Determinants of NEE and NBP of croplands in Europe

Data sets of biometeorological measurements of ecosystem CO2 flux, combined with harvest and manure data from several European cropland sites encompassing different climatic regions and contrasting crop rotations were integrated to provide an assessment of the carbon budget. The influence of different crops and management practices was also assessed to identify some of the major factors contributing to the cropland carbon balance. In a first analysis, it was shown that NEE mainly correlated with length of the cropping period. In a second analysis NBP sums were correlated to soil properties. Finally, the data were combined with additional data to derive a mean annual GHG balance for the European cropland sites under consideration. Five crop rotations and two monocultures were integrated over 4 years. During 4 years the average annual NEP was -240 ± 178 g C m-2 y-1. On average, 382 ± 185 g C m-2 y-1 were harvested, while average carbon inputs by manure and seeding was 47 ± 50 g C m-2 y-1. The average NBP of the seven sites under consideration was estimated to be a carbon loss of 95 ± 203 g C m-2 y-1. The full GHG balance of the considered sites was estimated to be 160 g C m-2 y-1 in CO2-equivalents These results challenge current good practice guidelines that predict neutral carbon budgets for systems where the inputs of manure and crop residues are of comparable magnitude to those associated with the sites examined in this study. Ongoing humus loss in spite of good practice is mainly related to soils with high carbon concentrations which are not in equilibrium but may also be a result of already ongoing climate change. A modification in the good practice guidelines to increase carbon inputs may be required. Uncertainty and the representativeness of the results are discussed. Results from a representativeness analysis suggest that more than 50 sites are necessary for a European cropland flux network to adequately represent the variability of climate, soil and management within the European continent. Thus, the uncertainties due to the network design are currently bigger than the uncertainty intrinsic in the measurement method.