research pro-ject 3837/B/P01/201039 sponsored by the Polish Ministry of Scienceand Higher Education, and statutory activities of the Institute ofOceanology, Sopot‐theme 2.2.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2012.08.058.

References

Andersen S, Baron L, Gudbjerg J, Gregersen J, Chapellier D, Jakobsen R, et al. Dischargeof nitrate-containing groundwater into a coastal marine environment. J Hydrol2007:98-114.

Beck AJ, Tsukamato Y, Tovar-Sanchez A, Huerta-Diaz M, Bokuniewicz HJ, Sañudo-Wilhelmy SA. Importance of geochemical transformations in determining subma-rine groundwater discharge‐derived tracemetal and nutrient fluxes. Appl Geochem2007a;22:477–90.

Beck M, Dellwig O, Kolditz K, Freund H, Liebezeit G, Schnetger B, et al. In situ porewater sampling in deep intertidal flat sediments. Limnol Oceanogr 2007b;5:136–44.

Beck AJ, Cochran JK, Sañudo-Wilhelmy SA. The distribution and speciation of dissolvedtrace metals in a shallow subterranean estuary. Mar Chem 2010;121:145–56.

Bokuniewicz HJ. Analytical descriptions of subaqueous groundwater seepage. Estuaries1992;15:458–64.

Burnett WC, Bokuniewicz H, Huettel M, Moore WS, Taniguchi M. Groundwater andpore water inputs to the coastal zone (review). Biogeochemistry 2003;66:3-33.

Burnett WC, Aggarwal PK, Aureli A, Bokuniewicz H, Cable JE, Charettte MC, et al. Quan-tifying submarine groundwater discharge in the coastal zone via multiplemethods. Sci Total Environ 2006;367:498–543.

Bussmann I, Suess E. Groundwater seepage in Eckernförde Bay (Western Baltic Sea):effect on methane and salinity distribution of the water column. Cont Shelf Res1998;18:1795–806.

Carlton J. Wetland hydrology. Water quality and associated function. USGS. NationalWater Summary on Wetland ResourcesUSGS Water Supply 2006;2425:35–48.

Cyberski J. River outflow from Poland. Oceanolog Stud 1992;26:65-23.Eamus D, Froend G, Lommes R, Hose G, Murray B. A functional methodology for deter-

mining the groundwater regime needed to maintain health of groundwater-dependent vegetation. Aust J Bot 2006;54:97-114.

Falkowska L, Piekarek-Jankowska H. The submarine seepage of the fresh water: distur-bance in hydrological chemical structure of the water column in the Gdansk Deep.J Mar Sci 1999;56:153–60.

Helcom. Baltic sea environment proceedings 100,101. Helsinki: Helcom; 2005.Hosono T, Ono M, Burnett WC, Tokunaga T, Taniguchi M, Akimichi T. Spatial distribu-

tion of submarine groundwater discharge and associated nutrients within a localcoastal area. Environ Sci Technol 2012;46:5319–26.

Kaleris V, Lagas G, Marczinek S, Piotrowski JA. Modelling submarine groundwaterdischarge: an example from the western Baltic Sea. J Hydrol 2002;265:76–99.

Korzeniowski K, Zatoka Pucka. Instytut Oceanografii Uniwersytetu Gdańskiego, Gdynia2003.

Kryza J, Kryza H. The analytic and model estimation of the direct groundwater flow toBaltic Sea on the territory of Poland. Geologos 2006;10:153–65.

Lee Y, Hwang D, Kim G, Lee W, Oh H. Nutrients inputs from submarine groundwaterdischarge (SGD) in Masan Bay, an embayment surrounded by heavily industrialized,Korea. Sci Total Environ 2009;407:3181–8.

Leote C, Ibanez JS, Rocha C. Submarine groundwater discharge as a nitrogen source tothe Ria Formosa studied with seepage meters. Biogeochemistry 2008;88:185–94.

Li L, Barry DA, Stagnitti F, Parlange J-Y. Submarine groundwater discharge and associ-ated chemical input to a coastal sea. Water Resour Res 1999;35:3253–9.

McCoy CA, Corbett DR. Review of submarine groundwater discharge (SGD) in coastalzones of the Southern and Gulf Coast regions of United States with managementsimplications. J Environ Manag 2009;90:644–51.

Millham NP, Howes BL. Freshwater flow into a coastal embayment: groundwater andsurface water inputs. Limnol Oceanogr 1994;39:1928–44.

Mulligan AE, Charette MA. Groundwater flow to the coastal ocean. Encycl Ocean Sci2009;32:3842–51.

Nienchesky LFH, Windom HL, Moore WS, Jahnke RA. Submarine groundwaterdischarge of nutrients to the ocean along a coastal lagoon barrier. Southern Brazil.Mar Chem 2007;106:546–61.

Nowacki J. The morphometry of the Bay of Puck. Gdynia: Fundacja RozwojuUniwersytetu Gdańskiego; 2003.

Oberdorfer JA, Valentino MA, Smith SB. Groundwater contribution to the nutrient bud-get of Tomales Bay, California. Biogeochemistry 1999;10:199–216.

Peltonen K. Direct groundwater flow to the Baltic Sea. Copenhagen: Nordic Council ofMinisters; 2002.

Pempkowiak J, Szymczycha B, Kotwicki L. Submarine groundwater discharge (SGD) tothe Baltic Sea. Rocznik Ochrony Środowiska 2010;12:17–32.

Piekarek-Jankowska H. Hydrochemical effects of submarine groundwater discharge tothe Bay of Puck. Geogr Pol 1996;67:103–19.

Piekarek-Jankowska H, Matciak M, Nowacki J. Salinity variations as an effect of ground-water seepage through the seabed (Bay of Puck, Poland). Oceanologia 1994;36:33–46.

Przewłócka M. Gdansk Aquifer, PhD Dissertation. Gdańsk: Wydawnictwo PolitechnikiGdanskiej Gdańsk; 2007.

Rapaglia JP. Using Direct Measurements of Submarine Groundwater Discharge toInvestigate the Coupling Between Surface and Pore waters. PhD Dissertation,Stony Brook: Marine and Atmospheric Science, 2007.

Salley BA, Bradshaw JB, Neillo BJ. Results of comparative studies of preservation tech-niques for nutrient analysis on water samples. Virginia Institute of Marine Science:Gloucester Point; 1986.

Schlüter M, Sauter EJ, Andersen CA, Dahlgaard H, Dando PR. Spatial distribution andbudget for submarine groundwater discharge in Eckernförde Bay (Western BalticSea). Limnol Oceanogr 2004;49:157–67.

Slomp CP, Van Cappellen P. Nutrient inputs to the coastal ocean through submarinegroundwater discharge: controls and potential impact. J Hydrol 2004;295:64–86.

Street JH, Knee KL, Grossman EE, Paytan A. Submarine groundwater discharge andnutrient addition to the coastal zone and coral reefs of leeward Hawai'i. MarChem 2008;109:355–76.

Strickland JDH, Parsons TR. A practical handbook of seawater analysis. Ottawa: FisheriesResearch Board of Canada Bulletin; 1967.

Szczepańska A, Zaborska A, Maciejewska A, Kuliński K, Pempkowiak J. Distribution andorigin of organic matter in the Baltic sea sediments dated with 210Pb and 137Cs.Geochronometria 2012;39:1–9.

Turner RK, Georgiou S, Gren IM, Wulff F, Barrett S, Soderqvist T, et al. Managing nutri-ent fluxes and pollution in the Baltic: an interdisciplinary simulation study. EcolEcon 1999;30:333–52.

Ullman WJ, Chang B, Miller DC, Madse JA. Groundwater mixing, nutrient diagenesis,and discharges across a sandy beachface, Cape Henlopen, Delaware (USA). Estua-rine Coastal Shelf Sci 2003;57:539–52.

Uścinowicz Sz. Geochemistry of the Baltic Sea surface sediments. Warsaw: PolishGeological Institute-National Research Institute; 2011. p. 141–5.

Valiela I, Bowen JL, Kroeger KD. Assessment of models for estimation of land-derivednitrogen loads to shallow estuaries. Appl Geochem 2002;17:935–53.

Viventsova EA, Voronov AN. Groundwater discharge to the Gulf of Finland (Baltic Sea):ecological aspects. Environ Ecol 2003;45:221–5.

Zekster IS. Groundwater discharge into the Baltic Sea and methods of estimating. J Hydrol1973;20:105–18.

Zohdy AAR, Jackson DB. Application of deep electrical soundings for groundwaterexploration in Hawaii. Geophysics 1969;34:584–600.