本研究目的為發展河川水系中超越機率百分之八十五之流量(Q85)之推估方法,同時探討資料年限長度對Q85之影響,希冀可藉此提供地面水水權管理之參考及應用。研究中,引入無因次流量延時曲線(Dimensionless Flow Duration Curves, DFDC),消除面積參數之影響後,提出適合之累進年限長度,再進一步引入比平均流量法(Specific Mean Discharge, SMD),提出Q85之簡易推求方法。 本研究選定石門水庫地區上游流量站為本文之研究區域,利用區域內4個流量站之歷史日流量紀錄資料,計算各站各月之無因次 ,採用累進分析法,即可得各站各月累進年限結果。另外,採用總年分析法,計算總無因次 與計算SMD值,分成全年分析與二期分析,其中二期分析再分成兩種型態,建立各種分析之DFDC與SMD點繪圖,以迴歸模式探討各站之迴歸曲線,進一步利用調整型相關係數(adjusted R-square)研判各站之最適曲線,即可得各站之最適曲線結果。 由研究結果顯示,一般而言資料長度於10至20年間,各月之 可漸趨穩定。此外,各站在最適迴歸模式之採用上,並無固定之單一模式,實際應用上可分別採用其最適之迴歸方程式結果。故本研究提出,未來在實務應用上僅需利用10至20年之資料即可先建立計畫地區之相關無因次及比平均流量之迴歸式。當需推求某一位置之Q85值時,則代入推估點上游集水面積與比平均流量值,即可簡易及快速推估其位置之Q85值,以供地面水水權管理之參考及應用。 關鍵字:流量延時曲線、比流量法、調整型相關係數。 The purpose of this study was to introduce a simple method for estimating stream flow with the exceedance probability of 85% (Q85) from readily available mean specific discharge data and catchment area. First of all, for each of the four streaflow stations located in Shihmen watershed, the dimensionless Q85 (normalized using monthly? mean stream flow) for each month of the year was estimated from past daily stream flow records. An accurate estimation of Q85 required a minimal time series length of 10 to 20. The dimensionless Q85 for different months of the year were then plotted against the respective specific mean discharges. A best fit curve was then determined based on adjusted R-square values. The correlation between Q85 and specific mean discharge for each station is described ay a unique curve. Once the best-fit curve for a specific stream gage is determined, Q85 for that specific location in the future then can be estimated from its updated specific mean discharge and the catchment area.