Identification et contribution des sources de HCNM en zone rurale

It has been well established that non-methane hydrocarbons (NMHCs) play an important role as precursors of ozone and other secondary pollutants like Oxygenated Volatile Organics Compounds (OVOC), organic nitrates or particulate organics matter. Forty-six non-methane hydrocarbon (NMHC) concentrations have been measured between 1997 and 2006 at three MERA sites (Mesure des Retombées Atmosphériques, French acronym for background air pollution monitoring) according to the EMEP programme. VOCs monitoring has been carried out firstly to estimate their background concentration levels and trends, and secondly to provide observation data for model validation and quality check of VOC emission inventories. The objective of this work was to use this extensive dataset in order to identify NMHC source types and to apportion their contributions.

Data analysis was performed using Positive Matrix Factorization (PMF). The general concept of this receptor model being mass conservation, the computed factors may not be directly related to emission profiles when reactive species are measured far from the sources but can be interpreted as aged emission profiles. Besides, in this work the uncertainty taken into account as PMF input data combined uncertainty of the measurement method and photochemical reactivity of each considered species. The PMF parameters and the uncertainty model were established and discussed by examining the sensibility of the model.

Six factors were considered for the 3 sites. The model results explain on average 86.5% of the total concentrations of the 46 hydrocarbons measured for the 3 sites. About 80% of the variability of the VOC species concentrations could be explained by the model for all sites. The calculated factors have been identified as potential source categories by comparing them to known source profiles from the literature, taking into account species lifetimes and examining time series of source contributions. Five of the six PMF source profiles are common to all sites and all factors are identified as known and aged plume profiles: vehicle exhaust, evaporative sources, residential heating, biogenic and remote sources. Contribution of evaporative sources (29.5%) remains predominant for the southern site while residential heating (28.0% and 21.7%) is the main factor for the two others. The resultant information allowed comparison with emission inventories usually given by activity sectors. Contributions of vehicle exhaust and evaporative sources (36.6 ± 8.4% on average for the 3 sites) are in agreement with emissions coming from road transport and extraction & distribution of fossil fuels (31.6% on average over the period 2002-2006 period), and so is residential heating (19.4 ± 6.4%) which is close to nonindustrial combustion emissions (13.7%). Biogenic and background factors contribute respectively at 10.5 ± 3.5% and 12.7 ± 3.9% on average for the 3 sites.

This work was completed with an air mass trajectory analysis in order to apportion source contribution depending on air mass origins. Each sample was associated with its 3-day back-trajectory arriving at the sampling station at 12:00 GMT (P = 950 hPa) and calculated by the British Atmospheric Data Centre. The K-means clustering algorithm was used to group these trajectories depending on their direction, their speed and their altitude. Two main groups have been distinguished: (1) older and diluted air masses from an oceanic origin where “remote sources” present the main contribution; and (2) anthropogenic and closer sources indicating continental influence. In this group, some clusters represent trajectories coming from the north and distant urban areas with high wind speed, and are marked by high contributions of residential heating and evaporative sources. Trajectories with short pathways corresponding to low wind speed allow to discern more local influences from urban areas with vehicle exhaust and residential heating profiles as main contributions.

Further works concerning the potential source locations of NMHC observed in these French rural sites will be published.