Evaluation of the AIRS near‐real‐time channel selection for application to numerical weather prediction

Abstract

AbstractThe Atmospheric Infrared Sounder (AIRS) on board the National Aeronautics and Space Administration (NASA) Earth Observing System (EOS) Aqua satellite provides 2378 channels for each field of view of the instrument. As it is neither feasible nor efficient to assimilate all the channels in a numerical weather‐prediction system, a policy of channel selection has to be designed in this context. This paper attempts to assess the optimality of the selection of the AIRS radiance channels that are made available to the scientific community in near real time (hereafter called AIRS NRT) by the National Oceanic and Atmospheric Administration (NOAA) National Environmental Satellite Data and Information Service. This assessment is done by comparing this channel selection with a method preserving the information content of the instrument, the so‐called ‘global’ method. It turns out that although the selected channels are different and the information content as measured by the entropy reduction (ER) and the degrees of freedom for signal (DFS) is slightly smaller for the AIRS NRT channel set than for the ‘global’ set, both channel selections give similar results in terms of analysis error for temperature, humidity and ozone. The robustness of the results is then evaluated by varying the range of input parameters to the channel‐selection scheme, in particular the atmospheric training dataset on which the channel selection is based, and the background‐error covariance matrix. It is found that the performance of the ‘global’ channel selection is sensitive to the training dataset, while the AIRS NRT channel selection remains robust, even, to some extent, for the retrieval of key analysis‐error structures. Altogether, the ‘manually selected’ AIRS NRT channels provide a good compromise between robustness and quality. Copyright © 2003 Royal Meteorological Society.