Radar Coverage Prediction Using Atmospheric Conditions and Observed Track Data

This study proposes a method for predicting radar coverage by integrating actual radar track data with atmospheric conditions at the time of detection. To incorporate atmospheric refractivity, altitude-dependent refractive index data for the Korean Peninsula were generated using rawinsonde data from the Korea Meteorological Administration, standard meteorological formulas, and the Kriging interpolation technique. The radar track and atmospheric data were preprocessed to extract features such as azimuth, elevation, and detection distance, and were merged into a training dataset. Two machine learning models—a random forest regression model and a neural network model—were trained using the processed data. Performance evaluation showed that the random forest model achieved better prediction accuracy than the neural network model. Additionally, a sensitivity analysis based on the equivalent Earth radius factor confirmed that the proposed method effectively reflects the influence of atmospheric refractivity on radar coverage estimation.