Retrieving Vertical Air Motions of Tibetan Plateau Clouds Using Ka-band Cloud Radar Spectra

Clouds on TP (Tibetan Plateau) have important influences on weather and climate of the East Asian region, deriving vertical air motions can help us to understand the cloud dynamic and microphysical processes. During June to August in 2014, CAMS (Chinese academy of Meteorological Sciences) launched the third TIPMEX (Tibetan Plateau Meteorological Science Experiment) with several other departments, a Ka-band millimeter-wave cloud radar system which designed with solid-state transmitter and multiple detection modes was deployed at Naqu of TAR (Tibet Autonomous Region) for cloud and precipitation observation. This paper proposed the postprocessing method for this cloud radar spectrum, containing a velocity dealiasing method based on spectra characteristics of different modes; combining data of multiple modes, small particle trace method was utilized to derive vertical air velocity in TP's clouds. As crucial parts of this work, inversion results of Ac, As and Cb cases were analyzed, and deviation factors caused by noise level estimation methods and radar sensitivity descent were mainly discussed. The results show: 1) For this radar system, compromise among radar sensitivity, maximum detectable range and Nyquist velocity could be solved by using three different operation modes that applying pulse-compression technique, different pulse-width and different number of coherent accumulation. Rain mode with a higher Nyquist velocity are helpful for velocity dealiasing of cirrus mode; algorithm success rate reach 99.41%. 2) Qualitative analysis demonstate that the retrieval results are basically reliable; the updrafts in Ac and Cb are corresponded to the echo area of reflectivity enhancement, spectral width increasement and positive Doppler velocity. As general phenomenons on TP, updrafts in clouds (As, Ci) are weak before noon, but increase obviously in the afternoon as a result of ground thermodynamic effects. 3) For spectra noise level estimation, simulation results suggest that the segment method is more accurate than the objective method. The former is nearly unbiased, while, the later which has been used by major researchers will overvalue the noise level approximately 0-2 dB, leading to shift the retrieval results toward to the direction of downdraft and cause a non-ignorable retrieval bias for Cb and Ns clouds.4) Retrieval results may be overrated caused by the decline of radar sensitivity; the bias can increase along with the probe heights and concern with cloud type, spectral width and turbulence. Deviations can be neglected only in weak clouds such as Ci, As and Ac, but not in Cb or other stronger clouds.

Key Words：the 3th TIPMEX (Tibetan Plateau Meteorological Science Experiment), Ka-band state cloud radar, spectra postprocessing, vertical air motion speed retrievals