Ground Clutter Filtering pada Radar Cuaca Manado

Abstract

This study aims to improve the accuracy of Manado weather radar data by filtering ground clutter interference originating from reflections of surface objects such as mountains and buildings. The scope of the study includes the analysis of Manado weather radar reflectivity data in 2024, focusing on the North Sulawesi region. The methods used include thresholding techniques to separate weather signals from ground clutter, data processing using the Python programming and the wradlib module, and validation based on Himawari-9 satellite imagery and Automatic Weather Station (AWS) data. Radar data in VOL format is converted to geospatial projections to produce CMAX (Column Maximum Reflectivity) products and integrated with SRTM30 topographic data. The results show that thresholding is effective in reducing ground clutter, with a correlation value between radar rainfall estimates and AWS observations reaching 0.859 (MAE: 0.306 mm/10 minutes; RMSE: 1.283 mm/10 minutes). Mapping the results via Web GIS facilitates visualization of corrected data. The study's conclusion states that this method improves the accuracy of radar interpretation and rainfall estimates. This study provides operational contributions to weather information-based disaster mitigation in the Manado region.