State-of-the-art operational analyses and reanalyses, on time scales ranging from weather to seasonal, provide a substantially different depiction of prominent components of the tropical circulation. For example, differences on the order of 20% can be noted in the speed of the African Easterly Jet (AEJ), which controls the meteorology of the tropical northern Atlantic. Large discrepancies between analyses can also be detected in the atmospheric circulation over the tropical Pacific and in the analysis of tropical cyclones (TCs). In particular, observed TCs can be completely missed in the operational analysis.
Improvements stemming from a more accurate representation of the thermal structure of the atmosphere, as provided by the NASA Atmospheric Infrared Sounder (AIRS), are discussed. Results suggest that a greater impact from AIRS is obtained if data in presence of clouds are not entirely rejected, as happens when assimilating clear-sky radiances only. Assimilation of cloudy AIRS retrievals can consistently produce better TC analysis and forecast track than clear-sky radiances, at any given resolution, and on all basins. Analysis and forecast of extreme precipitation originated from TCs or monsoonal depressions are also affected by the strategy adopted to assimilate AIRS data. In particular, the effect of AIRS data on the low-level moist flow which is an essential pre-requisite for any type of flood-producing precipitation, will be discussed.