1. INTRODUCTION In this paper we present a method for enhancing the clear air signal in archived upper-level velocity spectra. Typically the quality of VHF wind profiler radial velocity estimates deteriorate above about 10 km as a result of low signal to noise ratio. This method attempts to enhance the signal through spectral averaging (e.g. Merrit, 1995), half plane subtraction (Passarelli et al., 1981), and the combination of spectra from coplanar beams.

2. DATA AND TECHNIQUES The primary aim of this technique is to improve the signal in VHF spectra in the height range from 10-20 km. However, to provide a quantitative test for this method, we apply the method to UHF spectra for heights between 2 and 6 km, and compare the results to wind estimates from collocated VHF profilers. VHF profilers provide good velocity estimates in this height range and provide a reference to which the modified UHF spectra can be compared.

The spectral averaging phase of the method involves averaging velocity spectra over various averaging periods, combining spectra from coplanar beams with an assumption that the wind is uniform between the two beams, and averaging spectra over multiple heights. The half plane subtraction then removes symmetric signals such as ground and sea clutter, and interference.

3. RESULTS The UHF and VHF spectra are often found to contain ground clutter and sea clutter signals. At lower heights, the ground clutter signal is often stronger than the UHF clear air signal, while a sea clutter signal is often stronger than the VHF clear air signal. The ground clutter signal is significantly reduced using a half plane subtraction. The sea clutter signal is reduced by combining a particular beams spectrum with the reverse of the opposite coplanar beams spectrum and then performing a half plane subtraction. Combining a beams spectrum with the reverse of the opposite beams spectrum enhances any clear air signal that is uniform across the beam separation, while introducing a symmetric sea clutter and ground clutter signal that can be removed by a half plane subtraction.

The method increases the signal to noise ratio for spectra where there is already good clear air signal (e.g. below 3 km for the UHF spectra and below 10 km for the VHF spectra). It also typically improves the height coverage of good signal to noise ratio spectra by 1-2 range gates. For higher range gates, however, the method only occasionally enhances a clear air signal.

4. CONCLUSION This spectral averaging method has been shown to increase the number of spectra from which a clear air signal can be extracted. However, this method only succeeds if at least some of the averaged spectra contain a clear air signal. It should be combined with a quality control routine that can determine if a peak in the modified spectra is an actual clear air velocity peak.

REFERENCES Merritt, David A., 1995: A Statistical Averaging Method for Wind Profiler Doppler Spectra. Journal of Atmospheric and Oceanic Technology: Vol. 12, No. 5, pp. 985–995.

Passarelli, R. E., P. Romanik, S. G. Geotis, and A. D. Siggia, 1981: Ground clutter rejection in the frequency domain. Proc. 20th Conf. On Radar Meteorology, Boston, MA, Amer. Meteor. Soc., 295-300.