Avaliação de desempenho dos modelos AERMOD e CALPUFF associados ao modelo PRIME

Abstract

The presence of houses and buildings and other structures near emission sources affects the pattern of air flow and dispersion of contaminants in surface boundary layer. Moreover, the average time for which the average concentrations of the contaminant are determined depends on the chemical composition of the contaminant and the time of impact. For example, for compounds odorants, the average time should be related to the time interval of an inspiration (1 to 5s) or the time interval for which the compounds causing odor nuisance effectively. One of the tools used in studies of these impacts are mathematical models that have the ability to include the effect of the presence of obstacles and represent concentrations of short-term in the flow. Because of the ease and speed in implementation, the Gaussian models are widely employed with adaptations including the effect of obstacles and represent concentrations of short-term. This study aims to evaluate the models AERMOD and CALPUFF using the PRIME model to consider the effects of the presence of the obstacle, and two methodologies to obtain peak concentrations, one that applies an adjustment factor in the coefficient of dispersion and other applying an adjustment factor directly in the concentrations to small time intervals. The results of modeling were compared with data from wind tunnel experiments, and showed that the models tended to underestimate the concentration values near to those obstacles, with the AERMOD model overestimating their results in relation to CALPUFF. Furthermore, it was possible to infer that the CALPUFF model improves its performance as the distance to the obstacle increases. With respect to the AERMOD model, it was found that best results occur in regions near of the obstacle, higher than CALPUFF in most cases. However, over longer distances, both models estimated similar results. The analysis of mean peak concentrations for short periods of time suggested that the adjustment in the concentrations applied directly on predictions of the models AERMOD and CALPUFF not differ substantially. But when the two methodologies are reviewed in CALPUFF, the highest peak concentrations are found with the adjustment factor applied directly to the concentrations, with a decrease in the difference between the methodologies as they have long time intervals.