An operational method for assessing traffic-related air pollution in urban streets

Urban air quality has been a topic of major public concern and scientific research in recent years. Several theoretical and experimental studies have focused on the assessment of air quality within street canyons and other microenvironments (intersections, motorways, parking spaces, etc.), where population exposure to traffic-related pollutants is relatively high. The aim of this study was to develop a practical methodology for assessing traffic-related air pollution in urban streets, after testing available monitoring and modelling techniques. To meet this objective, a large amount of original air quality, meteorological and traffic data were collected during four intensive short-term and one long-term monitoring campaigns carried out in the region of Paris from December 1998 to December 2001. These campaigns covered three representative street canyon sites (Bd. Voltaire, Rue de Rennes, Av. Leclerc - PI. Basch) as well as a motorway service station (RN10 petrol station). Passive and active monitoring techniques were used to sample a wide range of inorganic (CO, NO X and Os) and organic gases (benzene, toluene, xylene, ethylbenzene, formaldehyde, acetaldehyde, etc.) at different heights and distances from the kerb. Indicative background measurements were also taken during the same sampling periods. Furthermore, relevant meteorological (synoptic and local) and traffic information was obtained on each site. The analysis of the data gave insights into the dispersion and transformation processes taking place within the streets. Channelling effects induced by parallel to the road axis winds gave rise to relatively high kerbside pollution levels. On the other hand, perpendicular synoptic winds generated air vortices within the canyons, which resulted in steep crossroad concentration gradients. In that case, higher pollution levels were observed on the leeward than on the windward side of the streets. A significant reduction of concentrations with height above the ground was also observed within two of the street canyons (Bd. Voltaire and Av. Leclerc). In all cases, roadside concentrations were several times higher than the corresponding urban background values. This spatial variability indicates a strong transport effect on the pollutant distribution within urban canyons, caused by the synoptic wind and influenced by the geometry of the street. That may have serious implications in terms of population exposure and compliance with air quality legislation. In this context, the siting of permanent monitoring equipment becomes crucial. A relationship between CO and benzene as well as an exponential expression linking pollutant concentrations at different heights within the canyons were empirically deduced. Five dispersion models of different levels of complexity (STREET-SRI, OSPM, AEOLIUS, CAR-International, and CALINE4) were used to calculate CO and benzene concentrations at the campaign sites. The Computational Fluid Dynamic code PHOENICS was also tested for one location. The comparison between observed and predicted values revealed the advantages and drawbacks of each model in association with the configuration of the street and the meteorological conditions. Furthermore, a sensitivity and uncertainty analysis involving three of the available models (STREET-SRI, OSPM and AEOLIUS) was carried out. OSPM was slightly modified in order to allow user access to certain internally coded parameters. An operational method combining multi-site sampling and dispersion modelling was finally proposed for assessing air quality in urban streets, taking into account the pronounced spatial and temporal variability of traffic-related air pollution, the modelling uncertainty, the practical constraints related to measurements and models, and the needs of decision makers. This methodology may find wider application in air quality management, urban and transport planning, and population exposure studies