High frequency Barium profiles in shells of the Great Scallop Pecten maximus : a methodical long-term and multi-site survey in Western Europe

Abstract. Skeletal barium/calcium ([Ba]/[Ca])shell ratios were measured every third daily striae in 39 flat valves of the Great Scallop Pecten maximus collected in temperate coastal environments of Western Europe. A methodical evaluation of the ([Ba]/[Ca])shell ratio was performed for the first time and demonstrates that ([Ba]/[Ca])shell profiles are reproducible for several scallop individuals from the same population (2-year old; 3 shells/year), over a 7-year period (1998–2004), and from different coastal environments in France (42–49° N). As previously determined in the shells of other bivalve species, ([Ba]/[Ca])shell profiles generally exhibited a background ratio punctuated by two transient maxima occurring in early and late summer. Background partition coefficient (DBa=0.11±0.03, in 2000) was similar to that previously reported in P. maximus shells, suggesting a direct shell uptake of dissolved seawater Ba (Gillikin et al., 2008). The 7-year survey in the Bay of Brest of the high frequency ([Ba]/[Ca])shell profiles in the scallop's shell was exploited to better constrain both the occurrence and the amplitude of the summer Ba relative enrichments as influenced by environmental processes. Seawater Ba contents in 2000 underlined significant particulate Ba inputs at the sediment water interface (SWI) during ([Ba]/[Ca])shell peak events. These Ba inputs are thus suggested to be subsequently induced by a pelagic biogenic process, which mainly occurs under summer post-bloom conditions in relationship to the cycling of particulate organic matter and associated Ba. The long term survey reveals that such pelagic Ba cycling processes are responsible for particulate Ba inputs to the sediment water interface (SWI). Subsequent indirect Ba uptake by the bivalve results in higher ([Ba]/[Ca])shell ratios, in that archived Ba within the shell cannot be used as a direct paleo productivity tracer. Our methodical approach, based on a multi-year and multi-site-survey of ([Ba]/[Ca])shell ratio in Scallop bivalves, allows us to establish the potential application of such high frequency archives for further biogeochemical and ecological investigations of bivalves in the coastal environment.


Introduction
Since 1965, trace elements in mollusc shells were investigated because bivalves form in their shells successive calcium carbonate layers, potential archives of conditions experienced by the organism during its life time (Dodd, 1965;Lorens and Bender, 1980;Klein et al., 1996aKlein et al., , 1996bStecher et al., 1996). The main aim of these studies was to use trace element records in shells as a bio indicator of environmental conditions. A recent increasing number of studies have investigated Ba content in biogenic carbonates, such as corals, foraminifera, and otoliths (Lea et al., 1989;Lea and Martin, 1996;McCulloch et al., 2003;Sinclair and McCulloch, 2004;de Vries et al., 2005;Sinclair, 2005). Coastal waters are enriched in Ba in the low to mid salinity ranges during estuarine mixing by either freshwater inputs of dissolved Ba or Ba release from river-born particulate phases (Coffey et al., 1997;Shaw et al., 1998;McCulloch et al., 2003). Alternatively Ba input can occur from the exchange of Ba-rich ground waters and pore waters within the tidal prism (Shaw et al., 1998). Coralline Ba content was reported as a proxy for discharge and/or sediment load in riverine systems (Sinclair and McCulloch, 2004;McCulloch et al., 2003). Some sharp Ba peaks in corals cannot however be assigned to any tracers of river discharge (Sinclair, 2005). Barium incorporation into carbonate parts of larval protoconchs or statoliths was also demonstrated to be promoted at lower temperature (Zacherl et al., 2003). In recent studies, there was much attention to the skeletal Ba content in mussels (Mytilus edulis; (Vander Putten et al., 2000;Gillikin et al., 2006), Isognomon ephippium; (Lazareth et al., 2003)), clams (Mercenaria mercenaria, Spisula solidissima, Arctica islandica, Saxidomus giganteus, Asiatic Corbicula fluminea; Fritz et al., 1990;Stecher et al., 1996;Epplé, 2004;Gillikin et al., 2005Gillikin et al., , 2008, and scallops (Pecten maximus, Comptopallium radula, Argopecten purpuratus; Thébault, 2005;Gillikin et al., 2008). In all these studies, [Ba/Ca] shell profiles were usually characterized by a relatively stable background ratio interrupted by sharp intense maxima. A simple interpretation achieved with several investigated bivalves (Mercenaria mercenaria, Mytilus edulis, Isognomon ephippium, Ensis siliqua) is also limited by an approximate dating of the shell record. In both lab and field experiments, the background [Ba/Ca] shell in the common mussel shell (Mytilus edulis) was demonstrated to track the [Ba/Ca] water aqueous concentration ratio (Gillikin et al., 2006). Considering the inverse relationship between [Ba/Ca] water and salinity, background [Ba/Ca] shell was considered as a specific indicator of salinity in estuarine environments (Gillikin et al., 2006). Barium was also established as a paleo productivity tracer in marine waters, suspended particles and sediments (Dymond and Collier, 1996;Paytan and Kastner, 1996). Stecher et al. (1996) postulated first that, in Mercenaria mercenaria and Spisula solidissima shells, these [Ba/Ca] shell peaks might be induced by the ingestion of high levels of particulate Ba in estuarine waters. Bivalves are nonselective filter feeders and can assimilate these Ba-rich particles. Once inside the digestive tract of the bivalve, Ba may be metabolised, at least in part, by being shunted to the extrapalleal fluid (EPF) and sequestered into the shell (Stecher et al., 1996;Vander Putten et al., 2000;Gillikin et al., 2006). Time resolved [Ba/Ca] shell peaks were suggested to be related to phytoplankton blooms (Stecher et al., 1996;Vander Putten et al., 2000;Lazareth et al., 2003). These Ba-rich particles were supposed to originate from diatom blooms as either biogenic matter, or barite precipitates (Stecher et al., 1996). Gillikin et al. (2006Gillikin et al. ( , 2008 indicated that these [Ba/Ca] shell peaks could not be used as a direct proxy of [Ba/Ca] water or phytoplankton production. These peaks were rather induced by ingestion of barite particles inputs to the sediment water interface (SWI) during the phytoplankton decay (Gillikin et al., 2006). Barium is recognized to be accumulated at high levels in phytoplankton, both in diatoms and in dinoflagellates (Fisher et al., 1991). A relatively large pool of labile Ba can be rapidly released during plankton decomposition and acts as the main source of Ba for barite formation in supersaturated microenvironments (Bishop, 1988;Ganeshram et al., 2003). Elevated levels of Ba, both as suspended particulate Ba and sedimentary barite occurrs in high primary productivity oceanic regions (Dehairs et al., 1980;Bishop, 1988;Dehairs et al., 1991Dehairs et al., , 1997. Most probably, barite formation results from its passive precipitation through sulphate enrichment and/or release in barium containing biogenic material (Bishop, 1988;Stecher and Kogut, 1999;Jeandel et al., 2000;Ganeshram et al., 2003;Dehairs et al., 2000). The formation of specific aggregates or Ba-rich micro-environments is responsible for the Ba enrichment in biogenic particles (Bishop, 1988). These aggregates are related to the diatom extracellular polymeric substances (EPS) produced during summer blooms and their settling (Stecher and Kogut, 1999;Thornton, 2002). The subsequent vertical mixing and diatom decay provide a rapid flux and large abundance of biogenic particles transported down to the SWI (Sarthou et al., 2005). This was suggested to induce a sudden pulse of Ba to the epibenthic environment, subsequently ingested by molluscs to be incorporated into shells and recorded by ([Ba]/[Ca]) shell maxima (Gillikin et al., 2006). Phytoplankton Ba uptake and barite formation are however not sufficient to explain the vertical flux of Ba in oceanic waters (Sternberg et al., 2005). Barium is also scavenged in the presence of diatom biogenic particles or by adsorption onto mineral oxides (i.e. Fe, Mn) that contribute to significant Ba inputs to the SWI (Sternberg et al., 2005).
This study focuses on Ba content in the Great Scallop shell Pecten maximus (L.) to further establish how these bivalves can provide time resolved ([Ba]/[Ca]) shell profiles (daily scale) and valuable archives of environmental changes in temperate environments (Lorrain et al., 2000;Chauvaud et al., 1998Chauvaud et al., , 2005Lorrain et al., 2005;Barats, 2006;Barats et al., 2007Barats et al., , 2008Gillikin et al., 2008 ). High frequency profiles of Ba content in the calcitic shell can be determined for each individual striae using laser ablation -inductively coupled plasma -mass spectrometry (LA-ICP-MS) (Barats et al., 2007). In this paper, the significance of ([Ba]/[Ca]) shell profiles in P. maximus is first evaluated for both background concentration and episodic sharp peaks in a single scallop population from the Bay of Brest (northwest France), among different coastal sites in western Europe, and over a 7-year period (1998)(1999)(2000)(2001)(2002)(2003)(2004) in the Bay of Brest. Second, the demonstrated reproducibility, recurrence and ubiquity of ([Ba]/[Ca]) shell profiles is further utilized to clarify the biogeochemical processes influencing both the background and episodic sharp peaks of ([Ba]/[Ca]) shell content in scallop bivalves.

Study sites and scallop sampling
Different sampling areas in Western Europe (Bay of Seine, Bay of Brest, Quiberon peninsula and Belle Ile, Ria de Vigo) were considered because they represent ecosystems with their own ecological characteristics (Barats, 2006 (Fig. 1): a shallow embayment with more than a half of its surface (180-km 2 ) and with depth of less than 5 m. The hydrodynamics of this ecosystem is mainly influenced by tidal exchange with the Iroise Sea, but also by freshwater input mainly by two small riverine watersheds (80%): the Aulne (1842 km 2 ) with monthly flow rate ranging from 2 to 52 m 3 /s and averaging 22 m 3 /s, and the Elorn (402 km 2 ) with monthly flow rate ranging from 1 to 12 m 3 /s and averaging 5.6 m 3 /s. The sampling area (Roscanvel) is 30 m deep where the bottom consists of mixed sediments (mud, sand) and selected because scallop density is one of the highest in the Bay of Brest .
Three replicates of live juvenile P. maximus specimens were collected for each year and site (except in 2002 in the Bay of Brest, n=2) during the late autumn period by SCUBA diving in the Bay of Brest and by dredging for other sites. Only the juvenile stage of the shells (the second year of growth) was examined because it exhibits the longest annual growth period. An exception was made for a three-year old P. maximus specimen collected in the Bay of Brest in 2003 to examine its third year of growth. The shells were isolated and  cleaned by submerging in glacial acetic acid (Merck, analytical grade) for 60 s to remove bio-fouling, rinsed with Milli-Q water (R=18.2 M .cm) and dried. For convenience in visualizing the growth striae, only the flat valve was considered for LA-ICP-MS analysis. A 45×10 mm 2 cross section was cut from each shell using a diamond saw to fit into an ablation chamber.

Determination of Ba concentrations in shells by
Laser Ablation -Inductively Coupled Plasma -Mass Spectrometry (LA-ICP-MS) Quantitative analyses of 39 shells were performed by LA-ICP-MS method consisting of coupling a UV laser ablation unit (LSX 100 UV 266 nm, Cetac Tech.) to an ICP-MS (X7 serie, Thermo Fisher). The methodology and validation were described elsewhere and summarized here (Barats et al., 2007). Samples were analysed during 2 min at a scan speed rate of 5 µm/s. For shell analyses, a pre cleaning ablation of the calcite surface was conducted to avoid surface contamination (Stecher et al., 1996;Vander Putten et al., 1999;Lazareth et al., 2003;Wyndham et al., 2004). This pre cleaning step consisted in a quick (around 20 s) pre-ablation of shell surface at a rate of 50 µm/s. Twenty-eight isotopes including 135,137,138 Ba, 55 Mn and 43 Ca were analysed. The intensity of each isotope signal was systematically normalized against the 43 Ca signal to compensate for instrumental drift and instability. An external calibration was performed with lab prepared CaCO 3 standards containing Ba concentrations ranging from 0.005 to 19.6 µg/g (Barats et al., 2007). The Ba calibration curve displayed a good linearity with a regression coefficient r 2 above 0.998 (Barats et al., 2007). The detection limit was about 3.3 ng/g, and the relative standard deviations obtained for both analytical repeatability (5 integration zones during 1 analysis) and reproducibility (5 successive independent analyses) were below 5% (Barats et al., 2007).
Analytical performances obtained for Mn were previously presented (Barats et al., 2007(Barats et al., , 2008 and expressed in µmol/mol. Mollusc shells were recognized to contain less than 5 % of organic matrix in calcite or aragonite shells (Carroll et al., 2006;Levi-Kalisman et al., 2001). The shell is thus composed by minimum 95% of CaCO 3 , i.e. a minimum shell Ca content of 38%. In comparison with the Ca concentration of 40%, the error is only of 5%. This study thus used a shell Ca content of 40% because scallop shells have a calcite structure. While the exact percentage of organic matter remains unknown, yet likely less than 5%, the analytical error will be negligible. A recent study of Takesue et al. (2008) revealed higher organic content in clams (19%), but they demonstrated also that Ba assay in shell content was unchanged by the removal of organic matter. This conclusion underscores that while Ba occurs exclusively in shell aragonite of clams, the analytical result agrees with the use of Ca concentration about 40% to calculate our ([Ba]/[Ca]) shell ratios. Shell analyses were performed each third striae during the shell growth period (from April to November) to limit the analysis time to approximately 5 h per shell. A date of formation was assigned to each ablated sample, by backdating from the harvest date, based on the daily periodicity of the striae formation in P. maximus. An evaluation of the shell growth rate (dorso-ventral linear extension of the shell per unit time), expressed in µm/d, was also performed for each shell by measuring distances between successive striae (growth increment width) using an image analysis technique (Chauvaud, 1998). For each shell analysis, a profile of shell growth rate and a profile of ([Ba]/[Ca]) shell ratio were obtained. For each year and site, a mean shell growth rate profile and a mean ([Ba]/[Ca]) shell were then defined averaging the results of 3 shells from a same scallop population. Due to the uncertainties from the backdating of analysed striae and those from the intershell comparison, the uncertainty in the date for mean shell profiles was estimated to range from 1 to 7 days.

Environmental monitoring database ([Ba]/[Ca]
) shell time series were compared with environmental variables that may influence Ba biogeochemical cycle at the SWI. The Bay of Brest (Roscanvel station) was mainly investigated because of the regular environmental monitoring and the long time series obtained by shell analyses (1998 to 2004). Regular measurements (2-3 days resolution) of Ba and Mn concentrations in dissolved (<0.6µm, Nucleopore) and particulate phases from bottom seawater (1m above the SWI) were also performed at the sampling site (Roscanvel) in 2000 (from February to December). Filtrated samples (dissolved seawater phase) were acidified in 2% HNO 3 (69-70% Suprapur, Merck) and diluted 30 times to determine concen-trations of dissolved elements. Two internal standards were added (Y and Bi) in diluted samples. Elemental concentrations (Ba, Mn, . . . ) were then determined by ICP-MS (X7 series, Thermo Fisher) by an external and internal calibration. Particulate samples (filters) were digested in closed vials (Savilex, PFA) by an acid mixture (1.5 ml HCl, 1 ml HNO 3 , 0.5 ml HF; suprapur quality) at 95 • C during one night. The digested samples were evaporated to dryness at 110 • C under a laminar hood to eliminate the matrix. The residues were dissolved in 2.5 ml of nitric acid solution (HNO 3 2%) and the volumes were adjusted to 11 ml. These samples were analyzed by ICP-AES (THERMO Optek Iris Advantage, Thermo Fisher) at the Royal Museum for Central Africa (Belgium). Samples were spiked with a mixture of internal standards (Au, Y) before its introduction into the spectrometer to compensate instrumental drift and instability. All these experiments were performed under clean conditions (clean vials, laminar flood, blank procedure. . . ).
([Ba]/[Ca]) shell time series were also compared with an environmental database: hydrological (S, T, suspended particulate matter), chemical (O 2 , NO − 3 , NO − 2 , NH + 4 , PO 3− 4 , Si(OH) 4 , particulate organic carbon (POC) or nitrogen (PON)) and biological parameters. Biological parameters include total phytoplankton biomass, as reflected by chlorophyll-a and abundance of phytoplankton species. These measurements were performed in the Bay of Brest at the "Sainte Anne" site near to Roscanvel sampling zone, as part of monitoring activities lead by the Intitut Universitaire Européen de la Mer (http://www.univ-brest.fr/IUEM/ observation/observation iroise.htm; Fig. 1). All these measurements were performed at a weekly resolution. Nutrient dynamics in SOMLIT station were previously demonstrated to reflect environmental conditions in Roscanvel . Phytoplankton composition (a hundred species) were monitored from 1998 to 2002 at Lanvéoc (48 • 18 N, 4 • 27 W), located near to Roscanvel, as a part of the littoral environment monitoring program lead by the Institut Français de Recherche pour l'Exploitation de la MER (IFREMER database, http://www.ifremer.fr/envlit/), and from 2003 to 2004, at the seawater surface (0-1 m) in a reference SOMLIT station (Service d'Observation en Milieu LITtoral: 48 • 22 N, 4 • 33 W) (Fig. 1). Phytoplankton identifications were performed in surface seawater (0-1 m) and every 15 days, except in 2004 (weekly resolution).

Statistical analyses
A statistical data treatment was performed to highlight environmental parameters influencing the occurrence and the amplitude of ([Ba]/[Ca]) shell peaks. ([Ba]/[Ca]) shell maximum events from their beginning to their end, usually last 20 days, whatever year studied. Averaged environmental parameters were thus examined over different periods: 4 weeks before-1 week after (−4 W+1,W), 3 weeks before-1 week after (−3 W+1 W), 2 weeks before-1 week after ) shell profiles exhibits a background ratio of 0.535±0.134 µmol/mol, and two significant enrichments from July to August (respectively, 3.80±0.74 µmol/mol the 20th of July, and 1.72±0.12 µmol/mol the 26th of August) (Fig. 2b, Table 1). This ([Ba]/[Ca]) shell profile demonstrates a high inter individual reproducibility among scallops of the same age for both background content and summer peak events with similar occurrence dates and amplitudes (Fig. 2a). The reproducibility of ([Ba]/[Ca]) shell ratios was previously demonstrated for 2-year old scallops from the Bay of Brest (Barats et al., 2007). Analyses of a 3-year old scallop from the same population (shells collected the same year and site) were also performed for the third year of the shell growth.      Differences in calibration steps can also induce differences in ([Ba]/[Ca]) shell ratios. Quantitative Ba analyses were performed with LA-ICP-MS using external calibrations with NIST glass standard reference materials at a minimum of 41 µg/g of Ba (Carré et al., 2006;Lazareth et al., 2003;Pearce and Mann, 2006;Stecher et al., 1996;Vander Putten et al., 1999. Such external calibration is probably not suitable to accurately measure shell Ba concentrations ranging from 1 to 10 µg/g, especially considering the different matrix properties and response between glasses and CaCO 3 shells when using some laser ablation units (Belloto and Mikeley, 2000;Barats et al., 2007). Based on these conclusions, external calibrations with CaCO 3 standards was thus preferred in this study and others (Belloto and Mikeley, 2000;Thébault, 2005;Gillikin et al., 2006;Barats et al., 2007). Finally, taking into account the accurate dating of this study, and differences in analytical methods and in bivalve species, all these studies agree with a common profile of ([Ba]/[Ca]) shell ratio in bivalves; widespread over the world a background ratio punctuated by sharp summer episodic peaks. This observation thus supports specific and ubiquitous processes involved in the increase of ([Ba]/[Ca]) shell ratio, which must be directly related to changing conditions in the scallop environment.  were compared with variations of Ba concentrations in seawater ( Fig. 4a-b). Variations of dissolved and particulate Ba exhibit concentrations ranging, respectively, from 45 to 100 nmol/l and 0.7 to 9 nmol/l, and averaging 50 and 2 nmol/l. The background ([Ba]/[Ca]) ratio in dissolved seawater in 2000 is about 5.2±0.5 µmol/mol  which is slightly different than the value reported in 2003 (3.8 µmol/mol) (Gillikin et al., 2008) . The ratio calculated in 2003 corresponds to only 3 measurements, which is considerably less than in 2000 (n=43) and may explain such difference. The major pool of Ba in the seawater at the SWI originates from the dissolved phase. Background ([Ba]/[Ca]) shell ratios in Mytilus edulis mussel shells were previously demonstrated to be directly related to the ([Ba]/[Ca]) sw ratios of the water in which they grew (Gillikin et al., 2006). For the same scallop species (Pecten maximus) and in the same sampling site (Bay of Brest), the partition coefficient D Ba was reported to be 0.18 in 2003 and similar to those obtained for Mytilus edulis mussels or Saxidomus giganteus clams in other coastal temperate ecosystems (Gillikin et al., 2006(Gillikin et al., , 2008 (Fig. 4ab) . This maximum of ([Ba]/[Ca]) shell ratio is 2.6 times higher than the background one, whereas dissolved seawater Ba concentration is 1.5 times higher than the background one. The increase of dissolved Ba concentration is probably not sufficient to explain the increase in skeletal ([Ba]/[Ca]) shell content. At the end of July, particulate Biogeosciences, 6, 157-170, 2009 www.biogeosciences.net/6/157/2009/ Ba concentration increases up to 5 times the background content. This significant increase of particulate Ba may provide additional inputs of Ba at the SWI. As bivalves are non specific filter feeders, these additional inputs of Ba-rich particles at the SWI, in the surrounding scallop environment, are supposedly ingested as food, in part digested, transferred to the extrapalleal fluid, and finally archived in the shell (Stecher et al., 1996;Vander Putten et al., 2000;Gillikin et al., 2006).

A proposed pelagic biogenic process as the initial cause of Ba-rich particles at the SWI and subsequent ([Ba]/[Ca]) shell maxima
The occurrence and the amplitude of summer ([Ba]/[Ca]) shell peaks were intensively examined in the Bay of Brest over 7 years (1998)(1999)(2000)(2001)(2002)(2003)(2004) because of a regular and complete monitoring of physiological (growth) and environmental parameters (hydrological, biological, chemical).

([Ba]/[Ca]) shell variation as a result of environmental changes:
The influence of the shell growth rate on ([Ba]/[Ca]) shell peaks is studied to take in account a potential external environmental control. Decreases of shell growth rate were previously considered to be mainly induced by lower seawater temperature or the occurrence of specific phytoplankton blooms Lorrain et al., 2000). The specific survey in 2000 supports rather supplementary dissolved and particulate Ba inputs at the SWI in summer, subsequently taken up by the bivalve to explain ([Ba]/[Ca]) shell maxima. There are different potential sources providing these Ba inputs at the SWI: either related to a benthic Ba recycling and remobilization processes, or pelagic Ba enrichments originating from anomalous ecological changes, or a coupling of both phenomena. (

[Ba]/[Ca]) shell variation as a result of a biogenic pelagic process:
First, the examination of hydrological conditions (seawater temperature and salinity) reveals no significant similarity neither with the occurrence, nor with the amplitude of maximum ([Ba]/[Ca]) shell ratios. In the Bay of Brest, the role of hydroclimatic events (flood and/or resuspension) was previously investigated in detail  and did not exhibit any influence on the shell growth and chemistry. Ba inputs due to such hydroclimatic events are thus improbable to explain these Ba particulate inputs at the SWI. Benthic release due to more reducing conditions at the SWI can also occur, but these should preferentially release dissolved components. More reductive condition at the SWI, as observed in a eutrophicated estuarine bay (Seine Bay), promote the benthic release of dissolved Mn which was demonstrated  to induce increasing shell Mn content in summer (Barats et al., 2008). In 2000, skeletal Mn concentrations are relatively constant, concurrent with constant seawater particulate Mn content and a low increase of Mn concentration in the dissolved phase (Fig. 4c). The dissolved oxygen concentrations (annual average: 6.1±0.5 ml/l; and summer average from mid-May to mid-September: 6.0±0.3 ml/l) are constant in the Bay of Brest, suggesting well oxygenated seawater. For other years, even if seawater Mn measurements are not available, archived shell Mn supports stable content during ([Ba]/[Ca]) shell maxima, contrary to benthic Ba remobilisation due to more reductive conditions. The origin of Ba inputs at the SWI is thus rather initiated by a pelagic biogenic process .
([Ba]/[Ca]) shell variation as a result of phytoplankton biomass dynamic: The phytoplankton dynamics are  hairs et al., 1980Bishop, 1988). The Chl-a concentrations in the seawater that are supposed to reflect phytoplankton biomass, exhibited a general pattern with an intense maximum occurring in spring and later ones somewhat smaller in late summer (Fig. 3) ratios. In 1998ratios. In , 1999ratios. In , 2001ratios. In and 2002, the first maximum of ([Ba]/[Ca]) shell occurs simultaneously or within a short time lag of dominant Chaetoceros blooms (Fig. 5). The second maxima are in turn found to mainly occur during Gymnodynium blooms (Fig. 5). In 2000, Chaetoceros exhibit 3 blooms, the most intense over the 7-year period (Fig. 5). This second maxima can only be associated to a maximum of ([Ba]/[Ca]) shell ratios, the lowest one over the 7-year period. In 2001, a Gymnodynium bloom, the most intense over 7-year period, occurs concurrent to ([Ba]/[Ca]) shell ratios close to the background concentration (Fig. 5). These 2 specific blooms cannot thus influence the amplitude or the occurrence of maximum ([Ba]/[Ca]) shell ratios. Neither specific species, nor these two genera (Chaetoceros and Gymnodynium), nor these two largest group of eukaryotic algae (diatoms or dinoflagellates), nor the total phytoplankton abundance are directly correlated with these (   in the seawater (r 2 >0.70, p<0.30, n=16) (  (Carter et al., 2005). Nutrient availability governs the amplitude of phytoplankton blooms and its composition Le Pape et al., 1996;Ragueneau et al., 2002). However, higher dissolved inorganic nitrogen in the seawater during the month preceding ([Ba]/[Ca]) shell maxima promotes the primary productivity (Carter et al., 2005). Higher phytoplankton biomass induces then the production of higher inputs of PON in seawater. Periods of higher summer productivity and post-bloom conditions are prone to amplify the maximum ([Ba]/[Ca]) shell ratio.

Processes involved in Ba-enrichment at the SWI that evidence subsequent ([Ba]/[Ca]) shell maxima
This study underlines the influence of an initial pelagic biological bloom process on the occurrence and the amplitude of maximum ([Ba]/[Ca]) shell . These processes are nonspecific to either phytoplankton species or genera, and rather related to post-bloom summer conditions. The episodic ([Ba]/[Ca]) shell maxima in P. maximus shells are supposed to be induced by a trophic uptake of supplementary particulate Ba inputs at the SWI. These Ba-enriched particles may originate either from scavenging of phytoplankton-derived particles or from a benthic post-bloom remobilization. A pathway leading to particulate Ba enrichment within the water column and scavenged to the SWI is most plausible (Ganeshram et al., 2003;Sternberg et al., 2005). Different pathways were proposed to explain Ba enrichment in particles at the SWI: such as Ba adsorption in phytoplankton cells, barite formation during phytoplankton decay, Ba enrichment in exopolymeric substances (EPS), or Ba adsorption onto mineral oxides formed within diatom biogenic particles.
Chaetoceros spp. blooms are usually dominant during the month preceding maximum ([Ba]/[Ca]) shell . The agglomeration of EPS is specifically reported during bacterial decomposition of these Chaetoceros spp. blooms, as promoted in response to nutritional stress. These conditions are indeed collated in this study, which systematically display depleted concentrations of nitrates and nitrites in seawater before the most significant maximum [Ba]/[Ca]) shell (Passow and Alldredge, 1995;Stecher and Kogut, 1999;Alldredge et al., 1995;Thornton, 2002). Both colloidal organic material and EPS are particularly enriched in Ba by five orders of magnitude (Quigley et al., 2002). Ba-rich EPS may thus support inputs of Ba-enriched particles at the SWI.
Among the different pathways such as Ba adsorption in phytoplankton cells, barite formation during phytoplankton decay, Ba enrichment in EPS, or Ba adsorption onto mineral oxides formed within diatom biogenic particles, the exact processes inducing Ba enrichment in particles at the SWI cannot be clearly identified. This study however supports the concept that a pelagic biogenic process initiates the delivery and subsequent bivalve uptake of Ba-enriched particles leading to incorporation into the individual striae of scallop shells.

Conclusions
Barium shell profiles from Great Scallops obtained in this methodical survey in temperate waters agree with those of other bivalve species previously investigated and demonstrate their reproducibility, recurrence and ubiquity in such coastal environments. This study confirms that ([Ba]/[Ca]) shell profiles are characterized by a relatively constant background mainly governed by the ambient seawater dissolved Ba, and the occurrence of distinct summer ([Ba]/[Ca]) shell maxima. In 2000, Ba measurements in seawater allowed identifying that particulate Ba inputs can be the dominant pathway explaining ([Ba]/[Ca]) shell maxima. Pelagic biogenic processes are supposed to initiate seawater Ba enrichment at the SWI, subsequently taken up by scallops and translated by increased ([Ba]/[Ca]) shell ratios. Examination of the complete dataset (1998)(1999)(2000)(2001)(2002)(2003)(2004) demonstrates that ([Ba]/[Ca]) shell maxima occurred under summer post-bloom conditions, such that their amplitude depends on nitrogen cycle: both particulate organic nitrogen and the turnover of dissolved nitrogen species. This whole dataset demonstrates that: (1) records of maximum ([Ba]/[Ca]) shell ratio cannot be used directly as a relevant paleo productive tracer, and (2) complex processes occur in the pelagic/benthic Ba cycle and are responsible for significant Ba inputs at the SWI. If these processes are better constrained, scallop ([Ba]/[Ca]) shell records could provide a proxy of Ba biogeochemistry with high temporal resolution in coastal environments.