Bias errors of pulse pair estimation in a weather radar

High resolution wind speed profile measurements are needed in a weather radar to provide reliable information of rapidly changing weather conditions. The commonly used pulse pair method is quite attractive when processing an enormous amount of weather radar data in real time since it is considered the fastest algorithm available. The pulse pair estimator calculates the first two moments of the Doppler spectrum from estimates of the complex autocorrelation function at lag T/sub s/. The goodness of this estimator is typically determined by examination of the bias and the variance of the moment estimates. To analyze the bias in the pulse pair estimates, we consider the process autocorrelation function R(T/sub s/) expressed in terms of the true mean Doppler frequency f/sub d/. Unbiasedness of pulse pair estimates is based on the assumption that the spectrum is symmetric or narrow. However, the weather return Doppler spectrum is often broad and not symmetric thus causing a bias. This bias effect is analyzed here using a skewed Gaussian spectrum model with various spectrum widths.