Page 10 - 201905

JOUR NAL O F E NVI R ONM E NT AL S C I E NC ES 79 (2019 ) 1 – 10 5

Fig. 3 – Scatterplots of MVK and MACR during (a) nighttime and (b) daytime on different sampling days. The black lines in the
figure indicate the 1.4/1 line predicted by the yields of MACR and MVK from isoprene oxidation.

formation and solvent usage contributed significantly to Factor 2 was distinguished by high percentages of methyl
MACR and MVK abundances. Furthermore, in addition to chloride (CH 3 Cl), ethane, and benzene, as well as certain
high correlations with MVK, propanal and acrolein, MACR amounts of other combustion tracers (e.g., ethyne and C 3 –C 4
2
correlated well with aromatics (i.e., C 6 –C 8 aromatics, R = alkanes). Furthermore, this factor made the highest contri-
2
0.56~0.71), and MTBE (methyl tert-butyl ether) (R = 0.64), bution to ACN, a typical tracer for biomass burning plumes in
typical species arising from gasoline vehicular emissions on the PRD region (Yuan et al., 2010). Therefore, this factor was
“comparable” days, suggesting that gasoline vehicular emis- assigned as biomass burning. Previous studies have demon-
sions were another source affecting the abundances of MACR strated that MACR and MVK (MACR + MVK or MACR and
and MVK at the sampling site. MVK individually,) were the major OVOC components in the
emission profiles of biomass burning in China and other
2.3. Source apportionments areas (Inomata et al., 2014; Wang et al., 2014; Kudo et al.,
2013; Yuan et al., 2010; Ralf, 2007). Furthermore, a previous
The above-mentioned observed ratios of MACR and MVK as study based on data measured by the proton transfer
well as their correlations with other species (Section 2.2) reaction-mass spectrometer (PTR-MS) and fire hotspots
suggest that other sources, apart from the oxidation of from MODIS reported that biomass burning occasionally
isoprene, contributed to the mixing ratios of MACR and influenced the levels of VOCs in Jiangmen (Yuan et al.,
MVK, thus resulting in variations of MACR and MVK at the 2010). The contribution of biomass burning to MACR and
HS. To quantitatively apportion the sources of MACR and MVK were 2% and 5%, respectively, which is consistent with
MVK, the PMF model was applied to the dataset at the HS the higher emission rate of MVK than MACR as indicated by
site. A five-factor simulation that best reproduced the the biomass burning emission-based measurement study in
observed concentrations was chosen based on the calculated the US (Gillman et al., 2015).
statistical parameters and prior-knowledge about emission The third factor was dominated by MACR, MVK, and PAN.
source profiles specific to the PRD region. The five factors The dominant contribution of PAN, a secondary product, in
included biogenic emissions, biomass burning, secondary the profile suggested that this source was related to secondary
formation, gasoline, and diesel vehicular emissions. Fig. 4 formation. Additionally, its contribution to MACR and MVK
illustrates the explained variations of the individual appor- was high at 45% and 70%, respectively.
tioned sources as well as their corresponding profiles, i.e., the Factor 4 was rich in ethene, propene, and 1-butene, as
relative contribution of each source to the individual species well as certain amounts of ethyne and aromatics, consis-
at the HS. tent with the profiles of diesel vehicular exhaust in the PRD
The first factor is identified as biogenic emission, which region (HKEPD, 2015; Liu et al., 2008). Therefore, this source
is solely dominated by isoprene and accounts for ~83% of was identified as diesel vehicular emissions. The contribu-
the measured concentration of isoprene at the HS site. tion of this source to MACR and MVK was approximately
About 26% of the MACR was attributed to this factor, 11% and 2%, respectively. The presence of high levels of
whereas its contribution to MVK was negligible. Though butanes, pentanes, methyl pentane and aromatics in Factor
these results are consistent with previous measurement 5aswellascomparisons with source profiles of gasoline
studies, which reported that MACR was one of the most vehicular exhaust obtained from emission-based measure-
abundant OVOCs from biogenic emissions (Jardine et al., ments (HKEPD, 2015; Liu et al., 2008) suggested that this
2011; Carvalho et al., 2005), it should be noted that the source was related to gasoline vehicular emissions. Addi-
OVOCs from biogenic emissions are dependent on the tionally, it contributed nearly 17% and 24% of MACR and
vegetation type and meteorological conditions (i.e., temper- MVK, respectively, at the HS. The tunnel and vehicular
ature and solar radiation). Therefore, further studies on the exhaust emissions measurements as well as the compari-
profiles of OVOCs primarily emitted from vegetation in the sons between the emission inventory and the observed
PRD region are needed. data have indicated that MACR and MVK could be emitted