Tellus: Series A, Dynamic Meteorology and Oceanography (Oct 2015)

Analysis of tropospheric ozone and carbon monoxide profiles over South America based on MOZAIC/IAGOS database and model simulations

  • Marcia A. Yamasoe,
  • Bastien Sauvage,
  • Valerie Thouret,
  • Philippe Nédélec,
  • Eric Le Flochmoen,
  • Brice Barret

DOI
https://doi.org/10.3402/tellusb.v67.27884
Journal volume & issue
Vol. 67, no. 0
pp. 1 – 29

Abstract

Read online

We analysed ozone and carbon monoxide profiles measured by commercial aircrafts from the MOZAIC/IAGOS fleet, during ascending and descending flights over Caracas, in Venezuela, from August 1994 to December 2009, over Rio de Janeiro, from 1994 to 2004 and from July 2012 to June 2013, and over São Paulo, in Brazil, from August 1994 to 2005. For ozone, results showed a clean atmosphere over Caracas presenting the highest seasonal mean in March, April and May. Backward trajectory analyses with FLEXPART, of case studies for which the measured concentrations were high, showed that contributions from local, Central and North America, the Caribbean and Africa either from anthropogenic emissions, biomass burning or lightning were possible. Satellite products as fire counts from MODIS, lightning flash rates from LIS, and CO and O3 from Infrared Atmospheric Sounding Interferometer and wind maps at different levels helped corroborate previous findings. Sensitivity studies performed with the chemical transport model GEOS-Chem captured the effect of anthropogenic emissions but underestimated the influence of biomass burning, which could be due to an underestimation of GFEDv2 emission inventory. The model detected the contribution of lightning from Africa in JJA and SON and from South America in DJF, possibly from the northeast of Brazil. Over São Paulo and Rio de Janeiro, GEOS-Chem captured the seasonal variability of lightning produced in South America and attributed this source as the most important in this region, except in JJA, when anthropogenic emissions were addressed as the more impacting source of ozone precursors. However, comparison with the measurements indicated that the model overestimated ozone formation, which could be due to the convective parameterisation or the stratospheric influence. The highest ozone concentration was observed during September to November, but the model attributed only a small influence of biomass burning from South America, with no contribution of long-range transport from Africa.

Keywords