Assessment of source contributions to air pollution in Beirut, Lebanon: a comparison of source-based and tracer-based modeling approaches

Abstract

A chemical-transport model (CTM), Polyphemus/Polair3D, is used to investigate the contributions of various anthropogenic and biogenic sources to total organic carbon (OC) in PM2.5 in Beirut, Lebanon, during the summer of 2011. Those results are compared to a tracer-based source apportionment of OC at an ambient site in Beirut where a measurement campaign was conducted in July 2011. The results obtained from the CTM in the base simulation S1 suggest contributions to total simulated OC mass (3.24 μg/m3) of 66 % (2.14 ± 1.07 μg/m3) from fossil fuel burning (FFB) and 8 % (0.27 ± 0.135 μg/m3) from biogenic secondary organic carbon (BSOC). The tracer-based approach leads to contribution estimates to total measured OC mass (5.6 μg/m3) of 16 % (0.9 μg/m3 ± 0.22) from FFB, 53 % (2.9 ± 1.7 μg/m3) from BSOC, and 32 % (1.8 ± 0.88 μg/m3) from cooking activities. In a second CTM simulation S2, emissions related to cooking activities were added to the emission inventory, monoterpene and sesquiterpene secondary organic aerosol (SOA) surrogate species were added to the boundary conditions, and a lower ratio of semi-volatile organic compounds to primary organic aerosols (SVOC/POA) was used. The S2 results obtained showed contributions to total simulated OC mass (3.01 μg/m3) of 33 % (0.98 ± 0.49 μg/m3) from FFB, 18 % (0.53 ± 0.27 μg/m3) from BSOC, and 39 % (1.2 ± 0.6 μg/m3) from cooking activities. The differences between these two methods are discussed in terms of their uncertainties and biases. The comparison of both approaches showed that the model underestimates the secondary fraction of OC, which may be due to underestimations of some biogenic volatile organic compound (VOC) emissions and/or boundary concentrations as well as the use of SOA yields that may not be representative of the eastern Mediterranean region. Concerning the tracer-based approach, the use of tracer/OC ratios that are not specific to Lebanon because of a lack of data could represent a limitation of this methodology. Nevertheless, this comparative analysis suggests that on-road transportation and diesel generators used for electricity production are major sources of atmospheric PM and should be targeted for emission reduction. Finally, cooking activities, which were identified as a significant source of PM with the tracer-based approach, should be studied further.