本研究利用時間序列方法探討水文時序補 遺問題,研究中以定常性合成資料驗證Ljung(1989) ,Pourahmadi(1989),Abraham(1981)所提方法之精確度,同 時並探討其於實測資料之適用性。研究結果顯 示,上述方法各有優劣,然而對於低階定常性之 補遺皆有令人滿意之精確度;其中又以Ljung方法 最為穩定。另,研究中亦探討具遺失值時間序 列模式研判之正確性,具遺失值時間序列模式 之判定以BIG準則較優。 In the present study, the interpolation of missing data in the hydrological time series was investigated. The synthetic data were employed to verify the accuracy of the Ljung (1989), Pourahmadi (1989) and Abraham (1981) methods. The real data were also investigated the aptness of those methods. The results indicated that the accuracy of interpolation was unexpected high, especially for the low order of stationary process. Meanwhile, the results indicated that Ljung method gave the best performance. Besides, the BIC criterion can successfully identify the true model for a given time series with missing data.