Air quality models include representations of pollutant emissions, which necessarily entail spatial averaging to reflect the model grid size; such averaging may result in significant uncertainties and/or systematic biases in the model output. This study investigates such uncertainties, considering ozone concentrations in idealised street canyons within the urban canopy. A photochemical model with grid-averaged emissions of street canyons is compared with a multiple-box model considering each canyon independently. The results reveal that the averaged, 'one-box' model may significantly underestimate true (independent canyon mean) ozone concentrations for typical urban areas, and that the performance of the averaged model is improved for more 'green' and/or less trafficked areas. Our findings also suggest that the trends of 2005-2020 in emissions, in isolation, reduce the error inherent in the averaged-emissions treatment. These new findings may be used to evaluate uncertainties in modelled urban ozone concentrations when grid-averaged emissions are adopted.
Keywords: Emission heterogeneity; Photochemical box model; Segregation effect; Urban ozone concentrations; Urban street canyon.
Copyright © 2014 Elsevier Ltd. All rights reserved.