Phytoplankton gathers organisms living in aquatic and marine ecosystems that can perform photosynthesis, the process transforming the energy included in the sunlight into organic matter. In marine environment, the limiting factors for phytoplankton growth are generally the necessary nutrients (for instance nitrate, phosphate, or micro-nutrients such as iron) and the availability of light (which explains their presence in the upper layers of the water column). Phytoplankton is at the basis of the food chain in the oceans, with zooplankton, fishes and predators composing the upper trophic levels. Thus, understanding phytoplanktonic cycles is important for marine resources issues. Photosynthesis also consumes carbon dioxide dissolved in the water, which has to be taken into account in the current context of climate change to estimate the role of the biology in the oceans in terms of CO₂ sources and sinks. Photosynthesis is performed by a variety of organisms (for instance, their size may vary by two orders of magnitude) and there are large spatial variations since the oceans experiment diverse conditions for organisms (upwelling systems rich in nutrients, subtropical gyres poor in nutrients, polar regions with dramatic changes in light availability). The surface of the oceans is very large and oceanographic scientific cruises largely undersample them with respect to the full understanding of the primary production processes and variability. Satellite ocean colour sensors, such as MERIS (launched by ESA in 2002), are a unique way of observing the phytoplankton in the ocean, through its proxy, the chlorophyll a concentration ([chl a] in mg m^-3), which is a ubiquitous pigment in phytoplanktonic organisms. They provide us with maps of [chl a] with a global coverage and they regularly revisit the same area.

Oceanic 3D coupled physical-biogeochemical models are powerful tools to study the behaviour of phytoplankton. The physical part of the model deals with the circulation of the ocean, and provides a realistic description of the currents, the salinity and temperature of the water column relying on physical equations. The biogeochemical part of the model deals with the description of the nutrient and light availability for the phytoplankton, the grazing pressure of zooplankton and other living pools. It is based on biological equations, mainly derived from laboratory experiments, reason why there is a need to calibrate the models. Modellers performed this parameterization for a certain number of models with available in situ data and their expertise.

The purpose of the ASSOCO project is to take advantage of both sources of knowledge: the model and the observations to get a deeper knowledge of the modelling of phytoplankton cycles. In a 3D ocean coupled physical-biogeochemical model, implemented on the North Atlantic at 1/4° and including six biogeochemical variables, three parameters (zooplankton grazing g, phytoplankton growth µ and mortality m, all expressed in day^-1) are assumed to be stochastic and have regional variations. Ensemble simulations (200 members lasting 30 days during the spring bloom) show that the phytoplankton concentration is sensitive to the parameterization, with strong spatial heterogeneity, combined to a nonlinear and non-Gaussian behaviour. Within the Kalman filter theory, parameter estimation can be done, in the framework of optimal estimate with Gaussian assumptions and reduced rank approximation, when the state vector is augmented with the uncertain parameters. Twin data assimilation experiments, using surface phytoplankton as observations, were performed either in the linear framework or introducing a nonlinear local transformation (anamorphosis). Nonlinear parameter estimation performed better than linear estimation: on the 39 estimated parameters, there is a reduction in the variance obtained with the nonlinear analysis, compared to the variance obtained with the linear analysis, except for 2 parameters. The variance reduction is better than 60% in 80% of these cases. The anamorphosis is also useful to define an objective error norm for the biogeochemical variables.

Next steps of the ASSOCO project will use real ocean colour data, measured with satellite sensors. Data from three main ocean colour sensors are merged in the ESA Globcolour database (www.globcolour.info) and provide maps of [chl a] since 1998. The methodology previously developed to estimate the regional values of three key biogeochemical parameters will be applied. Further work will include the comparison of values of estimated parameters with the literature and the impact of regional parameterisation on the modelling of primary production.