Use of MODIS data to assess atmospheric aerosol before, during, and after community evacuations related to wildfire smoke

Wildfires can pose direct and indirect threats to public health via fire and smoke, respectively. Evacuations are used to manage both types of threat given situational information about fires and their behaviour. However, there has been limited documented use of objective remote sensing data to inform the decision-making process. We conduct a retrospective assessment of 41 smoke-related evacuations in Canada between 2000 and 2007 using smoke and cloud data from the Terra MODIS. Daily mean aerosol optical depth (AOD) measurements were used to represent smoke levels in the area surrounding each evacuation site. Time series of these data were plotted for the four days prior to evacuation, the evacuation date, and the four days following the evacuation. Daily AOD values were scaled using the daily cloud mask to crudely account for the fact that heavy aerosol can be classified as cloud. Time series plots and the percentage differences between pre- and post-evacuation AOD were used to contextualize each evacuation. Out of 41 cases, approximately half met each of the criteria used to assess public health protection: (1) post-evacuation AOD values over the risk threshold for sensitive populations; (2) peak AOD exposures occurring post-evacuation; and (3) mean AOD higher post-evacuation than pre-evacuation. Some evacuations did not meet any of these criteria. Although evacuation decisions would never be made in the absence of situational intelligence, we conclude that systematic use of objective and readily available remote sensing data can help to retrospectively evaluate smoke-related evacuations, and may be useful for prospective, real-time application with further development.