Developing receptor-oriented methods for non-methane hydrocarbon characterisation in urban air. Part II: source apportionment

The methods and the results of non-methane hydrocarbon (NMHC) source apportionment are described at urban scale on a spatial and temporal basis. Here, hourly ambient concentrations of nearly 40 C-2-C-9 NMHC are used. Methods are based on the knowledge of the wintertime NMHC vehicle-exhaust emission ratio generally determined by simple regression analysis taking acetylene, ethylene, propene and 1,3-butadiene as auto-exhaust tracers. The RSD of the estimated source contributions is lower than 20%; the developed receptor-oriented methods are flexible and easily transposable to other areas. In winter, vehicle-exhaust emissions explain 100% of the NMHC majority levels and even isoprene. From May to November, our models revealed the temperature-dependent contribution of additional sources (0.71 < r < 0.90). On the one hand, the evaporation of fuel and solvent affects the whole C-4-C-9 NMHC fraction, and fluctuates between 20% and 50%, even for a northern France urban area. On the other hand, both vehicle-exhaust and biogenic emissions control the highly photoreactive isoprene distribution whatever the site; the traffic is responsible for a third of its levels in summer. Finally, the particular behaviour of the C-2-C-4 compounds pointed out dominant contributions, generally other than traffic. Suspected sources are numerous: natural gas leakage for ethane and propane, wintertime fuel evaporation for butanes and butenes, non-automotive combustion for ethylene and acetylene. Ethane and propane also showed that long-range advective transport, responsible for background concentrations, could significantly contribute to the hydrocarbon levels with a high atmospheric residence time (from 20% to 50%). (C) 2003 Elsevier Ltd. All rights reserved.