A Fuzzy Logic Based Algorithm for Performing Automatic Quality Control of Airborne Doppler Radar Observations

The removal of spurious echo from airborne Doppler radar observations is time-consuming and difficult. It is estimated that the manual editing of 10 minutes of airborne data can take as long as 240 hours. Radar data editing also requires experience to be able to identify non-weather radar echoes such as side lobes, ground return, and second trip echoes. These echoes are difficult to remove for airborne radars because the sensor is moving and the features are continuously changing. A rule-based automatic QC algorithm was presented in Bell et al. (2013), but the rule-based procedure revealed several limitations in the dichotomous classification of radar echoes. An improved algorithm using fuzzy logic concepts is currently under development to address these limitations. The new algorithm is based on a probabilistic approach, moving beyond a yes/no designation to provide the probability that a given volume contains weather, non-weather, or a combination of echoes. The weather probability field is calculated by combining interest maps from different data characteristics that suggest the likelihood a volume contains weather or non-weather echo. The user can then determine which threshold to use as valid radar echo in Doppler wind syntheses and other applications. The fields used in the weather probability calculation and their membership functions will be described, along with the methodology for using them to derive the final product. The algorithm is being tested on a variety of cases and the results compared with radar data that has been manually edited in order to evaluate its performance. The algorithm will be applicable for a wide range of meteorological conditions and different airborne radar configurations.