New Study Evaluates the Association Between Extreme Heat and Mortality in Urban Southwestern Ontario Using Different Temperature Data Sources
A team of researchers led by Kristin Clemens with Alexandra Ouedraogo, Lihua Li, James Voogt, Jason Gilliland, E. Scott Krayenhoff, Sylvie Leroyer, and Salimah Shariff published a study entitled: “Evaluating the Association Between Extreme Heat and Mortality in Urban Southwestern Ontario Using Different Temperature Data Sources.”
With rising global temperatures, there has been increased attention placed on the association between ambient temperature, morbidity, and mortality. Extremes in temperature have been linked with cardiovascular and respiratory events, heat illness, and even death. The health impact of extreme temperature on cities requires special attention. Cities have high population densities and complex microclimates. They have higher surface and near-surface atmospheric temperature than non-urbanized areas due to modifications to the cover, form and materials used in these regions (urban heat island effect).
To study this the research team examined the association between extreme temperature and mortality in urban Ontario, using two temperature data sources: high-resolution and weather station data. They used distributed lag non-linear Poisson models to examine census division-specific temperature–mortality associations between May and September 2005–2012. They used random-effect multivariate meta-analysis to pool results, adjusted for air pollution and temporal trends, and presented risks at the 99th percentile compared to minimum mortality temperature. As additional analyses, they varied knots, examined associations using different temperature metrics (humidex and minimum temperature), and explored relationships using different referent values (most frequent temperature, 75th percentile of temperature distribution).
Weather stations yielded lower temperatures across study months. U-shaped associations between temperature and mortality were observed using both high-resolution and weather station data. Temperature–mortality relationships were not statistically significant; however, weather stations yielded estimates with wider confidence intervals. Similar findings were noted in additional analyses. In urban environmental health studies, high-resolution temperature data is ideal where station observations do not fully capture population exposure or where the magnitude of exposure at a local level is important. If focused upon temperature–mortality associations using time series, either source produces similar temperature–mortality relationships.
Citation: Kristin Clemens, Alexandra Ouedraogo, Lihua Li, James Voogt, Jason Gilliland, E. Scott Krayenhoff, Sylvie Leroyer, and Salimah Shariff. “Evaluating the Association Between Extreme Heat and Mortality in Urban Southwestern Ontario Using Different Temperature Data Sources.” Scientific Reports 11, 8153 (2021); https://doi.org/10.1038/s41598-021-87203-0