中 文 摘 要
台灣在氣象條件上屬亞熱帶海島型氣候，雖屬多雨地區，惟降雨在空間與時間上分佈極為不均，地理條件上又因河川流短、地勢陡峻，可資直接利用之水資源相當有限。為能抒解供水不足之壓力以滿足目前或未來用水成長之需求，實有賴既有水資源之統籌管理與調配。因此，為能達成水資源調配及整體運用之管理目標、或河川最低生態需水量之需求、水中泥砂運移啟動速率之探討，甚至缺水指標及河川水權之研擬等，河川基流量基本資料之取得與其相關研究實不可或缺。
河川之基流量為入滲雨水經深層滲漏，流入地下含水層形成地下水蓄水量，而在無降雨情況，由地下蓄容所排出，變動量甚小之河川流量，亦即河川在乾旱情況下之流量。而河川之基流分離方法大致上可分為地化法、圖解法、濾波法及分析法，但這些方法或過於主觀、或不具物理概念或與實際水文狀況不符。因此本研究嘗試以水文模擬常被應用之修正型水筒模式，利用1994~1997年之雨量及流量資料建模，建立一適合松茂流量站之修正型水筒模式，其中參數的決定係使用參數優選法中之適應逢機搜尋法；完成建模後隨即利用1998年之雨量及流量數據來做模式之驗証，經逕流之預測評估得效率係數(CE)為0.764，証明所建置之模式為具代表性之降雨~逕流模式。模式中第一層水筒下層之側孔流出量(Q3)及第二層水筒之側孔流出量(Q4)結合即為河川之基流量。最後，本研究推估之基流量與傳統圖形法及改良濾波法推估之結果作比較，並以平均枯水、低水流量（相當程度反應基流）及乾季分析為基準比較其結果，其中修正型水筒模式推估之基流量有86.5%~92.5%落在於低水流量及枯水流量間之範圍，且與乾季分析之平均流量差僅為57.9cms，故可証明修正型水筒模式足以反應實際之水文現象，其應用於河川基流量之分析上具有良好之實用性。

Abstract
Taiwan’s climate belongs to subtropics type with rainy and humid weather during the summer season. Because of the unequally rainfall distribution in space and time, the surface water resource is restricted for short stream length and steep slope condition. Appropriately assigning water resources is necessary for water supplying at present and demand for the future. In order to achieve this object, basic base flow data of streams are important as well as relative research of minimum biological demand in stream, starting velocity of send/gravel transported in river, water shortage index and water right studies.
Base flow comes from the storage of groundwater which has the small variation of discharge during drought and/or sunny days. There are four base flow separation method which are geochemical method, graphical method, modified filter and analytical method. Because these methods are arbitrary or possess no physical means or not meet the real hydrologic situations, this paper try to use the concept of modified tank model to separate the base flow. Rainfall and discharge data of Song Mao station in the year 1994~1997 were collected to build the modified tank model and the adaptive random search technique is used for parameters optimization analysis. This model then simulated the rainfall~runoff relation with the data of 1998 and got the CE (coefficient of efficiency) value of 0.764. The simulation result showed that the modified tank model of Song Mao station can be accepted as a reasonable model. The base flow is the sum of Q3 and Q4 value analyzed by this model. At last, base flow of same period which has been separated by graphical method and modified filter were compared with the result analyzed by modified tank model. Mean dried and mean low stream discharges (respond base flow effect) as well as dried season analysis value are considered as the quantitative tools to identify the analysis error of base flow separation. In the result, there are 86.5%~92.5% base flow computed by modify tank model in between the mean low stream and dried stream discharges, and it is only 57.9 cms variance with the mean flow value calculated by dried season analysis. So the modify tank model can be response real hydrologic phenomenon with physical means and has a good practicality in the way of base flow separation.

參考文獻
1. 楊宗翰、顏沛華，「地震引致河川基流量變化分析～以集集大地震為例」，碩士論文，國立成功大學水利及海洋工程研究所，(2002)。
2. 李德河、謝正倫、顏沛華、詹錢登、林慶偉、游鎮烽、饒瑞鈞、徐國錦、黃崇明，「地震發生前後地下水水位異常變化之研究(2/5)」，國立成功大學防災研究中心，經濟部水利署，第247~254頁，(2002)。
3. 蔡宜華，「三角形不規則網格之分佈型降雨　逕流模式」，碩士論文，國立臺灣大學農業工程學研究所，(1999)。
4. 游保杉、鄭玉荻、蔡長泰，「應用全域最佳化技巧於分佈型降雨逕流模式」，台灣水利，第43卷第2期，pp.45-53，(1995)。
5. 徐義人，應用水文學，國立編譯館主編，大中國圖書公司印行，台北，(1995)。
6. 王盈欽、周乃昉，「修正型水筒模式對洪水事件模擬特性之探討」，碩士論文，國立成功大學水利及海洋工程研究所，(1994)。
7. 王如意、易任，應用水文學，上、下冊，國立編譯館出版，茂昌圖書有限公司發行，台北，(1992)。
8. 王如意、譚智宏，「修正型水筒模式之研究及其應用於流域逕流量之預測」，台灣水利，第三十九卷，第三期，第1~23 頁，(1991)。
9. 譚智宏，「修正型水筒模式之研究及其應用於長短程之水文模擬」，碩士論文，國立台灣大學農業工程學研究所，(1990)。
10. 徐享崑，「水筒模式法簡介及應用」，國立成功大學水利及海洋工程研究所，台南(1985)。
11. 菅原正巳，「水文水櫃模式分析技術研習會講義」，國立台灣大學土木工程學研究所，台北，(1985)。
12. 葉在民，「水筒模式法應用於長期逕流推估之研究」，碩士論文，國立成功大學水利及海洋工程研究所(1979)。
13. Birtles, A. B.，（1978），「Identification and separation of major base flow components from a stream hydrograph」，Water Resour. Res., 14(5), 791-803, 1978.
14. Lyne, V., and M. Hollick，（1979），「Stochastic time-variable rainfall-runoff modeling」， I.E. Aust. Natl. Conf. Publ. 79/10, pp. 89-91, Inst. of Eng. Aust., Canberra, ACT.
15. Hino, M. and M. Hasebe，（1986）「Separation of a storm hydrograph into runoff components by both filter-separation AR method and environmental isotope tracer」，J. Hydrol., 85, 251-264.
16. Kadoya M. , Tanakamaru H.(1988) ,「Flood Runoff Forecasting with Long and TermsRunoff Model」, 6th APRD International Association for Hydraulic Research ,Kyoto , Japan , pp183~190.
17. Nathan, R. J., and T. A. McMahon，（1990），「Evaluation of automated techniques for base flow and recession analyses」Water Resour. Res., 26(7), 1465-1473.
18. Chapman, T.,(1991) Comment on「Evaluation of automated techniques for base flow and recession analyses」by R. J. Nathan and T. A. McMahon, Water Resour. Res., 27(7), 1783-1784, 1991.
19. Boughton, W. C.，（1993），「A hydrograph-based model for estimating the water yield of ungauged catchments」，paper presented at Hydrology and Water Resources Symposium, Inst. of Eng. Aust., Newcastle, N.S.W.
20. Jakeman, A. J., and G. M. Hornberger，（1993），「How much complexity is warranted in a rainfall-runoff model? 」，Water Resour. Res., 29, 2637-2649.
21. Chapman, T. G., and A. I. Maxwell，（1996），「Baseflow separation – Comparison of numerical methods with tracer experiments」，paper presented at Hydrology and Water Resources Symposium, Inst. of Eng. Aust., Hobart, Tasmania.
22. Rutledge, A. T.，（1998），「Computer programs for describing the recession of groundwater discharge and estimating mean ground-water recharge and discharge from streamflow records – Update」U.S. Geol. Surv. Water Resour. Invest. Rep., 98-4148.
23. Peter R. Furey、Vijay K. Gupta，（2001），「A physically based filter for separating base flow from streamflow time series」，Water Resources Research, Vol.37, No.11, p.2709-2722, November 2001。