Automatic tropical cyclone eye fix using genetic algorithm

Tropical cyclones (TCs) are weather systems with vast destructive power. To forecast TC tracks, forecasters need to locate their circulation centers, or eyes. This eye fix process is often done manually in practice. Since subjective elements are involved in the process, forecasters could disagree on the results even when multiple factors are considered. This paper presents an objective TC eye fix method that utilizes genetic algorithm to search for the optimal values of a six-parameter TC model. The results of this study indicate that the proposed method gives the best average error of 0.127° in latitude/longitude on Mercator projected map with respect to the best track data. This is well within the relative errors of about 0.3° from the results of different TC warning centers. The method can process 6 min of radar data in about 10 s when implemented on a notebook computer, meeting practical real time constraints. It provides a practical, independent and objective source of information to assist forecasters to fix TC eye centers.