Evaluation of TROPOMI observations for estimating surface NO2 concentrations over Europe using XGBoost Model 

<p>Nitrogen dioxide (NO₂) is among the major air pollutants in Europe posing severe hazard to environmental and human health. The concentrations of surface NO₂ are measured by ground monitoring stations which are fairly limited in representation and distribution. While NO₂ estimates from chemical transport models are realistic, their complexity makes them computationally intensive. Satellite observations from instruments such as TROPOMI provide high spatiotemporal distribution of NO₂. However, these instruments capture NO₂ density only along the tropospheric column and not on the surface. Exploiting the availability of ground station measurements and spatially continuous information from TROPOMI, this study estimates surface NO₂ concentrations over Europe at 1km spatial resolution for 2019-2021 using XGBoost machine learning model. While ground measurements are used as target reference features, satellite observations such as tropospheric column density of NO₂ (from TROPOMI), night light radiance (from VIIRS), NDVI (from MODIS) and modelled meteorological parameters such as planetary boundary layer height, wind velocity, temperature are used as input features to the model. We find an overall mean absolute error of 7.87&#181;g/m3, mean bias of -3.13&#181;g/m3 and spearman correlation of 0.61 during model validation. We found that the performance of the model is influenced by NO₂ concentration levels and is most reliable for predictions at concentration levels <40&#181;g/m3 with a relative bias of <40%. The spatial error analysis also indicates the spatial robustness of the model across the study area. The importance of input features is evaluated using SHapley Additive exPlanations (SHAP), which shows TROPOMI NO₂ being the most important source for the modelled NO₂ predictions. Furthermore, SHAP values also highlight the role of VIIRS night light radiance in deriving finer detailed spatial patterns of surface NO₂ estimates. Despite the complex non-linear relationship of the input features, the trained XGBoost model requires an average of 570 seconds to predict single day surface NO₂ concentrations for the large study area of continental scale. Thus, this work evaluates the importance of TROPOMI data and reliability of machine learning models for estimating surface NO₂ concentrations on a larger spatial scale.</p>