The application of CALMET/CALPUFF modelling system is well known, and several validation tests were performed until now; however, most of them were based on experiments with a large compilation of surface and aloft meteorological measurements, not always available. Also, the use of an operational large smokestack as tracer source is not so usual. In this work, CALPUFF model is applied to simulate the local dispersion of SO
(as a tracer) from the large smokestack (with four parallel liners) of a coal-fired power plant emitting SO
(as a tracer), considering both different stack configurations (one single point source vs. one point source per liner) and meteorological inputs (WRF model output vs. measurements). Comparison of CALPUFF results against glc measurements along three different periods shows that the best model performance was obtained by using WRF model output; better results, but not so significant, are obtained considering one point source per liner.
Keywords: air pollution, plume dispersion, CALPUFF, stack configuration, PBL modelling, model validation