Air pollution exposure monitoring and estimation VII. Estimation of population exposure in a central European airshed

In order to clarify the local variation in exposure and source-receptor rel ationships, a dispersion model for estimating air pollution concentrations was developed for a polluted area in the Czech Republic. Three models chara cterized by different spatial resolution were integrated into one modelling tool. A regional-scale dispersion model accounted for pollution contributi on from sources outside the modelling area. Local- and urban-scale dispersi on models were used to calculate local concentration distributions. Calculated concentration distributions were evaluated. Deviations between o bserved and calculated concentrations were not correlated in space, except in episodes, and concentrations measured at spatially representative statio ns were assimilated into the model results using statistical interpolation (simple kriging). The results indicated that centralized heating plants and local home heatin g were the most important sources for sulfur dioxide (SO2) pollution. Both high and low level sources may contribute to the accumulation of pollution concentrations in episodes. The measured concentrations were important for the description of distributions in episodes characterized by complex wind and dispersion conditions. The applicability of source oriented model calcu lations to correctly represent measured concentrations in the pollution epi sodes was limited due to the fact that meteorological conditions representa tive of high concentration episodes were characterized by very low wind spe ed and variable wind directions. About 8000 individuals were given an exposure estimate representing contrib ution from local emissions, based on the estimated hourly outdoor exposure to SO2 at their home/work addresses in the 3 month study period in the autu mn of 1991. The results showed that, for 5% of participants, the maximum ho urly contribution of local emissions was over 380 mu g m(-3). For the 3 mon th average, both large-scale and local-scale pollution contribute significa ntly. For primary compounds, such as SO2, steep gradients are observed in t he vicinity of strong local sources. These gradients are important for expo sure characteristics and health effect quantification, and often will not b e captured by an existing monitoring network. The calculations can be exten ded to other periods or to different compounds.