Traffic Flow Characteristics and Impacts on Air Quality at Urban Roundabout Junction

Abstract

Rapid urbanization and the unprecedented growth in vehicles have resulted in profound deterioration of air quality in urban centers. Traffic junctions of urban centers often attract large number of vehicles creating congested conditions which, as a result, generate higher carbon monoxide (CO) emissions. Since traffic flow characteristics generally observed at junctions complicate dispersion phenomena, understandings on pollutant dispersions in the close vicinity of traffic junctions are important for accurate air quality assessment. Traffic junctions in particular of type conventional roundabouts, where vehicles spiral in and out change lanes and vary speeds in circular pattern, tend to impose behavioural changes on drivers to utilize the free space on road while maneuvering. This generates a zone in which emissions are recirculated within road-width leading to the canyon-type effects between the continuous moving vehicles. Several dispersion models exist to evaluate air quality near roadways and traffic junctions. However, complex pollutant dispersion at a conventional non-signalized roundabout cannot be described accurately by either intersection or open-terrain line source models alone. In order to simulate such a complex dispersion pattern at a nonsignalized roundabout, it is proposed that a line source model is combined with streetcanyon effects. This is demonstrated by estimating 1-min average CO concentration for a period of 30 min by coupling an open-terrain line source model, GFLSM with a streetcanyon (SC) model, STREET to capture the combined (SC-GFLSM) effects to describe the dispersion pattern. This research involves the development of detailed methodologies to quantify traffic flow characteristics and emissions pertinent to local conditions as inputs to the SC-GFLSM model. A time-width occupancy model was developed to simulate traffic flow characteristics and semi-empirical model was developed to simulate emission pattern. The results show that the GFLSM predicted measured concentrations almost three times higher, while the results of the SC-GFLSM matched well with the measurements and the prediction errors reduced by about 50 %. The research, further, demonstrated this with traffic characteristics and emissions calculated by field and semi-empirical methods. The SC-GFLSM model was also validated for another 30 min data set for which the results were equally promising.....