Submarine groundwater discharge as a major source of nutrients to the Mediterranean Sea Valentí Rodellasa,1, Jordi Garcia-Orellanaa,b,1, Pere Masquéa,b,c,d, Mor Feldmane, and Yishai Weinsteine a Institut de Ciència i Tecnologia Ambientals and bDepartament de Física, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain; cOceans Institute and School of Physics, The University of Western Australia, Crawley, WA 6009, Australia; dSchool of Natural Sciences and Centre for Marine Ecosystems Research, Edith Cowan University, Joondalup, WA 6027, Australia; and eDepartment of Geography and Environment, Bar-Ilan University, Ramat-Gan 52900, Israel
Edited by Edward A. Boyle, Massachusetts Institute of Technology, Cambridge, MA, and approved February 18, 2015 (received for review October 4, 2014)
Mediterranean Sea radium
| submarine groundwater discharge | nutrients |
T
he Mediterranean Sea (MS) is one of the most oligotrophic seas in the world (1), resulting from an antiestuarine circulation by which Mediterranean intermediate waters export nutrients to the Atlantic Ocean. In such an oligotrophic system, external inputs of nutrients play a significant role in sustaining marine productivity (2–6). Among these inputs, atmospheric deposition and riverine runoff have been traditionally considered the major sources of nutrients to the MS (5, 7, 8). Even though submarine groundwater discharge (SGD) has been recognized as an important source of dissolved compounds (e.g., nutrients, metals, carbon) to the ocean (9–11), its role as a conveyor of dissolved chemicals to the MS has been largely overlooked. SGD into the coastal ocean is composed of fresh meteoric groundwater, which may enter the ocean directly, and former seawater recirculating through permeable sediments (12, 13). Both types of SGD are important in oceanic chemical budgets, as SGD may become enriched in various solutes (i.e., nutrients, metals, carbon) during its pathway through the coastal aquifer (14). Previous studies have estimated the fresh groundwater discharge to the MS, using hydrologic approaches (15, 16), but none of them has evaluated the total SGD flow and the associated total contribution of dissolved compounds to the MS. Because recirculated SGD usually dominates the water flow (12), its quantification is key to fully understanding the role of SGD in the biogeochemical cycles of nutrients and other chemical species in the MS. Radium isotopes have been widely used as tracers of SGD, mainly because they are highly enriched in SGD relative to seawater and because they behave conservatively once released into the sea (17). The approach we followed to estimate the magnitude of SGD into the entire MS is based on a mass balance of 228Ra (18). Assuming steady state, 228Ra outputs must be www.pnas.org/cgi/doi/10.1073/pnas.1419049112
equal to its inputs, most of which originate from continental margins. By evaluating all of the sink and source terms, the 228Ra supplied by SGD can be determined as the difference between its inputs and outputs, and this flux can be converted to an SGD flow by characterizing the 228Ra concentration in the SGD fluids. We used 228Ra over other Ra isotopes (i.e., 223Ra, 224Ra, and 226Ra) because its half-life (T1/2= 5.75 y) is considerably lower than the residence time of the MS waters [∼100 y (19)]; hence, its radioactive decay is the primary sink of 228Ra in the MS and allows us to accurately constrain the 228Ra removal terms (18). Results and Discussion Estimating all the terms of the 228Ra mass balance requires the characterization of the distribution of 228Ra in the water column of the MS (Fig. 1; see Methods). The highest 228Ra concentrations are observed in coastal areas, consistent with radium inputs from continental margins. 228Ra concentrations measured along the MS mainly reflect the antiestuarine thermohaline circulation of the basin: Surface 228Ra concentrations generally increase from west to east (from 17 to 34 dpm·m−3), implying that waters moving eastward receive 228Ra from the continental margins at a greater rate than radioactive decay or other forms of removal, and reflecting the evaporation of water (solute concentration). Because Levantine Intermediate Water (LIW), which is regularly formed in the Eastern Mediterranean, flows westward to the Strait of Gibraltar, and there are no other major sources of 228Ra at intermediate depths, 228Ra concentrations in LIW generally decrease from east to west (from 29 to 12 dpm·m−3). 228Ra concentrations in deep waters (DW; depth > 600 m) are considerably lower (5–13 dpm·m−3; SI Appendix), confirming that the main 228Ra inputs occur at the uppermost layers. 228 228 Ra Mass Balance for the Upper MS. The loss of Ra resulting from radioactive decay can be determined from the total 228Ra
Significance The Mediterranean Sea (MS) is one of the most oligotrophic seas in the world, and external inputs of nutrients are especially relevant to sustaining primary productivity in this basin. Here we evaluate the role of submarine groundwater discharge (SGD) as a source of nutrients to the entire MS, a pathway that has been largely overlooked. This study demonstrates that SGD is a volumetrically important process in the MS, is of a larger magnitude than riverine discharge, and also represents a major source of dissolved inorganic nitrogen, phosphorous, and silica to the MS. Author contributions: V.R., J.G.-O., P.M., and Y.W. designed research; V.R., J.G.-O., P.M., and Y.W. performed research; V.R. and M.F. analyzed data; and V.R., J.G.-O., P.M., M.F., and Y.W. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1
To whom correspondence may be addressed. Email: [email protected] or jordi. [email protected]
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1419049112/-/DCSupplemental.
PNAS Early Edition | 1 of 5
ENVIRONMENTAL SCIENCES
The Mediterranean Sea (MS) is a semienclosed basin that is considered one of the most oligotrophic seas in the world. In such an environment, inputs of allochthonous nutrients and micronutrients play an important role in sustaining primary productivity. Atmospheric deposition and riverine runoff have been traditionally considered the main external sources of nutrients to the MS, whereas the role of submarine groundwater discharge (SGD) has been largely ignored. However, given the large Mediterranean shore length relative to its surface area, SGD may be a major conveyor of dissolved compounds to the MS. Here, we used a 228Ra mass balance to demonstrate that the total SGD contributes up to (0.3–4.8)·1012 m3·y−1 to the MS, which appears to be equal or larger by a factor of 16 to the riverine discharge. SGD is also a major source of dissolved inorganic nutrients to the MS, with median annual fluxes of 190·109, 0.7·109, and 110·109 mol for nitrogen, phosphorous, and silica, respectively, which are comparable to riverine and atmospheric inputs. This corroborates the profound implications that SGD may have for the biogeochemical cycles of the MS. Inputs of other dissolved compounds (e.g., iron, carbon) via SGD could also be significant and should be investigated.
43 - 48 32 - 34 26 - 34 15 - 25
16 - 20 11 - 12 15
104
22 - 24 20 - 21
34 - 36 28 - 34
19 - 23 15 - 16 26 - 28 19 - 23
29 22 - 29 28 19 - 27
SW LIW
Fig. 1. Distribution of 228Ra concentrations (dpm·m−3) in surface waters of the MS. Dots indicate sampled stations. The subbasin division and the weighted average 228Ra concentration in each subbasin for surface (SW) and intermediate (LIW) waters are indicated (a range is reported to include the concentrations derived from the different approaches) (see SI Appendix for details).
inventory multiplied by its decay constant (λ = 0.120 y−1). To calculate the total 228Ra inventory, the MS was divided into nine subbasins (Fig. 1), defined according to their physiography and general circulation patterns (20, 21). Each subbasin was split into three water layers [surface water (SW), LIW, and DW]. The total 228 Ra inventory for the upper water column of the MS (including SW and LIW; i.e., from 0 to 600 m depth) was obtained from the integration of the box inventories, which were calculated using three different approaches (uniform 228Ra concentrations with depth, vertical linear interpolation between measured 228Ra concentrations, and inferring 228Ra concentrations based on salinity profiles; see Methods and the SI Appendix). The three approaches used resulted in similar total 228Ra inventories for the upper MS [(37 ± 4)·1015 dpm, (34 ± 4)·1015 dpm, and (33 ± 4)·1015 dpm], allowing us to constrain the total 228Ra inventory in (35 ± 6)·1015 dpm. The loss of 228Ra by radioactive decay is thus of (4.2 ± 0.7)·1015 dpm·y−1. The DW layer was excluded from the total inventory because the main 228Ra inputs to the water column occur at the uppermost layers (18). In doing so, 228Ra released from deep sediments can be neglected, but vertical exchange fluxes need to be considered. The migration of 228Ra from upper waters to DW was estimated from the average 228Ra box concentrations and from water transfer rates obtained in previous studies (20, 21). This results in a total 228Ra loss of (2.0 ± 0.1)·1015 dpm·y−1 as a result of downward water flow (SI Appendix). In addition, 228Ra could also be transported downward because of particle scavenging. Assuming a radium residence time with respect to scavenging of ∼500 y (18), the 228Ra removal resulting from particle scavenging would only be of (0.07 ± 0.01)·1015 dpm·y−1. 228 Ra is also lost from the MS because of the water outflow to the Atlantic Ocean (25·10 12 m3·y −1) (22) and the Black Sea (1.0·1012 m3·y −1) (23), through the straits of Gibraltar and Dardanelles, respectively. From the characterization of the 228Ra concentrations in outflowing waters, a total loss of 228Ra of (0.36 ± 0.03)·1015 dpm·y−1 through the straits is obtained. The total loss of 228Ra from the upper MS [(6.6 ± 0.7)·1015 dpm·y−1; Table 1] must be balanced by continuous new inputs of 228Ra that can consist of riverine discharge, atmospheric input, release from sediments, deep water vertical advection, water input through the straits, and SGD. The input of 228Ra from river discharge includes Ra in the dissolved fraction and desorption of Ra from river-borne particles (24). A concentration of 130 ± 50 dpm·m−3 captures all the dissolved 228Ra concentrations measured in rivers discharging into the MS (SI Appendix). A desorbed 228Ra flux of 0.5 ± 0.4 dpm per 2 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1419049112
gram of sediment includes most of the 228Ra desorption estimates reported throughout the world (SI Appendix). Considering the river flow (0.31·1012 m3·y−1) (5) and the riverine particle flux into the MS (730·1012 g·y−1) (25), the total supply of 228Ra from river discharge would be (0.40 ± 0.29)·1015 dpm·y−1. The atmospheric dust input to the entire MS is estimated at 41·1012 g·y−1 (8). Assuming a desorption of 228Ra from dust of 2 dpm·g−1, which is close to the maximum reported in the literature (18, 24), the atmospheric input of 228Ra to the entire basin is (0.08 ± 0.03)·1015 dpm·y−1. The flux of 228Ra released from shelf sediments as a result of diffusion and bioturbation might be highly variable and, in large-scale studies, can represent a major source of 228Ra resulting from the large extension of the continental shelf (18). Moore et al. (18) conducted a detailed review of the literature to determine the 228Ra fluxes from fine-grained shelf sediments (11·103 ± 5·103 dpm·m−2·y−1), coarse-grained shelf sediments (230 ± 110 dpm·m−2·y−1), and slope sediments (2.3·103 ± 1.1·103 dpm·m−2·y−1). Using the areas of continental shelf and slope of the MS above 600 m [490·109 and 310·109 m2, respectively (26)], and considering that fine-grained sediments represent about 35% of the shelf area (27), the total flux of 228 Ra from sediments would account for (2.7 ± 0.9)·1015 dpm·y−1. Table 1. Summary of the
228
Ra mass balance for the upper MS 228
Ra flux
(·1015) dpm·y−1
Percentage total flux
4.2 1.95 0.07 0.30 0.06 6.6
± ± ± ± ± ±
0.7 0.13 0.01 0.03 0.01 0.7
51–77 26–34 ∼1 4–5 ∼1 84–116
0.40 0.08 2.7 1.1 0.37 0.14 4.7 1.8
± ± ± ± ± ± ± ±
0.29 0.03 0.9 0.2 0.03 0.01 1.0 1.2
2–11 1–2 26–56 13–19 5–6 ∼2 55–89 9–47
228
Ra outputs Decay Advection to DW Particle Scavenging Outflow (Gibraltar) Outflow (Dardanelles) Total outputs 228 Ra inputs Rivers Atmospheric dust Sediments Advection from DW Inflow (Gibraltar) Inflow (Dardanelles) Total inputs (without SGD) 228 Ra from SGD
Rodellas et al.
SGD Flow to the MS. The SGD-derived
228
Ra flux can be converted to a SGD flow by characterizing the 228Ra concentration in the SGD fluids inflowing to the MS. Several studies have reported 228Ra data in SGD along the MS (45 different locations, mainly located on the Northwestern MS), with concentrations spanning a wide range (130–72,000 dpm·m−3; Fig. 2). The highest 228 Ra concentrations are found mainly in SGD dominated by recirculated seawater, as radium mobilization from the solid matrix strongly depends on water salinity. SGD at karstic areas usually shows lower activities than that at granular coastal aquifers. This is for several reasons, including the relatively low 232Th (228Ra parent) levels in carbonates and the typical point-sourced discharge at karstic coastal terrains, which results in a low water–rock interaction in the coastal aquifer and allows the direct discharge of fresh groundwater with a low 228Ra content. A 228Ra concentration of 640–2,200 dpm·m−3 covers the range between the first and third quartiles of the groundwater dataset, accounting for the variability of 228 Ra in SGD and excluding extreme values. Using this range, the resulting SGD flow to the MS ranges from 0.3·1012 to 4.8·1012 m3·y−1 (median, 1.4·1012 m3·y−1). The SGD flow estimated here appears to be equal or larger by a factor of 16 to the riverine discharge to the MS (0.3·1012 m3·y−1) (5). A recent study based on an integrated hydrologic–hydrogeologic approach estimated the fresh SGD flow to the MS to be on the order of 0.07·1012 m3·y−1 (15). Therefore, fresh groundwater inputs
would represent 1–25% of the total SGD inputs to the MS, revealing that seawater recirculating through coastal sediments constitutes the main fraction of SGD inflowing to the MS. As detailed earlier, 228Ra may be low in fresh groundwater inflowing to the MS (28); therefore, using 228Ra as a tracer could underestimate fresh (and consequently total) SGD inputs. However, given that significant 228Ra enrichments (higher than 1,000 dpm·m−3) may be also found in entirely fresh groundwater springs [e.g., Badum (29), Sa Nau (30)], and fresh groundwater may be mixed with saline waters before inflowing to the sea, the 228Ra approach followed in this study is likely capturing most of the fresh groundwater inputs. In any case, such underestimation would not compromise the range of total SGD calculated here. Normalizing the total SGD to the Mediterranean shore length (64,000 km) (31) results in a flow ranging from 6·106 to 100·106 m3·km−1·y−1. This shore-normalized SGD estimate is in good agreement with local studies conducted along the MS (Fig. 3), reinforcing the consistency of our estimates. The only other SGD estimate for a basin-wide scale is that of Moore et al. (18) for the Atlantic Ocean [(230–470)·106 m3·km−1·y−1]. The shore-normalized SGD to the MS is lower than the SGD flow to the Atlantic Ocean, likely as a consequence of the arid/semiarid conditions on the Eastern and Southern Mediterranean coasts (low groundwater recharge rates) and the lack of significant tidal ranges (low tidally pumped SGD). However, when normalizing to the total ocean-sea surface area, SGD to the MS [(110–1,900) ·103 m3·km−2·y−1] turns out to be comparable or even higher than the flow to the Atlantic Ocean [(200–410)·103 m3·km−2·y−1), highlighting the potential importance of SGD to the MS biogeochemical cycles. SGD-Derived Nutrient Inputs. Considering the estimated magnitude of SGD to the entire basin, SGD-driven inputs of dissolved compounds (e.g., nutrients, metals, carbon) to the MS could be of great importance and need to be evaluated. Given the oligotrophic nature of the MS, external inputs of essential nutrients for marine productivity (e.g., nitrogen, phosphorous, silica, iron) are particularly important for this basin. The flux of bioavailable nutrients [dissolved inorganic nitrogen (DIN), phosphorous (DIP), and silica (DSi)] supplied by SGD can be estimated from the SGD flow and nutrient concentrations reported in SGD studies along the MS coastline (SI Appendix). This results in SGDdriven fluxes of (20–1,500)·109 mol·y−1 DIN, (0.1–4.6)·109 mol·y−1 DIP, and (20–650)·109 mol·y−1 DSi (Fig. 4). These figures could be
5
10
4
3
10
228
Ra
10
Q3 Median Q1
2
10
Fig. 2. 228Ra concentrations in SGD and coastal groundwater throughout the MS. The median and the first and third quartiles of the dataset are also shown. The range between the first and third quartiles is used to estimate the 228Ra concentration in SGD (references in the SI Appendix).
Rodellas et al.
PNAS Early Edition | 3 of 5
ENVIRONMENTAL SCIENCES
Vertical advection of DW can also be a source of 228Ra to the upper MS. A flux of (1.1 ± 0.2)·1015 dpm·y−1 of 228Ra from DW is estimated from the 228Ra concentration box average and the water transfer rates between layers (20, 21) (SI Appendix). Water inflows through the straits of Gibraltar and Dardanelles are 26·1012 m3·y−1 (22) and 1.3·1012 m3·y−1 (23), respectively. Using 228Ra concentrations measured in the inflowing waters, the total 228Ra input through both straits amounts to (0.51 ± 0.03)·1015 dpm·y−1. Fluxes from the sources and sinks of 228Ra are summarized in Table 1. Total outputs and inputs are largely dominated by the radioactive decay and the sediment flux from the continental shelf and slope, respectively. The results show a difference of (1.8 ± 1.2)·1015 dpm·y−1 between the total loss of 228Ra [(6.6 ± 0.7)·1015 dpm·y−1] and the inputs from evaluated sources [(4.7 ± 1.0)·1015 dpm·y−1], which must be balanced by a continual input of 2 2 8 Ra from the remaining source; that is, SGD.
9
10
8
10
7
SGD
10
6
10
5
10
Fig. 3. Estimates of SGD normalized to shore length in different studies along the MS. The gray band and the dashed line represent the range and median of SGD flow estimated for the entire MS, respectively (references in the SI Appendix).
overestimated because some of the recirculated seawater may be nutrient-poor, whereas most of the data used were measured at coastal sites containing a fraction of fresh groundwater. Nevertheless, the enrichment in nutrients of recirculated seawater may be relevant for long-scale recirculation processes (32, 33), which are the only recirculation processes captured using the 228Ra approach (because of the long regeneration time of 228Ra). A conservative estimate of nutrient inputs from SGD can be obtained from the fresh SGD calculated by Zektser et al. (15) and the concentration of nutrients measured in fresh SGD (SI Appendix). Nutrient inputs derived from this conservative approach (30·109, 0.02·109, and 10·109 mol·y−1 for DIN, DIP, and DSi, respectively) are at the low end of the estimates from total SGD (Fig. 4). The significance of these nutrient fluxes becomes apparent compared with the nutrient supply from rivers (5) and atmospheric deposition (34, 35) to the entire MS (Fig. 4). Fluxes of nutrients from SGD are comparable to or even higher than those supplied by atmospheric and riverine sources, revealing that SGD is likely a major external source of DIN, DIP, and especially DSi to the MS. Unlike atmospheric inputs (distributed throughout the
DSi
DIP
DIN
12
10
MS) and riverine discharge (restricted to river mouth surroundings), SGD is essentially ubiquitous in coastal areas. Thus, the relative significance of SGD as a source of nutrients may be even higher for most of the Mediterranean coastal areas, where SGD can represent by far the dominant source of dissolved compounds and the major sustainment for coastal ecosystems. SGD-derived nutrient inputs to the MS have high DIN:DIP ratios (80:1–430:1), according to the first and third quartiles of the set of data collected (see SI Appendix), mainly because DIP is rapidly removed from groundwater (10). DIN:DIP ratios from SGD are comparable to those from riverine [40:1–140:1 (5)] and atmospheric [60:1–120:1 (34)] inputs, all of which are well above the average N:P ratio in Mediterranean waters [24:1 in the western basin and 38:1 in the eastern one (6)]. Thus, nutrient inputs from SGD could provide an additional source for the actual P limitation in the MS (36). The influence of SGD on the nutrient cycling in both coastal areas and open waters should be further studied because a fraction of the nutrient inputs driven by SGD might be removed by biological activity in areas proximal to the discharge sites.
)
Nutrient Flux (
11
10 10
10
9
10
8
10
7
10
SGD
RIV
ATM
SGD
RIV
ATM
SGD
RIV
ATM
Fig. 4. Comparison of nutrient inputs to the MS derived from SGD to inputs from rivers (RIV) and atmospheric deposition (ATM). Nutrient fluxes via SGD were calculated by multiplying the estimated SGD flow range by the range of nutrient concentrations in SGD obtained from different studies conducted in the MS. Dashed lines represent nutrient inputs derived exclusively from fresh SGD. Fluxes from rivers and atmosphere were obtained from the literature (5, 34, 35) (SI Appendix).
4 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1419049112
Rodellas et al.
Methods
depth profiles with 228Ra discrete samples (see details in the SI Appendix). For all of the three approaches used, weighted average 228Ra concentrations for SW (0–220 m depth) and LIW (220–600 m) were estimated for all of the subbasins. These average 228Ra concentrations were multiplied by the box volume to estimate the inventories of 228Ra in each box (SI Appendix). Water samples (n = 38) were also collected to determine the concentrations of 228Ra in SGD fluids or coastal groundwater. 228Ra samples were filtered as detailed earlier and analyzed through radium delayed coincidence counter measurements (via 224Ra ingrowth from 228Ra) (41) and gamma spectrometry. All data on 228Ra concentrations in the MS used in this study are available in the SI Appendix, including data obtained from the literature.
Ra measurements were made on more than 80 seawater samples collected throughout the MS during several oceanographic cruises; namely, M84/3 (April 2011, onboard the R/V Meteor), GA04S-MedSeA (May 2013, B/O Ángeles Alvariño), 65PE370-MedBlack (June 2013, R/V Pelagia), FAMOSO I (March 2009, B/O Sarmiento de Gamboa), and EDASMAR (May 2008, B/O Garcia del Cid). Large-volume seawater samples (100–360 L) were filtered through columns loaded with MnO2-impregnated acrylic fiber, which quantitatively adsorbs radium isotopes from seawater (40). The fibers containing 228Ra were subsequently ashed, ground, and transferred to sealed counting vials to determine 228Ra concentrations by gamma spectrometry, using a well-type high-purity Ge detector. Additional data available from the literature (n = 28) were also used to complement the dataset. The 228Ra data include samples collected at surface (SW, n = 80), intermediate (LIW, n = 15), and deep (DW, n = 13) waters. Three different approaches were used to infer the vertical variability of 228Ra with depth for all of the stations: assuming uniform 228Ra concentrations with depth for each layer, linearly interpolating discrete 228Ra samples collected at different depths, and combining salinity
ACKNOWLEDGMENTS. We are grateful to M. Rutgers van der Loeff, E. Garcia-Solsona, W. C. Burnett, W. S. Moore, C. R. Benitez-Nelson, O. Radakovitch, J. M. Gasol, P. van Beek, J. C. Scholten, M. Guida, P. Puig, and M. Rodilla for their comments on an early draft of this manuscript and/or for providing us with samples, unpublished data, and references. We also thank T. Tanhua, M. Rijkenberg, P. Ziveri, M. Latasa, and the captains and crews of the R/V Meteor, R/V Pelagia, B/O Ángeles Alvariño, B/O Sarmiento de Gamboa, and B/O Garcia del Cid. We also thank our colleagues at the Laboratori de Radioactivitat Ambiental (Universitat Autònoma de Barcelona) for their help and assistance during field and laboratory work. The manuscript was improved by comments from two anonymous reviewers. This research was partially funded by the Spanish Government (CGL2011-13341-E; CGL2006-09274/HID), Consejo de Seguridad Nuclear (2686-SRA), and a grant from US Aid–Middle East Regional Cooperation (M29-073). V.R. acknowledges financial support through a PhD fellowship (AP-2008-03044) from the Spanish Government. J.G.-O. and P.M. are grateful for the support of the Generalitat de Catalunya (2014 SGR1356). Support for the research of P.M. was received through the prize ICREA Academia (Institució Catalana de Recerca i Estudis Avançats), funded by the Generalitat de Catalunya, and by a Gledden Visiting Fellowship awarded by the Institute of Advanced Studies at the University of Western Australia.
1. Bethoux J, Morin P, Chaumery C, Connan O (1998) Nutrients in the Mediterranean Sea, mass balance and statistical analysis of concentrations with respect to environmental change. Mar Chem 63:155–169. 2. Markaki Z, et al. (2003) Atmospheric deposition of inorganic phosphorus in the Levantine Basin, eastern Mediterranean: Spatial and temporal variability and its role in seawater productivity. Limnol Oceanogr 48:1557–1568. 3. Bonnet S, Guieu C (2006) Atmospheric forcing on the annual iron cycle in the western Mediterranean Sea: A 1-year survey. J Geophys Res Atmospheres 111:C09010. 4. Herut B, Krom MD, Pan G, Mortimer R (1999) Atmospheric input of nitrogen and phosphorus to the Southeast Mediterranean: Sources, fluxes, and possible impact. Limnol Oceanogr 44:1683–1692. 5. Ludwig W, Dumont E, Meybeck M, Heussner S (2009) River discharges of water and nutrients to the Mediterranean and Black Sea: Major drivers for ecosystem changes during past and future decades? Prog Oceanogr 80:199–217. 6. Pujo-Pay M, et al. (2011) Integrated survey of elemental stoichiometry (C, N, P) from the western to eastern Mediterranean Sea. Biogeosciences 8:883–899. 7. Krom MD, Herut B, Mantoura RFC (2004) Nutrient budget for the Eastern Mediterranean: Implications for phosphorus limitation. Limnol Oceanogr 49:1582–1592. 8. Guerzoni S, et al. (1999) The role of atmospheric deposition in the biogeochemistry of the Mediterranean Sea. Prog Oceanogr 44:147–190. 9. Cai W-J, Wang Y, Krest J, Moore WS (2003) The geochemistry of dissolved inorganic carbon in a surficial groundwater aquifer in North Inlet, South Carolina, and the carbon fluxes to the coastal ocean. Geochim Cosmochim Acta 67:631–639. 10. Slomp CP, Van Cappellen P (2004) Nutrient inputs to the coastal ocean through submarine groundwater discharge: Controls and potential impact. J Hydrol (Amst) 295:64–86. 11. Windom HL, Moore WS, Niencheski LFH, Jahnke RA (2006) Submarine groundwater discharge: A large, previously unrecognized source of dissolved iron to the South Atlantic Ocean. Mar Chem 102:252–266. 12. Burnett WC, Bokuniewicz H, Huettel M, Moore WS, Taniguchi M (2003) Groundwater and pore water inputs to the coastal zone. Biogeochemistry 66:3–33. 13. Taniguchi M, Burnett WC, Cable JE, Turner JV (2002) Investigation of submarine groundwater discharge. Hydrol Processes 16:2115–2129. 14. Moore WS (1999) The subterranean estuary: A reaction zone of ground water and sea water. Mar Chem 65:111–125. 15. Zektser IS, Everett LG, Dzhamalov RG (2007) Submarine Groundwater (CRC Press, Boca Raton, FL). 16. PNUE/PAM/PLAN BLEU (2004) L’eau des Méditerranéens: Situation et perspectives. No. 158 de la Série des rapports techniques du PAM (PNUE/PAM, Athens). 17. Burnett WC, et al. (2006) Quantifying submarine groundwater discharge in the coastal zone via multiple methods. Sci Total Environ 367(2-3):498–543. 18. Moore WS, Sarmiento JL, Key RM (2008) Submarine groundwater discharge revealed by 228Ra distribution in the upper Atlantic Ocean. Nat Geosci 1:309–311. 19. Lacombe H, Richez C (1982) Hydrodinamics of semi-enclosed seas, ed Nihoul JCJ (Elsevier, Amsterdam), pp 13–74. 20. Sanchez-Cabeza JA, Ortega M, Merino J, Masqué P (2002) Long-term box modelling of 137Cs in the Mediterranean Sea. J Mar Syst 33-34:457–472. 21. Bethoux JP, Gentili B (1996) The Mediterranean Sea, coastal and deep-sea signatures of climatic and environmental changes. J Mar Syst 7:383–394.
22. Soto-Navarro J, Criado-Aldeanueva F, García-Lafuente J, Sánchez-Román A (2010) Estimation of the Atlantic inflow through the Strait of Gibraltar from climatological and in situ data. J Geophys Res 115:C10023. 23. Tugrul S, Besiktepe T, Salihoglu I (2002) Nutrient exchange fluxes between the Aegean and Black Seas through the Marmara Sea. Mediterr Mar Sci 3:33–42. 24. Moore WS, Shaw TJ (2008) Fluxes and behavior of radium isotopes, barium, and uranium in seven Southeastern US rivers and estuaries. Mar Chem 108:236–254. 25. UNEP/MAP/MED POL (2003) Riverine transport of water, sediments and pollutants to the Mediterranean Sea (UNEP/MAP, Athens). 26. Intergovernmental Oceanographic Commission (1981) International Bathymetric Chart of the Mediterranean (Head Department of Navigation and Oceanography, Leningrad, Russia). 27. Emelyanov EM, Shimkus KM, Kuprin PN (1996) Unconsolidated Bottom Surface Sediments of the Mediterranean and Black Seas. IBCM Geol-Geoph Series [Intergovernmental Oceanographic Commission (UNESCO), St. Petersburg, Russia]. 28. Schubert M, et al. (2014) Submarine Groundwater Discharge at a Single Spot Location: Evaluation of Different Detection Approaches. Water 6:584–601. 29. Garcia-Solsona E, et al. (2010) Groundwater and nutrient discharge through karstic coastal springs (Castelló, Spain). Biogeosciences 7:2625–2638. 30. Tovar-Sánchez A, et al. (2014) Contribution of groundwater discharge to the coastal dissolved nutrients and trace metal concentrations in Majorca Island: Karstic vs detrital systems. Environ Sci Technol 48(20):11819–11827. 31. Stewart I, Morhange C (2009) The physical geography of the Mediterranean, ed Woodward J (Oxford University Press Inc., New York), pp 385–413. 32. Santos IR, Eyre BD, Huettel M (2012) The driving forces of porewater and groundwater flow in permeable coastal sediments: A review. Estuar Coast Shelf Sci 98:1–15. 33. Weinstein Y, et al. (2011) What is the role of fresh groundwater and recirculated seawater in conveying nutrients to the coastal ocean? Environ Sci Technol 45(12):5195–5200. 34. Markaki Z, Loÿe-Pilot MD, Violaki K, Benyahya L, Mihalopoulos N (2010) Variability of atmospheric deposition of dissolved nitrogen and phosphorus in the Mediterranean and possible link to the anomalous seawater N/P ratio. Mar Chem 120:187–194. 35. Koçak M, Kubilay N, Tugrul S, Mihalopoulos N (2010) Atmospheric nutrient inputs to the northern levantine basin from a long-term observation: Sources and comparison with riverine inputs. Biogeosciences 7:4037–4050. 36. Krom MD, Emeis K-C, Van Cappellen P (2010) Why is the Eastern Mediterranean phosphorus limited? Prog Oceanogr 85:236–244. 37. Rodellas V, et al. (2014) Submarine groundwater discharge as a source of nutrients and trace metals in a Mediterranean Bay (Palma Beach, Balearic Islands). Mar Chem 160:56–66. 38. Bone SE, Charette MA, Lamborg CH, Gonneea ME (2007) Has submarine groundwater discharge been overlooked as a source of mercury to coastal waters? Environ Sci Technol 41(9):3090–3095. 39. Kim I, Kim G (2014) Submarine groundwater discharge as a main source of rare earth elements in coastal waters. Mar Chem 160:11–17. 40. Moore WS, Reid DF (1973) Extraction of radium from natural waters using manganese-impregnated acrylic fibers. J Geophys Res 78:8880–8886. 41. Moore WS, Arnold R (1996) Measurement of223Ra and224Ra in coastal waters using a delayed coincidence counter. J Geophys Res, C, Oceans 101:1321–1329.
228
Rodellas et al.
PNAS Early Edition | 5 of 5
ENVIRONMENTAL SCIENCES
This study demonstrates the importance of SGD to the coastal biogeochemical cycles of the MS and emphasizes the need for a better characterization of the SGD-related inputs of dissolved compounds and its inclusion in coastal and basin-wide material balances. Considering the calculated magnitude of the SGD flow, the SGD-driven supply of iron could also rival riverine and atmospheric iron inputs to the MS, as previously suggested for a Western Mediterranean bay (37). SGD could also represent a relevant source of other compounds to the MS, such as carbon (9), toxic metals [e.g., Hg (38)], or rare earth elements (39).
Edited by Edward A. Boyle, Massachusetts Institute of Technology, Cambridge, MA, and approved February 18, 2015 (received for review October 4, 2014)
Mediterranean Sea radium
| submarine groundwater discharge | nutrients |
T
he Mediterranean Sea (MS) is one of the most oligotrophic seas in the world (1), resulting from an antiestuarine circulation by which Mediterranean intermediate waters export nutrients to the Atlantic Ocean. In such an oligotrophic system, external inputs of nutrients play a significant role in sustaining marine productivity (2–6). Among these inputs, atmospheric deposition and riverine runoff have been traditionally considered the major sources of nutrients to the MS (5, 7, 8). Even though submarine groundwater discharge (SGD) has been recognized as an important source of dissolved compounds (e.g., nutrients, metals, carbon) to the ocean (9–11), its role as a conveyor of dissolved chemicals to the MS has been largely overlooked. SGD into the coastal ocean is composed of fresh meteoric groundwater, which may enter the ocean directly, and former seawater recirculating through permeable sediments (12, 13). Both types of SGD are important in oceanic chemical budgets, as SGD may become enriched in various solutes (i.e., nutrients, metals, carbon) during its pathway through the coastal aquifer (14). Previous studies have estimated the fresh groundwater discharge to the MS, using hydrologic approaches (15, 16), but none of them has evaluated the total SGD flow and the associated total contribution of dissolved compounds to the MS. Because recirculated SGD usually dominates the water flow (12), its quantification is key to fully understanding the role of SGD in the biogeochemical cycles of nutrients and other chemical species in the MS. Radium isotopes have been widely used as tracers of SGD, mainly because they are highly enriched in SGD relative to seawater and because they behave conservatively once released into the sea (17). The approach we followed to estimate the magnitude of SGD into the entire MS is based on a mass balance of 228Ra (18). Assuming steady state, 228Ra outputs must be www.pnas.org/cgi/doi/10.1073/pnas.1419049112
equal to its inputs, most of which originate from continental margins. By evaluating all of the sink and source terms, the 228Ra supplied by SGD can be determined as the difference between its inputs and outputs, and this flux can be converted to an SGD flow by characterizing the 228Ra concentration in the SGD fluids. We used 228Ra over other Ra isotopes (i.e., 223Ra, 224Ra, and 226Ra) because its half-life (T1/2= 5.75 y) is considerably lower than the residence time of the MS waters [∼100 y (19)]; hence, its radioactive decay is the primary sink of 228Ra in the MS and allows us to accurately constrain the 228Ra removal terms (18). Results and Discussion Estimating all the terms of the 228Ra mass balance requires the characterization of the distribution of 228Ra in the water column of the MS (Fig. 1; see Methods). The highest 228Ra concentrations are observed in coastal areas, consistent with radium inputs from continental margins. 228Ra concentrations measured along the MS mainly reflect the antiestuarine thermohaline circulation of the basin: Surface 228Ra concentrations generally increase from west to east (from 17 to 34 dpm·m−3), implying that waters moving eastward receive 228Ra from the continental margins at a greater rate than radioactive decay or other forms of removal, and reflecting the evaporation of water (solute concentration). Because Levantine Intermediate Water (LIW), which is regularly formed in the Eastern Mediterranean, flows westward to the Strait of Gibraltar, and there are no other major sources of 228Ra at intermediate depths, 228Ra concentrations in LIW generally decrease from east to west (from 29 to 12 dpm·m−3). 228Ra concentrations in deep waters (DW; depth > 600 m) are considerably lower (5–13 dpm·m−3; SI Appendix), confirming that the main 228Ra inputs occur at the uppermost layers. 228 228 Ra Mass Balance for the Upper MS. The loss of Ra resulting from radioactive decay can be determined from the total 228Ra
Significance The Mediterranean Sea (MS) is one of the most oligotrophic seas in the world, and external inputs of nutrients are especially relevant to sustaining primary productivity in this basin. Here we evaluate the role of submarine groundwater discharge (SGD) as a source of nutrients to the entire MS, a pathway that has been largely overlooked. This study demonstrates that SGD is a volumetrically important process in the MS, is of a larger magnitude than riverine discharge, and also represents a major source of dissolved inorganic nitrogen, phosphorous, and silica to the MS. Author contributions: V.R., J.G.-O., P.M., and Y.W. designed research; V.R., J.G.-O., P.M., and Y.W. performed research; V.R. and M.F. analyzed data; and V.R., J.G.-O., P.M., M.F., and Y.W. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1
To whom correspondence may be addressed. Email: [email protected] or jordi. [email protected]
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1419049112/-/DCSupplemental.
PNAS Early Edition | 1 of 5
ENVIRONMENTAL SCIENCES
The Mediterranean Sea (MS) is a semienclosed basin that is considered one of the most oligotrophic seas in the world. In such an environment, inputs of allochthonous nutrients and micronutrients play an important role in sustaining primary productivity. Atmospheric deposition and riverine runoff have been traditionally considered the main external sources of nutrients to the MS, whereas the role of submarine groundwater discharge (SGD) has been largely ignored. However, given the large Mediterranean shore length relative to its surface area, SGD may be a major conveyor of dissolved compounds to the MS. Here, we used a 228Ra mass balance to demonstrate that the total SGD contributes up to (0.3–4.8)·1012 m3·y−1 to the MS, which appears to be equal or larger by a factor of 16 to the riverine discharge. SGD is also a major source of dissolved inorganic nutrients to the MS, with median annual fluxes of 190·109, 0.7·109, and 110·109 mol for nitrogen, phosphorous, and silica, respectively, which are comparable to riverine and atmospheric inputs. This corroborates the profound implications that SGD may have for the biogeochemical cycles of the MS. Inputs of other dissolved compounds (e.g., iron, carbon) via SGD could also be significant and should be investigated.
43 - 48 32 - 34 26 - 34 15 - 25
16 - 20 11 - 12 15
104
22 - 24 20 - 21
34 - 36 28 - 34
19 - 23 15 - 16 26 - 28 19 - 23
29 22 - 29 28 19 - 27
SW LIW
Fig. 1. Distribution of 228Ra concentrations (dpm·m−3) in surface waters of the MS. Dots indicate sampled stations. The subbasin division and the weighted average 228Ra concentration in each subbasin for surface (SW) and intermediate (LIW) waters are indicated (a range is reported to include the concentrations derived from the different approaches) (see SI Appendix for details).
inventory multiplied by its decay constant (λ = 0.120 y−1). To calculate the total 228Ra inventory, the MS was divided into nine subbasins (Fig. 1), defined according to their physiography and general circulation patterns (20, 21). Each subbasin was split into three water layers [surface water (SW), LIW, and DW]. The total 228 Ra inventory for the upper water column of the MS (including SW and LIW; i.e., from 0 to 600 m depth) was obtained from the integration of the box inventories, which were calculated using three different approaches (uniform 228Ra concentrations with depth, vertical linear interpolation between measured 228Ra concentrations, and inferring 228Ra concentrations based on salinity profiles; see Methods and the SI Appendix). The three approaches used resulted in similar total 228Ra inventories for the upper MS [(37 ± 4)·1015 dpm, (34 ± 4)·1015 dpm, and (33 ± 4)·1015 dpm], allowing us to constrain the total 228Ra inventory in (35 ± 6)·1015 dpm. The loss of 228Ra by radioactive decay is thus of (4.2 ± 0.7)·1015 dpm·y−1. The DW layer was excluded from the total inventory because the main 228Ra inputs to the water column occur at the uppermost layers (18). In doing so, 228Ra released from deep sediments can be neglected, but vertical exchange fluxes need to be considered. The migration of 228Ra from upper waters to DW was estimated from the average 228Ra box concentrations and from water transfer rates obtained in previous studies (20, 21). This results in a total 228Ra loss of (2.0 ± 0.1)·1015 dpm·y−1 as a result of downward water flow (SI Appendix). In addition, 228Ra could also be transported downward because of particle scavenging. Assuming a radium residence time with respect to scavenging of ∼500 y (18), the 228Ra removal resulting from particle scavenging would only be of (0.07 ± 0.01)·1015 dpm·y−1. 228 Ra is also lost from the MS because of the water outflow to the Atlantic Ocean (25·10 12 m3·y −1) (22) and the Black Sea (1.0·1012 m3·y −1) (23), through the straits of Gibraltar and Dardanelles, respectively. From the characterization of the 228Ra concentrations in outflowing waters, a total loss of 228Ra of (0.36 ± 0.03)·1015 dpm·y−1 through the straits is obtained. The total loss of 228Ra from the upper MS [(6.6 ± 0.7)·1015 dpm·y−1; Table 1] must be balanced by continuous new inputs of 228Ra that can consist of riverine discharge, atmospheric input, release from sediments, deep water vertical advection, water input through the straits, and SGD. The input of 228Ra from river discharge includes Ra in the dissolved fraction and desorption of Ra from river-borne particles (24). A concentration of 130 ± 50 dpm·m−3 captures all the dissolved 228Ra concentrations measured in rivers discharging into the MS (SI Appendix). A desorbed 228Ra flux of 0.5 ± 0.4 dpm per 2 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1419049112
gram of sediment includes most of the 228Ra desorption estimates reported throughout the world (SI Appendix). Considering the river flow (0.31·1012 m3·y−1) (5) and the riverine particle flux into the MS (730·1012 g·y−1) (25), the total supply of 228Ra from river discharge would be (0.40 ± 0.29)·1015 dpm·y−1. The atmospheric dust input to the entire MS is estimated at 41·1012 g·y−1 (8). Assuming a desorption of 228Ra from dust of 2 dpm·g−1, which is close to the maximum reported in the literature (18, 24), the atmospheric input of 228Ra to the entire basin is (0.08 ± 0.03)·1015 dpm·y−1. The flux of 228Ra released from shelf sediments as a result of diffusion and bioturbation might be highly variable and, in large-scale studies, can represent a major source of 228Ra resulting from the large extension of the continental shelf (18). Moore et al. (18) conducted a detailed review of the literature to determine the 228Ra fluxes from fine-grained shelf sediments (11·103 ± 5·103 dpm·m−2·y−1), coarse-grained shelf sediments (230 ± 110 dpm·m−2·y−1), and slope sediments (2.3·103 ± 1.1·103 dpm·m−2·y−1). Using the areas of continental shelf and slope of the MS above 600 m [490·109 and 310·109 m2, respectively (26)], and considering that fine-grained sediments represent about 35% of the shelf area (27), the total flux of 228 Ra from sediments would account for (2.7 ± 0.9)·1015 dpm·y−1. Table 1. Summary of the
228
Ra mass balance for the upper MS 228
Ra flux
(·1015) dpm·y−1
Percentage total flux
4.2 1.95 0.07 0.30 0.06 6.6
± ± ± ± ± ±
0.7 0.13 0.01 0.03 0.01 0.7
51–77 26–34 ∼1 4–5 ∼1 84–116
0.40 0.08 2.7 1.1 0.37 0.14 4.7 1.8
± ± ± ± ± ± ± ±
0.29 0.03 0.9 0.2 0.03 0.01 1.0 1.2
2–11 1–2 26–56 13–19 5–6 ∼2 55–89 9–47
228
Ra outputs Decay Advection to DW Particle Scavenging Outflow (Gibraltar) Outflow (Dardanelles) Total outputs 228 Ra inputs Rivers Atmospheric dust Sediments Advection from DW Inflow (Gibraltar) Inflow (Dardanelles) Total inputs (without SGD) 228 Ra from SGD
Rodellas et al.
SGD Flow to the MS. The SGD-derived
228
Ra flux can be converted to a SGD flow by characterizing the 228Ra concentration in the SGD fluids inflowing to the MS. Several studies have reported 228Ra data in SGD along the MS (45 different locations, mainly located on the Northwestern MS), with concentrations spanning a wide range (130–72,000 dpm·m−3; Fig. 2). The highest 228 Ra concentrations are found mainly in SGD dominated by recirculated seawater, as radium mobilization from the solid matrix strongly depends on water salinity. SGD at karstic areas usually shows lower activities than that at granular coastal aquifers. This is for several reasons, including the relatively low 232Th (228Ra parent) levels in carbonates and the typical point-sourced discharge at karstic coastal terrains, which results in a low water–rock interaction in the coastal aquifer and allows the direct discharge of fresh groundwater with a low 228Ra content. A 228Ra concentration of 640–2,200 dpm·m−3 covers the range between the first and third quartiles of the groundwater dataset, accounting for the variability of 228 Ra in SGD and excluding extreme values. Using this range, the resulting SGD flow to the MS ranges from 0.3·1012 to 4.8·1012 m3·y−1 (median, 1.4·1012 m3·y−1). The SGD flow estimated here appears to be equal or larger by a factor of 16 to the riverine discharge to the MS (0.3·1012 m3·y−1) (5). A recent study based on an integrated hydrologic–hydrogeologic approach estimated the fresh SGD flow to the MS to be on the order of 0.07·1012 m3·y−1 (15). Therefore, fresh groundwater inputs
would represent 1–25% of the total SGD inputs to the MS, revealing that seawater recirculating through coastal sediments constitutes the main fraction of SGD inflowing to the MS. As detailed earlier, 228Ra may be low in fresh groundwater inflowing to the MS (28); therefore, using 228Ra as a tracer could underestimate fresh (and consequently total) SGD inputs. However, given that significant 228Ra enrichments (higher than 1,000 dpm·m−3) may be also found in entirely fresh groundwater springs [e.g., Badum (29), Sa Nau (30)], and fresh groundwater may be mixed with saline waters before inflowing to the sea, the 228Ra approach followed in this study is likely capturing most of the fresh groundwater inputs. In any case, such underestimation would not compromise the range of total SGD calculated here. Normalizing the total SGD to the Mediterranean shore length (64,000 km) (31) results in a flow ranging from 6·106 to 100·106 m3·km−1·y−1. This shore-normalized SGD estimate is in good agreement with local studies conducted along the MS (Fig. 3), reinforcing the consistency of our estimates. The only other SGD estimate for a basin-wide scale is that of Moore et al. (18) for the Atlantic Ocean [(230–470)·106 m3·km−1·y−1]. The shore-normalized SGD to the MS is lower than the SGD flow to the Atlantic Ocean, likely as a consequence of the arid/semiarid conditions on the Eastern and Southern Mediterranean coasts (low groundwater recharge rates) and the lack of significant tidal ranges (low tidally pumped SGD). However, when normalizing to the total ocean-sea surface area, SGD to the MS [(110–1,900) ·103 m3·km−2·y−1] turns out to be comparable or even higher than the flow to the Atlantic Ocean [(200–410)·103 m3·km−2·y−1), highlighting the potential importance of SGD to the MS biogeochemical cycles. SGD-Derived Nutrient Inputs. Considering the estimated magnitude of SGD to the entire basin, SGD-driven inputs of dissolved compounds (e.g., nutrients, metals, carbon) to the MS could be of great importance and need to be evaluated. Given the oligotrophic nature of the MS, external inputs of essential nutrients for marine productivity (e.g., nitrogen, phosphorous, silica, iron) are particularly important for this basin. The flux of bioavailable nutrients [dissolved inorganic nitrogen (DIN), phosphorous (DIP), and silica (DSi)] supplied by SGD can be estimated from the SGD flow and nutrient concentrations reported in SGD studies along the MS coastline (SI Appendix). This results in SGDdriven fluxes of (20–1,500)·109 mol·y−1 DIN, (0.1–4.6)·109 mol·y−1 DIP, and (20–650)·109 mol·y−1 DSi (Fig. 4). These figures could be
5
10
4
3
10
228
Ra
10
Q3 Median Q1
2
10
Fig. 2. 228Ra concentrations in SGD and coastal groundwater throughout the MS. The median and the first and third quartiles of the dataset are also shown. The range between the first and third quartiles is used to estimate the 228Ra concentration in SGD (references in the SI Appendix).
Rodellas et al.
PNAS Early Edition | 3 of 5
ENVIRONMENTAL SCIENCES
Vertical advection of DW can also be a source of 228Ra to the upper MS. A flux of (1.1 ± 0.2)·1015 dpm·y−1 of 228Ra from DW is estimated from the 228Ra concentration box average and the water transfer rates between layers (20, 21) (SI Appendix). Water inflows through the straits of Gibraltar and Dardanelles are 26·1012 m3·y−1 (22) and 1.3·1012 m3·y−1 (23), respectively. Using 228Ra concentrations measured in the inflowing waters, the total 228Ra input through both straits amounts to (0.51 ± 0.03)·1015 dpm·y−1. Fluxes from the sources and sinks of 228Ra are summarized in Table 1. Total outputs and inputs are largely dominated by the radioactive decay and the sediment flux from the continental shelf and slope, respectively. The results show a difference of (1.8 ± 1.2)·1015 dpm·y−1 between the total loss of 228Ra [(6.6 ± 0.7)·1015 dpm·y−1] and the inputs from evaluated sources [(4.7 ± 1.0)·1015 dpm·y−1], which must be balanced by a continual input of 2 2 8 Ra from the remaining source; that is, SGD.
9
10
8
10
7
SGD
10
6
10
5
10
Fig. 3. Estimates of SGD normalized to shore length in different studies along the MS. The gray band and the dashed line represent the range and median of SGD flow estimated for the entire MS, respectively (references in the SI Appendix).
overestimated because some of the recirculated seawater may be nutrient-poor, whereas most of the data used were measured at coastal sites containing a fraction of fresh groundwater. Nevertheless, the enrichment in nutrients of recirculated seawater may be relevant for long-scale recirculation processes (32, 33), which are the only recirculation processes captured using the 228Ra approach (because of the long regeneration time of 228Ra). A conservative estimate of nutrient inputs from SGD can be obtained from the fresh SGD calculated by Zektser et al. (15) and the concentration of nutrients measured in fresh SGD (SI Appendix). Nutrient inputs derived from this conservative approach (30·109, 0.02·109, and 10·109 mol·y−1 for DIN, DIP, and DSi, respectively) are at the low end of the estimates from total SGD (Fig. 4). The significance of these nutrient fluxes becomes apparent compared with the nutrient supply from rivers (5) and atmospheric deposition (34, 35) to the entire MS (Fig. 4). Fluxes of nutrients from SGD are comparable to or even higher than those supplied by atmospheric and riverine sources, revealing that SGD is likely a major external source of DIN, DIP, and especially DSi to the MS. Unlike atmospheric inputs (distributed throughout the
DSi
DIP
DIN
12
10
MS) and riverine discharge (restricted to river mouth surroundings), SGD is essentially ubiquitous in coastal areas. Thus, the relative significance of SGD as a source of nutrients may be even higher for most of the Mediterranean coastal areas, where SGD can represent by far the dominant source of dissolved compounds and the major sustainment for coastal ecosystems. SGD-derived nutrient inputs to the MS have high DIN:DIP ratios (80:1–430:1), according to the first and third quartiles of the set of data collected (see SI Appendix), mainly because DIP is rapidly removed from groundwater (10). DIN:DIP ratios from SGD are comparable to those from riverine [40:1–140:1 (5)] and atmospheric [60:1–120:1 (34)] inputs, all of which are well above the average N:P ratio in Mediterranean waters [24:1 in the western basin and 38:1 in the eastern one (6)]. Thus, nutrient inputs from SGD could provide an additional source for the actual P limitation in the MS (36). The influence of SGD on the nutrient cycling in both coastal areas and open waters should be further studied because a fraction of the nutrient inputs driven by SGD might be removed by biological activity in areas proximal to the discharge sites.
)
Nutrient Flux (
11
10 10
10
9
10
8
10
7
10
SGD
RIV
ATM
SGD
RIV
ATM
SGD
RIV
ATM
Fig. 4. Comparison of nutrient inputs to the MS derived from SGD to inputs from rivers (RIV) and atmospheric deposition (ATM). Nutrient fluxes via SGD were calculated by multiplying the estimated SGD flow range by the range of nutrient concentrations in SGD obtained from different studies conducted in the MS. Dashed lines represent nutrient inputs derived exclusively from fresh SGD. Fluxes from rivers and atmosphere were obtained from the literature (5, 34, 35) (SI Appendix).
4 of 5 | www.pnas.org/cgi/doi/10.1073/pnas.1419049112
Rodellas et al.
Methods
depth profiles with 228Ra discrete samples (see details in the SI Appendix). For all of the three approaches used, weighted average 228Ra concentrations for SW (0–220 m depth) and LIW (220–600 m) were estimated for all of the subbasins. These average 228Ra concentrations were multiplied by the box volume to estimate the inventories of 228Ra in each box (SI Appendix). Water samples (n = 38) were also collected to determine the concentrations of 228Ra in SGD fluids or coastal groundwater. 228Ra samples were filtered as detailed earlier and analyzed through radium delayed coincidence counter measurements (via 224Ra ingrowth from 228Ra) (41) and gamma spectrometry. All data on 228Ra concentrations in the MS used in this study are available in the SI Appendix, including data obtained from the literature.
Ra measurements were made on more than 80 seawater samples collected throughout the MS during several oceanographic cruises; namely, M84/3 (April 2011, onboard the R/V Meteor), GA04S-MedSeA (May 2013, B/O Ángeles Alvariño), 65PE370-MedBlack (June 2013, R/V Pelagia), FAMOSO I (March 2009, B/O Sarmiento de Gamboa), and EDASMAR (May 2008, B/O Garcia del Cid). Large-volume seawater samples (100–360 L) were filtered through columns loaded with MnO2-impregnated acrylic fiber, which quantitatively adsorbs radium isotopes from seawater (40). The fibers containing 228Ra were subsequently ashed, ground, and transferred to sealed counting vials to determine 228Ra concentrations by gamma spectrometry, using a well-type high-purity Ge detector. Additional data available from the literature (n = 28) were also used to complement the dataset. The 228Ra data include samples collected at surface (SW, n = 80), intermediate (LIW, n = 15), and deep (DW, n = 13) waters. Three different approaches were used to infer the vertical variability of 228Ra with depth for all of the stations: assuming uniform 228Ra concentrations with depth for each layer, linearly interpolating discrete 228Ra samples collected at different depths, and combining salinity
ACKNOWLEDGMENTS. We are grateful to M. Rutgers van der Loeff, E. Garcia-Solsona, W. C. Burnett, W. S. Moore, C. R. Benitez-Nelson, O. Radakovitch, J. M. Gasol, P. van Beek, J. C. Scholten, M. Guida, P. Puig, and M. Rodilla for their comments on an early draft of this manuscript and/or for providing us with samples, unpublished data, and references. We also thank T. Tanhua, M. Rijkenberg, P. Ziveri, M. Latasa, and the captains and crews of the R/V Meteor, R/V Pelagia, B/O Ángeles Alvariño, B/O Sarmiento de Gamboa, and B/O Garcia del Cid. We also thank our colleagues at the Laboratori de Radioactivitat Ambiental (Universitat Autònoma de Barcelona) for their help and assistance during field and laboratory work. The manuscript was improved by comments from two anonymous reviewers. This research was partially funded by the Spanish Government (CGL2011-13341-E; CGL2006-09274/HID), Consejo de Seguridad Nuclear (2686-SRA), and a grant from US Aid–Middle East Regional Cooperation (M29-073). V.R. acknowledges financial support through a PhD fellowship (AP-2008-03044) from the Spanish Government. J.G.-O. and P.M. are grateful for the support of the Generalitat de Catalunya (2014 SGR1356). Support for the research of P.M. was received through the prize ICREA Academia (Institució Catalana de Recerca i Estudis Avançats), funded by the Generalitat de Catalunya, and by a Gledden Visiting Fellowship awarded by the Institute of Advanced Studies at the University of Western Australia.
1. Bethoux J, Morin P, Chaumery C, Connan O (1998) Nutrients in the Mediterranean Sea, mass balance and statistical analysis of concentrations with respect to environmental change. Mar Chem 63:155–169. 2. Markaki Z, et al. (2003) Atmospheric deposition of inorganic phosphorus in the Levantine Basin, eastern Mediterranean: Spatial and temporal variability and its role in seawater productivity. Limnol Oceanogr 48:1557–1568. 3. Bonnet S, Guieu C (2006) Atmospheric forcing on the annual iron cycle in the western Mediterranean Sea: A 1-year survey. J Geophys Res Atmospheres 111:C09010. 4. Herut B, Krom MD, Pan G, Mortimer R (1999) Atmospheric input of nitrogen and phosphorus to the Southeast Mediterranean: Sources, fluxes, and possible impact. Limnol Oceanogr 44:1683–1692. 5. Ludwig W, Dumont E, Meybeck M, Heussner S (2009) River discharges of water and nutrients to the Mediterranean and Black Sea: Major drivers for ecosystem changes during past and future decades? Prog Oceanogr 80:199–217. 6. Pujo-Pay M, et al. (2011) Integrated survey of elemental stoichiometry (C, N, P) from the western to eastern Mediterranean Sea. Biogeosciences 8:883–899. 7. Krom MD, Herut B, Mantoura RFC (2004) Nutrient budget for the Eastern Mediterranean: Implications for phosphorus limitation. Limnol Oceanogr 49:1582–1592. 8. Guerzoni S, et al. (1999) The role of atmospheric deposition in the biogeochemistry of the Mediterranean Sea. Prog Oceanogr 44:147–190. 9. Cai W-J, Wang Y, Krest J, Moore WS (2003) The geochemistry of dissolved inorganic carbon in a surficial groundwater aquifer in North Inlet, South Carolina, and the carbon fluxes to the coastal ocean. Geochim Cosmochim Acta 67:631–639. 10. Slomp CP, Van Cappellen P (2004) Nutrient inputs to the coastal ocean through submarine groundwater discharge: Controls and potential impact. J Hydrol (Amst) 295:64–86. 11. Windom HL, Moore WS, Niencheski LFH, Jahnke RA (2006) Submarine groundwater discharge: A large, previously unrecognized source of dissolved iron to the South Atlantic Ocean. Mar Chem 102:252–266. 12. Burnett WC, Bokuniewicz H, Huettel M, Moore WS, Taniguchi M (2003) Groundwater and pore water inputs to the coastal zone. Biogeochemistry 66:3–33. 13. Taniguchi M, Burnett WC, Cable JE, Turner JV (2002) Investigation of submarine groundwater discharge. Hydrol Processes 16:2115–2129. 14. Moore WS (1999) The subterranean estuary: A reaction zone of ground water and sea water. Mar Chem 65:111–125. 15. Zektser IS, Everett LG, Dzhamalov RG (2007) Submarine Groundwater (CRC Press, Boca Raton, FL). 16. PNUE/PAM/PLAN BLEU (2004) L’eau des Méditerranéens: Situation et perspectives. No. 158 de la Série des rapports techniques du PAM (PNUE/PAM, Athens). 17. Burnett WC, et al. (2006) Quantifying submarine groundwater discharge in the coastal zone via multiple methods. Sci Total Environ 367(2-3):498–543. 18. Moore WS, Sarmiento JL, Key RM (2008) Submarine groundwater discharge revealed by 228Ra distribution in the upper Atlantic Ocean. Nat Geosci 1:309–311. 19. Lacombe H, Richez C (1982) Hydrodinamics of semi-enclosed seas, ed Nihoul JCJ (Elsevier, Amsterdam), pp 13–74. 20. Sanchez-Cabeza JA, Ortega M, Merino J, Masqué P (2002) Long-term box modelling of 137Cs in the Mediterranean Sea. J Mar Syst 33-34:457–472. 21. Bethoux JP, Gentili B (1996) The Mediterranean Sea, coastal and deep-sea signatures of climatic and environmental changes. J Mar Syst 7:383–394.
22. Soto-Navarro J, Criado-Aldeanueva F, García-Lafuente J, Sánchez-Román A (2010) Estimation of the Atlantic inflow through the Strait of Gibraltar from climatological and in situ data. J Geophys Res 115:C10023. 23. Tugrul S, Besiktepe T, Salihoglu I (2002) Nutrient exchange fluxes between the Aegean and Black Seas through the Marmara Sea. Mediterr Mar Sci 3:33–42. 24. Moore WS, Shaw TJ (2008) Fluxes and behavior of radium isotopes, barium, and uranium in seven Southeastern US rivers and estuaries. Mar Chem 108:236–254. 25. UNEP/MAP/MED POL (2003) Riverine transport of water, sediments and pollutants to the Mediterranean Sea (UNEP/MAP, Athens). 26. Intergovernmental Oceanographic Commission (1981) International Bathymetric Chart of the Mediterranean (Head Department of Navigation and Oceanography, Leningrad, Russia). 27. Emelyanov EM, Shimkus KM, Kuprin PN (1996) Unconsolidated Bottom Surface Sediments of the Mediterranean and Black Seas. IBCM Geol-Geoph Series [Intergovernmental Oceanographic Commission (UNESCO), St. Petersburg, Russia]. 28. Schubert M, et al. (2014) Submarine Groundwater Discharge at a Single Spot Location: Evaluation of Different Detection Approaches. Water 6:584–601. 29. Garcia-Solsona E, et al. (2010) Groundwater and nutrient discharge through karstic coastal springs (Castelló, Spain). Biogeosciences 7:2625–2638. 30. Tovar-Sánchez A, et al. (2014) Contribution of groundwater discharge to the coastal dissolved nutrients and trace metal concentrations in Majorca Island: Karstic vs detrital systems. Environ Sci Technol 48(20):11819–11827. 31. Stewart I, Morhange C (2009) The physical geography of the Mediterranean, ed Woodward J (Oxford University Press Inc., New York), pp 385–413. 32. Santos IR, Eyre BD, Huettel M (2012) The driving forces of porewater and groundwater flow in permeable coastal sediments: A review. Estuar Coast Shelf Sci 98:1–15. 33. Weinstein Y, et al. (2011) What is the role of fresh groundwater and recirculated seawater in conveying nutrients to the coastal ocean? Environ Sci Technol 45(12):5195–5200. 34. Markaki Z, Loÿe-Pilot MD, Violaki K, Benyahya L, Mihalopoulos N (2010) Variability of atmospheric deposition of dissolved nitrogen and phosphorus in the Mediterranean and possible link to the anomalous seawater N/P ratio. Mar Chem 120:187–194. 35. Koçak M, Kubilay N, Tugrul S, Mihalopoulos N (2010) Atmospheric nutrient inputs to the northern levantine basin from a long-term observation: Sources and comparison with riverine inputs. Biogeosciences 7:4037–4050. 36. Krom MD, Emeis K-C, Van Cappellen P (2010) Why is the Eastern Mediterranean phosphorus limited? Prog Oceanogr 85:236–244. 37. Rodellas V, et al. (2014) Submarine groundwater discharge as a source of nutrients and trace metals in a Mediterranean Bay (Palma Beach, Balearic Islands). Mar Chem 160:56–66. 38. Bone SE, Charette MA, Lamborg CH, Gonneea ME (2007) Has submarine groundwater discharge been overlooked as a source of mercury to coastal waters? Environ Sci Technol 41(9):3090–3095. 39. Kim I, Kim G (2014) Submarine groundwater discharge as a main source of rare earth elements in coastal waters. Mar Chem 160:11–17. 40. Moore WS, Reid DF (1973) Extraction of radium from natural waters using manganese-impregnated acrylic fibers. J Geophys Res 78:8880–8886. 41. Moore WS, Arnold R (1996) Measurement of223Ra and224Ra in coastal waters using a delayed coincidence counter. J Geophys Res, C, Oceans 101:1321–1329.
228
Rodellas et al.
PNAS Early Edition | 5 of 5
ENVIRONMENTAL SCIENCES
This study demonstrates the importance of SGD to the coastal biogeochemical cycles of the MS and emphasizes the need for a better characterization of the SGD-related inputs of dissolved compounds and its inclusion in coastal and basin-wide material balances. Considering the calculated magnitude of the SGD flow, the SGD-driven supply of iron could also rival riverine and atmospheric iron inputs to the MS, as previously suggested for a Western Mediterranean bay (37). SGD could also represent a relevant source of other compounds to the MS, such as carbon (9), toxic metals [e.g., Hg (38)], or rare earth elements (39).
