Identification of biological and anomalous propagation echoes in weather radar observations—an imaging processing approach

Radar echoes from migratory birds/moving insects, and those from the ground due to the anomalous propagation (AP) are usually considered as non-meteorological contaminations for radar precipitation estimations. Separation of biological and AP echoes from precipitation is often challenging. The intensities and textures of biological echoes are similar to those of shallow stratiform rain, and deep AP echoes can have similar spatial characteristics to convective cells. Removing these non-precipitation echoes is very important for hydrological applications.

Based on digital imaging processing techniques and the large-scale physical characteristics of the biological and AP echoes, an automated technique is developed in this work to effectively identify and remove these two types of clutter when they are not mixed with precipitation. The new scheme includes multiple steps of tilt test, entity test and spectrum test to classify and clean up biological and AP echoes from stratiform and convective precipitation using base level radar reflectivity data. The scheme was evaluated using over 1600 volume scans of data from 60+ radars from different regions/countries and seasons, and was found to be very effective in removing the aforementioned non-precipitation echoes. The proposed algorithm is fully automated and computationally efficient, and it can be very useful for real-time hydrological operations.