Search

Satellite observations from instruments such as the TROPOspheric Monitoring Instrument (TROPOMI) show significant potential for monitoring the spatiotemporal variability of NO 2 , however they typically provide vertically integrated measurements over the tropospheric column. In this study, we introduce a machine learning approach entitled 'S-MESH' (Satellite and ML-based Estimation of Surface air quality at High resolution) that allows for estimating daily surface NO 2 concentrations over Europe at 1 km spatial resolution based on eXtreme gradient boost (XGBoost) model using primarily observation-based datasets over the period 2019–2021. Spatiotemporal datasets used by the model include TROPOMI NO 2 tropospheric vertical column density, night light radiance from the Visible Infrared Imaging Radiometer Suite (VIIRS), Normalized Difference Vegetation Index from the Moderate Resolution Imaging Spectroradiometer (MODIS), observations of air quality monitoring stations from the European Environment Agency database and modeled meteorological parameters such as planetary boundary layer height, wind velocity, temperature. The overall model evaluation shows a mean absolute error of 7.77 μg/m3, a median bias of 0.6 μg/m3 and a Spearman rank correlation of 0.66. The model performance is found to be influenced by NO 2 concentration levels, with the most reliable predictions at concentration levels of 10–40 μg/m3 with a bias of <40%. The spatial and temporal error analyses indicate the spatial robustness of the model across the study area, with better prediction accuracy during the winter months and the associated higher NO 2 concentrations. Despite the complexity and the continental scale of the study area, the XGBoost-based model shows fast execution potential in providing daily estimates of surface NO 2 concentrations over Europe. The Shapley Additive exPlanations (SHAP) value analysis highlights TROPOMI NO 2 tropospheric column density as the main source of information in deriving surface NO 2 concentrations, indicating its significant potential for such studies. The SHAP values also indicate the importance of anthropogenic emission proxy inputs such as VIIRS night lights, in complementing TROPOMI NO 2 values for deriving higher resolution and detailed spatial patterns of NO 2 variations. • Surface NO2 concentrations at 1 km resolution is estimated over Europe using XGBoost. • Satellite-based input features' potential contribution is observed. • SHAP values indicate highest importance of Sentinel-5P TROPOMI observation. • Finer spatial patterns can be derived using VIIRS nightlight. • XGBoost is a good candidate for continental-scale study areas. [ABSTRACT FROM AUTHOR]