Signal processing for ZLC-configuration insect-monitoring radars: Yields and sample biases

Abstract

A radar design that has proved effective for monitoring insect migratory flights employs a vertical beam in which rotating linear polarization is combined with a very narrow-angle conical scan - the “ZLC configuration”. The rather complicated echo-intensity time series produced when an individual insect traverses the radar's beam contains information about the insect's size, shape, orientation, and trajectory. However, retrieval of the various parameters is not always successful and some types of targets and target traverses are more vulnerable to processing failure than others. Variations in the ability of a profiling radar to detect targets of different sizes and at different heights are understood and can be corrected for if the target sizes and speeds are known (as is the case with a fully analyzed ZLC echo), but differences in processing success rates represent an additional source of bias. This preliminary study identifies height, the peak signal intensity, and target numbers as factors affecting processing yield, and presents some examples to illustrate how they act. These three quantities are available for all detected targets, as is required for use in any scheme for taking account of processing losses when generating profiles of target density (and other measures of insect activity of value to entomologists).