A Study on Algal Bloom Forecasting Methodology Based on Hydro-meteorological Factors Using Machine Learning in the Mainstream of Nakdong River, South Korea

Tuesday, May 20, 2025

2:00 PM – 2:15 PM Alaska Daylight Time (AKDT)

Room: Room 7 & 8

Author(s)

Taewoo Lee (he/him/his)

Ph.D. student
Inha University / Program in Smart City Engineering, Republic of Korea

Co-Author(s)

Soojun Kim

Inha University / Department of Civil Engineering, Inch'on-jikhalsi, Republic of Korea
Hung Soo Kim (he/him/his)

Inha University / Department of Civil Engineering, Kyonggi-do, Republic of Korea
Yujin Kang

Inha University / Department of Smart City Engineering, Inha University, Republic of Korea
Seonuk Baek

Inha University / Department of Smart City Engineering, Inha University, Republic of Korea
Seungmin Lee

Inha University / Department of Smart City Engineering, Inha University, Republic of Korea

Abstract Submission: In the freshwater ecosystem, blue-green algal bloom, or harmful algal bloom(HAB) has a negative impact on the aquatic ecosystem and purified water supply system due to oxygen depletion in the water body, odor, and secretion of toxic substance. This Blue-green algal bloom is expected to increase in intensity and frequency due to the increase in algae's hydraulic retention time(HRT) in the water body after the construction of the Nakdong River weir, as well as the increase in surface temperature due to climate change. In this study, an algal bloom forecasting methodology based on hydro-meteorological factors was presented for preemptive response before issuing a algal bloom warning. Through polyserial correlation analysis, the preceding influence periods of temperature and discharge according to the algal bloom forecast level were derived. Using the decision tree classification, a machine learning technique, classification models for the algal bloom forecast levels based on temperature and discharge of the preceding period were derived. And a algal bloom forecasting methodology based on hydro-meteorological factors was derived based on the results of the decision tree classification models. The proposed algae forecasting methodology based on hydro-meteorological factors can be used as basic research for preemptive response before blue-green algal blooms.

Learning Objectives/Expected Outcome (Optional) :