New adaptive algorithm for radar turbulence detection in clouds and precipitation

The relationship between back-scattered signal parameters and singularities of microstructure and dynamics of the scatterers in a single radar resolution volume is a physical basis of the inverse problem solution for radar turbulence detection and measuring in clouds and precipitation. Usually, known algorithms for radar turbulence detection in clouds and precipitation are based on clear physical principles. They measure specific informative parameters of signals or their combinations and test them by threshold. There are some reasons limiting significantly the detection reliability. The time of measuring is always limited. Statistical characteristics of sample estimations essentially depend on a size of sample. Insufficient sample size of initial echo-signals for deriving consistent estimates of informative parameters is a significant factor frequently limiting the reliability of turbulence detection. Noise and interference, which can be changed during the time of measuring, influence the reliability of obtained information. The threshold of decision-making in parametrical detection algorithms should be varied according to the distance for false alarm probability stabilization; other way is the use of various automatic gain controls, which are, basically, heuristic solutions. Algorithms, constructed because of physical reasons, do not always appear as the best in real-life situations. We are less interested into real physical processes but more in statistical relationships between signal parameters and weather object characteristics. The purpose of this paper is to synthesize an adaptive algorithm for inverse problem solution, which should be invariant to noise power and operate with short samples, in order to increase turbulence detection effectiveness.