Comparison of model-simulated atmospheric carbon dioxide with GOSAT retrievals

Abstract

Global atmospheric CO2 distributions were simulated with a chemical transport model (GEOS-Chem) and compared with space-borne observations of CO2 column density by GOSAT from April 2009 to January 2010. The GEOS-Chem model simulated 3-D global atmospheric CO2 at 2o×2.5ohorizontal resolution using global CO2 surface sources/sinks as well as 3-D emissions from aviation and the atmospheric oxidation of other carbon species. The seasonal cycle and spatial distribution of GEOS-Chem CO2 columns were generally comparable with GOSAT columns over each continent with a systematic positive bias of ~1.0%. Data from the World Data Center for Greenhouse Gases (WDCGG) from twelve ground stations spanning 90oS-82oN were also compared with the modeled data for the period of 2004-2009 inclusive. The ground-based data show high correlations with the GEOS-Chem simulation (0.66≤R2≤0.99) but the model data have a negative bias of ~1.0%, which is primarily due to the model initial conditions. Together these two comparisons can be used to infer that GOSAT CO2 retrievals underestimate CO2 column concentration by ~2.0%, as demonstrated in recent validation work using other methods. We further estimated individual source/sink contributions to the global atmospheric CO2 budget and trends through 7 tagged CO2 tracers (fossil fuels, ocean exchanges, biomass burning, biofuel burning, net terrestrial exchange, shipping, aviation, and CO oxidation) over 2004-2009. The global CO2 trend over this period (2.1 ppmv/year) has been mainly driven by fossil fuel combustion and cement production (3.2 ppmv/year), reinforcing the fact that rigorous CO2 reductions from human activities are necessary in order to stabilize atmospheric CO2 levels.