Assessment of seasonal variations in the air quality index (2019–2022) in Al-Jahra city, Kuwait

Abstract

The daily air quality indices (AQIs) for pollutants, including particulate matter (PM₁₀ and PM_2.5), carbon monoxide (CO), nitrogen oxides (NO₂), ozone (O₃), and sulfur dioxide (SO₂), were evaluated for the period of 2019–2022 in Al-Jahra City, Kuwait. This study is designed to (1) evaluate overall and seasonal changes in pollutants, (2) investigate the correlation for PM₁₀ and PM_2.5 and other pollutants during each season, and (3) examine the best model for prediction of air pollutant concentration of PM₁₀ and PM_2.5. An assessment of air quality indices was carried out by using different algorithms models, including random forest (RF), an artificial neural network (ANN), and extreme gradient boosting (XGBoost). The overall level of PM₁₀, PM_2.5, and SO₂ pollutants shows an increasing trend from 2019 to 2022, reaching their highest in 2022 with a significant decrease in 2020 during COVID-19 lockdown restrictions. The pollutants CO and O₃ reached their peak in 2021.The obtained results showed that the PM₁₀ and O₃ levels are higher in the summer, whereas PM_2.5, NO₂, and CO were recorded at the highest levels during spring, autumn, and winter respectively due to a variation in the meteorologic condition. Furthermore, during the winter season, the PM₁₀ is positively correlated with CO (r_p = 0.401), NO₂ (r_p = 0.121), and SO₂ (r_p = 0.119) AQIs while PM_2.5 is significantly positively correlated with the CO (r_p = 0.198), O₃ (r_p = 0.310), and SO₂ (r_p = 0.129) AQIs. In contrast, the performance of XGBoost is a more reliable predictor for PM₁₀ and PM_2.5 levels.