中文摘要 台灣於春、夏兩季是降雨最豐沛的時候,其降雨比例高達78%,但是,由於台灣的地勢十分狹長且陡高,所以河流湍急導致大部分的雨水都隨著河水流入海洋中;由於季節降雨分布極為不均以及地域的問題,所以若不處理很可能會造成非降雨季節無水可用的情況,所以我們必須採取有效的手段來儲蓄水源。 由於,台灣是屬於海島型氣候,雨量主要是集中在夏秋兩季,而且河川川短流急,入流量的預測非常的困難。如果能更準確地預測入流量,對操作水資源的運用更有幫助。 本文利用統計迴歸與時間序列中的多變量自迴歸模型(Auto-Regressive eXogeneous,簡稱ARX)與多變量自迴歸移動平均模型(Auto-Regressive Moving Average eXogeneous ,簡稱ARMAX)建立石們水庫流量與雨量間之關係,以進行模擬與預測。
Abstract In Taiwan,roughly 78% of its yearly rainfall concentrates in the summer and autumn. As the distribution of rainfall is extremely uneven and regional issues, If we can more predict the flow accurately,and it may be ues to estimate inflow for reservoir operator. In this paper, multivariate of time series autoregressive model(Auto-Regressive eXogeneous, called ARX)and multivariate autoregressive moving average model (Auto-Regressive Moving Average eXogeneous, referred to as ARMAX) to establish the relationship between flow and rainfall, for simulation and prediction.