Water Quality Predictions for Urban Streams Using Machine Learning

作者：Lokesh Jalagam, Nathaniel Shepherd, Jingyi Qi, Nicole Barclay, Michael Smith · 年份：2023 · DOI：10.1109/southeastcon51012.2023.10115154 · 被引用次数：19 · 研究领域：Hydrological Forecasting Using AI、Water Quality and Pollution Assessment、Water Quality Monitoring Technologies

The purpose of this study is to examine the impact of land use and rainfall on water quality in urban streams. This research includes analysis and prediction of a water quality indicator for streams in Mecklenburg County, the most urbanized county in North Carolina, United States. The analysis helps to determine future pollutant levels based on past trends using machine learning models. Land use data from the Multi-Resolution Land Characteristics Consortium that details land development in the county by classes (e.g., urban, forest) was used, along with monthly average precipitation data for the county. The work explores the use of eight regression models to predict levels of the total suspended solids (TSS) pollutant. The accuracy of the prediction is measured by statistical methods and compared among the models. Based on the results, the proposed approach shows viability for water quality prediction.