Evidence for anthropogenic 14C-enrichment in estuarine waters adjacent to the North Sea



[1] The isotopic composition (δ13C and Δ14C) of high molecular weight dissolved organic carbon (HMW DOC) was studied in the Tyne and Tweed estuaries, NE England. Despite significant removal of terrigenous HMW DOC in the low salinity regions (S < 15), Δ14C remained modern with little variation around 115‰. This lack of apparent age discrimination was attributed to either non-oxidative removal or the absence of a significant proportion of old refractory C in the HMW DOC pool. At S > 15, we observed seaward increases in δ13C and Δ14C. With no documented local 14C inputs, we attributed non-bomb related 14C-enrichment at S > 15 to a possible ‘lingering effect’ of distal anthropogenic sources in near-coastal North Sea HMW DOC. Given the global distribution of potential sources, we propose that anthropogenic 14C should be considered in assigning ages of DOC pools in near-coastal waters and suggest its possible use as a tracer for DOC transformations.

1. Introduction

[2] The 14C content of estuarine and near-coastal dissolved organic carbon (DOC) may be affected by aged riverine DOC, modern DOC which includes 14C from atmospheric nuclear weapons testing (post-AD 1955), and radioactive discharges from biomedical and nuclear fuel cycle related establishments. However, understanding the relative contributions of these inputs to 14C-DOC in near-coastal waters remains challenging. While some rivers are dominated by DOC with apparent radiocarbon ages of 100s to 1000s of years, most rivers discharging into the Atlantic Ocean have freshwater Δ14C–DOC signatures suggestive of modern, decadal-old DOC [Raymond and Bauer, 2001a, 2001b]. This modern Δ14C-DOC signature is believed to contain both labile and refractory components. Preferential microbial utilization of the younger, more labile DOC fraction during estuarine transport may therefore lower the Δ14C of bulk DOC, resulting in the net export of 14C-depleted, degraded terrigenous DOC into the ocean [Raymond and Bauer, 2001a; 2001b].

[3] The Δ14C of atmospheric CO2 has been declining steadily after reaching a maximum in the N Hemisphere of ∼1000‰ at the height of nuclear weapons tests in the mid 1960s [Nydal, 1998]. Consequently, Δ14C signatures of terrestrially fixed C that are significantly above the modern atmospheric Δ14C-CO2 value [∼100‰ in AD 1998; Levin and Kromer, 2004] but <1000‰ can be attributed to a sample age between AD 1955 and the present [Mayorga et al., 2005; Raymond and Bauer, 2001a]. Weapons testing related 14C-enrichment in surface ocean dissolved inorganic carbon (DIC) did not reach atmospheric levels because of atmospheric exchange and mixing with the large reservoir of 14C-depleted deep ocean DIC. As a result, surface ocean Δ14C-DIC reached bomb 14C maxima of only ∼270‰ and have since declined to modern values <100‰ [Nydal, 1998].

[4] In some coastal areas, however, recurring anthropogenic 14C releases may lead to Δ14C values which are significantly higher than the modern background signal. In this study we report elevated Δ14C signatures in high molecular weight (HMW) DOC collected in two North Sea estuaries, the Tyne and Tweed (NE England), with no previously documented sources of local anthropogenic 14C inputs. We discuss the behavior of DOC during estuarine mixing and potential sources and consequences of this 14C-enrichment for coastal C cycling research.

2. Study Sites and Methods

[5] The Tyne and Tweed (Figure 1) are mesotidal estuaries receiving humic-rich water from forested and moorland catchments at mean discharge rates of 48 m3s−1 and 84 m3s−1, respectively. The area around the Tyne estuary is densely populated. In contrast, the Tweed drains a rural, sparsely populated catchment dominated by agricultural activity. Near-surface water from a submersible pump (1–2 m depth) was collected into 25 L carboys on single day excursions to the Tyne (Jul-02, Mar-03, Jul-03) and Tweed (Jul-03, Dec-03) estuaries. Water samples were sequentially filtered through 0.7 μm glass fibre (Whatman, GF/F) and 0.2 μm capsule filters (Whatman, Polycap 75 TF) to remove particulates, algae and most bacteria. HMW DOC was 30-fold concentrated via tangential-flow ultrafiltration (UF) using a cassette filter system with stainless steel impeller pump (Millipore, Pellicon-2) and regenerated cellulose membrane (Millipore, Cat. P2PLACC05). The molecular size cut-off (90% retention) of the membrane was determined at 15.5 kDa. Mass balances indicated DOC recovery of 90.2–107.5%, signifying minimal loss or contamination. The HMW component comprised ∼3–5% and ∼20–50% of the total DOC in coastal North Sea and riverine samples, respectively. Aliquots for HMW DOC analysis were acidified (pH 2), stored frozen in 20 ml glass vials and measured on a Shimadzu TOC-5000 analyzer after shaking thoroughly. The coefficient of variance of replicates was better than ±2%.

Figure 1.

Map of the North Sea showing the location of the Tyne and Tweed estuaries and several known or potential sources of anthropogenic 14C discharge.

[6] After UF, HMW DOC samples were diafiltered with Milli-Q water to remove salts, lyophilized, homogenized, re-wetted, decarbonated with HCl, and dried prior to isotopic analysis. δ13C signatures were determined using a Carlo Erba NA2500 elemental analyzer coupled to a Micromass PRISM III isotope ratio mass spectrometer. Based on replicate sample analyses the 1σ precision for δ13C values was ±0.3‰. Radiocarbon analyses were carried out at the SUERC AMS facility (5MV NEC). Samples were combusted to CO2 at 900°C using CuO for 8 h in sealed quartz tubes or at 600°C using high purity O2 for 4 h. In the case of the latter, the resultant gases were then passed through a Pt/Ag furnace at 900°C. CO2 was cryogenically separated and converted to graphite targets by Fe/Zn reduction [Slota et al., 1987]. Standard material of known ages added to ultrafiltered Milli-Q water and subjected to the sample protocol described above (lyophilization, etc.) showed minimal 14C contamination. Replicate analyses of riverine end-members, salinity (S) ≤0.2, from the Tyne (Mar-03) and Tweed (Dec-03) indicated a 1σ precision for Δ14C values [Stuiver and Polach, 1977] of ±7‰. For further details on the HMW DOC component and a complete list of data reported here please refer to the auxiliary material.

3. Results and Discussion

[7] HMW DOC concentrations varied by more than two orders of magnitude, ranging from 5.2–261.1 μM and 1.3–354.0 μM in the Tyne and Tweed estuaries, respectively (Figure 2). All concentration profiles were concavely shaped at low S (<10) and showed an approximately linear decrease at S > 10, indicating ∼30–70% removal of terrigenous HMW DOC in the upper reaches of both estuaries. This was reflected in the total DOC as losses of up to ∼10%. Estuarine DOC removal was previously attributed to flocculation [Sholkovitz et al., 1978], adsorption onto suspended particulates [Uher et al., 2001], and bacterial and photochemical oxidation [Amon and Benner, 1996]. Regardless of the mechanisms responsible, our data clearly indicate significant, non-conservative reductions in terrigenous HMW DOC, favoring the predominance of marine over terrestrial HMW DOC in near-coastal waters. In contrast, low molecular weight (LMW) DOC showed more or less conservative mixing behavior in these estuaries (auxiliary material). This suggests that the reactivity of terrestrial organic matter (OM) during estuarine mixing decreases along a size continuum from HMW to LMW DOC, as inferred in open ocean settings [Skoog and Benner, 1997; Loh et al., 2004].

Figure 2.

Concentrations [μM], δ13C and Δ14C (‰) of HMW DOC in the Tyne and Tweed estuaries plotted versus salinity (S).

[8] The prevalence of North Sea HMW DOC in the outer Tyne and Tweed estuaries is substantiated by axial profiles of δ13C (Figure 2). All transects showed progressive 13C-enrichment with increasing S, consistent with mixing between peat- and moorland-derived HMW DOC (River Tyne: δ13C = −27.6 to −28.4‰; River Tweed: δ13C = −26.7 to −28.6‰) and heavier marine-derived HMW DOC. The δ13C of marine end-members (−25.7‰, Tyne Mar-03; −24.0‰, Tyne Jul-03; −23.9‰, Tweed Jul-03; −23.3‰, Tweed Dec-03) fell toward the lower end of the range reported for North Sea DOC [−18 to −24‰; Le Clercq et al., 1997].

[9] The Δ14C of riverine HMW DOC (S ≤ 0.2) ranged between 76–121‰ (Figure 2). Our data therefore suggest that UK rivers draining peat and moorland catchments are a source of modern HMW DOC, as reported for various DOC-rich Arctic [Benner et al., 2004] and Amazonian [Mayorga et al., 2005] rivers. Despite substantial HMW DOC removal Δ14C showed little variation (115 ± 29‰, n = 16) at low to mid S (<15). This is in contrast to the behavior of Δ14C of total DOC reported in the York River estuary (Virginia, USA) by Raymond and Bauer [2001a]. In this case a downriver decrease in Δ14C-DOC of ∼200‰ across the low S range was attributed to the preferential utilization of younger, more labile fractions during bacterial degradation. This contrasting behavior of the two size classes of DOC again illustrates the size dependency of age and reactivity for various C pools. The higher size fraction of terrigenous DOC considered here is likely to be much more homogenous in age than total DOC examined elsewhere, and thus may have contained only minor amounts of older C, masking any age discrimination effects in the residual fraction. It is likely that the older components were contained mainly within the LMW DOC pool, as was observed in DOC from the open Atlantic and Pacific oceans [Loh et al., 2004]. It is also possible that the large removals of HMW DOC found in the Tyne and Tweed were caused mainly by non-oxidative processes such as flocculation and adsorption, which may not discriminate against age.

[10] Δ14C of HMW DOC broadly increased toward the North Sea (Figure 2), reaching values of up to 471‰ (Tyne) and 811‰ (Tweed) at mid to high S (>15). These Δ14C values were much greater than those previously reported (−200 to ∼100‰) for DOC sampled across similar salinities in other temperate estuaries, and were contrary to the general trend of decreasing Δ14C with increasing S attributed to mixing with older marine DOC or the preferential utilization of younger DOC fractions [Guo and Santschi, 1997; Raymond and Bauer, 2001a]. Adopting the following two lines of reasoning, our estuarine Δ14C data may be attributed to anthropogenic 14C inputs into coastal North Sea waters that are unrelated to nuclear weapons testing. Firstly, the values of our near-coastal HMW DOC in most cases were significantly higher than the peak weapons testing DIC values in the N Atlantic and cannot be explained either by modern DIC or potential bomb 14C memory effects. Secondly, estuarine HMW DOC removal in conjunction with seaward increases in δ13C and Δ14C precludes a riverine origin of 14C-enriched HMW DOC in the near-coastal North Sea.

[11] To confirm the marine origin of 14C-enrichment, we plotted δ13C against Δ14C (Figure 3). These plots yielded statistically significant positive linear relationships during the Jul-03 surveys to the Tyne (r2 = 0.86, n = 9) and Tweed (r2 = 0.74, n = 5), indicating mixing between a 13C-, 14C-enriched (marine) and a 13C-depleted modern (terrestrial) end-member. A weaker correlation was obtained for Mar-03 (r2 = 0.32, n = 5), perhaps due to the low S bias for these samples. The Dec-03 Tweed survey is subject to one anomalously high Δ14C value (811‰, S = 19.38; denoted by question mark) which has not been included in the regression (r2 = 0.85, n = 7) plotted in Figure 3. But given that the low S samples during this survey did not exhibit any significant 14C-enrichment, it is unlikely that terrigenous OM contributed to this sample's elevated Δ14C. Ignoring this one anomalous value, the mixing trends shown in Figure 3 collectively indicate a marine origin of 14C-enriched HMW DOC.

Figure 3.

δ13C and Δ14C (‰) of HMW DOC in the Tyne and Tweed estuaries plotted with their linear regressions. The one anomalously 14C-enriched sample (?) is not included in the Dec-03 regression.

[12] The largest documented source of anthropogenic 14C to the marine environment in the UK is the Sellafield nuclear fuel reprocessing plant (Cumbria, NW England; Figure 1), whose discharges are primarily in the form of DIC [Cook et al., 2004]. As a result of this 14C-enriched DIC [Δ14C ∼ 18000‰; Cook et al., 2004] entering the food chain via algal photosynthesis, significantly elevated 14C levels (100s to 1000s Bq kg−1 C) were reported for particulate organic carbon (POC) and marine biota collected from the NE Irish Sea [Cook et al., 1998, 1995]. Sellafield discharge spreads along the NW British coast and reaches the waters off eastern England after transport times of 3–5 a [Otto et al., 1990]. It may therefore be plausible to assume a diluted, Sellafield-derived 14C contribution to the near-coastal North Sea.

[13] An additional source of anthropogenic 14C into the North Sea is discharge from nuclear reprocessing at the La Hague plant in Normandy, France (Figure 1). This discharge enters the North Sea via the Straits of Dover and reaches SE England after 1–4 a [Otto et al., 1990]. La Hague discharge was suggested as the likely source for 14C-enriched POC sampled off the Dutch coast in 1994 [Megens et al., 2001], consistent with recent reports of elevated Δ14C (up to 700‰) in DIC and POC collected near La Hague [Douville et al., 2004]. Localized but recurring 14C-enrichments have also been found in DIC (450–496‰), mussels (2534–4048‰), and seaweed (320–339‰) collected near the Hartlepool advanced gas-cooled reactor (AGR) nuclear power station, located ∼40 km south of the Tyne estuary (Figure 1), in 1998/9 surveys [Gulliver et al., 2004]. Similarly, samples collected near Torness AGR, a reactor with a similar design to Hartlepool AGR and located ∼35 km NW of the Tweed (Figure 1), also showed enhanced 14C activities, albeit only marginally above modern levels [Gulliver et al., 2004].

[14] Most anthropogenic 14C-enrichments previously reported in the literature occurred in the vicinity of known or potential sources. One notable exception is the value of 210‰ reported for HMW DOC sampled at the Frisian Front (southern North Sea) in 1996 [van Heemst, 2000]. The 14C-enrichments of marine HMW DOC observed by us occurred in two distinct estuaries with no documented 14C releases over two seasons and hence may not be the sole result of a local, single point source. Furthermore, POC collected during these surveys was significantly older [Δ14C = −465 to −71‰; Ahad, 2005], suggesting that DOC release from the local POC pool was not the source of elevated 14C in HMW DOC. This consideration and the mixing behavior in Figure 3 allude to an external marine source for enhanced 14C activities.

[15] One feature of the elevated 14C found in samples near Hartlepool and in the Irish Sea is the general pattern of greater enrichment in marine biota compared to DIC. An exception is seaweed, in which lower 14C activities can be attributed to its partial incorporation of atmospheric CO2 [Cook et al., 2004, 1998]. Higher 14C activities in OM relative to concurrent DIC may be explained by the ‘lingering effect’ of the 14C signal once it enters the food chain. Filter feeders such as mussels attain elevated Δ14C from their food source (i.e., planktonic detritus), maintaining this signal through episodic 14C-enriched inorganic releases [Cook et al., 1998]. In contrast, the 14C-enriched signal in the DIC is lost fairly quickly by mixing and dilution with the large pool of older marine DIC and via atmospheric equilibration.

[16] This ‘lingering effect’ of anthropogenic 14C may also explain the 14C-enrichment in marine HMW DOC encountered in the Tyne and Tweed estuaries. Firstly, unlike POC, which can settle out, DOC in general can be advected over larger distances. Recent estimates indicate that the turnover time of the bulk of HMW DOC is likely to be in the order of a few decades [Loh et al., 2004], although some labile components produced during spring phytoplankton blooms can break down much more rapidly [e.g., Amon and Benner, 1994]. The anthropogenic 14C signal could therefore remain over these time scales on entering the HMW DOC pool. Secondly, this signal is less prone to dilution by mixing with aged HMW DOC from deeper waters because HMW DOC concentrations typically exhibit decline with depth [e.g., Loh et al., 2004]. This combination of factors – propensity to advect, decadal turnover time, less susceptibility to dilution – may allow anthropogenic 14C incorporated in HMW DOC to travel relatively long distances away from the initial source.

[17] In the North Sea for instance, Sellafield discharge can reach the mouth of the Tyne and Tweed estuaries in 3–5 a [Otto et al., 1990], which is shorter than the presumed turnover time of the bulk of marine HMW DOC. The advection time is expected to be even shorter for other potential 14C sources due to their relative proximity (Figure 1). Therefore, it is conceivable that the coastal North Sea HMW DOC pool integrated its 14C signal from multiple non-bomb sources. On entering the Tyne and Tweed estuaries, this signal remained elevated above background 14C levels despite incurring partial losses by mixing and turnover during advection. Our finding that the HMW DOC pool may act as the vector for transport of anthropogenic 14C to areas distal from its initial source has broad implications. The diffuse signal resulting from anthropogenic 14C releases may lead to an underestimation of DOC ages far from the point of discharge. It may also be possible to exploit elevated 14C activities in affected environments as a tracer for C transfer between various DOC size fractions.


[18] We thank A. Anestis, S. Mowbray, A. Pike and J. Smith for field assistance, C. Chilcott for help with δ13C analyses (Edinburgh), and staff at the SUERC AMS and NERC Radiocarbon Laboratory for 14C support. We also thank G. Cook for comments on an initial draft, G. Cowie, J. Barth and R. Upstill-Goddard for constructive discussions, and E-S. Badr and E. Achterberg for assistance with HTCO (Plymouth). This research was funded by NERC (NER/T/S/2000/01369 and 14C allocation 1034 and 1084). Support was also provided by the Schools of Marine Science & Technology (Newcastle) and GeoSciences (Edinburgh).