"Inter annual variations of net ecosystem carbon exchange (NEE), driven by weather and climate, are known to contribute to variation in the annual growth rate of atmospheric CO2 globally, but little is known about the underlying processes in each region. Better ability to predict future atmospheric CO2 levels and climate change should thus be gained by understanding the current inter annual variation in the biotic response to environmental forcing. In this context, I propose to assess, using a state of the art data assimilation approach, the impact of extreme climate spells, such as the summer 2003, versus the one of “normal” year-to-year climate variations on the processes which control the European carbon balance. This knowledge will be propagated into better future projections of the European carbon balance. Overall, the scientific objectives of the project will be primarily investigated using an parameter optimisation approach. Estimation of the process-based ecosystem parameters using radiation, sensible heat, latent heat, and CO2 exchanges at several ecosystem sites (forest) will help to improve the simulation of inter annual variations in terrestrial CO2 fluxes and to assess the potential of the ORCHIDEE model. The overall objectives of this proposal are threefold: (1) Optimisation of all critical parameters of ORCHIDEE and assessment of the limits of this process-based model to simulate eddy covariance measurements. (2) Identification of drivers and processes determining inter annual variations of forest carbon fluxes. (3) Improved model future projections of Europe’s carbon balance, with uncertainty estimates, using the optimized parameters."