Predicting Dominance of Toxic-Producing Cyanobacteria Genus by Machine Learning in a Eutrophic River in South Korea

• Published in: 2023 International Conference on Sustainable Technology and Engineering (i-COSTE) pp. 1 - 5
• Main Authors: Kim, Jayun, Jung, Woosik, Yeon, Seungjae, An, Jusuk, Oh, Hyunje, Park, Joonhong
• Format: Conference Proceeding
• Language: English
• Published: IEEE 04.12.2023
• Subjects: algae; Biological system modeling; Compounds; cyanobacteria; Ecosystems; eutrophic river; Graphical models; Machine learning; Nitrogen; nutrients; Rivers; water quality
• DOI: 10.1109/i-COSTE60462.2023.10500770

Climate crisis is aggravating harmful algal blooms globally. Especially, cyanobacteria harmful algal blooms (CBs) produce taste and odorous compounds and cyanotoxins (microcystins), thereby threatens human health and ecosystem. Cyanotoxins are known to be produced by mainly Microcystis, Anabaena, Oscillatoria, and Aphanizomenon, In the Republic of Korea, extreme CBs occur at the Nakdong River every summer. This study examined spatial distribution of cyanobacterial composition, characterizes major environmental factors influencing microcystins by using machine learning models in the Nakdong River. It was shown that only Microcystis was significant for microcystin and higher ratio of nitrogen to phosphorous contributes to lower microcystin.