Biomass burning contributions to urban PM<sub>2.5</sub> along the coastal lines of southeastern China

Levoglucosan (LG), water soluble organic carbon (WSOC) and potassium (K+), and the light absorption at 365 nm (Abs365) of the extracted WSOC are measured in PM2.5 samples collected from November 2011 to July 2013 at four coastal urban sites in southeast China (Fuzhou, Putian, Quanzhou and Xiamen). These species are markers of biomass burning and used to determine the contributions of biomass burning to the PM2.5 burden in these locations. LG and WSOC concentrations exhibited a clear seasonal pattern, with a large enhancement in winter and spring and a minimum in summer, and annual means across all sites of 59.2±46.8 ng m−3 and 2.69±1.21 µg C m−3, respectively. The distinctive seasonal patterns of LG and WSOC are more explained by the East Asian monsoon than the upwind varying emission sources according to the HYSPLIT backward trajectories and MODIS fire spots. Observations produced significant correlation (at the p<0.01 level) between LG and non-sea salt K+ (nss-K+) at each site, but the correlations exhibited no clear seasonal trend. The LG/nss-K+ ratios ranged from 0.03±0.01 to 0.24±0.13 which lay within the limits for the crop residues and/or grass combustion smoke. Stronger correlations were found between WSOC or Abs365 and sulphate than between WSOC and LG. This observation is consistent with the fact that biomass burning is a less important contributor to WSOC and/or brown carbon than is secondary organic aerosol formation and oxidation. The average relative contributions of biomass burning to OC and WSOC in PM2.5 were 8.3 and 15.2 %, respectively, estimated by the measured LG to OC and WSOC (LG/OC and LG/WSOC) ratios in comparison to literature-derived LG/OC and LG/WSOC values for biomass burning smoke. Using the reported conversion factor of LG to PM2.5 for crop straw burning smoke, the LG-estimated PM2.5 contributions from biomass burning exhibited minimum values in summer and higher values in winter and spring. Positive Matrix Factorisation was used to analyse PM2.5 sources, and seven major source factors including biomass burning emissions were resolved. The biomass burning sources contributed 3.75–8.08 % to PM2.5 mass as an annual average while exhibiting a seasonal variability similar to and higher than those of LG-estimated contributions. These results demonstrate that the contribution from biomass burning smoke to PM2.5 – while small – is non-negligible, especially in the winter and spring sampling periods, that is, northeastern monsoon season.