Assimilation of AIRS Radiances Affected by Mid- to Low-Level Clouds

An approach to make use of Atmospheric Infrared Sounder (AIRS) cloud-affected infrared radiances has been developed at Meteo-France in the context of the global numerical weather prediction model. The method is based on (i) the detection and the characterization of clouds by the CO2-slicing algorithm and (ii) the identification of clear-cloudy channels using the ECMWF cloud-detection scheme. Once a hypo- thetical cloud-affected pixel is detected by the CO2-slicing scheme, the cloud-top pressure and the effective cloudfractionareprovidedtotheradiativetransfermodelsimultaneouslywithotheratmosphericvariablesto simulate cloud-affected radiances. Furthermore, the ECMWF scheme flags each channel of the pixel as clear or cloudy. In the current configuration of the assimilation scheme, channels affected by clouds whose cloud- top pressurerangesbetween 600 and950 hPa are assimilatedoversea in addition toclear channels. Resultsof assimilation experiments are presented. On average, 3.5% of additional pixels are assimilated over the globe but additional assimilated channels are much more numerous for mid- to high latitudes (10% of additional assimilated channels on average). Encouraging results are found in the quality of the analyses: background departures of AIRS observations are reduced, especially for surface channels, which are globally 4 times smaller, and the analysis better fits some conventional and satellite data. Global forecasts are slightly im- proved for the geopotential field. These improvements are significant up to the 72-h forecast range. Pre- dictability improvements have been obtained for a case study: a low pressure system that affected the southeastern part of Italy in September 2006. The trajectory, intensity, and the whole development of the cyclogenesis are better predicted, whatever the forecast range, for this case study.