ESTIMATING PM2.5 CONCENTRATIONS IN BRITISH COLUMBIA, CANADA DURING WILDFIRE SEASON USING SATELLITE OPTICAL DATA
- 1Mobile Sensing and Geodata Science Group, Department of Geography and Environmental Management, University of Waterloo, Waterloo, ON N2L 3G1, Canada
- 2Fujian Key Laboratory of Sensing and Computing for Smart Cities, School of Informatics, Xiamen University, Xiamen, FJ361005, China
Keywords: Air pollution, PM2.5, Wildfires, MODIS, Geographically weighted regression
Abstract. British Columba, Canada experienced its record-breaking wildfire season in 2017. The wildfire smoke is one of the main sources of fine particles with diameters smaller than 2.5 μm (PM2.5). The rising level of PM2.5 concentrations during the wildfire season would considerable increase the risk of premature death, especially for people with weak immune systems. In this study, the satellite optical data collected from 3 km MODIS aerosol optical depth (AOD) products were adopted to estimate PM2.5 concentration levels derived from wildfires in British Columbia, Canada from July to September 2017. The satellite optical data were combined with ground station measurements, meteorological and supplementary data to estimate PM2.5 concentrations using the geographically weighted regression (GWR) model. Our results demonstrated that PM2.5 concentrations were the highest in July and August based on the estimation results of seasonal and monthly GWR models. It indicated that the application feasibility of MODIS AOD products in predicting PM2.5 concentrations during the wildfire season in British Columbia.