Using Tropospheric Emission Spectrometer CO2 Observations for Inverse Modeling Estimates of Carbon Fluxes
S. Kulawik1, R. Nassar2, D. Jones2, J. Worden1, F. Irion1, K. Bowman1, T. Machida3, H. Matsueda4, R. Andres5, P. Suntharalingam6 and J.M. Chen7
1Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA 91109; 818-313-7123, E-mail: Susan.Kulawik@jpl.nasa.gov
2University of Toronto, Department of Physics, Toronto, Ontario, Canada
3National Institute for Environmental Studies, Tsukuba, Ibaraki, Japan
4Meteorological Research Institute, Tsukuba, Japan
5Oak Ridge National Laboratory, Oak Ridge, TN 37831
6University of East Anglia, Norwich, England
7University of Toronto, Department of Geography, Toronto, Ontario, Canada
We present carbon dioxide estimates and characterization from the Tropospheric Emission Spectrometer (TES) aboard the NASA Aura spacecraft, launched in 2004, with comparisons to aircraft and surface in situ data. TES CO2 is sensitive between about 200 and 800 hPa, with peak sensitivity in the mid-troposphere (511 hPa) and ~1.5 ppm accuracy for regional monthly averages. We compute terrestrial and ocean flux estimates with TES data and the GEOS-Chem chemical transport model, using a Bayesian inversion approach, and find that TES satellite observations of CO2 provide important constraints on flux estimates, particularly in the tropics.
Figure 1. TES comparison to validation data and versus GLOBALVIEW (circles) for March – May, 2006-2007. TES data monthly averaged over 15x15 degrees shows similar patterns to surface data and has about 1.5 ppm error.