This paper describes a numerical interpretation of the April 2007, CarboEurope Regional Experiment Strategy (CERES) campaign, devoted to the study of the CO2 cycle at the regional scale. Four consecutive clear sky days with intensive observations of CO2 concentration, fluxes at the surface and in the boundary layer have been simulated with the Meso-NH mesoscale model, coupled to ISBA-A-gs land surface model. The main result of this paper is to show how aircraft observations of CO2 concentration have been used to identify surface model errors and to calibrate the CO2 driving component of the surface model. In fact, the comparisons between modelled and observed CO2 concentrations within the Atmospheric Boundary Layer (ABL) allow to calibrate and correct not only the parameterization of respired CO2 fluxes by the ecosystem but also the Leaf Area Index (LAI) of the dominating land cover. After this calibration, the paper describes systematic comparisons of the model outputs with numerous data collected during the CERES campaign, in April 2007. For instance, the originality of this paper is the spatial integration of the comparisons. In fact, the aircraft observations of CO2 concentration and fluxes and energy fluxes are used for the model validation from the local to the regional scale. As a conclusion, the CO2 budgeting approach from the mesoscale model shows that the winter croplands are assimilating more CO2 than the pine forest, at this stage of the year and this case study.