Observational and modelling analysis of pollutant concentrations and meteorological factors in Andorra: A thermal inversion case study

Abstract

The behavior of pollutants in urbanized mountainous areas is complex due to the interactions between the atmosphere and the orography, and frequent thermal inversions inside the valleys can lead to critical contamination episodes that are detrimental to the environment. This work aims to analyze the behavior of pollutants in a complex orography area like the Central Valley of Andorra and to test the interpretability of WRF in an observed persistent thermal inversion case in the area. Anthropogenic impact on primary pollutants like NO2 and PM10 and the higher dependence of the secondary contaminant O3 on meteorological variables are observed. The Random Forest Regression machine learning model shows that thermal inversions have great importance in the levels of pollutants at the bottom of the valley and also that O3 has a high dependence on relative humidity. The WRF model at higher horizontal resolutions is capable of resolving slope and valley winds but can not interpret well the observed thermal inversions. Increased vertical resolution appears to have little effect on outputs and the need for improved model parameterizations formountain areas is highlighted