The ocean is the largest active reservoir of carbon on Earth and absorbs about 2.6 ± 0.5 Pg C yr-1 (Le Quéré et al., 2013). Carbon uptake is controlled by two mechanisms: the solubility and biological pumps (Figure I-4). The solubility pump is the process mentioned above: CO2 dissolves in the ocean and is sequestered to the ocean interior by water masses sinking at high latitudes. The biological pump is the transport of organic matter from the surface to the deep sea. It is considered that about 50% of the global Earth primary production occurs in the ocean (Field et al., 1998), despite the fact that it represents less than 1% of the global photosynthetic biomass (Antoine et al., 1996; Behrenfeld and Falkowski, 1997). Oceanic primary production has, therefore, a key role in carbon cycle and climate regulation.
The majority of carbon fixation in the surface layer is performed by pelagic phytoplankton, which uses CO2 and converts it to organic matter (OM) through photosynthesis (~50 PgC yr-1; Field et al., 1998). This reaction is powered by light and requires nutrients (nitrogen and phosphate being the main macro-nutrients), following the simplified photosynthetic reaction:
Nutrients + light + CO2,aq + H2O -> O2 + CH2O
In addition, the production of calcareous structures by many planktonic or benthic species in the ocean counteracts the CO2 sequestration during photosynthesis and calcium carbonate production represent about 0.8-1.4 PgCaCO3 yr-1 (Feely et al., 2004). Indeed, calcifying organisms use bicarbonate ions to build their skeleton and the production of calcium carbonate releases CO2:
Ca2+ + 2 HCO3- -> CaCO3 + H2O + CO2
The organic matter produced in the surface layers can be exported to the deep sea. However, heterotrophs (e.g. bacteria, flagellates) remineralise this OM in surface layers through respiration, consuming O2 and releasing CO2 back to seawater. Although most of the community respiration is due to bacteria, it must be stressed that part of the respiration is realised by autotrophs during both light and dark periods. It has been estimated that about 70% of the OM produced in the mixed layer is recycled while 30% is exported to the deep sea (Falkowski et al., 1998) where it is partially remineralised by bacteria. Finally, only 1-3% of the OM produced in the surface layer is definitely buried in the sediments (De La Rocha and Passow, 2007) while approximately 13-30% of the CaCO3 produced is ultimately stored in the sediment (Feely et al., 2004; Sarmiento and Gruber, 2006).
The surface ocean is not homogeneous in terms CO2 exchanges with the atmosphere. The solubility capacity depends on the ocean surface temperature as CO2 dissolves more in cold than warm waters. In addition, the potential for carbon sequestration also depends on the metabolic status of the plankton community in the surface mixed layer, controlled by the balance between community gross primary production (GPP) and respiration (CR), i.e. the net community production (NCP) defined as the production of organic matter after it has been respired by all plankton communities (NCP = GPP – CR, if CR is expressed as a positive process). An ecosystem is autotrophic, and potentially a CO2 sink for the atmosphere, when GPP exceeds CR (NCP > 0). Conversely, in a heterotrophic system, CR exceeds GPP (NCP < 0) with potentially a source of CO2 for the atmosphere. Primary production and respiration can push ecosystems towards being CO2 sinks or sources. However, the ecosystem metabolic state does not always imply a air-sea CO2 flux as it depends on the CO2 partial pressure at the air-sea interface (Gattuso et al., 1998).
The capacity for producing organic matter in the surface layer depends on environmental conditions such as temperature, water-column structure (mixed VS. stratified), irradiance and nutrient availability. In the classical plankton food-web, described as the production of phytoplankton species grazed by zooplankton which are subsequently consumed by higher trophic levels, high nutrient concentrations are required. In addition to this food web, the microbial loop describes the use of dissolved organic carbon (DOC) released by phytoplankton (about 10-15 % of the particulate primary production; Baines and Pace, 1991) and zooplankton, as a substrate for bacterial growth and leading to the recycling of nutrients. Classic food-web and microbial loop exist as a continuum of trophic structure and the predominance of one path relative to the other depends on the nutrient availability (Legendre and Rassoulzadegan, 1995) that also influence the metabolic balance of the ecosystem. The remineralisation by bacteria is also subject to temperature control by an inverse function (Rivkin and Legendre, 2001) and therefore influences scope for carbon export to the deep-sea. Finally, the capacity for carbon sink also depends on the phytoplankton community composition as phytoplankton species with calcium carbonate (e.g. coccolithophores) or silicate (e.g. diatoms) structures have better ballast properties and are fast-sinking particles (Klaas and Archer, 2002).
The evolution of plankton community in the Anthropocene
The ongoing environmental perturbations such as ocean acidification and warming could have profound effects on the functioning of plankton community. As a result, the strength of the biological pump could be affected, thereby altering the carbon storage capacity of the ocean Temperature exerts a positive effect on phytoplankton metabolic rates as observed in laboratory culture (Eppley, 1972) and at sea (Regaudie-de-Gioux and Duarte, 2012), although a recent study suggests that phytoplankton carbon-specific production rates mostly depend on nutrient supply rather than on temperature (Marañon et al., 2014). However, photosynthesis and respiration exhibit different sensitivities to temperature. Phytoplankton growth and photosynthesis are less affected by a temperature increase (irradiance and nutrient availability exert tighter controls) than bacterial and heterotrophic respiration (López-Urrutia et al., 2006).
Thus, warmer conditions should increase respiration and therefore altering carbon cycling by promoting heterotrophy.
However, experimental evidence for this process is still missing and contradictory results have been reported. For instance, in Kiel Fjord (Germany) the effect of temperature has been investigated during two mesocosm experiments. During the first experiment, enhanced respiration was measured in warmer treatments, diminishing the CT drawdown. Additionally, a shift toward a larger accumulation of dissolved organic carbon (DOC) (Wohlers et al., 2009) and higher C:N ratio of the dissolved organic matter (DOM) in warmer treatments (Engel et al., 2011) were found. In contrast, in the second experiment, CT uptake as well as particulate organic carbon (POC) and DOC increased in the warmer treatments (Taucher et al., 2012). The differences in CT drawdown during these two experiments were attributed to the different species of diatoms present in the community (Skeletonema costatum vs. Dactyliosolen fragilissimus), but could also be due to differing irradiance and temperature levels.
Plankton species have different metabolic thermal optima and a rapid change of average temperature could cause shifts in the community structure with some species benefiting from warmer conditions and adapting better than others (Lürling et al., 2013). Mesocosms and in situ data show that small species are favoured under warmer conditions (Sommer and Lengfellner, 2008; Morán et al., 2010; Peter and Sommer, 2012; Daufresne et al., 2009). This would have consequences on carbon export efficiency, as some phytoplankton species (e.g. diatoms and coccolithophores) have better ballasting properties than others. In addition, not all phytoplankton have the same food quality (lipids content and stoichiometric ratios) and therefore energy transfer capacity to higher trophic levels (zooplankton, fishes) differs (Dickman et al., 2008) with, for example, diatoms (large species) presenting better food quality than cyanobacteria (Müller-Navarra et al., 2000).
A full understanding of the effect of ocean warming on the plankton community, requires to consider both the direct effect of increased temperature on metabolic rates, and the indirect effect due to nutrient depletion in surface layers, as well as increased irradiance at high latitudes, caused by stronger stratification (Behrenfeld et al., 2006; Lewandowska et al., 2014). Indeed, satellite observations reported a decline of ~ 1% of the global median per year in surface plankton biomass during the last decade (Boyce et al., 2010). However, faster nutrient remineralisation by bacteria could offset the decrease in phytoplankton biomass by earlier bacterial activity peak after phytoplankton bloom, that tighten the coupling between phytoplankton and bacteria (Hoppe et al., 2008). Temperature is recognized as a major parameter controlling plankton community structure and dynamics and there are still uncertainties on how the plankton community will evolve in the future warmer ocean. Finally, it is very likely that, as the efficiency of the carbon pump and its evolution in a warmer ocean seems to be closely related to nutrient regime and community composition (Boyce et al., 2010; Taucher and Oschlies, 2011), important regional variations will be observed in the coming decades.
Effect of ocean acidification
Higher levels of CO2 in seawater lead to an ocean acidification, an environmental perturbation that could also affect phytoplankton metabolism. Phytoplankton acquires dissolved inorganic carbon to produce organic matter, using CO2 as a substrate for photosynthesis despite the low affinity for CO2 of rubilose-1,5-bisphosphate carboxylase oxygenase (RuBisCO), a key enzyme involved in photosynthetic carbon fixation (Raven and Johnston, 1991) and limited diffusion capacities of CO2 in water. To overcome these limitations, marine phytoplankton have developed carbon concentrating mechanisms (CCMs; Giordano et al., 2005) to raise CO2 concentrations in the vicinity of RuBisCO involving carbonic anhydrase activities or bicarbonate transport through the cell for example (Reinfelder, 2011). An increase in CO2 would increase the diffusion of CO2 in the cells and may therefore benefit primary producers by lowering the energy cost of carbon acquisition, increasing CO2 diffusion through the cell membranes (Raven et al., 2005) and reducing CO2 leakage (Rost et al., 2006). Different CCMs have been developed by phytoplankton species varying within and among groups, it is therefore expected that organisms will be differently affected by CO2 increase, potentially causing shifts in the plankton community composition (e.g. Rost et al., 2008).
Single cells cultures
The majority of early laboratory experiments performed on single species have shown enhanced carbon fixation (e.g. Buitenhuis et al., 1999; see Riebesell and Tortell, 2011 for comprehensive review on this section) however some species, such as coccolithophore strains, have shown a neutral (e.g. Langer et al., 2006) or inhibitive effect under nitrate limitation (e.g. Sciandra et al., 2003). Coccolithophores have been more studied than other species for the impact of ocean acidification on calcification, with decreases in calcification rates (e.g. Riebesell et al., 2000) observed in most of the studies, although contradictory results showing neutral or enhanced calcification have also been reported (e.g. Iglesias-Rodriguez et al., 2008). Differences in the strains used have been shown to be critical and repetitions of experiments on some strains have not always led to the same results (see Riebesell and Tortell, 2011 for details).
The contradictory results obtained at the species level indicated that extrapolation from monocultures to assemblages is not straightforward. The investigation of the effect of ocean acidification at community level is therefore necessary.
Initial experimental work at the community level has reported an increase (~ 15 %) of 14C fixation under high PCO2 conditions (Hein and Sand-Jensen, 1997) in the South Atlantic Ocean. The first large mesocosm experiments (> 10 m3) performed in the North Sea (PeECE 2001, 2003 and 2005) have shown different responses. Only one of these three experiments have shown an increase in primary production under high PCO2 conditions (Egge et al., 2009) while no change in primary production was found in the first and second experiments (Delille et al., 2005 for PeECE I; unpublished data for PeECE II, see Egge et al., 2009). Other experiments carried out in different oceanic regions, with different incubation volume, have led to increased primary production (e.g. Tortell et al., 2008) or to no effect (e.g. Tortell et al., 2002; Yoshimura et al., 2013). The recent Svalbard mesocosm experiment showed no clear trend in net community production and community respiration over the whole period (Silyakova et al., 2013; Tanaka et al., 2013), but found decreased NCP when considering only the post-bloom period. However, 14C carbon fixation increased with increasing PCO2 levels (Engel et al., 2013). In this experiment, the community response was not straightforward and possibly related to the change in community composition over the different phases (before and after nutrient addition). However, the different methods for measuring primary production did not show the same results, so the conclusions of this experiment should be regarded with caution, although they reveal the complexity of the community metabolic response to ocean acidification under different physiological states conditioned by nutrient availability.
Table of contents :
Chapter I-Introduction to the plankton community in the Anthropocene
1. The Anthropocene
2. Carbon pump
3. The evolution of plankton community in the Anthropocene
3.1 Effect of ocean warming
3.2 Effect of ocean acidification
3.2.1 Single cells cultures
3.2.2 Community studies
3.3 Combined effect of warming and acidification
4. Oligotrophic areas under anthropogenic perturbation
5. Objectives and experimental approaches followed in this thesis
Chapter II-Ocean acidification and plankton metabolism in LNLC
1. Context of mesocosm experiments
1.1 Mesocosms acidification and sampling
1.2 Main results of Corsica mesocosm experiment
1.3 Main results of Villefranche mesocosm experiment
2. No effect of ocean acidification on planktonic metabolism in the NW oligotrophic Mediterranean Sea: results from two mesocosm studies
2.2 Material and Method
2.2.1 Study sites and experimental set-up
2.2.3 Oxygen light-dark method
2.2.4 GPP-18O method
2.2.5 14C primary production
2.2.6 Data analysis, statistics and data availability
2.3.1 Summer conditions (Bay of Calvi)
2.3.2 Winter-spring conditions (Bay of Villefranche)
Chapter III-Carbon 13 labelling studies and biomarkers analysis on Mediterranean plankton communities
2. Carbon-13 labelling studies show no effect of ocean acidification on Mediterranean plankton communities
2.2 Material and Method
2.2.1 Study sites, experimental set-up and sampling
2.2.2 Laboratory analysis
2.2.3 Data analysis
2.3.1 Bay of Calvi
2.3.2 Bay of Villefranche
Chapter IV-Combined effects of temperature and pCO2 increase on a plankton community
Effect of ocean warming and acidification on a plankton community in the NW
Chapter V-Synthesis and general discussion
Conclusions and perspectives