The different observations shall be adequately combined during the analysis and data assimilation phase (there will be both near-real time data assimilation, and a reanalysis later on, when all the data will be available) to provide a continuous and realistic mapping at the eddy scale during the course of the seasonal cycle, and in particular in the late winter and spring season.
For physical oceanography, monitoring the currents requires on the order of 100 simultaneous velocity measurements in the domain to constrain the eddy circulation (but not the fronts). This can rely mostly on floats, but mooring data will be interesting to obtain a complementary Eulerian perspective in a few spots, while current profiles during cruises will provide snapshots of the spatial variability at finer scales, in particular near fronts. The temperature and density fields of the upper ocean, sampled at the eddy scale during the cruises, will be used to initialize numerical simulations. Finer surveys of the upper ocean, both of the density, T, S, current and fluorescence fields shall be carried to understand the 3-D dynamics at the finer scales, in particular in eddies or in the transition region. Time series of the stratification at a few locations are required to provide a time continuity, although it is not necessary to have them at the eddy scale, except in one or two well sampled structures. The evolution of spatially integrated temperature, for example, from a combination of altimetric and tomographic measurements, is required to provide an important comparison with the heat and buoyancy surface forcing.
Assimilation into eddy-resolving models
The evolution of the eddy structure during each cruise shall be analyzed and simulated using data assimilation in models (altimetry and n situ data), in order to provide a realistic view of the large scale and mesoscale circulation. Initialized with the fields measured and mapped during the first leg, these simulations shall be coupled with different biological models to study the biogeochemical tracer distributions, using estimates of parameters and/or fluxes obtained during the second leg with the time series stations. These simulations will be validated using data not considered in the implementation of the model (like surface data obtained during transit), including satellite data (sea color). Budgets and major fluxes (primary production, export production, air-sea gas exchanges) will then be estimated and the role of the coupling between dynamics and biological activity will be studied at meso-scale and regional scale. It is planned to use eddy-resolving basin scale models as well, which will integrate the information obtained during the cruises on an annual time scale for the North Atlantic ocean : in that respect, sea color data assimilation techniques will be developed.
The assimilation technique of Gaillard et al (1997, 1999) (and see in this volume the paper by Gaillard), based on a Kalman filter type of procedure, will be implemented to combine the various data sets and to obtain an optimal estimation of the evolving fields. The complementarity of the tomography technique and the profiling floats will be an essential component of the observing network.
The observational network
Tomography and eulerian measurements
Five tomography moorings will be deployed as shown on figure 3. The moorings in S1 and S2 will have the autonomous SARA receivers; JH1 and JH2 will be the new
