Extraction of cloud parameters using multipeak analysis of cloud radar data

An advanced real time retrieval algorithm for cloud radar target classification has been applied to radar data. The Doppler spectra are decomposed into multiple peaks by a peak detection algorithm. Every peak is characterized by 4 parameters: its first three moments (power, Doppler velocity and width) and the peak-specific LDR (Linear Depolarization Ratio).

Based on this data a coarse separation into hydrometeors and "non-meteorological" targets is done. That means basically that echoes from insects, seeds (so called atmospheric plankton) and birds are distinguished from clouds and precipitation. This is a necessary task to detect the cloud base level in the atmospheric boundary layer particularly in the hot season, during which atmospheric plankton echoes can be the dominating signal. Reliable individuation of plankton is necessary for a K-Band cloud radar, which has been used here, while higher frequency bands are less affected by it, but at the same time they suffer from more attenuation in clouds.

The hydrometeor echoes are assigned either to "cloud/drizzle" or "rain" depending on a threshold fall velocity. In case of "rain" the DSD, LWC, and rain rate is calculated using the well known fall velocity versus size relation assuming zero vertical wind velocity. More details of this preliminary classification algorithm are described by Bauer and Goersdorf (this conference).

The preliminary analysis of data from heights above the melting layer showed the presence of multiple peaks characterized by well separated values of LDR. That can be explained by mixed phase hydrometeors. Every type of hydrometeor has specific fall velocity and LDR depending on its phase, size and shape. For example coexisting cloud droplets and snow flakes would create the observed signal characteristics. Further steps towards unambiguous interpretation is to integrate this enhanced analysis in a multi-sensor system including lidar and microwave radiometer measurements and a cloud resolving model.

In order to assess the feasibility of this hydrometeor classification for inferring mixed phase ratios, measurements will be continued together with other remote sensing systems as well as in-situ aircraft measurements of cloud microphysics within the upcoming COPS experiment (Convective and Orographically-induced Precipitation Study) taking place in Southern Germany in summer 2007 (http://www.uni-hohenheim.de/spp-iop).