On the application of multiprocessor computer systems for modelling of regional atmospheric pollutant transport

Development of an effective strategy of the reduction of air pollution in the regional scale requires fast estimation of both existing pollution level and that resulting from the different possible sets of measures affecting the pollutant emissions from industrial sources. In order to accelerate the necessary calculations the Euleran-Lagrangian scheme for regional air pollution modelling and the corresponding computer code were modified for multiprocessor systems. The code was tested on three systems with local memory and different software environment. The results demonstrated the effectiveness of high performance computing in the field of environmental management.