臺灣水庫優養化之成因分析

Identifying the factors affecting eutrophication level in Taiwan major reservoirs

卜維拉(Marsha Savira Agatha Putri)

指導教授 ： 王玉純

中原大學/工學院/環境工程學系/碩士(2019年)

https://doi.org/10.6840/cycu201900015

摘要

氣候變遷將使得水溫較高並可能會導致水體污染濃度上升及優養化的風險。 而近年來，優養化已成為水庫管理的其中一項關鍵議題，尤其是當水庫用為重要的供水來源. 本研究旨在評估卡爾森指數，因此選定四個臺灣主要水庫，即石門水庫、鯉魚潭水庫、烏山頭水庫和澄清湖水庫作為研究對象。應用主成分分析-絕對主成分分析(Principal Component Analysis - Absolute Principal Component Scores, PCA - APCS)以確認導致優養化的關鍵因素，並利用多元線性迴歸(Multiple Linear Regression, MLR) 及分類與迴歸樹(Classification and Regression Tree, CART)以預測未來卡爾森指數的趨勢。
由交通部中央氣象局及行政院環境保護署分別收集2000年至2017年的相對濕度、總降水量、日照百分比和雲量5項氣象參數及葉綠素-a(Chl-a)、總磷酸鹽(TP)、透明度(SD)、pH值、化學需氧量(COD)、懸浮固體(SS)、氨、總硬度、硝酸鹽、亞硝酸鹽和水温(WT) 11項水質參數之數據。另外由科技部臺灣氣候變遷推估資訊與調適知識平台(Taiwan Climate Change Projection and Information Platform, TCCIP)選用代表濃度途徑(Representative Concentration Pathways, RCPs)的RCP 2.6、RCP 4.5、RCP 6.0、RCP 8.5 四種情境下及臺灣地區適合的全球氣候模式(Global Climate Model, GCM)所預測之平均氣溫數據與降雨量，來推估21世紀中每一水庫於近未來(2016-2035)、中未來(2045-2065)與遠未來(2081-2100)三個時段的未來溫度。
研究發現依照2000年至2017年四個水庫的卡爾森指數大多可將水庫區分為貧養狀態的鯉魚潭水庫及烏山頭水庫、普養狀態的石門水庫、優養狀態的澄清湖水庫。其中利用PCA/APCS模式解析水庫優養化的來源分配如下：石門水庫分別以營養鹽因素(16%)、降雨強度因素(51%)、溫度因素(3%)；鯉魚潭水庫分別以營養鹽因素(35%)、降雨強度因素(38%)、溫度因素(27%)；烏山頭水庫分別以營養鹽因素(19%)、降雨強度因素(41%)、溫度因素(40%)；澄清湖水庫則分別以營養鹽因素(25%)、降雨強度因素(58%)、溫度因素(17%)。依據21世紀近未來、中未來與遠未來的卡爾森指數之程度，優養化的情形將會出現在石門水庫、 烏山頭水庫及澄清湖水庫。而鯉魚潭水庫未來則可能會發生中度優養化的狀況，在下列條件下將使水庫產生優養化：石門水庫(Chl-a > 4.5、 μg/L、 TP > 0.0309 mg/L與SD < 1 m)；鯉魚灘水庫(TP > 5.9 mg / L、Chl-a > 15.6 μg/ L 與 SD < 0.8 m) ；烏山頭水庫(Chl-a > 1.24μg/ L 、SD < 1.2 m 與 TP > 4.97 mg/L)；澄清湖水庫(TP > 22.86 mg / L、Chl-a > 5.2μg/ L與SD < 1.1 m)。
本研究認為對於臺灣主要水庫來說氣候變遷與營養鹽攝取量可能是影響優養化的因素之一，並證明極端高溫、人為活動及自然因素將會提升整個主要水庫歷史及預測的卡爾森指數的潛在水平。因此，建議可透過持續監測評估、適當水處理措施、建置表面流人工溼地及嚴格規範以控管臺灣水庫的優養化程度。

關鍵字

優養化 ； 預測 ； 氣候變遷 ； 水質 ； 臺灣主要水庫

並列摘要

Climate change may increase the risk of eutrophication in water bodies due to higher water temperatures. Eutrophication has become a critical issue for reservoir management in recent years especially when the reservoir serves as a major water supply source. This study aims to evaluate the Carlson’s trophic states index (CTSI), identify the key factors affecting eutrophication and their source apportionment using principal component analysis-absolute principal component scores (PCA-APCS), to predict the Carlson’s trophic states index (CTSI) tendency using linear regression prediction and classification and regression tree (CART) and suggest the appropriate water management strategy to control the eutrophication in four Taiwan major reservoirs, namely Shihmen, Liyutan, Wushantou, and Chengchinghu Reservoirs.
This study used 5 weather parameters (air temperature, relative humidity, total precipitation, sunshine percentage and cloud amount) obtaining from Taiwan Central Weather Bureau and 11 water quality parameters datasets (chlorophyll-a (Chl-a), total phosphorus (TP), transparency (SD), pH, chemical oxygen demand (COD), suspended solid (SS), ammonia, total hardness, nitrate, nitrite and water temperature) obtaining from Taiwan Environmental Protection Administration from 2000 to 2017. The mean air temperature and rainfall intensity prediction dataset under the representative concentration pathways (RCP) 2.6, 4.5, 6.0 and 8.5 with 5 general circulation models (GCMs) for each reservoir for early (2016-2035), middle (2046-2065), and end (2081-2100) of 21st century were collected from Taiwan Climate Change Projection and Information Platform Project (TCCIP).
This study found the trophic states index status were mostly defined as the mesotrophic reservoir in Shihmen, Liyutan, Wushantou Reservoir from 2000 to 2017, and Chengchinghu Reservoir was defined as the eutrophic reservoir. The eutrophication source apportionment in Shihmen, Liyutan, Wushantou and Chengchinghu Reservoir are 16%, 35%, 19% and 25% from nutrient factor; 51%, 38%, 41% and 58% from rainfall intensity factor; 33%, 27%, 40% and 17% from temperature factor, respectively. The eutrophic condition would be potentially occurred in Shihmen, Wushantou, and Chengchinghu Reservoir, and mesotrophic condition would be in Liyutan Reservoir in association with ambient temperature effect in the future. However, the eutrophic condition would be potentially occurred in all reservoirs in association with mean rainfall intensity in the future. The eutrophic condition would be happened if Chl-a > 4.5 μg/L or TP > 0.0309 mg/L or SD < 1 m in Shihmen Reservoir; TP > 5.9 mg/L or Chl-a > 15.6 μg/L or SD < 0.8 m in Liyutan Reservoir; Chl-a > 1.24 μg/L or SD < 1.2 m or TP > 4.97 mg/L in Wushantou Reservoir; and TP > 22.86 mg/L or Chl-a > 5.2 μg/L or SD < 1.1 m in Chengchinghu Reservoir.
This study concludes that a lot of nutrient intake amounts and climate change could be the factors affecting the eutrophication in the Taiwan major reservoirs. The potential increment of CTSI tendency in whole major reservoirs would be caused by the anthropogenic activity, extreme high-temperature and rainfall in the future. Thus, the continuous monitoring assessment, appropriate water treatment process, construct the free water surface wetlands, and strict regulation were highly recommended to control the trophic state level in Taiwan major reservoirs.

並列關鍵字

Eutrophication ； prediction ； climate change ； water quality ； Taiwan major reservoirs

參考文獻

[1] R. Moss et al., "Towards new scenarios for the analysis of emissions: Climate change, impacts and response strategies," ed: Intergovernmental Panel on Climate Change Secretariat (IPCC), 2008.

[2] N. V., "Integrated water quality management," Water Science and Technology, vol. 33, no. 4-5, 1996.

[3] L. Somlyody, "Water quality management: Can we improve integration to face future problems?," Water Science and Technology, vol. 31, no. 8, 1995.

[4] J.-S. Chou, C.-C. Ho, and H.-S. Hoang, "Determining quality of water in reservoir using machine learning," Ecological Informatics, vol. 44, pp. 57-75, 2018/03 2018.

[5] P. G. Whitehead, R. L. Wilby, R. W. Battarbee, M. Kernan, and A. J. Wade, "A review of the potential impacts of climate change on surface water quality," Hydrological Sciences Journal, vol. 54, no. 1, pp. 101-123, 2009/02/01 2009.