碩士 / 國立臺灣海洋大學 / 海洋環境資訊系 / 105 / The purpose, the water quality data provided by the Environmental Protection Department of the Executive Yuan from 2002 to 2017 were collected from the east coast of Taiwan, Yilan Lanyang River Basin, Hualien Hualien River Basin and Taitung Beinan River Basin. Class of neural model prediction, carried out in the assessment of river water quality comprehensive indicators for the "river pollution index, River Pollution Index" referred to as "RPI" model.
Due to the terrain position of the east coast, the typhoon is dominated by the eastern region, and the flow characteristics are short, so the concentration of suspended solids (SS) often increases abruptly. Therefore, the model is found to be affected by the concentration of suspended solids (SS) Resulting in increased bias.
Yilan Lan Yangxi River Basin the RMSE value is 0.444, while the measured value and the four water quality parameters are converted to RPI with MAE value of 0.214 and RMSE value of 0.463. It is found that the accuracy of pollution degree is estimated to be 42.85% by the estimated value of pollution degree. The accuracy of pollution index is 78.57% for the RPI index.
Hualien Hualien River Basin the RMSE value was 1.069, while the measured value and the four water quality parameters were converted to RPI with the MAE value of 0.771 and the RMSE value was 0.988. The results showed that the MAE value was 0.906. The accuracy of the pollution degree is estimated to be 50% by the forecast value. The accuracy of the four indicators is estimated to be 78.57% for the RPI index.
Taitung Beinan River Basin the results of the measured values and estimated values of the MAE were 0.685 and the RMSE values were 0.900. ; And the measured value and four water quality parameters converted to RPI its MAE value of 0.300; RMSE value of 0.577. The accuracy of the pollution degree is estimated to be 42.85% by the forecast value. The accuracy of the four indicators is 64.28% for the RPI index.
The results of this paper can be learned that the four water quality parameters are converted into RPI single index, and the prediction result is good and can be provided to the scholars and related units as a reference.
| Identifer | oai:union.ndltd.org:TW/105NTOU5276022 |
| Date | January 2017 |
| Creators | Hsiang, Hua-Wei, 向華葳 |
| Contributors | Wei, Chih-Chiang, 魏志強 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 92 |
Page generated in 0.0014 seconds