What can we learn from European continuous atmospheric CO2 measurements to quantify regional fluxes - Part 1: Potential of the 2001 network

An inverse model using atmospheric CO₂ observations from a European network of stations to reconstruct daily CO₂ fluxes and their uncertainties over Europe at 40 km resolution has been developed within a Bayesian framework. In this first part, a pseudo-data experiment is performed to assess the potential of continuous measurements over Europe using a network of 10 stations of the AEROCARB project such as in 2001 (<a href="http://www.aerocarb.cnrs-gif.fr/"target="_blank">http://www.aerocarb.cnrs-gif.fr/</a>). Under the assumptions of a small observation noise and a perfect atmospheric transport model, the reconstruction of daily CO₂ fluxes and in particular of their synoptic variability is best over Western Europe where the network is the densest. At least a 10 days temporal and a 1000 km spatial averaging of the inverted daily/40 km fluxes is required in order to obtain a good agreement between the estimated and the "true" fluxes in terms of correlation and variability. The performance of the inversion system rapidly degrades when fluxes are sought for a smaller temporal or spatial averaging.