Precision of tropospheric/stratospheric winds measured by the Chatanika radar

Abstract

A method for reducing tropospheric/stratospheric radar backscatter measurements to wind components and for relating uncertainties in these components to uncertainties in Doppler backscatter moments estimation is described. The centroid of the Doppler spectrum is used to estimate the wind component along the radar line of sight, while the variance of this estimator is used to judge the precision that can be attached to the measurement. Error estimates are given in geographical coordinates so that errors in resolved velocities can be related directly to radar measurements. To assess the validity of the procedure, wind profiles derived from Chatanika radar measurements are compared with wind profiles obtained from a series of 12 windsondes. These windsondes were launched nearby (at approximately 45‐min intervals) on April 6 and 7, 1978. Radar measurements include selected horizontal wind profiles (from about 4‐ to 14‐km altitude) and winds at 9.05 km extending over 15 hours. Windsonde and radar‐derived velocities generally agree within the precision of the measurements, but in regions of strong wind shear, differences exceeding the calculated error bars are found. These discrepancies are shown to be caused by several independent scattering layers that often exist within the scattering volume. When these layers have sufficiently different velocities, causing multiple peaks to appear in the Doppler spectrum, and when the relative strengths of these spectral peaks change with azimuth, the centroid no longer accurately reflects the mean motion of the scatterers. Significant errors in resolved velocities can thus occur. Because regions of high reflectivity apparently exist in thin layers separated by regions of low reflectivity, radar backscatter will provide an accurate sampling of the wind field, only if the scattering layers can be accurately located. To overcome these limitations in accuracy, radars that are used to measure tropospheric/stratospheric winds should have range resolutions comparable with the thickness of the scattering layers.