전국 시군구 PM2.5 연평균 농도 추정 방법 제안: 전국규모 국가대기오염측정망 측정 자료가 없는 2001~2014년을 중심으로

Abstract

Previous national cohort studies in Europe and North America have reported the relationship between long- term exposure to fine particulate matter air pollution and human health. However, there have been few studies in South Korea, because nationwide PM2.5 regulatory air quality monitoring data were not available before 2015. This study aimed to develop an approach to predicting nationwide annual average PM2.5 concentrations for 2001 similar to 2014 using the ratios of PM2.5 to PM10. We used regulatory air quality monitoring data for PM2.5 available only in Seoul for 2001 similar to 2014 and for the entire country in 2015, as well as national-scale data for PM10 for 2001 similar to 2015. Our approach is based on multiplication of the district-specific annual means of PM10 for 2001 similar to 2014 predicted from a previously-developed approach by ratios of annual average PM2.5 to PM10. Given the limited data availability to match PM2.5 and PM10 data over time and space, we explored five approaches for computing ratios of PM2.5 to PM10: 1) an overall ratio based on national data in 2015; 2) a ratio for 2001 similar to 2014 in Seoul; 3) annual ratios for 2001 similar to 2014 in Seoul; 4) regional ratios in 2015; 5) regional ratios in 2015 with a spatial adjustment factor. We validated the models using PM2.5 data for 2011 similar to 2014 collected at 24 regulatory monitoring sties in the second largest metropolitan city in South Korea, Busan. Whereas R-2 was computed over all sites and all years, spatial R-2 was computed by using 4-year averages at each site focusing on spatial prediction ability. In our validation, the approach 5 gave the largest overall and spatial R-2 of 0.79 and 0.84. Our highly predictive approach for estimating national-scale annual-average PM2.5 concentrations for 15 years will allow future cohort studies to investigate health effects of long-term exposure to PM2.5 in South Korea.