Inputs, losses, dynamics and the chemical forms of macro- and micronutrients determine nutrient availability and influence both autotrophic and heterotrophic organisms in the ocean (Bruland et al., 2001). Metabolic processes, population and community dynamics and food web structures are all strongly dependent on nutrient availability. For example, a nutrient may be required for the functioning of a specific enzyme and/or metabolic pathway and thus exert considerable control on the species composition of marine communities. The marine osmotrophs (Karp-Boss et al., 1996) have specific characteristics enabling them to take up nutrients from the surrounding water and thus the concentrations in nutrients will strongly influence the composition of the species. This is particularly important where nutrient concentrations are low compared to the needs of the organisms especially in oligotrophic waters. Despite the complexity of the planktonic foodweb in the pelagic ecosystem, the ultimate constraints for autotrophic primary production are the abiotic factors: light and inorganic nutrients (De Baar, 1994). Indisputably, light variation is an important factor controlling the growth rate of algae. Light conditions at the sea surface may be approximated with simple calculations (Kirk, 1994), and changes in light conditions with depth may be seen as a consequence of biological activity in the open ocean. Assessing nutrient availability appears to be a key factor in understanding and eventually predicting the biological response (in terms of trophic structure and diversity, and/or in terms of carbon fluxes) in the upper photic zone (Moutin & Raimbault, 2002).

Reciprocally, planktonic organisms modify nutrient availability. Organisms can exhaust nutrients or introduce new nutrients. For example, the N2-fixing organisms di-azotrophes, quantitatively introduce “new” nitrogen (Capone, 2001) which could account for 50% of the total allochtonous nitrogen input into the ocean (Gruber and Sarmiento, 1997). Organisms can also produce organic ligands able to alter nutrient availability. Most of the iron in the iron-depleted environment is taken up in the organic form (Achilles et al., 2003). Organisms are responsible for the major nutrient fluxes in oligotrophic surface waters through organic matter recycling processes, remineralisation, excretion and sloppy feeding (Jumars et al. 1989). The continuous transfer and transformation from inorganic to organic substrates and back again, explains why biological processes drive almost all biogeochemical cycles. A better understanding of marine life requires the simultaneous investigation of biogeochemistry and marine food web structures. Whilst our understanding of the links between biological, physical and chemical factors that influence nutrient uptake and remineralisation in the ocean, is improving, it is not sufficient for producing realistic predictive models. A holistic approach on the impact of macro- and micronutrients on food web structure and function in different ocean regimes is therefore necessary (IMBER, 2005).

Understanding the transformation of organic matter in marine food webs is also essential for assessing the global carbon cycle and subsequently evaluating the impact of the anthropogenic CO2 input on climate change. Assessing the export of carbon from the photic zone to the deep layer is central in the debate on the oceans ability to absorb the anthropogenic CO2 excess (Longhurst, 1991). Oceanic oligotrophic areas represent more than 50 % of the global ocean and about 40 % of the total oceanic production (Antoine et al., 1996). However, the functioning and productivity of oligotrophic systems and particularly the balance between production and mineralization in these areas, is still the subject of much debate (Karl et al., 2003; Williams et al., in press). The role of oligotrophic areas in overall export is probably not very important, as the greatest proportion of photosynthesised carbon is recycled in the surface layer and rapidly re-exchanged with the atmosphere (Fig. 1). Nevertheless, recycling may lead to the accumulation of dissolved organic carbon (DOC) (Thingstad et al., 1997) and needs to be considered since DOC accumulation and export in oligotrophic areas, is potentially more important than transport via sinking particles (Copin-Montegut & Avril, 1993; Carlson et al., 1994; Avril, 2002). In the context of global warming, an increase in stratification has been predicted and thus the role of DOC in carbon export may change.

Fig. 1. Schematic representation of the major carbon fluxes in the photic zone of the ocean and the nutrients control of primary production (Moutin, 2000).

The Mediterranean Sea and the oligotrophic ocean (an overview of previous cruises):

Oligotrophic marine areas are characterized by a more or less, pronounced thermal stratification of the water column, which delimits (1) a warm surface mixed layer with high light intensity but depleted in nutrients and (2) a sub-superficial layer with low light levels and more nutrients. Tropical areas, as well as large anticyclonic gyres, the Sargasso Sea and the Mediterranean Sea, have long been considered as typical oligotrophic systems (Herbland & Voituriez, 1977). The depth where nitrate concentration approaches zero is around 10 m in the Alboran Sea, during the stratified period, and can reach more than 150 m in the Levantine basin of the Mediterranean Sea (Fig. 2). This is related to hydrological conditions and to, two major external sources of nutrients, the Rhône river input and the entry of the nutrient rich Atlantic surface waters, in the Western part of the Mediterranean Sea. The great depth of the Levantine Basin nitracline is only found elsewhere in ultra-oligotrophic conditions, for example, in the centre of the South Pacific gyre. It can be said that the Mediterranean Sea presents, on a regional scale, the main oceanographic features of contrasting environments in the oligotrophic ocean.

Fig. 2. Integrated Primary Production vs depth of the top of the nitracline. MS = Mediterranean Sea: data from Moutin & Raimbault (2002), MINOS cruise (Western basin : 7 stations, Ionian basin: 8 stations, Levantin basin: 4 stations). PO = Pacific Ocean: Unpublished data from the central station in the Gyre during the BIOSOPE cruise. AO = Atlantic Ocean: data from Steinberg et al. (2001) at BATS station.

The distribution of primary production, particulate carbon export, from the photic zone to deeper layer and nutrient concentrations were investigated in the first trans-Mediterranean cruise during May-June 1996 (MINOS cruise : Mediterranean INvestigation of Oligotrophic Systems). A decrease in integrated primary production, particulate carbon export and nutrient availability towards the eastern side of the Mediterranean Sea was observed, while integrated chlorophyll a remained constant. Integrated primary production reached 150 mgC m-2 d-1 in the Levantine basin, a value considered as a limit for primary production rates under strong oligotrophic conditions (Moutin & Raimbault, 2002).

It has long been suspected that photosynthetic production is limited by phosphate availability, in the Mediterranean Sea. Bioassays have shown that phosphate enrichments stimulate photosynthesis (Berland et al., 1980; Diaz et al. 2001). However, not only primary production, but also bacterial production may be controlled by phosphate availability. During the PROSOPE (PROductivité des Systèmes Océaniques PElagiques) and TMC (Trans Mediterranean Cruise) of September and June 1999, it was demonstrated that phosphate limitation on bacterial production, already observed in several locations (Thingstad 1998; Zohary and Robarts, 1998), was a general feature of the western and eastern Mediterranean Sea (Van Wambeke et al, 2002). Dissolved inorganic phosphate concentrations in the upper photic zone were shown to decrease from west to east reaching levels well below 1 nM (Moutin et al, 2002). Synechococcus spp., the most abundant phytoplankton in surface waters of the Mediterranean Sea during summer (Vaulot et al. 1996), were shown to have specific advantages concerning dissolved inorganic phosphate uptake that may explain their abundance in P depleted environments (Moutin et al, 2002). Adaptation to P limitation has also been demonstrated among higher trophic levels. For instance mixotrophic nanoflagellates obtain organic phosphate from their bacterial ingested prey using phagotrophy, stopping as soon as inorganic phosphate becomes available (Christaki et al. 1999). The study of tintinnid during the PROSOPE cruise showed that the diversity of this group reflects resource diversity rather than competitive interactions or predation (Dolan et al., 2002). Recent work has shown an increase in copepod egg abundance following phosphate addition to surface waters in the eastern MS (CYCLOPS), implying that they may be coupled to lower trophic levels through interactions which are not usually taken into account (Thingstad et al., 2005). These mechanisms are not fundamentally new, however there are very few studies particularily concerning the consequences for the oceanic carbon cycle when there is a possible short cut of primary production between nutrients and exported carbon. In the past, research on marine food webs tended to focus on either, the phytoplankton and microbial food web, or on zooplankton, fish and top predators. Marine food webs must be considered as integrated systems because perturbations at any point in these systems can propagate both up and down through the trophic levels. Nutrient limitation of organic production has been largely studied in the MS and although there is a consensus on the major control exerted by phosphate availability, nitrogen is scarce as well and the availability of silicic acid may play a central role in controlling the export of production (Leblanc et al., 2003). Biological diversity may reflect multiple organic production limitations, thus a multi-element approach is necessary to increase our understanding of marine food webs. Nutrient dynamics and its role in the variability of the stoechiometry of organic matter pools will be the central aspect of our biogeochemical study. There is abundant evidence of the uncoupling between nitrogen and phosphate cycles in the MS. (1) The Nitrate:Phosphate ratio is higher in the deep Mediterranean waters than in other oceans (2) The highest Nitrate:Phosphate ratios, which are higher than the Redfield ratio, are found in the sub-surface waters of both the western (Mc Gill, 1961, Mc Gill,1965, Raimbault and Coste, 1990) and the eastern Mediterranean (Krom et al. 1991, Moutin & Raimbault, 2002). (3) The high N:P ratios in the particulate fraction (Krom et al., 2005). (4) The phosphacline is deeper than the nitracline in the eastern part of the MS (Moutin & Raimbault, 2002). At the present time, there is no definitive explanation for this particular feature of the MS, also observed in the Sargasso Sea. Two largely different processes have been proposed to explain the typical NO3:PO4 ratios observed in deep waters. Firstly, there is the biological process of nitrogen (N2) fixation (Bethoux & Copin-Montegut, 1986; Bonin et al., 1989; Sachs & Repeta, 1999, Kerhervé et al., 2001; Pantoja et al., 2002) which may lead to nitrogen accumulation in deep waters, and secondly, chemical processes such as phosphate adsorption onto iron rich particles which lead to further P depletion in the Mediterranean Sea (Krom et al., 1991). As phosphate removal by adsorption from the water column did not represent a significant sink for phosphate in the MS (Herut et al., 1999; Ridame et al. 2003), nitrogen fixation appears to be the key factor in explaining the high NO3:PO4 ratios. Never-the-less very few measurements are available. Nitrogen fixation rates have been measured recently at the DYFAMED station in the western Mediterranean (MELISSA program). Whilst these rates were typically low, this biological process supplies significant new nitrogen which can balance the nitrogen biogeochemical budget and explain the high nitrate/phosphate ratio in deep waters (Garcia et al., 2006). Indeed, the role of N2 fixation in the marine nitrogen cycle has been undergoing increasing scrutiny and re-evaluation over the last decade, leading to increased estimates of its role in supporting oceanic new production (Karl et al., 2002). The discovery of marine diazotrophs, other than Trichodesmium spp. (Zehr et al., 2001; Montoya et al. 2004), has given a new dimension to the significance of nitrogen fixation in the ocean. If significant amounts of new nitrogen are introduced by small organisms, previously thought to recycle nitrogen, our conception of the functioning of oligotrophic systems needs to be revised (Garcia et al., en révision). d15N data from fossilised chlorophyll (MINOS cruise) provides geochemical evidence for extensive nitrogen fixation in the eastern Mediterranean (Sachs & Repeta, 1999). Thus, it becomes of great interest to describe and quantify the nitrogen input by nitrogen fixation as well as understanding the organisms responsible for this biogeochemical function. As the quantity of dissolved atmospheric nitrogen is inexhaustable, it is important to understand the control of N2 fixation. It seems that phosphate or iron availabilities are key factors in controlling these fluxes on a global ocean scale (Falkowski 1997, Karl et al. 2002). If nitrogen availability by nitrogen fixation is important in the Mediterranean where low phosphate availability is thought to be the key factor controlling this flux, then the control of new production, initially defined as the fraction of production associated with new nutrients (generally nitrate), should be defined starting from new phosphate (Dugdale and Goering, 1967). It is important to further our understanding of the phosphate cycle in surface waters, which is sparsely studied, (Benitez-Nelson, 2000 ; Karl, ASLO meeting 2006) in order to improve our understanding of oceanic production particularly in oligotrophic areas. The chemical element phosphorous 31P only exists in the water in the form of phosphate, organic or mineral, particulate or dissolved, and is non reducable under natural conditions. Thus the many complex reactions of oxydoreduction found in the nitrogen cycles are not found in the phosphate cycle (Moutin, 2000), so it is possible to envisage coupling with production and the establishment of a budget from a different angle.

The Mediterranean Sea has a wide range of oligotrophic conditions suitable for studying the transformation of organic matter in marine food webs during low new nutrient availability and provides a case study for observing the links between the C, N, P, Si and Fe-cycles. Comparisons between different systems along a longitudinal gradient of trophic status will provide new insights for identifying and understanding fundamental interactions between marine biogeochemistry and ecosystems.