Prediction of Extreme Weather’s Frequency Based on Seasonal ARIMA and Correlation Analysis Methods

The extreme-weather events have exposed high risks to many aspects of life to the current world, especially for those places with more frequency like Japan and the southeastern coast of America. High frequency of such natural disasters will do lots of damages to every aspect of human’s life. However, the current research hasn’t paid enough attention to the seasonal and frequency characteristic of the extreme weather. Hence, to analyze and predict the occurrence of the extreme weather in those places, this paper introduces correlation analysis and seasonal ARIMA model to figure out the seasonal characteristic and future trends of the extreme weather’s frequency. It turns out that extreme weather in both sides is more frequent in 2nd and 3rd quarter of the year, and shows high similarity in the two places, which can be explained by the climate characteristic of both places. The research is expected to provide frequency information for prevention of extreme weather.