Improvements in Profiler Wind Estimates Using Smoothed Data in the Spectrum Finder Algorithm

摘要

Wind profilers obtain the component of wind velocity parallel to the radar beam by calculating the first moment of the Doppler spectrum. An effective data processing algorithm uses peak detection to locate the subset of the spectrum containing the Doppler signal before the mean frequency is calculated. Simulations showed that smoothing the data before finding the peak yields better accuracy when the signal-to-noise ratio is marginal. Sensitivity studies considered the dependence of the improvement on signal shape, smoother shape, number of spectral points, width of the signal spectrum, and number of spectra in the average, all as a function of signal-to-noise ratio. The effect of the smoother can also be interpreted as improved sensitivity with no loss of accuracy when signal-to-noise ratio is low. Simulations of consensus averaging demonstrated that the smoothed finder improves both the accuracy and effective sensitivity.