Microplastic pollution in the Yangtze River: Characterization, influencing factors, and scenario-based predictions using machine learning method

Microplastic (MP) pollution in the Yangtze River has emerged as a major environmental concern, because MPs are frequently detected and pose serious threats to ecosystems. Understanding the characteristics of MPs is essential for assessing their environmental behavior and associated risks. This paper investigated the current status of MP pollution in the Yangtze River, including the abundance, shape, polymer type, and color. It also explored key geographic and anthropogenic factors (such as road length, precipitation, mismanaged plastic waste, forest volume, cropland area, and altitude) that influence MPs abundance, along with their spatial influence radii. Based on these factors, machine learning models were developed to predict MP abundance under various scenarios, where input factors were increased or decreased by 50 %. The results revealed higher MPs concentrations in downstream areas of the river, with polyethylene (PE) and polypropylene (PP) as the most common polymer types, white/transparent as the predominant color, and fibers as the dominant shape. Among the models developed, the Random Forest model demonstrated the best performance, achieving an R² value of 0.712 and an MSE value of 90.8. Scenario-based predictions identified precipitation as the most influential factor, with decreased precipitation leading to a significant rise in MPs concentrations. Enhanced waste management, lower road density, and greater biodiversity were linked to reduced MP concentrations. These findings highlight the need for targeted management strategies to mitigate MP pollution, particularly during dry seasons.