An important European campaign (ESCOMPTE) took place in the vicinity of Marseilles (South-East of France) in June and July 2001. The main objectives of the field campaign were to analyze and document several photochemical episodes in this area, as well as to constitute a detailed chemical and meteorological database for a Chemical Transport Model (CTM) intercomparison at the regional scale. Within this frame the RAMS – CAMx modelling system has been applied for the period 21-26 June 2001. The first 3 days were characterized by a moderate wind blowing from NW to W and critical ozone levels over SE part of the domain. Differently, during the other 3 days clear local breeze circulation (mountain – sea) developed, inducing high ozone concentrations in suburban areas and in the valleys North of Marseilles. RAMS is a mesoscale atmospheric model used to simulate the mean flow and turbulence. In this frame RAMS simulated all the meteorological fields needed by the chemical model, using three nested grids of 38×49, 38×47, 47×56 cells with horizontal resolution of 30, 10, 3.3 km and vertical resolution of 40 m near the ground up to 1000 m at the top of the domain. The boundary conditions were provided by the ECMWF analysis fields. CAMx is an Eulerian 3D chemical and transport model. In this frame the SAPRC99 chemical mechanism has been chosen. The emission fields have been performed processing the ESCOMPTE inventory, that includes point stacks, area sources and ship emissions. The boundary conditions have been provided at the continental scale by the CHIMERE model. PUBBLICATO A5045869 (PAD – 706727)

The photochemical model has been applied over a 144×150 km2 area, subdivided according to a grid system having 3 km step size and 15 vertical layers. Different input configurations have been analyzed to reproduce the measured ozone concentrations. The great availability of observations has allowed to perform a thorough evaluation of model performances for both meteorological and chemical parameters. This work shows some of the obtained results, highlighting the O3 and NO2 system performances for the two different temporal period and for different geographical areas (urban or rural air quality monitoring stations). Model evaluation reveals the key role of meteorological input driving simulations of severe ozone episodes over complex domain. Finally, a change in the system has been tried out, substituting RAMS with MM5 as meteo driver, to investigate the related effect on CAMx results.