A modeling framework by linking air quality simulation with system optimization was presented in this paper to develop cost-effective urban air quality management strategies in Fengnan district of China. The relation between the total allowable emission and wind speed as well as the relation between the total allowable emission and air-quality-guideline satisfaction were quantified based on the simulation results of the Gaussian-box modeling system. The area-source emission reduction objective in each functional zone of the study city during the heating and non-heating seasons was calculated based on such relations. A linear programming model was then developed to optimize the emission abatement which was subject to a number of dust and SO 2 control measures. The economic objective of the air quality management strategy was to minimize the total emission control system cost while the environmental objective can still be satisfied. The environmental objective was reflected by the emission reduction objective of TSP, PM 10 and SO 2 corresponding to an air-quality-guideline satisfaction percentage of 80%. Consequently, the modeling system comprehensively took into account the information of emission reduction objectives, emission abatement alternatives, emission reduction cost, and related resources constraints. An optimal emission abatement strategy and the related cost were obtained for various pollution control measures. The results would provide sound bases for decision makers in terms of effective urban air quality management and ensuring healthy economic development in the study city.

A Gaussian-box modeling approach was presented in this paper to examine the urban air quality due to multiple point-and area-source emissions in the northern Chinese city of Fengnan, which is associated with a deteriorated air quality as a consequence of industrialization and rapid urban growth. A 3-D multibox (3DMB) air quality model was developed to predict air quality due to area-source emissions. It improved upon the conventional box models by allowing consideration of more details in spatial variations of emission sources and meteorological conditions. The modeling domain was divided into various layers within the mixing height, while each layer was associated with a number of sub-boxes. A multi-source and multi-grid Gaussian modeling approach was then applied to predict the air quality in different sub-boxes that are associated with multiple point-source emissions. Thus the Gaussian-box modeling approach could effectively simulate impacts of both area-and point-source emissions but also reflect more details of the spatial variations in source distributions and meteorological conditions. This modeling approach was employed to predict daily average SO 2 , TSP and PM 10 concentrations for each sub-box during the heating and non-heating seasons, respectively. The analysis of the mean normalized error of the modeling results demonstrates the feasibility and applicability of the developed method, and the presented method could provide more useful and scientific bases for developing effective urban air quality control and management strategies.