臺灣博碩士論文加值系統

本研究利用中央氣象局QPESUMS (quantitative precipitation estimation and segregation using multiple sensors)系統所提供之雷達估計雨量，以支撐向量回歸(support vector regression, SVR)及輻狀基底函數類神經網路(radial basis function neural network, RBFNN)兩種模式，結合地面觀測雨量及地理空間因子進行修正雷達估計雨量，其目的為期望提昇雷達估計降雨之準確性，並保有雷達雨量具高解析度及涵蓋完整區域之優點。本研究選取七場颱風事件作為模式率定與驗證，其中前四場事件作為模式之率定，後三場事件為驗證，以地面雨量站對應上方九個雷達估計雨量、地面雨量站位置之X 坐標與Y 坐標、地面雨量站高程以及地面雨量站與七股雷達站的直線距離為四種輸入因子，依不同因子的組合分別建立五種輸入向量，並討論不同的輸入向量對於模式建立之影響程度。經由分析結果可知隨著影響降雨特性之輸入因子資訊量增加，兩種模式的率定結果越好；本文以2007 年三場颱風事件進行台灣西南部之時雨量修，初步結果顯示兩種模式表現在總雨量方面，能將原雷達估計雨量之均方根誤差減少將近一半，且其相關係數能由0.7 提昇至0.8 以上，顯示SVR 模式以及RBFNN模式能夠有效利用地面觀測雨量訊息結合雷達估計雨量，尤其在山谷地區及靠近中央山脈之雷達估計降雨之修正結果，支撐向量回歸(SVR)比輻狀基底函數類神經網路(RBFNN)能有效地掌握雷達雨量於地理空間因子之相關訊息。

This study developed two machine learning methods, the support vector regression (SVR) and the radial basis function neural network (RBFNN), to improve radar rainfall from the high resolution QPESUMS (quantitative precipitation estimation and segregation using multiple sensors) system provided by the Central Weather Bureau in Taiwan. The goal of this research is to enhance the accuracy of radar rainfall estimation which are found largely underestimate in the mountain regions, but to preserve the advantage of high spatial and temporal resolution of radar rainfall. Various input vectors of two machine learning models were combinations of predictors consisting of the nine radar estimated rainfalls centered at the raingauge location, the coordinates of raingauge positions, the elevation of raingauge, and the distance between Chi-Guw weather radar station and the raingauge. This work found that the model using more predictors in the input vector performs better than that using less predictors. Seven typhoon rainfall events in the southwest Taiwan were used as calibration and validation data in this work. Analytic results show that the SVR and the RBFNN have the capability to adjust radar estimated rainfall, from the evidence that the root mean square error was reduced and the correlation coefficient increased. Moreover, the SVR outperformed the RBFNN, especially in the valley and in the mountain regions.

中文摘要 I
英文摘要 II
誌 謝 III
目 錄 IV
表 目 錄 VII
圖 目 錄 VIII
第一章 緒論 1
1-1 研究動機與目的 1
1-2 文獻回顧 2
1-2-1 雷達回波資料處理 2
1-2-2 雨量內插法 4
1-2-3 回歸模型估計法 6
1-3 本文組織 9
第二章 氣象雷達雨量估計方法 12
2-1 氣象雷達觀測原理 12
2-2 估計降水量方法 15
2-3 QPESUMS系統介紹 16
2-3-1 功能簡介 16
2-3-2 產品資訊 17
第三章 模式介紹 22
3-1 支撐向量機(SVM) 22
3-1-1 結構風險最小化 22
3-1-2 支撐向量回歸(SVR) 24
3-1-3 SVR之參數率定 30
3-2 輻狀基底函數類神經網路(RBFNN) 30
3-2-1 RBFNN網路架構 31
3-2-2 RBFNN學習演算法 34
第四章 研究區域與資料分析 36
4-1 研究區域 36
4-2 QPESUMS網格繪製 39
4-3 資料收集 41
4-4 降雨空間分布比較 47
4-4-1 單點雨量比較 47
4-4-2 雨量與地面雨量站高程比較 51
4-4-3 雨量與七股雷達站距離比較 55
第五章 雷達估計雨量之修正 61
5-1 輸入因子選擇與資料之前置處理流程 61
5-1-1 輸入因子選擇與組合 61
5-1-2 資料之前置處理流程 64
5-2 模式率定結果 65
5-2-1 模式率定方法與評鑑指標 65
5-2-2 SVR之率定結果 66
5-2-3 RBFNN之率定結果 71
5-3 SVR與RBFNN模式驗證結果 76
5-3-1 時雨量之驗證結果 76
5-3-2 雨量誤差之高程分布 84
5-3-3 雨量誤差之空間分布 88
第六章 結論與建議 94
6-1 結論 94
6-2 建議 96
參考文獻 97

1.王安翔、林李耀(2008). 子計畫一：流域集水區之雷達定量降雨預報技術發展(2/3). 財團法人國家實驗研究院國家地震工程研究中心, 行政院國家科學委員會專題研究計畫進度報告.
2.丘台光(2002). 梅雨季豪（大）雨預報技術之發展與整合(I)-子計畫：結合衛星、雷達與閃電資料發展豪(大)雨與定量降水預報之研究(I), 交通部中央氣象局.
3.丘台光、陳嘉榮、張保亮、林品芳(2003). 應用多重觀測工具之定量降雨估計之研究, 中央氣象局氣象衛星中心.
4.田璦菁(2003). 颱洪期間區域總雨量估計最佳化之研究. 土木工程研究所. 中壢, 國立中央大學. 碩士論文.
5.江衍銘(2003). 類神經網路於水文氣象-以雷達及數值天氣預報資訊建構洪水預測. 生物資源暨農學院生物環境系統工程學研究所. 台北, 國立台灣大學. 博士論文.
6.江衍銘、陳品安、張斐章(2008). 遞迴式類神經網路於降雨-逕流之探討, 農業工程報告.
7.李天浩、陳台琦(2002). 水文氣象資訊與水文水理系統模式整合之研發(1/2). 經濟部水利署計畫報告, 國立台灣大學水工試驗所.
8.李明軒(2008). 支撐向量機與模糊推論於流量預報即時誤差修正之研究. 水利及海洋工程研究所. 台南, 國立成功大學. 碩士論文.
9.張永欣 (2007). 以多變量地質統計方法進行雨量空間內插. 應用地質研究所. 中壢, 國立中央大學. 碩士論文.
10.張斐章 and E. H. Edwin (2006). 水資源最佳化管理於河川生態復育之研究, 經濟部水利署水利規劃試驗所.
11.張斐章、張麗秋(2005). 類神經網路(newural network). 台北市, 東華書局.
12.張智昌(2006). 整合氣象雷達與即時降雨資料於颱風降雨推估之研究. 地理環境資源研究所. 台北, 國立台灣大學. 碩士論文.
13.張逸凡(2005). 支撐向量機在即時河川水位預報之應用. 水利及海洋工程研究所. 台南, 國立成功大學. 碩士論文.
14.張麗秋、林永堂、張斐章(2005). 結合OLS 與SGA 建構輻狀基底類神經網路於洪水預測之研究, 台灣水利. 53(4).
15.陳國鴻(2005). 導入權重概念對支撐向量機分類正確率之影響. 工業與資訊管理學系. 台南, 國立成功大學. 碩士論文.
16.陳憲宗(2006). Real-time Porbabilistic Flood Stage Forecasting Using Support Vector Machines and Fuzzy Inference Model. 水利及海洋工程研究所. 台南, 國立成功大學. 博士論文.
17.陳憲宗、張逸凡、謝章廷、游保杉(2006). 支撐向量機制：洪水水位預報模式, 台灣水利. 54: 50-61.
18.陳憲宗、游保杉(2007). 洪水位之即時機率預報-結合支撐向量機與模糊推理, 農業工程學報. 53: 1-20.
19.曾忠一(1997). 氣象資料同化. 台北市, 渤海堂.
20.黃威雄(2000). 應用類神經網路於颱風期間雷達降雨模擬之研究. 土木工程學研究所. 台北, 國立台灣大學. 碩士論文.
21.楊政潭(2003). 雷達回波應用於颱風降雨空間分佈與總量之研究—以納莉颱風為例. 水文科學研究所. 中壢, 國立中央大學. 碩士論文.
22.劉清煌(2002). 雷達雨量估計與雨量即時預報技術-利用雷達及衛星資料與類神經網路之結合颱風降水估計之研究-類神經網路.
23.Andrieu, H., J. D. Creutin, et al. (1997). "Use of weather radar for the hydrology of mountainous area, Part I:Radar measurement interpretation." J. Hydrol. 193: 1-4, 1-25.
24.Andrieu, H. and C. J. D. (1995). "Identification of vertical profiles of radar reflectivities for hydrological applications using inverse method, Part 1: Formulation." J. Appl. Meteor. 34: 225-239.
25.Austin, P. M. (1987). "Relation between measured radar reflectivity and surface rainfall." Mon. Wea. Rev. 115: 1053-1070.
26.Baltas, E. (2007). "The combined use of weather radar and geographic information system techniques for flood forecasting." Advances in Geosciences 10: 117-123.
27.Borga, M. (2002). "Accuracy of radar rainfall estimates for streamflow simulation." Journal of Hydrology 267: 26-39.
28.Chang, F. J. and Y. C. Chen (2003). "Estuary water-stage forecasting by using radial basis function neural networks." Journal of Hydrological 270.
29.Chang, F. J., J. M. Liang, et al. (2001). " Flood Forecasting Using Radial Basis Function Neural Network." IEEE Trans: 530-535.
30.Chen, S.-T. and P.-S. Yu (2007). "Pruning of support vector networks on flood forecasting." Journal of Hydrology 347: 67-78.
31.Chiang, Y.-M., F.-J. Chang, et al. (2006). "Dynamic ANN for precipitation estimation and forecasting from radar observations." Journal of Hydrology 334: 250-261.
32.Cole, S. J. and R. J. Moore (2008). "Hydrological modelling using raingauge- and radar-based estimators of areal rainfall." Journal of Hydrology 358: 159-181.
33.Collier, C. G. (1996). "Applications of Weather Radar Systems: A Guide to Uses of Radar in Meteorology and Hydrology." John Wiley 2d ed.: 383
34.Cover, T. M. (1965). "Geometrical ans Statistical Properties of Systems of Linear Inequalities with Applications inPattern Recognition." IEEE Trans -actions on Electronic Computers 14: 326-334.
35.DeGaetano, A. T. and D. S. Wilks (2009). "Radar-guided interpolation of climatological precipitation data" International Journal of Climatology Int. J. Climatol. 29: 185-196.
36.Dibike, B. Y. B., S. Velickov, et al. (2001). "Model Induction with Support Vector Machines: Introduction and Applications." Journal of Computing in Civil Engineerning 15: 208-216.
37.Dinku, T., E. N. Anagnostou, et al. (2002). "Improving radar-based estimation of rainfall over complex terrain" Journal of Applied Meteorology 41(12): 1163-1178.
38.Fabry, F. (1997). Vertical profiles of reflectivity and precipitation intensity in: Weather radar technologies for water resources management. IRTCUD /University of Sao Paulo, Brazil and IHP-UNESCO.
39.Fulton, R. A., J. P. Breidenbach, et al. (1998). "The WSR-88D Rainfall Algor -ithm" Wea. Forecasting 13: 337-395.
40.Garc?’a-Pintado, J., G. G. Barbera, et al. (2009). "Rainfall estimation by rain gauge-radar combination: A concurrent multiplicative-additive approach." Water Resource Research 45: W01415.
41.Gjertsen, U. and J. I. Dahl (2002). "Challenges for precipitation estimation in mountainous regions." ERAD 2002: 250-254.
42.Gjertsen, U., M. Salek, et al. (2003). Gauge-adjustment of radar-based preci -pitation estimates-a review: 1-17.
43.Hong, W.-C. (2008). "Rainfall forecasting by technological machine learning models." Applied Mathematics and Computation 200: 41-57.
44.Hossain, F., E. N. Anagnostou, et al. (2004). "Hydrological model sensitivity to parameter and radar rainfall estimation uncertainty." Hydrological Process 18: 3277-3291.
45.Hsu, C. W., C. C. Chang, and C. J. Lin, “A Practical Guide to Support Vector Classification”, Techincal report, Department of Computer Science and Information Engineering, National Taiwan University, Avalible at : http://www.csie.ntu.edu.tw/~cjlin/papers/guide.pdf, 2003.
46.JORDAN, P. W., A. W. SEED, et al. (2003). "A Stochastic Model of Radar Measurement Errors in Rainfall Accumulations at Catchment Scale." American Meteorological Society 4: 841-855.
47.Joss, J. (1990). "A. Waldvogel Precipitation measurement and hydrology." Radar in Meteorology D. Atlas, Ed., Amer. Meteor. Soc: 577-606.
48.Joss, J. and R. Lee (1995). "The application of radar gauge comparisons to operational precipitation profile corrections."J. Appl. Meteor 34: 2612- 2630.
49.Joss, J., Waldvogel, A (1990). "Precipitation measurement and hydrology." In: Atlas, D. (Ed.), Radar in Meteorology American Meteorological Society: 577-606.
50.Kim, B. S., B. K. Kim, et al. (2008). "Flood simulation using the gauge-adjust -ed radar rainfall and physics-based distributed hydrologic model." Hydrological Process DOI: 10.1002: hyp.7043.
51.Kitchen, M., R. Brown, et al. (1994). "Real time correction of weather radar for the effects of bright band range and orographic growth in widespread precipitation." Q. J. R. Meteorol. Soc 120: 1231-1254.
52.Krajewski, W. F. (1987). "Cokriging radar-rainfall and rain-gauge data." Jour -nal of Geophysical Research Atmospheres 92 (D8): 9571-9580.
53.Li, W., V. Chandrasekar, et al. (2003). Investigations in radar rainfall estimation using neural networks. IGARSS.
54.Liong, S.-Y. and C. Sivapragasam (2002). "Flood Stage Forecasting with Support Vector Machine." Journal of The American Water Resources Associa -tion 38: 173-185.
55.Liong, S. Y., C. D. Doan, et al. (2009). Rainfall nowcast system Proceeding of the 8th International Conference on Hydroinformatics, Concepcion, Chile.
56.Liu, H., V. Chandrasekar, et al. (2001). "An adaptive neural network scheme for rainfall estimation from WSR-88D observations." Journal of Applied Meteorology 40(11): 2038–2050.
57.Marshall, J. S. and W. M. Palmer (1948). "The distribution of raindrops with size." J. Meteor. 5: pp.165-166.
58.Moore, R. J., May, B.C., Jones, D.A., Black, K.B. (1994b). "Local calibration of weather radar over London. In: Almeida-Teixeira,M.E., Fantechi, R., Moore, R., Silva, V.M. (Eds.)." Advances in Radar Hydrology Proceed -ings of an International Workshop,11-13 November, Lisbon, Portugal. Report EUR 14334 EN, European Commission: pp. 186–195.
59.Moore, R. J., Watson, B.C., Jones, D.A., Black, K.B. (1991). "Local recali -bration of weather radar. In: Cluckie, I.D., Collier, C.G. (Eds.)." Hydrological Applications of Weather Radar Ellis Horwood Chichester, UK: pp. 65-73.
60.Morin, E. and M. Gabella (2007). "Radar-based quantitative precipitation estimation over Mediterranean and dry climate regimes." Journal of Geophysical Research 112: D20108.
61.Navonne, H. D. and H. A. Ceccatto (1994). "Predicting Indian monsoon rainfall: a neural network approach." Climate Dynamics 10: 305-312.
62.Pai, P.-F. and W.-C. Hong (2007). "A recurrent support vector regression model in rainfall forecasting." Hydrological Processes 21: 819-827.
63.Rosenfeld, D., D. B. Wolff, et al. (1994). "The window probability matching method for rainfall measurements with radar." J. Appl. Meteor. 33: 682-693.
64.?alek, M., P. Novak, et al. (2004). "Operational application of combined radar and raingauges precipitation estimation at the CHMI." ERAD Publication Series 2: 16-20.
65.Shiiba, M., T. Takasao, et al. (1984). Investigation of short-term rainfall predict -tion method by a translation model. Proc. 28th Annual Conf, Hydraulics, JSCE.
66.Sivapragasam, C., Shie-Yui, et al. (2001). "Rainfall and runoff forecasting with SSA-SVM approach." Journal of Hydroinformatics 03.3: 141-151.
67.Suna, X., R. G. Meina, et al. (2000). "Flood estimation using radar and rain -gauge data." Journal of Hydrology 239: 4-18.
68.Takasao, T., M. Shiiba, et al. (1994). "A real-time estimation of the accuracy of short-term rainfall prediction using radar." Stochastic and Statistical Methods in Hydrology and Environmental Engineering Kluwer Academic Publishers, 2( ed. K. W. Hipel): 339-351.
69.Teschl, R., W. L. Randeu, et al. (2007). "Improving weather radar estimates of rainfall using feed-forward neural networks." Neural Networks 20: 519–527.
70.Tetzlaff, D. and S. Uhlenbrook (2005). "Significance of spatial variability in precipitation for process-oriented modelling: results from two nested catchments using radar and ground station data." Hydrology and Earth System Sciences 9: 29-41.
71.Trafalis, T. and B. Santosa (2001). "Predicting monthly flour prices through Neural Networks, RBFs and SVR" In: Dagli CH, Buczak AL, Ghosh J, Embrechts M, Ersoy O, Kercel S (eds) Intelligent Engineering Systems Through Artificial Neural Networks, ASME Pres 11: pp 745–750.
72.Trafalis, T., A. White, et al. (2002). "Data mining techniques for improved WSR-88D rainfall estimation." Computers in Industrial Engineering 43: 775-786.
73.Trafalis, T. B., B. Santosa, et al. (2005). "Learning networks in rainfall estima -tion." Computational Management Science 2: 229-251.
74.Tsintikidis, D., J. Haferman, et al. (1996). "A neural network approach to estimating rainfall from spaceborne microwave data." IEEE Trans. Geosci. Remote Sens 35: 1079-1093.
75.Umehara, S., T. Yamazaki, et al. (2006). "A Precipitation Estimation System Based on Support Vector Machine and Neural Network." Electronics and Communications in Japan, Part 3 89, No. 3(Translated from Denshi Joho Tsushin Gakkai Ronbunshi, Vol. J86-D-II, No. 7, July 2003, pp. 1090–1098): 38-47.
76.Vapnik, V.N., The Nature of Statistic Learning Theory, Spring-Verlag, New York, 1995.
77.Wood, S. J., Jones, D.A., Moore, R.J. (2000). "Static and dynamic calibration of radar data for hydrological use." Hydrology and Earth System Sciences 4 (4): 545–554.
78.Xiao, R. and V. Chandrasekar (1995). Multiparameter radar rainfall estimation using neural network techniques. Preprints, 27th Conf, Radar Meteorology, Vail, CO, Amer. Meteor. Soc.
79.Xiao, R. and V. Chandrasekar (1997). "Development of a neural network based algorithm for rainfall estimation from radar observations." IEEE, Transactions on Geoscience and Remote Sensing 35: 160–171.
80.Xin, L., G. Recuter and B. Larochelle (1997). "Reflectivity-rain rate relation -ship for convective rainshowers in Edmonton." Atmos. Ocean 35: 513 -521.
81.Xu, G. and V. Chandrasekar (2005). "Operational feasibility of neural-network based radar rainfall estimation." IEEE Geoscience and Remote Sensing Letters 2(1): 13–17.
82.Zbyne'k, S. (2003). "Utilization of regression models for rainfall estimates using radar-derived rainfall data and rain gauge data." Journal of Hydro -logy 278: 144-152.