The Tendency of Eutrophication Level Prediction in Chengchinghu Reservoir, Kaohsiung City, Taiwan

Abstract

Introduction: Reservoir management problems are increasing, and tools are needed to categorize and predict their eutrophication status in order to provide technical support for the government’s decision to protect drinking water resource. Thus, this study aims to predict and classify the tendency of eutrophication level in Chengchinghu Reservoir, Kaohsiung City, Taiwan as one of major water sources for industrial and domestical needs by supplying 109,170,00 m3 for Southern Taiwan. Method: The CTSI (Carlson’s Trophic States Index, which calculated from Chl-a, TP, and transparency) datasets in winter (December-February), spring (March-May), summer (June-August), and fall (September-November) from 2000 to 2017 was collected from Taiwan Environmental Protection Administration (EPA). This study used the Classification and Regresiion Tree (CART) model provides the explicit categorical rules for Chengchinghu Reservoir. Results and Discussion: The CART results for Chengchinghu Reservoir showed the good performance of prediction since the accuracy of the CART training process value reached 61.89%. According to the CART results, the eutrophic state condition is most probably occur in Chengchinghu Reservoir when the TP concentration is greater than 22.86 mg/L or Chl-a concentration is greater than 5.2 μg/L or SD is less than 1.1 m. Conclusion: The CART result may helps the local governments to understand the pollution conditions in Chengchinghu Reservoir and take responsibility for reservoir water management and conservation. Therefore, they could make policies to treat and manage water pollution sources in Chengchinghu Reservoir.