Global O3-CO Correlations in the Global Modeling Initiative Chemical Transport Model During July - August: Sensitivity to Emissions and Input Meteorological Data

As a useful indicator of O3 production efficiency, the correlation between O3 and CO has been used to evaluate the capability of chemical transport models (CTMs) to produce proper O3 levels from its precursors for right reasons. We evaluate here the capability of the Global Modeling Initiative (GMI) CTM to reproduce the O3-CO correlations observed by Tropospheric Emission Spectrometer (TES) aboard the NASA EOS Aura satellite during July-August and examine the sensitivity of model O3-CO correlations to emissions and input meteorological data. The July-August O3 concentrations simulated with GMI CTM driven by the fvGCM (1994), GEOS4 (2005), and MERRA (2005) meteorological data sets reasonably agree with ozone climatology from ozonesonde measurements and tropospheric ozone columns determined from satellite observations (TOMS/SBUV, OMI/MLS, and TES). The model O3-CO correlations show strong positive correlations in major continental outflow regions as observed by TES (at 618 hPa). However, the correlations are negative over the tropical Atlantic Ocean and Asian continent where deep convection and stratospheric influence occur, respectively, as indicated by radionuclide tracers (radon-222, lead-210, and beryllium-7). Preliminary results with GMI/MERRA simulations show that biomass burning emissions change the sign of correlations from negative to positive in the Southern Hemisphere mid-latitude regions. Fossil fuel emissions dominate the strong positive correlations over the North American and Asian continental outflow regions. Biogenic emissions have relatively small impact but reduce the correlations over the tropical western Pacific Ocean. Lightning NOx emissions change the correlations from weak positive to negative and significantly reduce the positive correlations over the Middle East, Northern Hemisphere mid-latitude, and Southern Hemisphere subtropics. Discrepancies in the O3-CO correlations simulated with the above three meteorological data sets will be discussed.