Post-Processing Methods for the Enhancement of Air Quality Forecasts in the Veneto Region Generated by CAMx Running on a Cloud Infrastructure

Air pollution is a crucial environmental parameter associated with alterations in the chemical composition of air. It is well-documented that elevated concentrations can be dangerous to health and damage the environment, biological resources and ecosystems. Measuring such concentration is an important task, especially in regions where, for morphological reasons, pollutants tend to persist longer. The Po Valley is an example where concentrations can reach the legal limits set by the European Union. The Regional Environmental Protection Agency Veneto (ARPAV) is the agency responsible for air quality control in Veneto: it measures the concentration of the principal pollutants (PM10, PM2.5, NO2, O3) through a station network. A deterministic Eulerian Chemical Transport Model (CAMx) is also used to get better estimates and forecasts of the whole territory, even far from stations. This thesis focuses on post-processing the model forecasts using different techniques: Mean Subtraction, Ratio Adjustment, Hybrid Forecast and Analogue Ensemble technique. The aim is improving their reliability over one-, two-, and three-day horizons. Statistical methods are applied to combine model outputs with observed data in order to correct systematic biases and enhance forecast accuracy in an operational context. After an introduction to the context where the problem arises, the work will focus on the data and statistics used for the analysis. First, an overview of air pollutant concentration data (PM10, PM2.5, NO2, O3) for the Veneto region is presented, describing how the data are gathered and organized. Afterwards, the deterministic model is also briefly described, with a particular interest in the verification with respect to the measurements. The main part of the work is an analysis of the technique to correct the model predictions using station measurements. Finally, the results of the analysis are discussed.