Carbon-Weather Data Assimilation: Progress and Outlook

Global observations of column CO2 mixing ratios are now available from satellites. As CO2 is mixed, albeit incompletely, by winds and convection, the inference of the temporal and geographic distributions of surface fluxes from these observations is contingent on the modeling of the atmospheric circulation. We present a carbon-weather data assimilation system, wherein raw weather observations together with satellite observations of the mixing ratio of column CO2 are assimilated every 6 hours into the NCAR carbon-climate model CAM5 coupled to the Ensemble Kalman Filter of DART. In an OSSE, we reduced the fossil fuel emissions from a large country, and estimated the emissions innovations demanded by the atmospheric CO2 observations. The uncertainties in the innovation are analyzed with respective to the uncertainties in the meteorology to determine the significance of the result. As CO2 mixing ratios are defined relative to dry air mass, lack of local and regional air mass conservation translate directly to artifacts in CO2 and surface fluxes. We propose a test case to ascertain mass conservation in DA experiments.