A Dual-Driven Intelligent Method for Clear-Sky Echo Identification in Millimeter-Wave Cloud Radar Observations

Clear-air echoes, observed in non-precipitation conditions, significantly impact radar observations. These echoes face challenges in identification due to confusion with meteorological targets and diverse interference signals. This study presents a novel model-data dual-driven method to enhance automatic recognition of clear-air echoes. By integrating fuzzy logic and deep learning, the method leverages prior physical information regarding the statistical distribution of meteorological parameters to improve accuracy and robustness. Training and validation conducted on Time-Height Indicator (THI) datasets from millimeter-wave cloud radar demonstrate the method's superior classification performance.