由於氣候衍生型商品快速的發展,每日平均天氣模型被廣泛的研究與討論。 Campbell and Diebold (2005) 利用包含季節性變動的GARCH模型來配適美國城市的天氣。 我們用不同的copula 做連結並應用在亞洲城市上來描述城市之間的天氣的相關性。 在本篇論文中,我們透過模擬去論證我們的預測方法以及亞洲城市的每日平均資料去做實例分析。Because of the rapid development of weather derivatives, models for daily average temperature have been extensively studied in the literature. citet{dat} provide a time series model with a GARCH model for the volatility to describe the features for modelling daily average temperature in U.S. cities. Motivated by Campbell and Diebold (2005), we apply this model in Asian cities and use trivariate fully nested Archimedean Gumbel and Clayton copula to describe the dependence structure for the error distribution.To show the superiority of our model, we construct the prediction interval for the one-year ahead daily average temperature data using eight-year historical data, and show the coverage rates are higher when the dependence structure is employed.