本研究針對鹽水溪流域集水區中8個雨量站來進行相關性分析以瞭解現行於集水區內各站之相對關係。除相關性分析之外,研究中亦進行抽站平均最大一日暴雨之計算並比較抽站前後集水區平均一日最大暴雨量之差異。為能進一步探討雨量站抽站後不同重現期不同機率分布集水區平均最大一日暴雨量的改變,乃進行頻率分析並比較不同重現期最大一日暴雨量的變化。 相關分析所得相關係數介於0.75~0.96之間,可見各測站間為高度相關。由徐昇氏多邊形網法得知流域內現行雨量站以中央氣象局和順站(P04)之權重最大,中央氣象局崎頂站(P05)為最小。另外以適合度檢定成果所建議之對數皮爾遜三型進行重現期25年、50年、100年與200年之鹽水溪流域平均年最大一日暴雨量之推估,由結果可知當重現期為25年時以P04站缺測時的最大一日暴雨量(484mm)為最大,而以水利署虎頭埤(P01)站時(463mm)次之,中央氣象局新市站( P07)(445mm) 為最小。 由本研究得知單以面積權重進行評估雨量站對於流域平均雨量是不足的,且雨量測站量測的合理性亦為水文分析成果可靠度很重要的一環,進一步思考此項結果又可發現,在同一鄉鎮區屬於中央氣象局的雨量測站紀錄均較水利署的測站紀錄高。
Reliable estimations of missing data is an important task for hydrologists, engineers and environmental researchers. This study applies the correlation analysis to assess the accuracy of the rainfall station network for Yenshui river basin. In addition, employ Thiessen polygons to compare maximum one-day precipitation variations in the basin. The objective is to study the effects on watershed rainfall estimation in case data is missing in one of the stations in the network. Then, we apply frequency analysis for comparing maximum one-day rainfall in different return periods. From the correlation analysis, relationship among the stations is high having correlation between 0.75 and 0.96. Thiessen polygons method shows that in Heshun rainfall station (P04) the weight is highest and in Qiding rainfall station (P05) is lowest. For frequency analysis based on the results, we suggest using the Log Pearson type III method to estimate the return periods of 25, 50, 100 and 200 year in Yenshui River. When P04 is omitted in a 25 year return period the maximum one-day precipitation is 484 mm, Hutoupi rainfall station (P01) is 463 mm and Xinshi rainfall station (P07) is 445mm. In this study it revealed that using Thiessen polygons analysis alone is insufficient and unreliable. Instead, more than one form of analysis should be taken into account in the adjustment and estimation of watershed rainfall under missing records condition.