Assessing particle dry deposition in an urban environment by using dispersion models

Özet

Deposited particles can potentially increase loads of nutrients or harmful compounds in the environment, and can be a source of nuisance to human populations. To estimate particle deposition flux, dispersion models usually include specific algorithms that consider meteorological conditions, land use, and characteristics of particles such as size distribution, density and diffusivity in the air. In this study, two Gaussian dispersion models, AERMOD and CALPUFF, were evaluated to determine their accuracy in predicting particle deposition fluxes. We compared their outcomes with one year of field experimental data obtained in an urban region with moderately complex terrain. As dispersion models rely on emission inventories which do not often include the contribution of fugitive sources, models validation by using field experimental data are compromised. This drawback to the validation procedure can be overcome by using receptor model results to identify the fugitive source apportionment and exclude their contribution from the observational data. With this model validation procedure, and by using the default parameters recommended for each model, AERMOD showed better agreement with observational data (0.35 FB, −0.05 FSD, 0.66 NMSE, and 0.47 COC) than CALPUFF (−0.60 FB, −1.33 FSD, 1.05 NMSE, and 0.38 COC). Sensitivity tests were performed to investigate model algorithms in order to estimate particle deposition flux, and revealed that deposition fluxes varied up to 220% in AERMOD and 16,376% in CALPUFF according to particle size distribution.