Automatic quality control of airborne Doppler radar data

Doppler radars mounted on aircraft platforms have proven useful for studying a variety of weather phenomena including tropical cyclones, convective initiation, tornadic thunderstorms, and mesoscale convective systems. In order to do a multi-Doppler synthesis with data from these platforms the data must be heavily edited to remove non-weather echoes such as second trip, sidelobe, ground, and low signal to noise returns. Some of these echoes can be removed automatically, but others require manual editing of individual sweeps and are time consuming to remove. This paper will present an algorithm to automatically flag spurious data collected from aircraft mounted radars using tunable interest maps and fuzzy logic and compute a probability of weather for each sampled volume. The method uses a combination of fields derived from the radar data including standard deviations, gradients, and derivatives of velocity and reflectivity, variation of spectral width, and an algorithm for edge detection. It will allow researchers to more quickly synthesize Dual Doppler radar data to derive the three dimensional wind fields and may someday be used in near real time to aid in making tactical decisions during research flights. Examples of this algorithm on data collected from a variety of field programs will be presented in order to show its flexibility.