Improving air quality assessment over complex terrain by optimizing meteorological and pollutant transport modeling

The Alpine region is a sensitive area to air pollution, as it presents specific characteristics, which expose it to a greater environmental burden with respect to flat areas. During the last decades, the scientific community has developed many different modeling tools to tackle the problem of air pollution. This issue demands at least three distinct procedures: the modeling of the meteorological fields, the modeling of the transport and dispersion of the pollutants and the modeling of the emitting sources. Each of these procedures performs differently across different space and time scales and carries its own strengths and weaknesses, which affect results in terms of pollutant dispersion patterns. The present work focuses on testing and improving different modeling tools at a local scale, over very complex topography, where most of them are brought to work at the limit of their applicability, but they are still the best available tools to face the problem. Different case studies are used in this research in order to evaluate strengths and deficiencies of the models and, where possible, to improve their performance. The experimental datasets used for this purpose come from both previously performed field campaigns and specifically designed campaigns, including meteorological and air quality observations. The performance of Land Surface Models within the Weather Research and Forecasting Model is evaluated and improved, focusing on their ability in reproducing near-ground variables, with specific attention to the frequent ground thermal inversion occurring in the mountainous areas. The performance of dispersion models recommended for applications over complex terrain is also tested and their results are compared with unique measurements (PM10 vertical profiles and tracer gas ground concentration), under challenging wintertime conditions. Atmospheric turbulence parameterizations are also analyzed, in order to understand their role and effects in a modeling chain for dispersion assessment purposes.