Spatio-Temporal Analysis of Convective Cloud Properties Deriving from Weather Radar Reflectivity during the Decaying Stage of Tropical Storm over the Lower Northern Thailand

Abstract

Disaster Management is one of the most important responsibilities of the governments all over the world. The systems of disaster prevention, preparedness and mitigation have been well established only in developed countries. For the developing countries including Thailand, people still have suffered from disaster since they are lacking of adequate information to cope with disaster. The weather radar is one of the tools that can provide spatio-temporal information for Nowcast which is useful for hydro-meteorological disasters warning and mitigation system.  The extremes of precipitation are usually detected by convective cloud where the updrafts and downdrafts have been strengthened in vertical wind motions These extremes are the vital threat to people during wet season over Thailand. In this study, therefore, we have developed algorithm to extract convective cloud information that is necessary for prediction of precipitation extremes. The algorithm is constructed in Python script using OpenCV library to extract the radar reflectivity of the 1^st Plan Position Indicator of radar image. The reflectivity data with the measurement frequency of once per hour has been obtained from the website of Thaiwater.net of which original data provided by Thai Meteorological Department (TMD).  This study has been done by using reflectivity data from Phitsanulok radar station which locates in Lower Northern Thailand.  The convective cloud regions have been detected by the developed algorithm. In addition, the fitting ellipses (Fitzgibbon and Fisher, 1995) have been applied to derive the properties of convective clouds during Sonca tropical storm in July of 2017. Results show that the number of convective cloud regions has abruptly increased during Sonca period compared to before the storm passing. The average number of convective cloud region has shown clearly of peak time at 17 LST. However, several peaks of total area on the convective region at hourly scale have also been found. The spatio-temporal analysis of the extracted storm information demonstrates the severe pattern of the high frequency convective storm during Sonca event over windward slope. The number of convective storm doubled increases comparing with period prior to Sonca storm.