Tropical Cyclone Hazardous Area Forecasting Using Self-adaptive Climatology and Persistence Model

Abstract

A tropical cyclone is one of the natural disaster which is the most economical and human losses in the world and tends to be more damaging and more frequent in the future due to climate change and human behaviors. To mitigate catastrophic phenomenon, a modern natural disaster management model (MNDM) is designed and the most important phase in MNDM is emphasis on process before the catastrophic phenomenon or preparing tropical cyclone track forecasting, intensity forecasting, and risk area identification. Although various tropical cyclone track and intensity forecasting techniques have been developing and improving for several years, the errors of the forecasting model still remain. Moreover, risk area assessment and uncertainty of the major model which are the most important phase in MNDM are excluded. To address these problems, this paper proposes an integrated short-range tropical cyclone hazardous area forecasting system which includes both track and hazardous area forecasting in system by using only 12 features which were extracted from satellite images with improvement of the traditional statistical methods. The performance of the model is satisfactory; the average error from the experimental results of R₃₄, R_50, and R₆₄ forecasting with unknown tropical cyclone data between years 2013–2015 on Mercator projection map is lower than traditional techniques by 28.99%, 22.81%, and 24.38%, respectively.