Estimating aerosol emissions by assimilating observed aerosol optical depth in a global aerosol model

• Main Authors: N. Huneeus, F. Chevallier, O. Boucher
• Format: Article
• Language: English
• Published: Copernicus Publications 2012-05-01
• Series: Atmospheric Chemistry and Physics
• Online Access: http://www.atmos-chem-phys.net/12/4585/2012/acp-12-4585-2012.pdf

This study estimates the emission fluxes of a range of aerosol species and one aerosol precursor at the global scale. These fluxes are estimated by assimilating daily total and fine mode aerosol optical depth (AOD) at 550 nm from the Moderate Resolution Imaging Spectroradiometer (MODIS) into a global aerosol model of intermediate complexity. Monthly emissions are fitted homogenously for each species over a set of predefined regions. The performance of the assimilation is evaluated by comparing the AOD after assimilation against the MODIS observations and against independent observations. The system is effective in forcing the model towards the observations, for both total and fine mode AOD. Significant improvements for the root mean square error and correlation coefficient against both the assimilated and independent datasets are observed as well as a significant decrease in the mean bias against the assimilated observations. These improvements are larger over land than over ocean. The impact of the assimilation of fine mode AOD over ocean demonstrates potential for further improvement by including fine mode AOD observations over continents. The Angström exponent is also improved in African, European and dusty stations. The estimated emission flux for black carbon is 15 Tg yr<sup>−1</sup>, 119 Tg yr<sup>−1</sup> for particulate organic matter, 17 Pg yr<sup>−1</sup> for sea salt, 83 TgS yr<sup>−1</sup> for SO<sub>2</sub> and 1383 Tg yr<sup>−1</sup> for desert dust. They represent a difference of +45 %, +40 %, +26 %, +13 % and −39 % respectively, with respect to the a priori values. The initial errors attributed to the emission fluxes are reduced for all estimated species.