Abstract
Recently, 95-GHz Doppler radars have been developed to document microphysical and dynamical properties of nonprecipitating clouds. RALI (airborne radar and lidar) plans to associate a dual-beam 95-GHz Doppler radar and a backscattering lidar on the same airborne platform. In optically thin clouds, data from both instruments (looking at nadir) will be available and can be combined to retrieve the microphysical characteristics of the cloud particles size distribution. But in dense clouds the lidar signal is extinguished over a few tenths of meters, and reflectivity may be biased by along-path attenuation. Thus the sampling strategy anticipated by RALI is the dual beam; that is, the radar operates along two viewing angles: nadir and 40° fore. The fact that the along-path attenuation acts in a different way along the two viewing angles is exploited by means of two algorithms (“stereoradar” and “dual beam”) allowing to retrieve the true (nonattenuated) reflectivity Z and the specific attenuation K.
In this paper, after recalling the principle of the two algorithms, the authors simulate the sampling of various typical clouds with RALI. The two algorithms are then applied to the simulations in order to evaluate their performances in the retrieval of true reflectivity and attenuation. The stereoradar algorithm retrieves Z fields very well and K fields less accurately. This is true even when a drizzle cell is embedded in the cloud. The dual-beam algorithm provides very good retrievals in the absence of drizzle, but its results are severely biased when there is drizzle. However, in such a situation the algorithm provides a diagnostic that two types of particles are present in the cloud.
It is argued that the retrieval of the true reflectivity and specific attenuation allows a determination of the liquid water content, the total droplet concentration, and the effective radius of the cloud.
Corresponding author address: Dr. Anne Guyot, CETP/CNRS/UVSQ, 10-12 av.de l’Europe, 78140 Vélizy, France.
Email: guyot@cetp.ipsl.fr
