Application of Spike Signal Detection and Error Reduction Techniques to Enhance the Efficiency of Radar-Based Rainfall Estimation

Abstract

Radar measurement errors significantly impact the accuracy of radar rainfall estimation, with spike errors being particularly challenging to detect due to their continuous variation in both space and time and their inconsistency patterns.  This study aims to enhance radar rainfall assessments by applying techniques for detecting and reducing errors from spike signals. Radar reflectivity data affected by spike signals from Samut Songkhram, Sattahip and Phimai radar stations were collected between 2020 and 2022. Techniques for detecting two common types of spikes found in the Samut Songkhram and Sattahip radar data, Narrow and Shorter Longitudinal, were applied in this study, alongside combined techniques to improve the quality of radar reflectivity data. The results showed that using these spike detection and error reduction techniques significantly improved data quality. When applied to radar rainfall data, the techniques increased accuracy by up to 72.9% compared to radar data before applying the techniques. This improvement in radar data quality could be further enhanced if combined with other error-reduction techniques, leading to the production of high-quality radar rainfall data, which would contribute to more effective weather monitoring and early warning systems.