Very high ground level concentrations of PM in winter and of ozone in summer often occur in Northern Italy, due to the high anthropogenic emissions and frequent stagnant meteorological conditions that characterize the area. These problems are not only related to urban, but also to suburban areas through the entire Po Valley. In such a situation it is important to use deterministic Chemical Transport Models, that allows to evaluate the effect of different air quality control policies on secondary pollution concentrations. Chemical Transport Models generally are part of more complex deterministic modelling systems, encompassing also emission models, meteorological models, and initial and boundary condition processors. Meteorological models are an important module of deterministic modelling systems and, due to their complexity, require high computational costs to perform simulations. In fact they solve a full set of non-hydrostatic equations that describe atmospheric dynamics and thermodynamics, and conservation equations, usually considering two-way interacting nested domains. Within the HPC-EUROPA (Pan-European Research Infrastructure on High Performance Computing) cooperation project, that allows to use clusters of CPUs all around Europe, the meteorological fields over Northern Italy were simulated using RAMS4.4 in parallel mode, creating a database for future air quality assessments. In the present work a CPUs cluster of the Italian computing centre CINECA have been used. The meteorological simulations have been performed considering three nested grids. The first grid covers an area that encompasses the entire Europe, the second grid is focused on Mediterranean sea, while the third one is limited to the Po Valley area. The spatial resolution of the three grids is respectively 128 km, 32 km and 8 km. The number of cells for the three grids is respectively 40×40, 86×86 and 102×102, with 33 vertical levels covering the domain from surface to roughly 20 km height. The entire 2004 year has been simulated through 72 simulations of 126 hours each, considering a spin-up time of 6 hours and 16 CPUs each simulation. In this paper the model configuration and the validation of the simulated meteorological fields are presented.
Validation of mesoscale meteorological simulation over Po Valley for air quality applications
Candiani Gabriele;
2008
Abstract
Very high ground level concentrations of PM in winter and of ozone in summer often occur in Northern Italy, due to the high anthropogenic emissions and frequent stagnant meteorological conditions that characterize the area. These problems are not only related to urban, but also to suburban areas through the entire Po Valley. In such a situation it is important to use deterministic Chemical Transport Models, that allows to evaluate the effect of different air quality control policies on secondary pollution concentrations. Chemical Transport Models generally are part of more complex deterministic modelling systems, encompassing also emission models, meteorological models, and initial and boundary condition processors. Meteorological models are an important module of deterministic modelling systems and, due to their complexity, require high computational costs to perform simulations. In fact they solve a full set of non-hydrostatic equations that describe atmospheric dynamics and thermodynamics, and conservation equations, usually considering two-way interacting nested domains. Within the HPC-EUROPA (Pan-European Research Infrastructure on High Performance Computing) cooperation project, that allows to use clusters of CPUs all around Europe, the meteorological fields over Northern Italy were simulated using RAMS4.4 in parallel mode, creating a database for future air quality assessments. In the present work a CPUs cluster of the Italian computing centre CINECA have been used. The meteorological simulations have been performed considering three nested grids. The first grid covers an area that encompasses the entire Europe, the second grid is focused on Mediterranean sea, while the third one is limited to the Po Valley area. The spatial resolution of the three grids is respectively 128 km, 32 km and 8 km. The number of cells for the three grids is respectively 40×40, 86×86 and 102×102, with 33 vertical levels covering the domain from surface to roughly 20 km height. The entire 2004 year has been simulated through 72 simulations of 126 hours each, considering a spin-up time of 6 hours and 16 CPUs each simulation. In this paper the model configuration and the validation of the simulated meteorological fields are presented.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
