Chapter 4.2 Application and sensitivity analysis of CAMx and CHIMERE air quality models in a coastal area

This work is dedicated to the evaluation of the influence of input data on the ability of chemical transport models to reconstruct a given pollution episode. Such a question remains critical in the frame of air pollution forecasting and management. The study relies on the large European ESCOMPTE campaign that took place in the Marseilles area (South-East of France) in summer 2001, and that was dedicated to the constitution of a detailed 3D chemical and meteorological database for a CTM intercomparison exercise. The huge amount of measurement data indeed allowed us to investigate the behaviour of CTMs in a complex environment. In fact, the domain is characterised by the presence of large urban and industrial poles along the coastline served by a dense road network. Submitted to land-sea breeze phenomena, it shows an increasing orographic altimetry which constrains the transport of polluted plumes inland. Simulations, carried out with CAMx and CHIMERE, were focused on two intense ozone episodes: 21-23 June 2001 (moderate synoptic wind) and 24-26 June 2001 (local sea-land breeze circulation). The specificity of this work is to compare the models outputs both in their original configurations and gradually changing the input configuration (meteorological data, boundary conditions) in order to quantify the related effect on the results. The model results were compared with the very rich set of dynamical parameters and chemical data of the campaign, obtained from ground stations, Lidar, aircrafts, boats, balloons, radiosoundings. Statistical indices and performance indicators were also set up and showed that the models could correctly reproduce photochemical smog phenomena over the Marseille area.