Nitrogen dioxide (NO2) Meteorology and predictability for air quality management using TROPOMI
Prince Junior Asilevi, Enock Nyameasem Dzidzorm, Patrick Boakye, Emmanuel Quansah
Abstract
Abstract Nitrogen dioxide (NO₂) is a critical air pollutant and key indicator for air quality. Due to limited monitoring, we leveraged TROPOMI NO₂ and NASA POWER meteorological datasets to evaluate the meteorological drivers on NO₂ tropospheric column concentrations and to develop predictive models for NO₂ levels over Ghana. Employing an 8:2 ratio for model training and testing, NO₂ and meteorology relationships were assessed by seasonality indices and correlation analyses. Results indicate marked seasonal variability in NO₂ columns, prominent during the dry season. Wind speed, relative humidity, and precipitation significantly reduce NO₂, whereas temperature correlated positively in the southern forested zone. Predictive models demonstrate varying efficacy across climatic zones, with mean percentage differences ranging 9.87 to 37.76% and agreement index up to 0.96. The Random Forest and XGBoost models showed outstanding performance, with correlation reaching 0.92. This results presents a scalable methodology for NO₂ monitoring providing insights for air quality management.