Scenario Analysis of Total Organic Carbon Changes in South Korea's Four Major Rivers under Climate Change Using Machine Learning

Total Organic Carbon (TOC) refers to the total amount of carbon contained in all organic matter present in water and is used as a key indicator of water pollution. Elevated TOC concentrations in water can lead to decreased dissolved oxygen levels and accelerated eutrophication, causing severe impacts on river and aquatic ecosystems. Moreover, the increase in toxic substances and pathogenic microorganisms may compromise the safety of drinking water sources.

Recent changes in rainfall patterns, rising water temperatures, and ecosystem shifts driven by climate change have further increased uncertainties in water quality monitoring and TOC prediction. To mitigate potential socio-economic damages caused by delays in greenhouse gas reduction and carbon neutrality policy implementation, this study aims to predict the TOC concentrations of Korea’s four major rivers—the Geum, Nakdong, Yeongsan, and Han Rivers—using various machine learning algorithms and climate change scenarios based on the IPCC Sixth Assessment Report’s RCP and SSP frameworks.

Water quality data from 2008 to 2022, including water temperature, DO, BOD, COD, chlorophyll-a, TN, TP, pH, conductivity, dissolved total phosphorus, dissolved total nitrogen, NH3-N, NO3-N, SS, and TOC, were combined with daily average temperature, background CO2 concentration, and precipitation data. Various machine learning algorithms, including CNN, ANN, Random Forest, and XGBoost, were employed to compare TOC prediction performance and identify the optimal model. Using the machine learning models trained on historical data, future water TOC concentrations were predicted by inputting scenario-based temperature and precipitation data. Climate change scenario data, specifically the SSP5-8.5 detailed daily data for South Korea, were utilized to predict and compare future TOC concentrations in water from 2023 to 2100 across different time periods.

Through this study, we aim to forecast the changing trends of TOC in Korea’s four major rivers and analyze the significance of TOC in achieving carbon neutrality. This research will contribute to the development of water quality management strategies aligned with climate change mitigation efforts.