Tracer-based estimation of secondary organic carbon in the Pearl River Delta, south China

Fine particles (PM2.5) were collected using filter-based high-volume samplers during summer-winter 2008 at a rural site in the central Pearl River Delta (PRD), south China, to determine typical secondary organic aerosol (SOA) tracers from significant biogenic (isoprene, monoterpenes, and sesquiterpenes) and anthropogenic (aromatics) precursors. Average isoprene SOA tracers were significantly higher during summer (126 ng m(-3)) than during fall-winter (25.1 ng m(-3)), owing largely to the higher isoprene emission and reaction rates in summer. Average monoterpene SOA tracers during summer (11.6 ng m(-3)) and fall-winter (16.4 ng m(-3)) showed much less difference compared to isoprene SOA tracers, probably resulting from the counteracting effects of temperature on the precursor emission/tracer formation and on gas/particle partitioning. The concentrations of the aromatics' SOA tracer (2,3-dihydroxy-4-oxopentanoic acid) ranged from 1.70 to 52.0 ng m(-3) with an average of 15.1 ng m(-3), which was the highest reported in ambient air. The secondary organic carbon (SOC) estimated by the SOA-tracer method averaged 3.07 mu g C m(-3) in summer and 2.00 mu g C m(-3) in fall-winter, contributing 38.4% and 8.7% to OC, respectively. During summer, aromatics-SOC and isoprene-SOC reached 2.25 +/- 1.5 mu g C m(-3) and 0.64 +/- 0.7 mu g C m(-3) and accounted for 76% and 18% of the estimated SOC, respectively, while during fall-winter, aromatics-SOC (1.64 +/- 1.4 mu g C m(-3)) was dominant with a share of 79% in total estimated SOC. These results indicated that anthropogenic aromatics were dominant SOC precursors in the highly industrialized and urbanized PRD region. During summer, SOC levels estimated by elemental carbon (EC) tracer method were not only consistent with but also correlated well with those by SOA-tracer method. During fall-winter, however, SOC by SOA-tracer method was only about one third of that by EC-tracer method. Their gaps were significantly correlated with the biomass burning tracer levoglucosan, indicating that input from biomass burning emission with very high ratios of OC/EC during fall-winter would result in an overestimate of SOC by EC-tracer method. Therefore cautions should be taken when estimating SOC by EC-tracer method, especially when biomass burning exhibits significant influences.