An AI Approach for Generating Instantaneous Rain Rates from Volumetric Radar Scans

Composite radar imagery is typically produced using all of the available elevations in a volumetric radar scan. The instantaneous timestamps of the individual radials in a given time bucket used for a composite image may correspond to several seconds or minutes before or after the nominal time for the composite image. In this study, an AI approach is applied to binned volumetric radar data in order to interpolate the intensity and position of radar returns between nominal timestamps to produce estimates of instantaneous rain rates. This approach is compared to a simpler interpolation scheme as well as other approaches.