結合衛星影像與模糊理論於水庫水質優養判釋與管理

Integration of Satellite Imagery and Fuzzy Theory into Eutrophication Interpretation and Management for Reservoir Water Quality

楊明德(Ming-Der Yang) ； 林佑昌(Yu-Chang Lin) ； 蔡婷鈺(Ting-Yu Yeh) ； 楊曄芬(Yeh-Fen Yang)

《中國土木水利工程學刊》 20卷2期 (2008/06) Pp. 205-215

https://doi.org/10.6652/JoCICHE.200806_20(2).0005

摘要

目前水庫水質優養程度的評估，主要使用卡爾森優養指標針對點狀的水質採樣數據進行評等。然而點狀採樣成果可能不足代表全水域優養程度，且既有指標對各優養程度的數值規範明確，無法全然詮釋氣候變化或人為觀點改變時水質潛在的優劣。本研究以翡翠水庫為研究區域，利用2004年二月SPOT4衛星影像萃取之水體影像，推求與葉綠素a、總磷及透明度等水質相關參數，以建立衛星影像轉換水質參數回歸模式，達成全水域優養評估；並引入模糊理論方法於傳統卡爾森指標，考慮使用者注重水質安全與管理者注重管理成本的立場設計對應的隸屬函數，以模糊綜合評價討論兩者對同一優養評等下的差異認定。研究成果顯示翡翠水庫在2004年二月時兩種觀點之綜合評價大多呈現貧養，僅有以總磷濃度認定之優養程度較有差異，於水庫上游與匯入支流處有部分水域為優養，其優養水域面積在使用者立場與管理者立場下分別佔總面積27%與16%，顯示總磷汙染為翡翠水庫的主要汙染型態。由於水源管理者對水質認定趨向寬鬆標準，顯示在管理者隸屬函數下所判定之16%優養水域之水質確實不佳，應優先致力於該區水質改善，此種模糊優養判釋模式未來可考慮季節水質變化趨勢於隸屬函數中，作為不同時節制定水質改善策略之依據。

關鍵字

模糊理論 ； 衛星影像 ； 影像處理 ； 水質管理

並列摘要

At present the eutrophic status of reservoir water quality in Taiwan is classified by Carlson index on point-basis sampling data. However, a limited number of water samples are insufficient to represent the entire waters, and the crisp eutrophic criteria can not reflect all aspect from different viewpoints varying with human's altitude and seasonal change. Taking a SPOT satellite image of Feitsui Reservoir in February 2004 for example, this paper addresses how to transform a satellite image into a twodimensional distribution of water quality variables, such as chlorophyll a, total phosphorus and Secchi depth, and subsequently to employ fuzzy criteria in Carlson index for eutrophic dispute between water suppliers and users. The Feitsui Reservoir in February 2004 is evaluated consonant as oligotrophe from both producer's and user's viewpoints. However, based on the criteria of total phosphorus there is 27% and 16% of total water area being considered as eutrophe from the user's and the producer's standpoints, respectively, that also implies that total phosphorus is the major factor of Feitsui's eutrophicaiton problem and those 16% water area should be paid great attention to in the first priority. Through a proper design of membership functions, the fuzzy evaluation model of eutrophication can also be adjusted for standing at different time points during a year while considering the seasonal eutrophication pattern in the water quality management strategy in the future.

並列關鍵字

fuzzy theory ； satellite imagery ； image process ； water quality management