Spatial variability of PM2.5 measured using compact filter-based sampler in Seoul, Korea

Abstract

There was good evidence of the association between exposure to PM2.5 and human health relying on spatial contrast. Because regulatory monitoring networks were not established for epidemiological studies, regulatory monitoring data may not well estimate exposures to air pollution across peoples residences for assessing health effects. To better characterize spatial variability of PM2.5, a two-week community monitoring campaign was performed three times during August 2015 to August 2016 in different seasons at 19 sampling sites of 5 district areas in Seoul using Deployable Particulate Impact Sampler (DPIS). Prior to apply DPIS to monitoring campaign, its performance was evaluated by comparing to Low-volume Cyclone Sampler (LCS) which has been consistently provided PM2.5 data since 2003 in the Seoul National University Yeongeon campus, Seoul, Korea. Using a duplicated DPIS, the reliability of DPIS was confirmed by computing relative precision and mean square error-based R squared value (R2). Relative precision was one minus the difference of measurements between two samplers relative to the sum. For accuracy, PM2.5 concentrations from two DPISs (DPIS_T and DPIS_Q) were compared to those of LCS. Two samplers included two types of collection filters (Teflon, T

quartz, Q). We assessed accuracy using accuracy value which is one minus the difference between DPIS and LCS PM2.5 relative to LCS PM2.5 in addition to MSE-based R2. DPIS showed high reliability (average precision = 97%, R2 = 0.98). Accuracy was generally high for all DPISs (average accuracy = 87-89%, R2 = 0.91). To determine the differences according to measurement method, LCS and beta-ray attenuation method (BAM) were compared and they were significantly different (t value = -9.0) especially in high-level mass concentrations. In the result of measuring PM2.5 by DPIS in community monitoring campaigns, the average mass concentrations ranged from 25.4 to 46.1 µg/m3 across the sampling sites and different seasons. Wintertime levels peaked at congested sites in southwest and central area under atmospheric stability conditions. Conversely, summertime concentrations were generally low relative to winter especially at sites nearby urban park in southwest and southeast area. Chemical component proportions of PM2.5 revealed that PM2.5 at 19 sampling sites in Seoul consisted of 18-22% OC, 3-4.1% EC, 8-9.5% nitrate, 17.5-20.6% sulfate and 9-10% ammonium, on an average basis of total sampling period. Organic carbon (OC), a major constituent of PM2.5, exhibited relatively high concentrations in winter particularly at sites affected by ambient sources such as incineration facility and charcoal fire emission of restaurants in northwest and southwest area. On the contrary, elemental carbon (EC) didnt show consistently high level sites over the season, indicating it was not contributed to stationary sources but traffic emissions. Among secondary ions, nitrate was extremely peaked in wintertime under the particle formation favored by condensation of cold temperature. Moreover, ionic balance indicated overall neutral PM2.5 aerosols close to slightly acidic. Correlation between components varied between sites, but generally high between ammonium and sulfate (average r = 0.76). Despite the high correlation between ammonium and nitrate (average r = 0.63), it varied across the sites (0.57-0.86) and some nitrate concentrations in several sites showed high correlation with OC rather than ammonium. For spatial variability of PM2.5 measurements in Seoul, coefficient of spatial variation (CV) and coefficient of divergence (COD) was computed across the sites. Consequently, EC and nitrate, which are mainly attributed to diesel emissions, were distributed unevenly in spatial patterns relative to other components. This study provides the knowledge of spatial characteristics of PM2.5 data measured at 19 sampling sites in Seoul, Korea. The standardized measurement of PM2.5 using evaluated sampler, DPIS, across Seoul will contribute to a consistent assessment of individual exposure level to PM2.5 in further health effect studies.