Marine Biology

, Volume 160, Issue 3, pp 541–552

δ13C and δ15N variations in organic matter pools, Mytilus spp. and Macoma balthica along the European Atlantic coast

Authors

    • Consiglio Nazionale delle Ricerche (CNR)IAMC Istituto per l’Ambiente Marino Costiero
    • CNR, ISMAR Istituto di Scienze Marine
  • S. Rajagopal
    • Department of Animal Ecology and Ecophysiology, Institute for Water and Wetland ResearchRadboud University Nijmegen
  • S. Como
    • Consiglio Nazionale delle Ricerche (CNR)IAMC Istituto per l’Ambiente Marino Costiero
  • J. M. Jansen
    • Royal Netherlands Institute for Sea Research (NIOZ)
  • G. van der Velde
    • Department of Animal Ecology and Ecophysiology, Institute for Water and Wetland ResearchRadboud University Nijmegen
    • Naturalis Biodiversity Center
  • H. Hummel
    • Royal Netherlands Institute for Sea Research (NIOZ)
Original Paper

DOI: 10.1007/s00227-012-2110-7

Cite this article as:
Magni, P., Rajagopal, S., Como, S. et al. Mar Biol (2013) 160: 541. doi:10.1007/s00227-012-2110-7

Abstract

Stable carbon (δ13C) and nitrogen (δ15N) isotope (SI) values of sedimentary organic matter (SOM), seston and two dominant bivalves, Mytilus spp. and Macoma balthica, were studied at 18 stations along the European coast in spring and autumn 2004. Three main regions, the Baltic Sea (BS), the North Sea and English Channel (NS), and the Bay of Biscay (BB), were tested for possible geographic (latitudinal) differences in the SI values. In spring, only BS showed lower δ13C values of seston and Mytilus spp., and higher δ15N values of SOM, than NS and BB. No significant differences between the 3 regions were found in autumn. Irrespective of season and regions, Mytilus spp. was more 13C-depleted than M. balthica. δ13C values of M. balthica, but not those of Mytilus spp., were significantly correlated with SOM. These results are consistent with differences in feeding behavior of Mytilus spp. and M. balthica, as the two species are known as obligatory-suspension and facultative-deposit feeders, respectively. In contrast, no differences in the δ15N values of Mytilus spp. and M. balthica were found at individual stations, indicating the same trophic level of the two bivalves within the food webs. At some stations, irrespective of geographic location, both bivalves showed δ15N values up to 18–20 ‰. These were two trophic levels higher than those found at the other stations, indicating local and/or episodic eutrophic conditions, probably due to waste water discharge, and the effectiveness of both Mytilus spp. and M. balthica as bio-indicators of anthropogenic eutrophication. Overall, our results suggest that pathways of energy flow from OM pools to dominant bivalves is more related to local environmental conditions than to geographic regions across the European coastline. This has implications for food web studies along the Atlantic coast because most of the values are consistent over a large area and show no significant differences. Therefore, the present study can be used twofold for the determination of trophic baselines and for the correction of the trophic position of consumers higher up in the food web in the case of differences in waste water discharge.

Introduction

Stable isotope (SI) analysis is an extremely powerful tool to study element cycles and various biologic, physical, and chemical processes in ecological studies. The two most used SIs in ecological research are carbon (δ13C) and nitrogen (δ15N). SI analysis has become an increasingly popular method to study both the origin of organic matter (OM) and the trophic dependency of faunal communities on different sources of OM (Fry and Sherr 1984; Peterson and Fry 1987; Michener and Schell 1994). In general, there is a predictable relationship between the isotopic composition of a consumer and its food source. Isotopic carbon ratios (δ13C) increase slightly, at an average of 1 ‰ per trophic level (De Niro and Epstein 1978), while isotopic nitrogen ratios (δ15N) change on a larger scale, with an average of 3–4 ‰ enrichment per trophic level (Michener and Schell 1994). This difference is caused by physiologic processes which discriminate between N atoms, while the majority of natural processes do not discriminate between C atoms based on mass, with the lighter (i.e., 15N) nitrogen being excreted preferentially (Wada and Hattori 1991; Lehmann et al. 2002; Nicholls and Trimmer 2009). Combining both carbon and nitrogen isotope data identifies the position of an organism in a food web, with δ13C indicating its carbon source and δ15N indicating its trophic level.

SI analysis helps to identify spatial and temporal variation in the isotopic composition of food sources and consumers (Connolly et al. 2005; Hill et al. 2008, Allan et al. 2010). Sources of OM are dependent on local environmental conditions, and thus may vary spatially, temporally, and in the amount of contribution they make to the OM pool, which are in turn borne out by the SI ratios of the OM (Deegan and Garritt 1997; Middelburg and Nieuwenhuize 1998; Svensson et al. 2007). Additionally, temporal variation in the isotopic values of the sources themselves may result in changes to the SI values of the OM (Cifuentes et al. 1988; Riera and Richard 1997; Zimmerman and Canuel 2001). Isotopic changes in each source may also contribute to variation in the isotopic composition of the pools. At temperate latitudes, for instance, primary producers often show lower δ13C values during winter than in the other seasons of the year (Riera and Richard 1997; Vizzini and Mazzola 2006; Magni et al. 2008), possibly due to low light intensity in winter (Hemminga and Mateo 1996). SI analysis has also proven to be a powerful tool for elucidating the importance of biogeography in relation to isotope values (Rau et al. 1982). However, despite earlier studies, there has been limited research on biogeographic variation in isotopic composition of marine consumers and their food, especially within intertidal and estuarine benthic communities, most of which has been done along the southern African coast (Hill et al. 2006; Hill and McQuaid 2008; Allan et al. 2010).

The European coast is extensive and therefore contains many different ecosystems from the Baltic to the Mediterranean Sea, with different weather and current patterns, different water temperatures and salinities and different organisms, different anthropogenic impacts, and different topographies. The number of marine species known in this part of the Atlantic Ocean is ca. 30,000 (Costello et al. 2001). Although the species richness of coastal ecosystems of the Baltic Sea (BS) (ca. 75 plant and animal species), the North Sea (more than 1,500 species), and the Mediterranean (more than 17,000 species) is different (Elmgren and Hill 1995; Hendriks et al. 2006; Coll et al. 2010), trophic relationships in all these systems are assumed to be similar (Elmgren and Hill 1995). However, little attention has been paid in Europe to understand whether variations in the composition of OM pools and consumers exist at large spatial scales across and among different geographic regions (Sokołowski et al. 2012).

In this study, we aimed to address the possible relationship between latitude/longitude across the European Atlantic coastline, including the BS, the North Sea and English Channel (NS), and the Bay of Biscay (BB), and the δ13C and δ15N values of OM pools, as well as of Mytilus spp. and Macoma balthica. Secondly, we aimed to test for differences in the isotopic composition of Mytilus spp. and M. balthica within each geographic region and between stations as a result of possible differences in their diet or trophic level within a food web. These species are known to have different feeding behavior, the former being an obligatory-suspension feeder, while the latter a facultative-deposit feeder (Hummel 1985a; Kamermans 1994; Rossi et al. 2004). As such, we would expect Mytilus spp. to rely on pelagic OM pools only, while M. balthica is expected to rely on both pelagic as well as sedimentary OM pools, thus displaying different isotopic values irrespective of biogeographic variation.

Materials and methods

Sampling stations

Samples were collected at 18 sampling stations along the European continental coast, including estuarine and coastal ecosystems from the BS (northern Sweden) to the southern coast of the BB, in spring and autumn 2004 (Fig. 1, Table 1). Stations were grouped into three main geographic regions corresponding to latitudinal/longitudinal gradients in terms of main environmental factors, such as water salinity and temperature (Kube et al. 2006; Jansen et al. 2007). Stations 1–6 were located in the BS, stations 7–12 in the NS, and stations 13–18 along the BB.
https://static-content.springer.com/image/art%3A10.1007%2Fs00227-012-2110-7/MediaObjects/227_2012_2110_Fig1_HTML.gif
Fig. 1

Location of sampling stations along the European coast. Circles indicate the three main geographic areas considered in this study: the Baltic Sea (1–6), the North Sea and the English Channel (7–12), the Bay of Biscay (13–18). Station number and code as given in Table 1

Table 1

Location of sampling stations, numbered (Num) and coded from north to south by study region and country, with GPS coordinates and general characteristics

Code

Num

Location

Study region

Country

Latitude

Longitude

Description

Max summer T (°C)

Min winter T (°C)

Max Sal (PSU)

Min Sal (PSU)

Ume

1

Umeå

Baltic Sea (BS)

Sweden

63.5636

19.8253

Subtidal, brackish

20.5

−0.2

4.5

0.4

Ask

2

Askö

Baltic Sea (BS)

Sweden

58.8299

17.6287

Subtidal, brackish

21.3

−0.5

6.8

5.8

Lom

3

Lomma

Baltic Sea (BS)

Sweden

55.7295

12.9580

Subtidal, brackish

20

0

17

6

Gda, Gdb

4, 5

Gulf of Gdansk

Baltic Sea (BS)

Poland

54.6500

18.5167

Subtidal, brackish

21.5

2

8

6

Mek

6

Mecklenburg Bight

Baltic Sea (BS)

Germany

54.0238

11.4830

Subtidal, brackish

21.7

2.2

18.7

10.8

Wil

7

Wilhelmshaven

North Sea (NS)

Germany

53.4229

8.0222

Intertidal, coastal

23

0

25

15

Bal

8

Balgzand

North Sea (NS)

The Netherlands

52.9301

4.7953

Intertidal, coastal

24

1

30

20

Gre

9

Grevelingen

North Sea (NS)

The Netherlands

51.7333

3.9833

Subtidal, marine

26

2

32.6

29.3

Wes

10

Westerschelde

North Sea (NS)

The Netherlands

51.3792

3.6272

Intertidal, estuary

26

2

32

15

Som

11

Somme

English Channel (NS)

France

50.2146

1.6227

Intertidal, estuary

25

3

32

25

Sei

12

Seine

English Channel (NS)

France

49.2570

0.1287

Intertidal, estuary

24

3

32

15

Loi

13

Loire

Bay of Biscay (BB)

France

47.2675

−2.2183

Intertidal, estuary

29

5

30

15

Aig

14

Aiguillon

Bay of Biscay (BB)

France

46.2791

−1.1596

Intertidal, estuary

30

6

32

20

Gir

15

Gironde

Bay of Biscay (BB)

France

45.5731

−1.0628

Intertidal, estuary

25

7

32

4

Arc

16

Arcachon

Bay of Biscay (BB)

France

44.6527

−1.1133

Intertidal, estuary

31

8

30

20

Bid

17

Bidasoa

Bay of Biscay (BB)

Spain

43.3627

−1.7734

Intertidal, estuary

30

9

35

5

Mun

18

Mundaka

Bay of Biscay (BB)

Spain

43.3863

−2.6833

Intertidal, estuary

30

9

34

15

They comprise three main geographic regions: the Baltic Sea (BS, 1–6), the North Sea and English Channel (NS, 7–12), and the Bay of Biscay (BB, 13–18). Station number and code are given as are reported in Fig. 1. Gda (4) and Gdb (5) indicate two different locations within the Gulf of Gdansk. Maximum (max) and minimum (min) values of temperature (T) and salinity (Sal) refer to the period 2003–2004 recorded within the BIOCOMBE project (EVK3-2001-00146; Hummel 2006). Temperature and salinity data for Gironde (45:31:000 N; 0:57:000 W) were obtained from SOMLIT Service d’Observation en Milieu LITtoral

Sample collection, treatment and stable isotope analysis

At each sampling station, the following components were sampled within a distance of ca. 200 m: sediment organic matter (SOM) (used as a proxy measure of microphytobenthos and sedimentary detritus), seston (used as a proxy measure of phytoplankton and suspended particulate OM), and two dominant benthic macro-invertebrates, the bivalves Mytilus spp. and M. balthica. For SOM, the uppermost few mm of surface sediment were scraped by hand or collected using a manual corer. Seston was collected by filtering 1–2 l of near-bottom water through pre-combusted (12 h at 450 °C) glass-fiber filters (Whatman GF/F). Mytilus spp. were collected from hard substrate mostly in rock pools, while M. balthica were sampled from the sediment using a manual corer and sieved through a mesh size of 1 mm at the same site. Animals were transported to the laboratory in containers with local water.

In the laboratory, bivalves were rinsed first in tap water and then in distilled water. The adductor muscles of 6–8 individuals were pooled for each sample and oven-dried at 60 °C for 3–5 days. All samples were subsequently ground to fine powder and stored in small glass bottles that were sealed with a plastic cap until weighing. In the case of filter samples for the seston, the top layer was scraped off and subsequently ground to a powder. Sediment samples for the analysis of SOM were decalcified using drops of 1 N HCl to remove carbonates, known to be 13C-enriched compared to organic material (Cloern et al. 2002), until no more bubbles appeared. The sample was then dried again until it was fully dehydrated. Acidification often also affects labile N in a sample by reducing the δ15N values of primary producers and SOM with respect to untreated samples (Bunn et al. 1995; Kennedy et al. 2005; Ng et al. 2007). However, we did not run some non-acidified samples for comparison and assumed that the change in δ15N should be negligible in our samples because of the use of a weak HCl solution (Bunn et al. 1995; Kennedy et al. 2005; Ng et al. 2007). For the seston, the total sample scraped off the filter was used in the measurements, unless the amount exceeded approximately 10 mg. For sediment samples, the tin cup was filled as much as possible, usually to ca. 30 mg. For bivalves, a subsample of pooled animals for each replicate at each station was taken from the total sample and placed in a small tin cup for SI measurements. All samples were kept in their tin cups until measured.

Carbon and nitrogen stable isotopic compositions were measured with a Carlo Erba NA 1500 elemental analyzer coupled online via a Finnigan Conflo III interface with a ThermoFinnigan DeltaPlus mass spectrometer. Carbon and nitrogen isotope ratios are expressed in the standard delta notation (δ13C, δ15N) relative to Vienna PDB and atmospheric nitrogen. Average reproducibility based on replicate measurements of internal standards sucrose (IAEA-CH-6) for δ13C and ammonium sulfate (IAEA-N-2) for δ15N was ca. 0.15 ‰. Acetanilide was used as the laboratory reference. For bivalves, there were two to four replicates per station, for SOM and seston, there were in most cases two to six replicates per station (suppl. Tables S1 and S2).

Statistical analysis

Due to differences in the dataset available within each geographic area (i.e., BS, NS, and BB), differences in the isotopic composition of SOM, seston, Mytilus spp. and M. balthica were analyzed, when possible, using a 1-way model of analyses of variance with unequal sample size (Zar 1991). Before analyses, the homogeneity of variances was evaluated using the Barlett’s test (Zar 1991). When significant differences among geographic areas were found, Bonferroni–Dunn tests were done as a posteriori comparisons. Before analysis, the homogeneity of variances was evaluated by using C-Cochran’s test (Winer et al. 1991).

In order to evaluate the relationships between Mytilus spp. and M. balthica and between bivalves and OM pools, correlation coefficients (R) between δ13C and δ15N values were calculated. To account for multiple simultaneous analyses, the level of significance was adjusted using the sequential Bonferroni technique (Rice 1989). The differences in the isotopic composition between Mytilus spp. and M. balthica were analyzed with a t test (Zar 1991). The mean values of δ13C and δ15N of Mytilus spp. and M. balthica at the stations where they were both present were used as replicates.

Results

δ13C and δ15N values in the three main geographic regions

On the basis of the individual data (suppl. Tables S1, S2), no clear geographic patterns (clines), nor consistent differences between spring and autumn could be observed. The δ13C values of SOM ranged from −24 to −19 ‰, while seston had a wider range from −25 to −13 ‰. For Mytilus spp. as well as M. balthica the δ13C values ranged mainly from −22 to −14 ‰. The δ15N values of SOM as well as seston ranged in spring as well as in autumn mainly from 5 to 10 ‰. For Mytilus spp. as well as M. balthica the δ15N values fluctuated strongly between stations and season, and ranged mainly from 5 to 18 ‰.

To better unravel overall patterns, the mean δ13C and δ15N values of sedimentary organic matter (SOM), seston, Mytilus spp. and M. balthica in the three main geographic regions (BS; NS; and BB) in spring and autumn were calculated (Fig. 2). The analyses of variance revealed significant differences (P < 0.05) among regions in δ13C values of both seston and Mytilus spp. in spring (Table 2). Seston and Mytilus spp. tended to be more 13C-depleted in the BS than North Sea–English Channel and Biscay Bay (Fig. 2a). The a posteriori comparisons, however, evidenced a clear-cut alternative hypothesis for the δ13C values of seston only. Significant differences among regions were also found in the δ15N values of SOM in spring. The a posteriori comparisons also revealed more 15N-enriched values in the BS than the North Sea–English Channel and Biscay Bay (Fig. 2b). No significant differences between the 3 regions were found in autumn (Table 2; Fig. 2c, d). A comparison between bivalves and OM pools at individual regions showed a clear 13C enrichment in M. balthica with respect to SOM in autumn in BB (~5 ‰) (Fig. 2c). Instead, 15N enrichment was strongly expressed in both bivalves and seasons, with respect to the OM pools, in NS (~4.5 ‰) (Fig. 2b, d).
https://static-content.springer.com/image/art%3A10.1007%2Fs00227-012-2110-7/MediaObjects/227_2012_2110_Fig2_HTML.gif
Fig. 2

The mean (±standard error, SE) of δ13C (left panelsa, c) and δ15N (right panelsb, d) values of sedimentary organic matter (SOM), seston, Mytilus spp. (Myt), and M. balthica (Mac) for each region (BS Baltic Sea, NS North Sea and English Channel, BB Bay of Biscay) in spring (top panelsa, b) and autumn (lower panelsc, d). Greek letters (α and β) indicate where significant differences (P < 0.05) between the 3 regions occur

Table 2

Summary of analyses of variance for differences in δ13C and δ15N values of sedimentary organic matter (SOM), seston, Mytilus spp. and M. balthica among the three regions (i.e., Baltic Sea [BS], North Sea and English Channel [NS], and Bay of Biscay [BB]) in Spring and Autumn

 

SOM

Seston

Mytilus spp.

M. balthica

δ13C

δ15N

δ13C

δ15N

δ13C

δ15N

δ13C

δ15N

df

MS

F

MS

F

df

MS

F

MS

F

df

MS

F

MS

F

df

MS

F

MS

F

Spring

 Regions

2

2.46

1.16

14.26

8.10*

2

62.61

16.27**

5.20

3.08

2

11.73

4.74*

20.53

0.80

1a

0.07

0.01

11.73

1.02

 Residual

9

2.12

 

1.76

 

12

3.85

 

1.69

 

8

2.47

 

25.58

 

8

5.27

 

11.46

 

Autumn

 Regions

2

3.43

0.81

4.29

0.83

2

38.35

2.60

13.98

1.48

2

0.61

0.21

11.46

1.99

2

0.33

0.04

17.94

4.02

 Residual

7

4.24

 

5.15

 

9

14.75

 

8.79

 

10

2.85

 

5.75

 

10

7.56

 

4.46

 

df degrees of freedom, MS mean squares, F Fischer’s F, P probability

P < 0.05; ** P < 0.001

aComparison between BS and NS only

Mytilus spp. versus Macoma balthica and bivalves versus OM pools

Mytilus spp. was more 13C-depleted than M. balthica of an average value of 2.6 ‰ (Fig. 3a, t test: t = −3.59, P < 0.01). In contrast, δ15N values of Mytilus spp. and M. balthica showed more equal values (and thus a close 1:1 linear relation; Fig. 3b), indicating the same trophic level within an individual station. No significant differences in the δ15N values between the two bivalves were found (t test: t = 1.12, P > 0.05).
https://static-content.springer.com/image/art%3A10.1007%2Fs00227-012-2110-7/MediaObjects/227_2012_2110_Fig3_HTML.gif
Fig. 3

Comparison of (a) δ13C and (b) δ15N values of Mytilus spp. and M. balthica. The dashed lines represent a 1:1 relationship

Among the various combinations between OM pools and bivalves (Table 3), the only significant (positive) correlation found was that between δ13C values of SOM and M. balthica (Fig. 4b), indicating that M. balthica also feeds on the sediment. For M. balthica an average 13C enrichment of 4 ‰ in relation to SOM is clear (Fig. 4b), but not for Mytilus spp. (Fig. 4a). However, when focusing on the data at individual stations, major differences in the δ15N values of both Mytilus spp. and M. balthica were found among individual stations irrespective of geographic location (Fig. 5). In general, in spring, the mean δ15N values of SOM and seston were averaging ca. 6–7 and 8–9 ‰, respectively (Fig. 5a, b). The δ15N values of the bivalves (Fig. 5c, d) were more variable than those of the OM pools, with varying degrees of enrichment, from almost no 15N enrichment up to 8 ‰ (equalling two trophic levels) higher than in SOM or seston. This was true for both species within the same location (Fig. 5c, d), indicating that the response has a common ground (see also suppl. Tables S1, S2). In autumn, the δ15N values of bivalves were in the range of 9–10 ‰ at several stations, indicating at least a one trophic level change between the OM pools and the bivalves.
Table 3

Linear correlation (Pearson’s R) between (a) δ13C or δ15N values of Mytilus spp. (Myt) or M. balthica (Mac) and sedimentary organic matter (SOM) and (b) δ13C or δ15N values of Mytilus spp. (Myt) or M. balthica (Mac) and seston (n number of samples)

 

δ13C-SOM

δ15N-SOM

n

R

P

 

n

R

P

(a)

 δ13C-Myt

20

0.38

0.10§

δ15N-Myt

20

−0.06

0.79

 δ13C-Mac

17

0.52

0.03§§

δ15N-Mac

17

−0.15

0.57

 

δ13C-seston

δ15N-seston

n

R

P

 

n

R

P

(b)

 δ13C-Myt

23

0.20

0.35

δ15N-Myt

23

0.11

0.61

 δ13C-Mac

21

−0.13

0.57

δ15N-Mac

21

0.04

0.86

Significant correlations after sequential Bonferroni correction (Rice 1989) are in bold

§As in Fig. 4a

§§As in Fig. 4b

https://static-content.springer.com/image/art%3A10.1007%2Fs00227-012-2110-7/MediaObjects/227_2012_2110_Fig4_HTML.gif
Fig. 4

Relationship between aδ13C values of M. balthica and SOM, and b δ13C values of Mytilus spp. and SOM (see also Table 3). In b the linear relationship indicates a 4 ‰ difference between M. balthica and SOM (R = 0.52; P < 0.03, Table 3)

https://static-content.springer.com/image/art%3A10.1007%2Fs00227-012-2110-7/MediaObjects/227_2012_2110_Fig5_HTML.gif
Fig. 5

Plots of the mean (±standard error, SE) δ13C and δ15N values of a sedimentary organic matter (SOM), b seston, cMytilus spp. (Myt), and dMacoma balthica (Mac) in the three regions (black squares Baltic Sea, gray triangles North Sea and English Channel, open circles Bay of Biscay). In Mytilus spp. and M. balthicaplots, numbers near the symbols indicate sampling locations with δ15N values above 12 ‰ as indicated by the dashed line (see Table 1 for the name of the sampling stations)

Discussion

Geographic (latitudinal) variations in the δ13C and δ15N values

There are relatively few studies assessing biogeographic variations in the δ13C and δ15N values of OM pools and primary consumers (Hill et al. 2006; Hill and McQuaid 2008; Allan et al. 2010). In our study, the δ13C and δ15N values of SOM and seston showed large variations across the wide geographic range considered. This large variability can be expected as SOM is made up, in varying proportions, of microphytobenthos, detritus (at various stages of decomposition), and bacteria, while seston can include a composite and variable mixture of nearshore primary production, plankton, and suspended particulate matter (Riera 1998; Marín Leal et al. 2008; Dubois et al. 2012). In spite of the large variability found across the sampling stations, the BS showed significantly 13C-depleted seston and 15N-enriched SOM in spring, compared to the North Sea and English Channel (NS), and the BB. 13C depletion in primary producers has been related to low light intensity at temperate latitudes (Cifuentes et al. 1988; Hemminga and Mateo 1996; Riera and Richard 1997), supporting the theory of enhanced 13C fixation with increasing irradiance and photosynthesis (Treignier et al. 2009). It may thus be possible that 13C-depleted values of seston in BS be related to a lower light intensity than in NS and BB. However, given the brackish nature of the BS with an influx of freshwater from many rivers, lower salinity waters could more likely be a factor for 13C depletion in the seston sampled in spring in BS. This is consistent with numerous studies showing that freshwater phytoplankton is far more 13C-depleted than that of marine phytoplankton (France 1995; Lefebvre et al. 2009a; Oczkowski et al. 2011). Alternatively, Hill et al. (2008) found at Kenton-on-Sea, on the south coast of South Africa, δ13C variation in seston to be related to hydrographic processes. On the other hand, 15N enrichment in SOM might be related to waste water discharge in bays and estuaries (McClelland et al. 1997; Cabana and Rasmussen 1996; Dailer et al. 2012). A large-scale study conducted in European waters showed that, regardless of geographic position and environmental/biocenotic conditions, the most 13C-depleted was seston compared to other OM pools, including SOM, macroalgae and vascular plants, while no differences in δ15N values between seston and SOM were found (Sokołowski et al. 2012). In our study, with the exception of some differences in BS, there was no consistent latitudinal variability with regard to δ13C or δ15N for either SOM or seston. We conclude that the OM pools analyzed in this study (SOM and seston) represented comparable food sources for the two bivalves, Mytilus spp. and M. balthica, owing to the difficulty in discerning the inherent variability existing at scales smaller than the biogeographic and latitudinal ones.

The studied bivalves, Mytilus spp. and M. balthica, showed some visible 13C enrichment, especially for M. balthica in autumn, and 15N enrichment both in spring and autumn. This is supported by the work of Sokołowski et al. (2012) suggesting that, irrespective of geographic location, the major source of energy to dominant suspension- and deposit-feeding bivalves originates from SOM and seston. Hill and McQuaid (2008) also showed that, along the southern African coast, nearshore seston was made up of ca. 60 % macroalgae (Ulva sp.) and contributed upwards of 40 % of the diet of all sampled filter feeders. Indeed, filter feeders such as Mytilus spp. and M. balthica are known to play an important role in benthic-pelagic coupling, contributing to the energy transfer to the benthic environment (Hummel 1985a, b; Hill et al. 2006). In our study, there was a slight variability in bivalves studied across different regions, indicating some variability in food source. Nevertheless, we conclude that in Europe, from Spain northward along the coast to northern Sweden, there is no relationship between the SI values of Mytilus spp. and M. balthica and latitude. Our results differ from the marked spatial patterns observed in the δ15N values of the mussel Perna perna along the southern African coast (Allan et al. 2010), which was related to upwelling areas associated with increased denitrification, resulting in 15N enrichment downstream from the upwelling. Another study conducted along the southern African coast (Hill and McQuaid, 2008) also showed biogeographic trends, with a east–west 15N enrichment, which was most apparent for the bivalves Mytilus galloprovincialis and P. perna.

Differences/similarities in the isotopic composition of Mytilus spp. and M. balthica

Mytilus spp. was more 13C-depleted than M. balthica, suggesting the two bivalve species had different diets. This is consistent with differences in feeding behavior of Mytilus spp. and M. balthica, known as obligatory-suspension and facultative-deposit feeders, respectively (Kamermans 1994; Rossi et al. 2004). However, Riera et al. (1999) showed that the isotopic composition of Mytilus edulis and M. balthica in Aiguillon Bay (France) could be mainly explained by a mixed diet of benthic diatoms and marine phytoplankton, in different proportions, making it difficult to identify one main source of OM in the bivalve diet. Another study conducted in Hiroshima Bay (Japan) showed that resuspension of 13C-enriched microphytobenthos increased the δ13C values of seston in the surf zone and consequently increased the δ13C values of the suspension feeders through their feeding on suspended matter (Takai et al. 2004). Similarly, Kang et al. (2003) on the southern coast of the Korean peninsula found that suspension feeders in both the intertidal and subtidal habitats equally fed on marine phytoplankton and benthic microalgae. We infer that the differences in δ13C values between the two bivalves at our sampling stations might also be related to the different substratum where Mytilus spp. and M. balthica were collected, that is, from hard substrate the former, from the sediment the latter.

With regard to δ15N, we did not find differences in M. balthica and Mytilus spp. values between the 3 main regions, nor at individual stations. This would indicate the same trophic level of the two bivalves within a food web. Our results are partially consistent with those of Hill and McQuaid (2008) who tested for interspecific differences in δ13C and δ15N values in mussels P. perna, and M. galloprovincialis and found no significant difference between species in mean values of isotopic carbon or nitrogen. The lack of enrichment in 13C and 15N in both bivalve species may indicate separate food sources in the BS. We cannot exclude, however, that the measurements of these broad categories of OM may be obscuring what is potentially a much narrower diet which represents a small fraction of the SOM and seston. Irrespective of geographic regions, high δ15N values were found at some stations (i.e., Lomma, Wilhelmshaven, Grevelingen, Loire, Arcachon, and Westerschelde), which may indicate local and/or episodic eutrophic conditions. Riera et al. (2000) for example, showed higher 15N enrichment in benthic primary producers and invertebrates from the Westerschelde estuary compared to the Oosterschelde. These authors suggested that this was most likely due to the incorporation of 15N-enriched dissolved inorganic nitrogen carried by the Scheldt River by benthic algae and then by benthic consumers. The 15N enrichment of bivalves at some of our stations would then also suggest higher isotope values for SOM and seston, which was not the case. Sewage effluent can be responsible for higher δ15N values in algae, bivalves, and other filter feeders (Mallela and Harrod 2008; Oczkowski et al. 2008; Thornber et al. 2008). McClelland et al. (1997) observed a strong correlation between percent wastewater and δ15N values in primary producers. The increase in δ15N values of primary producers with increasing wastewater loading was also passed on to consumers in the studied food web. Cabana and Rasmussen (1996) reported that δ15N of primary consumers was strongly correlated with the human population density in water sheds. They reported that this enrichment of 15N can be attributed not only to the high trophic level occupied by humans (and consequently their excreted nitrogen), but also fractionation during volatilization of ammonia gas, which leaves the remaining ammonium enriched in 15N. Subsequent nitrification produces nitrate with elevated δ15N values (Heaton 1986; Macko and Ostrom 1994). Enrichment is caused by the loss of 14N due to volatilization of ammonia, nitrification, and denitrification (Högberg 1990). Heaton (1986) showed that nitrate derived from animal or sewage water have δ15N values typically in the range of 10–20 ‰. Macko and Ostrom (1994) confirmed this and reported also that nitrate in barnyard soil have δ15N values in the range of 10–22 ‰. However, SI studies and other tracer techniques indicate alternative, microbially mediated processes of nitrate transformation such as dissimilarity (the reduction of nitrogen into other organic compounds coupled to energy producing processes), reduction of nitrate to ammonium, chemoautotrophic denitrification via sulfur or iron oxidation and anaerobic ammonium oxidation (anammox), as well as abiotic nitrogen removal processes (Burgin and Hamilton 2007). The elevated δ15N values of macrophytes, algae, and filter feeders (bivalves) can be used as an indicator of anthropogenic (domestic) eutrophication (Cole et al. 2004; Teichberg et al. 2010; Viana et al. 2011). Sewage water discharge can lead to algal blooms, and the derived detritus can be used by bivalves (Fertig et al. 2009; Lapointe et al. 2011). Wiedemeyer and Schwamborn (1996) found that phytoplankton production was the primary carbon source for M. edulis, with mean values of −19.9 to −19.5 ‰, in Kiel Fjord, BS, and concluded that 21 % contribution to the mussel’s tissue could be derived from benthic macroalgae.

We conclude that no major differences between the BS, NS, and BB regions were found in the δ13C and δ15N values of OM pools and the studied bivalves. Besides the lack of geographic/latitudinal/longitudinal trends found in this study, Mytilus spp. and M. balthica showed significant differences in their δ13C values, consistent with differences in their feeding behavior (Hummel 1985a; Kamermans 1994). In contrast, the δ15N values at individual stations were similar, indicating the same trophic level of the two bivalves within a food web. Thus, it is suggested that pathways of energy flow from OM pools to dominant bivalves is more related to local environmental conditions than to geographic regions across the European coastline. This conclusion has implications for food web studies along the Atlantic coast because most of the values are consistent over a large area and show no significant differences. In eutrophic conditions, δ15N levels of bivalves were elevated (up to 18–20 ‰ at some stations) with further consequences for the food web values. The latter results also indicate that both Mytilus spp. and M. balthica can be used as bio-indicators of anthropogenic eutrophication, as found for other bivalve species (McKinney et al. 2001; Lefebvre et al. 2009b; Watanabe et al. 2009). Overall, the present study can be used twofold for the determination of trophic baselines along the Atlantic coast and for the correction of the trophic position of consumers higher up in the food web in the case of differences in waste water discharge.

Acknowledgments

We are grateful to all BIOCOMBE colleagues who assisted in the fieldwork and kindly thank J. Eygensteyn for stable isotope analysis. We also gratefully acknowledge three anonymous reviewers for their insightful and detailed comments which greatly contributed to improve early versions of the manuscript. Temperature and salinity data for Gironde were kindly provided by SOMLIT Service d’Observation en Milieu LITtoral. The research leading to these results has received funding from the European Community’s Fifth Framework Programme under contract EVK3-2001-00146 for the project The Impact of BIOdiversity changes in COastal Marine Benthic Ecosystems (BIOCOMBE) and from the European Community’s Seventh Framework Programme (FP7/2007-2013) under Grant Agreement No. 266445 for the project Vectors of Change in Oceans and Seas Marine Life, Impact on Economic Sectors (VECTORS).

Supplementary material

227_2012_2110_MOESM1_ESM.doc (128 kb)
Supplementary material 1 (DOC 127 kb)

Copyright information

© Springer-Verlag Berlin Heidelberg 2012