A semi-empirical model for urban PM10 concentrations, and its evaluation against data from an urban measurement network

We have developed a model for evaluating the mass-based concentrations of u rban particulate matter. The basic model assumption is that local vehicular traffic is responsible for a substantial fraction of the street-level conc entrations of both PM10 and NOx either due to primary emissions or resuspen sion from street surfaces. The modelling system utilises the data from an a ir quality monitoring network in the Helsinki Metropolitan Area. We have de termined linear relationships between the measured urban PM10 data against those of NOx in various urban surroundings, based on continuously measured hourly concentration values. The data was obtained from two stations in cen tral Helsinki and one suburban station in the Helsinki Metropolitan Area du ring a period of 3 yr, from 1996 to 1998. The model also includes a treatme nt of the regional background concentrations, and resuspended particulate m atter. The model performance was evaluated against the measured PM10 data f rom the above-mentioned three stations and from two other stations, using d ata that was measured in 1999. We used two alternative model versions, one based on separate correlation parameters (PM10 vs. NOx) for each station, a nd another based on parameters averaged over the stations considered. We an alysed the agreement between the measured and predicted hourly concentratio n time series, utilising the values of the fractional bias (FB) and the so- called index of agreement (IA). As expected, the model predicts relatively well the yearly mean concentrations of PM10: the FB values range from -0.05 to +0.09. Model performance is also relatively good when predicting the ye arly mean values that are classified separately for each hour of the day: t he corresponding IA values range from 0.85 to 0.96. However, model performa nce is substantially worse in predicting the hourly time series of the year : the IA values using the station-specific parameters range from 0.46 to 0. 65. The model was applied in evaluating the yearly average spatial concentr ation distribution of PM10 in central Helsinki, based on the corresponding modelled NOx concentrations. With re-evaluation of a few parameters that ca n be determined empirically, the model could be evaluated, and most probabl y applied, in other urban areas as well. (C) 2001 Elsevier Science Ltd. All rights reserved.