SEPARATION OF CLEAR-AIR ECHOES FROM PRECIPITATION ECHOES IN UHF WIND PROFILER MEASUREMENTS

For the automatic distinction between clear-air echoes and precipitation echoes in UHF wind profiler measurements, a new real time method is presented. Its reliability is tested by comparing the echoes classified as precipitation with independent precipitation measurements on the ground and in the troposphere. The results show that, compared with two other methods tested in the same way, the new method is the best. When applied to a dataset of 8 days, 93.3% of the echoes, which were subjectively classified as precipitation, were also classified as precipitation by the method, and only 0.1% of those echoes, which were subjectively classified as clear air, were classified as precipitation by the method. The correlation coefficient between precipitation measured by the UHF profiler and detected by the new method, and precipitation measured on the ground is about the same as the corresponding correlation coefficient found in an experiment in Japan, where precipitation detected with an off-line method from VHF wind profiler measurements was compared to ground-based rain gauge measurements. Although several case studies show that suspicious-looking horizontal profiler winds are frequently found during situations of precipitation, the editing of all profiler data measured when precipitation was detected by the new method did not decrease the root mean square differences between the profiler winds and winds measured simultaneously by regular radio soundings. We suppose this to be due to the lack of radiosonde ascents during convective precipitation when profiler winds derived with the Doppler beam swing technique are at their worst.