Ground Clutter Detection and Correction Model for C-band Weather Radars

Abstract

This paper presents the results of a Ground Clutter detection and correction model for C- band weather radars. The model uses artificial intelligence techniques such as neural networks, decision trees and vector support machines for ground echo detection and neural networks for correction. For the validation of the model, data from two meteorological radars located in the Colombian territory are used, where through a controlled experiment, terrestrial echoes with partial and total obstruction of one beam and two beams were manually introduced, with the objective of being able to detect them and correct the information of the radar reflectivity of the shadow areas. The results showed a detection accuracy of 99.3 % using a redundant system and a 4.58 dBZ RMS error in the correction of the reflectivity.

Keywords: Weather Radar; Ground Clutter; Precipitation Echoes; Support Vector Machine; Artificial Neural Network; Decision Tree.