Blending forest fire smoke forecasts with observed data can improve their utility for public health applications. (November 2016)

• Record Type: Journal Article
• Authors: Yuchi, Weiran; Yao, Jiayun; McLean, Kathleen E.; Stull, Roland; Pavlovic, Radenko; Davignon, Didier; Moran, Michael D.; Henderson, Sarah B.
• Is Part Of: Atmospheric environment. Volume 145(2016)
• Page Start: 308
• Page End: 317
• Publication Date: 2016-11
• Subjects: Forest fire smoke -- Blended models -- Fine particulate matter -- Exposure assessment -- Epidemiology -- Public health; Air -- Pollution -- Periodicals; Air -- Pollution -- Meteorological aspects -- Periodicals; 551.51
• DOI: 10.1016/j.atmosenv.2016.09.049
• Languages: English
• ISSNs: 1352-2310

Abstract

Abstract: Fine particulate matter (PM2.5 ) generated by forest fires has been associated with a wide range of adverse health outcomes, including exacerbation of respiratory diseases and increased risk of mortality. Due to the unpredictable nature of forest fires, it is challenging for public health authorities to reliably evaluate the magnitude and duration of potential exposures before they occur. Smoke forecasting tools are a promising development from the public health perspective, but their widespread adoption is limited by their inherent uncertainties. Observed measurements from air quality monitoring networks and remote sensing platforms are more reliable, but they are inherently retrospective. It would be ideal to reduce the uncertainty in smoke forecasts by integrating any available observations. This study takes spatially resolved PM2.5 estimates from an empirical model that integrates air quality measurements with satellite data, and averages them with PM2.5 predictions from two smoke forecasting systems. Two different indicators of population respiratory health are then used to evaluate whether the blending improved the utility of the smoke forecasts. Among a total of six models, including two single forecasts and four blended forecasts, the blended estimates always performed better than the forecast values alone. Integrating measured observations into smoke forecasts could improve public health preparedness for smoke events, which are becoming more frequent and intense as the climate changes. Highlights: Forest fire smoke causes the worst air quality that many populations ever experience. Reliable tools are needed to predict population smoke exposures. Measured observations are accurate, but always retrospective. Smoke forecasts are prone to error, but necessary for smoke warning systems. Blending forecasts with measured data improves their public health utility.