Detection of centers of tropical cyclones through the synergistic use of geostationary and polar-orbiting satellite data

Abstract

Detecting tropical cyclone centers is crucial for better understanding the behavior and characteristics of tropical cyclones, and their development. In particular, early detection of centers of cyclone disturbances are important for tropical countries to prepare and respond to damages by tropical cyclones. There are several approaches for detecting the centers of tropical cyclones . Among them, the best track provided by Joint Typhoon Warning Center (JTWC) has been widely used as reference data of tropical cyclone center locations. However, JTWC uses multiple resources including geostationary satellite data and in situ measurements to determine the best track in a subjective way and makes it available to the public 6 months later after an event occurred. Thus, the best track data cannot be operationally used to identify the centers of tropical cyclones in real time. In this study, we proposed an automated approach for identifying the centers of tropical cyclones using both geostationary (i.e., Meteorological Imager of Communication, Ocean, and Meteorological Satellite by South Korea) and polar-orbiting (i.e., Windsat of Coriolis by Naval Research Laboratory and Air Force Research Laboratory of US) satellite data. Brightness temperatures and wind directions of sea surface wind field were extracted using COMS MI and Windsat data over the western North Pacific between June and August 2011, respectively. We adopted a spatial metric called circular variance to identify the centers of tropical cyclones. Circular variances were calculated from the surface data of brightness temperature gradients and wind field direction. The locations of the maximum circular variance were identified as the centers of tropical cyclones. Results were compared with the best track data, showing the distance between the detected center and the best track center up to 2 degrees. The distance between the centers became smaller as the tropical cyclones developed, which implies that the wind vorticity-based center and the cloud-based center tend to agree when a cyclone becomes tropical depression.