In this study, various remote sensing data, modeling data and emission inventories were integrated to analyze the tempo-spatial distribution of biomass burning in mainland Southeast Asia and its effects on the local ambient air quality from 2001 to 2016. Land cover changes have been considered in dividing the biomass burning into four types: forest fires, shrubland fires, crop residue burning and other fires. The results show that the monthly average number of fire spots peaked at 34,512 in March and that the monthly variation followed a seasonal pattern, which was closely related to precipitation and farming activities. The four types of biomass burning fires presented different tempo-spatial distributions. Moreover, the monthly Aerosol Optical Depth (AOD), concentration of particulate matter with a diameter less than 2.5 μm (PM2.5) and carbon monoxide (CO) total column also peaked in March with values of 0.62, 45 μg/m3 and 3.25 × 1018 molecules/cm2, respectively. There are significant correlations between the monthly means of AOD (r = 0.74, P < 0.001), PM2.5 concentration (r = 0.88, P < 0.001), and CO total column (r = 0.82, P < 0.001) and the number of fire spots in the fire season. We used Positive Matrix Factorization (PMF) model to resolve the sources of PM2.5 into 3 factors. The result indicated that the largest contribution (48%) to annual average concentration of PM2.5 was from Factor 1 (dominated by biomass burning), followed by 27% from Factor 3 (dominated by anthropogenic emission), and 25% from Factor 2 (long-range transport/local nature source). The annually anthropogenic emission of CO and PM2.5 from 2001 to 2012 and the monthly emission from the Emission Database for Global Atmosphere Research (EDGAR) were consistent with PMF analysis and further prove that biomass burning is the dominant cause of the variation in the local air quality in mainland Southeast Asia.
Keywords: AOD; CO; EDGAR; MODIS; PM(2.5); PMF.
Copyright © 2019 Elsevier Ltd. All rights reserved.