近年來,全球氣候變化極端,台灣時而受到嚴重水災又時而遭受嚴重乾旱之苦,水資源的過與不及對於農業、工業及民生都造成嚴重的影響,因此雨量的預測是一重要課題。本論文主要目的為藉由全球海水表面溫度的資料,預測台灣附近夏季之降雨量。我們首先以Permutation test檢定出海水表面溫度的變化確實與台灣附近的夏季雨量有關,在找出相關之海水區塊後,再以此區塊之海水表面溫度的資料有效的預測台灣附近夏季之降雨量。此研究中藉由統計分析所定義出的海洋區域與一般定義聖嬰現象的區域之地理位置相近,有部分區域重合。此外,所得之迴歸模式其預測能力略優於傳統的時序模型。
We proposed a regression model for predicting the summer rainfall around Taiwan area, in which the global sea surface temperature (SST) are incorporated as predictive variables. In constructing such a regression model, we first identify the effective spatial region based on Portmanteau test statistics which quantify the dependence between the rainfall series and the SST series for different locations. Then, the principal component analysis is applied to the SST data in the identified region. Finally, a regression model for predicting rainfall is built on the important principal components of past SST data. Empirical results show that our regression model performs better than the conventionally time series model in terms of smaller prediction mean square errors.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。