PNNL

Abstract

Global climate and chemistry models routinely simulate dust emission, transport, and deposition to provide estimates of the role dust plays in climate and ecosystems. However, differences between these model simulations are substantial, with 10 estimates of global dust aerosol optical depth (AOD) that vary by over a factor of 5. Here we develop an observationally-based estimate of the global dust AOD, using multiple satellite platforms, in-situ AOD observations and four state-of-the-science global models over 2004 - 2008. We estimate that the global dust AOD at 550 nm is 0.033 ± 0.006 (2s), higher than the AeroCom model median (0.023) and substantially narrowing the uncertainty. The methodology used provides regional, seasonal dust AOD and the associated statistical uncertainty for key dust regions around the globe with which model dust 15 schemes can be evaluated. Exploring the regional and seasonal differences in dust AOD between our observationally-based estimate and the four models in this study, we find that emissions in Africa are often overrepresented at the expense of Asian and Middle-Eastern emissions, and that dust removal appears to be too rapid in most models.