Estimating high latitude carbon fluxes with inversions of atmospheric CO₂

Atmospheric inversions have proven to be useful tools, showing for example the likely existence of a large terrestrial carbon sink in the northern mid-latitudes. However, as we go to smaller spatial scales the uncertainties in the inversions increase rapidly, and the task of finding the distribution of the sink between North America, Europe and Asia has been shown to be very difficult. The uncertainty in the fluxes due to network selection, transport model error and inversion set up tends to be too high for studying either net annual fluxes or interannual variability on spatial scales such as the North American Boreal or Eurasian Boreal regions. We discuss the path forward; to couple together the atmospheric inversions with process based terrestrial carbon models, creating carbon data assimilation systems. Such systems are being developed now and could prove to be very powerful. The multi-disciplinary nature of the data assimilation system requires information from flux towers, soil and above ground biomass inventories, remote sensed fields, atmospheric CO ₂ concentrations and climate data as well as model development and will need a massive community effort if it will succeed.