Ensemble Modeling of Atmospheric Pollutants: A Case Study with WRF-Chem and LOTOS-EUROS in Aburrá Valley, Colombia

An evaluation of an ensemble of atmospheric Chemical Transport Models (CTMs) was carried out using WRF-Chemand LOTOS-EUROS over the Aburrá Valley, Colombia, an interesting region for huge density of population and mountainous topography. The models were configured in the same spatiotemporal domain and the pollutants of interest were nitrogen dioxide (NO2) and ozone (O3) due to their significant impact on sensitive ecosystems. The ensemble concentrations were assessed by comparing them to the data collected by local air quality monitoring stations against the simulated surface concentration. Additionally, vertical profiles were compared between each model and the ensemble. The Weather Research and Forecasting (WRF) is utilised as the Numerical Weather Model (NWM) driver in the two CTMs with identical initial and boundary conditions, to let the chemical operators from each model to be the contributors to the differences incorporated with an ensemble. The statistical comparison to assess the ensemble includes various metrics expressed in Taylor diagrams. These metrics comprise the mean factorial bias (MFB), root mean square (RMSE) error, and correlation factor (Corr) to evaluate the models and its ensemble against the measurements. The ensemble perspective of the Chemical Transport Model (CTM) diminishes the drawbacks of each CTM and enables us to comprehend the impact of key dynamics over rough topography that has a direct influence on the vertical transport of pollutants. The findings reveal which model aligns better with surface observations in the Aburra Valley and enable a qualitative identification of the principal dynamics in pollutant transport across this region utilizing two CTMs and their ensemble.