Assimilation of EOS MLS ozone observations in the Met Office data-assimilation system

摘要

In this paper the impact of Earth Observing System Microwave Limb Sounder (EOS MLS) observations on ozone analyses is investigated using the Met Office data-assimilation system. EOS MLS was launched in 2004, and produces high-quality ozone observations in the upper troposphere and stratosphere at high vertical and horizontal resolution. The experiments shown here are run using 3D-Var and a forecast model that has 50 levels - from the surface to approximately 63 km - and a horizontal resolution of 2.5° latitude by 3.75° longitude. Most of the experiments are run for the period January-February 2005. The chief impact of assimilating EOS MLS data is a reduction of the mean analysis error, and of the standard deviation, in the lower stratosphere. Compared with control simulations, mean errors drop by 5%-25% in the Southern-Hemisphere extratropics, by around 10% in the Northern-Hemisphere extratropics, and by around 50% (with respect to ozonesondes) in the region of the tropical upper troposphere andlower stratosphere. Further investigation shows that the improved ozone analyses in the Southern Hemisphere largely result from a much better representation of summertime low-ozone events. These events have been documented in the northern summer stratosphere, but never before in the southern summer stratosphere. In the Northern Hemisphere the addition of EOS MLS to the assimilation system is shown to lead to a better representation of winter polar ozone depletion. At low latitudes, it appears that errors in the background model transport fields often lead to errors in the ozone analysis, but that the addition of high-density, good-quality EOS MLS data alleviates much of this error. It is noted that the EOS MLS data have a more beneficial impact on ozone analyses in this region than do data from the Michelson Interferometer for Passive Atmospheric Sounding, which has a similar resolution and observational error to EOS MLS.