Changes in atmospheric concentrations of polycyclic aromatic hydrocarbons and polychlorinated biphenyls between the 1990s and 2010s in an Australian city and the role of bushfires as a source.

Over recent decades, efforts have been made to reduce human exposure to atmospheric pollutants including polycyclic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs) through emission control and abatement. Along with the potential changes in their concentrations resulting from these efforts, profiles of emission sources may have also changed over such extended timeframes. However relevant data are quite limited in the Southern Hemisphere. We revisited two sampling sites in an Australian city, where the concentration data in 1994/5 for atmospheric PAHs and PCBs were available. Monthly air samples from July 2013 to June 2014 at the two sites were collected and analysed for these compounds, using similar protocols to the original study. A prominent seasonal pattern was observed for PAHs with elevated concentrations in cooler months whereas PCB levels showed little seasonal variation. Compared to two decades ago, atmospheric concentrations of ∑13 PAHs (gaseous + particle-associated) in this city have decreased by approximately one order of magnitude and the apparent halving time (t1/2) was estimated as 6.2 ± 0.56 years. ∑6iPCBs concentrations (median value; gaseous + particle-associated) have decreased by 80% with an estimated t1/2 of 11 ± 2.9 years. These trends and values are similar to those reported for comparable sites in the Northern Hemisphere. To characterise emission source profiles, samples were also collected from a bushfire event and within a vehicular tunnel. Emissions from bushfires are suggested to be an important contributor to the current atmospheric concentrations of PAHs in this city. This contribution is more important in cooler months, i.e. June, July and August, and its importance may have increased over the last two decades.


Introduction
in Brisbane, which has a subtropical climate with hot summers and moderately warm winters.  Self-designed active air samplers were used with a sampling rate of approximate 4 m 3 h -1 , 113 similar to the one typically used during the 1994/5 study (Mueller, 1997). The sampling 114 volume was recorded using a gas meter connected to the outflow of the pump. The particle-115 associated fraction of the samples was collected on a glass fibre filter (GFF) (Whatman™, 90 116 mm Ø, grade GF/A), followed by a cartridge containing 10 g of XAD-2 (styrene-117 divinylbenzene copolymer, Supelco ® , 90 Å mean pore size) to collect chemicals in the gas

124
A high-volume air sampler (Kimoto Electric Co., LTD.) was used with a typical sampling 125 rate of 60 m 3 h -1 . Particle-associated and gaseous chemicals were collected on a GFF 126 (Whatman™, 203×254 mm, grade GF/A) and a subsequent PUF plug (90 mm diameter and 127 40 mm thickness) respectively. The sampler was calibrated using an orifice plate prior to the 128 sampling campaign and the sampling volume was calculated based on the calibrated sampling 129 rate and sampling duration. A bypass gas meter installed on the sampler was used to monitor 130 any anomalous fluctuation of the sampling rate during the collection. ). An XAD-2 cartridge (1 g) was used to trap chemicals 137 (gaseous + particle-associated phases) and the flow rate was checked at the beginning and the 138 end of the sampling period to ensure its constancy.

139
Detailed information related to sample collection is provided as S1 in the supplementary   144 The collected GFFs, XAD and PUFs were extracted separately using an Accelerated Solvent 145 Extractor (ASE, Thermo Scientific™ Dionex™ ASE™ 350) after being spiked with a 146 solution containing 7 deuterated PAHs and 18 13 C 12 -PCB congeners at different levels as 147 internal standards for quantification purposes (Table S2). Concentrated extracts were divided 148 into three portions. The first portion (40% v/v) was cleaned up by neutral alumina and neutral 149 silica for PAH analysis, the second (40% v/v) was cleaned up by neutral alumina and acid 150 silica for PCB analysis and the third (20% v/v) was archived for future analytical 151 investigations. Eluants were carefully blown down to near dryness and refilled with 250 pg of  as also used in the 1994/5 study (Mueller, 1997), was used to quantify 13 PAH analytes and 160 18 PCB congeners comprising dioxin-like (dl-PCB) and indicator (iPCB) compounds (Table   161   S2). Details are given as S2 in the SM.   (Table S3). Breakthrough. The breakthrough percentage of chemicals was typically negligible (Table S3).  Reproducibility. Within the replicated QC samples (n = 15 in total), relative standard 199 deviation (RSD) of the analytical results was less than 15% for most (90%) analytes (Table   200 S3).   (Table S5).

244
As shown in Figure 3(b), the median concentration of ∑ 6 iPCBs at this site decreased by 80% 245 (from 75,000 to 15,000 fg m -3 ) over these two decades with the trichlorobiphenyl congener 246 PCB 28 achieving the greatest reduction (81%) (Table S5) Sun et al., 2006). In this work, the apparent 255 halving time ( ⁄ ; y) of an SVOC analyte in air was calculated from applying a first order 256 decay model using PAH and PCB data (gaseous + particle-associated) collated from studies 257 carried out at Site Gri or WG (Table S6)   PAHs. Mean concentrations (gaseous + particle-associated) of each compound were 273 consistently higher at Site WG (paired t test with P < 0.01). This is illustrated in Figure 2     Residential/commercial heating has been predicted to account for 50% of total PAH    PAHs. As seen in Table S7, for most PAHs, the average halving time estimated at Site WG  It was noted that the halving time of Ant was longer than Phe at Site Gri (Table S7). This was 351 unexpected given that Ant is generally less stable in the air and has an estimated lifetime 352 some 2 to 4 times shorter than its linear isomer (Phe) based on reaction with OH radicals

374
Over the last two decades, the profile of indicator PCBs has shifted slightly towards a higher 375 proportion of medium sized congeners. For example, the contribution of PCB 101 increased 376 slightly from 3.6 ± 1.4% to 10 ± 7.0%, again indicating re-volatilisation from reservoirs such 377 as soil have been acting as the main source for atmospheric PCBs in Brisbane.   (Table S10). As seen in Figures 5, S4 and Table S10,  Table S10) as well.