Seasonal Analysis and Machine Learning-Based Prediction of Air Pollutants in Relation to Meteorological Parameters: A Case Study from Sakarya, Türkiye

Air pollution poses significant environmental and public health challenges, especially in urban–industrial areas where pollutant dynamics are influenced by complex interactions with meteorological factors. This study examines the seasonal variations and correlations between air pollutants (PM10, NO, NO2, and CO) and meteorological parameters (wind speed, temperature, relative humidity, and rainfall) in Sakarya, Türkiye, in 2021–2023. Statistical analyses and predictive models, including multiple linear regression (MLR) and random forest (RF), were applied to evaluate the factors shaping pollutant levels and assess model effectiveness in forecasting air quality. The findings highlight wind speed and rainfall as critical in reducing PM10 and NO concentrations, with notable seasonal effects. RF outperformed MLR for PM10 predictions, while MLR better captured the linear relationships influencing NO and NO2 levels. Both models faced challenges in predicting CO due to its diverse sources and weak meteorological links. The dynamic effects of temperature and relative humidity further emphasize the complexity of pollutant behavior. This research underscores the necessity of integrating meteorological data into air quality strategies and provides actionable recommendations for policymakers and urban planners to advance sustainable urban development.