Impacts of forest fires on ambient near–real–time PM2.5 in Ontario, Canada: Meteorological analyses and source apportionment of the July 2011–2013 episodes

The complexity of analyzing and predicting smoke plumes that originate from forest fire events and impact populated regions of southern Ontario motivates the innovative application of analytical techniques including trajectory–based receptor modeling for spatial source apportionment of the observed near–real–time particulate matter (PM) impacts. PM2.5 was selected as an indicator of a pollutant emitted by fires that could be transported over long distances (when entrained into the transport layer above the planetary boundary layer (PBL), and subject to sink and transformation processes) and be monitored using the existing air quality monitoring network. The source term modeling technique of simplified Quantitative Transport Bias Analysis (sQTBA) was applied to several summertime forest fire events to identify the locations of sources affecting air quality in Ontario during these events. Complementary techniques that helped to understand the movement of smoke plumes included satellite remote sensing of carbon monoxide and aerosols. All of these techniques, along with meteorological analysis, jointly provide a means of identifying the forest fire events that resulted in noticeably higher pollutant levels in Ontario. Specifically, three forest fire events in July of 2011, 2012 and 2013 were analyzed, and source regions of near–real–time PM2.5 concentrations were revealed to be both within Ontario and across northern Canada from Quebec to Yukon. The sQTBA was found to successfully identify the relative importance of various source regions contributing plumes from forest fires and non–wildfire related sources that caused higher pollutant levels that were measured in Ontario. The use of near–real–time PM2.5 data in this study facilitates the identification of the exact periods with high pollution impacts across multiple receptor sites, thus improving the overall quality of the analyses. This work shows how trajectory–based receptor models can be integrated with meteorological analyses for thorough source apportionment of wildfire–related pollution events.