Dispersion modeling of particulate emissions from low-level point sources

Abstract

The use of dispersion modeling as a regulatory tool is continually increasing. In many cases, a source's right to operate hinges on the results of dispersion modeling. To provide fair regulation, it is critical that the models used provide concentration estimates that are as representative as possible. This research examines Gaussian modeling and the inputs and theory behind its application. Specifically, this research looks at the form and function of the horizontal stability parameter ([][y]). This parameter defines the degree of dispersion in the horizontal plane accounted for by the models. This degree of dispersion results from variation in the wind speed and direction over a given period of time. A methodology for estimating the most appropriate value of [][y], based on collected, short-interval meteorological data was developed. This new estimate of [][y] ([][yBKF]) is compared to the presently used Pasquill-Gifford (PG) estimate ([][yPG]). An estimate of the most appropriate time period of application of the [][yPG] is made by comparing [][yPG] to values of [][yBKF] developed for various time interval, downwind distance, stability class combinations. Regression analysis was performed to develop relationship between the hourly plume spread and the hourly observed meteorological variations. The new model, FTAM, retains the Gaussian basis but estimates [][y] using the developed regression fit in place of the PG estimate. Comparative modeling between FTAM, ISC3 and AERMOD is performed in order to demonstrate differences in model performance. Concentration predictions are compared for both 24-hour and 1-hour periods. Overall it was found that, on average, the ISC3 predicted 24-hour concentrations were 130% of FTAM predictions, and 170% of AERMOD predictions. Similarly, the 24-hour predicted concentration from AERMOD were, on average, 80% of FTAM predictions. Overall it was found that, on average, the ISC3 predicted 1-hour concentrations were 370% of FTAM predictions, and 700% of AERMOD predictions. Similarly, the 1-hour predicted concentration from AERMOD were, on average, 144% of FTAM predictions. Additionally, it was found that the FTAM 1-hour concentration predictions could range from 32% to 1460% of the ISC3 1-hour concentration predictions, depending on meteorological variations.