夏季日降雨量空間分佈特性之研究

Characterization of Spatially Distributed Summer Daily Rainfall

謝惠紅(Huey-Hong Hsieh) ； 鄭士仁(Shin-Jen Cheng) ； 劉璟燁(Jing-Ye Liou) ； 周世昌(Shin-Chang Chou) ； 蕭博仁(Bo-Ren Siao)

《農業工程學報》 52卷1期 (2006/03) Pp. 47-55

https://doi.org/10.29974/JTAE.200603.0005

摘要

集水區水資源之經營與管理，有賴於精確地推估集水區內水資源之輸入與輸出。在利用分散式水文、水理模式推估集水區內降雨－逕流之運動行為時，若能獲得精確之降雨空間分佈之輸入，則將有助於此類模式模擬之結果，進而改善集水區之經營與管理。
本研究選定石門水庫集水區民國84至89年夏季日降雨量紀錄，用差值方式推估集水區內之降雨量空間分佈。本研究所選用之差值方式為：反距離全重法(Inverse Distance weighting, IDW)以及克利金法(Kriging)。本研究採用交叉驗證法(Cross validation)以檢驗所推估之降雨平面圖。根據驗證所得之結果顯示以反距離權重法推求降雨之空間分佈優於克利金法（IDW法之平均誤差為0.04，Kriging法之平均誤差為0.54）。因此在雨量站分佈較少之區域，建議以反距離權重法推求該區域之降雨分佈。

關鍵字

差值 ； 降雨空間分佈 ； 反距離權重法 ； 克利金法

並列摘要

Precipitation is one of the most important factors for determining the hydrological response of a catchment. Watershed management, flood control, and water resources management all rely on accurate rainfall estimations for further analysis. With the increasing demands of distributed hydrological modeling, spatial-distributed rainfall input becomes necessary to utilize the full capabilities of distributed modeling.
The daily summer rainfall records (20 rain gague stations) of the Shih-men Watershed from 1999-2000 were used to predict the spatial rainfall distribution on Shih-men Watershed using two interpolation schemes inverse distance weighting (IDW) and ordinary kriging. From these data, 200-m grid subsets were extracted and interpolated. The different interpolation schemes were compared in terms of cross-validation statistics (six events). The results indicate that IDW (mean error 0.04) lead to more accurate representations of the spatial distribution of rainfall than ordinary kriging (mean errors 0.54). Therefore, with sparse rain gage network, IDW is suggested for spatial rainfall predictions.

並列關鍵字

Interpolation ； Spatial rainfall distribution ； Inverse distance weighting ； Ordinary kriging