台灣近年來因天候的異常變遷，以及在對自然生態諸多的人為破壞下，豪雨時往往造成土石流災害。例如：98年8月8日莫拉克颱風重創南台灣，造成六百多人傷亡與近二百億的重大損失，特別是小林村與甲仙鄉的土石流之災情最為重大。為防制類似災害的重複發生，除有賴於調查88水災造成土石流之真正肇因，並分析其影響性，進而建立一較為可靠及準確的預測模式，以提供相關單位參考或應用。
採用預報和警報措施，減輕土石流所帶來的危害，不僅可行性高，同時也能減少許多經濟上的損失。土石流監測系統可隨時掌控土石流潛勢、流動方向及其規模大小，再配合預警系統發布土石流預警訊息，告知土石流危險區域內居民，採取適當之因應或疏散措施，以減少土石流災害。準確的土石流發生之預測，有助於提高土石流監測預警系統的可靠度。
由於土石流與影響因子的量測存有誤差，以及往往有賴於專家的語意評估而具有模糊性，例如：土石流的潛勢為高或低；雨量很大等。傳統迴歸分析及其他方法無法有效處理此種模糊資料，而模糊迴歸分析已被認為可克服此種類模糊資料或模糊變數問題的有效方法，故本研究運用此法來預測土石流警戒雨量值，並以土石流發生歷史資料驗證。
本文蒐集影響土石流發生各種可能之地文因子，考慮其對土石流發生之影響，並利用模糊線性迴歸模式建立一土石流潛勢預測模式，最後以97~98年之土石流發生資料為案例，針對高雄以及整個南部地區做驗證，總合各因子之影響程度作綜合評估，用以預測土石流、探討土石流發生降雨與地文之關聯性，以及掌控各重要變數之影響，進而提供給相關單位作為土石流警戒發布及避難疏散時機的決策依據。

Due to abnormal climate changes and many man-made ecological damages, mudflows and landslides have been caused by cloudburst in Taiwan since recent years. For example, Typhoon Morakot hit southern Taiwan last year. It caused 600 casualties and nearly 20 billion NT dollars, especially Jiasian Township, which was the most seriously hit by mudflows and landslides. For the prevention of recurrence of similar disasters, we should investigate the real cause of mudflows and landslides by the survey of 88 flood disasters to analyze the influence on our environment. We also need to establish a more reliable and accurate prediction models to provide information for related organizations or applications.
Using forecast and warning measures not only decreases the harm caused by mudflows and landslides but also reduces economic losses. Mudflows and landslides monitoring system can be controlled to dominate any potential mudflows and landslides, flow direction and size anytime. We can also operate early warning system of mudflows and landslides in coordination to inform residents in dangerous regions. Besides, the early warning system can adopt appropriate response or evacuation measures to reduce mudflows and landslides. It can enhance the reliability of mudflows and landslides monitoring system by accurate prediction on the occurrence of mudflows and landslides.
It often depends on the assessment of expert because mudflows and landslides and its impact factor measurement errors exist. The assessment is usually a semantic ambiguity such as the potential of mudflows and landslides. Traditional regression analysis and other methods cannot effectively deal with such vague information. Fuzzy regression analysis has been considered to overcome this type of fuzzy data. Therefore, this paper uses the method to predict mudflows and landslides and verifies the occurrence of mudflows and landslides by historical data.
This paper collects several factors which may result in mudflows and landslides and considers rainfall and physiographic conditions on the impact of mudflows and landslides. We also use fuzzy linear regression model to establish a prediction model of potential mudflows and landslides. Then, we use 2008-2009 mudflows and landslides data in Kaohsiung and the entire southern region data to confirm the combined influence of all factors. We finally make a comprehensive assessment to predict the occurrence of mudflows and landslides. Thus, we controlled important variables influenced on the system to provide the decision-making basis of issuing warning system of mudflows and landslides and proper evacuation time for related organizations.

[1] Agostino, V.D. and Marchi L. “Debris Flow Magnitude in the Eastern ItalianAlps: Data Collection and Analysis”, Phys. Chem. Earth (C), 26(9), 2001 ,pp. 657-663.
[2] Chang, Y.H., “Hybrid fuzzy least-squares regression analysis and its reliability measures”, Fuzzy Sets and Systems, Vol. 119, 2001, pp. 225-246.
[3] Chang, L.M. and Lee, Y.J., “Evaluation of performance of bridge deck expansion joints”, Journal of Performance of Constructed Facilities, Vol.16, 2002, No.1, pp.3-9.
[4] Chang, T.C. and Chao, R.J., “Application of back-propagation networks in debris flow prediction”, Engineering Geology Vol. 85 , 2006, pp. 270–280.
[5] Conway, S.J., Decaulne, A. and Balme, M.R., ” A new approach to estimating hazard posed by debris flows in the Westfjords of Iceland”, Geomorphology Vol. 114, 2010, pp. 556–572.
[6] Di, B.F., Chen, N.S. and Cui, P., ” GIS-based risk analysis of debris flow:an application in Sichuan, southwest China” International Journal of Sediment Research , Vol. 23, 2008, pp. 138-148.
[7] D’Urso, P., “Linear regression analysis for fuzzy/crisp input and fuzzy/crisp output data”, Computational Statistics and Data Analysis, Vol. 42, 2003, pp. 47-72.
[8] Gartner J.E., Cannon S.H., Santi P.M., Dewolfe V.G., “Empirical models to predict the volumes of debris flows generated by recently burned basins in the western U.S.”, Geomorphology Vol. 96 , 2008, pp. 339–354.
[9] Hojati M., Bector C.R., Smimou K., “A simple method for computation of fuzzy linear regression”, European Journal of Operational Research Vol166 . pp. 172-184.
[10] Liu X., Lei J., “A method for assessing regional debris flow risk: an application in Zhaotong of Yunnan province (SW China).” , Geomorphology Vol. 52, 2003, pp. 181–191.
[11] Pan N.F., Lin T.C. and Pan N.H., “Estimating bridge performance based on a matrix-driven fuzzy linear regression model”, Automation in Construction, Vol. 18, No.5, 2009, pp. 578-586.
[12] Redden D. and Woodall W., “Properties of certain fuzzy regression methods”, Fuzzy Sets and systems, Vol. 64, 1994, pp.361-375.
[13] Rupert M.G., Cannon S.H., Gartner J.E., Michael J.A., and Helsel D.R., “Using Logistic Regression to Predict the Probability of Debris Flows in Areas Burned by Wildfires, Southern California, 2003–2006”, U.S. Geological Survey , 2008, Report 2008-1370.
[14] Sakawa, M. and Yano, H., “Multiobjective fuzzy linear regression analysis for fuzzy input-output data”, Fuzzy Sets and Systems, Vol. 47, 1992, pp. 173-181.
[15] Sakawa, M. and Yano, H., “Fuzzy linear regression analysis for fuzzy input-output data”, Information Sciences, Vol. 63, 1992, pp. 191-206.
[16] Shieh C. L., Chen Y. S., Tsai Y. J., and Wu J.H., “Variability in rainfall threshold for debris flow after the Chi-Chi earthquake in central Taiwan, China”, International Journal of Sediment Research Vol. 24 , 2009, pp. 177–188.
[17] Tanaka H., Uejima S. and Asai K., “Linear regression analysis with fuzzy model”, IEEE Transactions on System, Man and Cybernetics, Vol.12, No.6, 1982, pp. 903-907.
[18] Tanaka, H. and Watada, J., “Possibilistic linear systems and their application to the linear regression model”, Fuzzy Sets and Systems, Vol. 27, 1988,pp. 275-289.
[19] Wang H.F. and Tasur R.C., “Resolution of fuzzy regression model” European journal of operational research, Vol. 126, 2000, pp. 637-650.
[20] Wan S. and Lei T.C., ” A knowledge-based decision support system to analyze the debris-flow problems at Chen-Yu-Lan River, Taiwan”, Knowledge-Based Systems , Vol.22 , 2009,pp. 580-588.
[21] Yang, M.S. and Lin, T.S., “Fuzzy least-squares linear regression analysis for fuzzy input-output data”, Fuzzy Sets and Systems, Vol. 126, 2002,pp. 389-399.
[22] Zadeh L.A., “Fuzzy sets.” Information and Control, 8, 1965, 338-353.
[23] 詹錢登，“土石流概論”，科技圖書，2000
[24] “93年土石流年報”，行政院農業委員會水土保持局，2005
[25] “94年土石流年報”，行政院農業委員會水土保持局，2006
[26] “95年土石流年報”，行政院農業委員會水土保持局，2007
[27] “96年土石流年報”，行政院農業委員會水土保持局，2008
[28] “97年土石流年報”，行政院農業委員會水土保持局，2009
[29] “水土保持手冊”，行政院農業委員會水土保持局，2006
[30] 潘南飛，“模糊迴歸在工程應用之探討”，技術學刊，第22卷第1期，第1-11頁，2006。
[31] 范正成與吳明峰，“一級溪流土石流危險因子及其與降雨臨界線之關係”，中華水土保持學報，第32卷，第3期，第227-234 頁，2001。
[32] 傅裕盛，“模糊理論應用於土石流危險度分析之研究”，國立成功大學，碩士論文，2004。
[33] 李明熹，“土石流發生降雨警戒分析及其應用”，國立成功大學，博士論文，2006。
[34] 黃婷卉，“土石流發生降雨特性之研究”，國立成功大學，碩士論文，2002。
[35] 黃冠智，“模糊迴歸模式之建立及其應用”，國立高雄應用科技大學，碩士論文，2006。
[36] 謝正倫、江志浩與陳禮仁，“花東兩縣土石流現場調查與分析”，中華水土保持學報，第22卷，第2期，第109-122頁，1992。
[37] 徐惠亭，“應用數值地形模型及可拓工程方法進行土石流潛勢溪流之評估”，國立成功大學，碩士論文，2008。
[38] 許楷均，“以模糊迴歸分析評估柔性鋪面現況服務指標”，國立成功大學，碩士論文，2008。
[39] 詹錢登與陳晉琪，“土石流發生機率分析-以花蓮縣銅門村及東興部落為應用對象”，中華水土保持學報，第30卷第1期，第65頁至76頁，1999。
[40] 鄭胥智，“降雨及地文條件與土石流發生關係之研究”，國立成功大學，碩士論文，2003。
[41] 鄒采苓，“土石流災害風險評估之研究─以台南縣為例”，國立成功大學，碩士論文，2006。
[42] 曹銳勤，“模糊迴歸模式之建構與分析”，國立交通大學，博士論文，1999。