Cumulative and relativistic temperature metrics for public heat alerts: A new approach proposed for greater Vancouver, Canada

BackgroundMany heat alert systems rely on fixed absolute temperature thresholds that may not fully characterize risk. Furthermore, absolute temperatures can vary widely across some urban areas, leading to the same absolute thresholds being associated with different risks. The impending modernization of meteorological services in Canada provides an opportunity to address these challenges with heat alerts in greater Vancouver, British Columbia (BC). Objectives(1) To evaluate a cumulative heat alert indicator based on the sum of high temperatures over two consecutive days and the intervening overnight low (High + Low + High, or H+L+H) using mortality and heat-related morbidity data. (2) To compare the performance of absolute and relativistic H+L+H thresholds for generating heat alerts, including consideration of susceptible populations. MethodsTime-series analyses were used to examine the relationships between all-cause mortality (2008-2024), heat-related emergency department (ED) visits (2014-2023), and H+L+H temperature observations from three weather stations across greater Vancouver. For each station, the outcomes were modelled using absolute and relativistic H+L+H thresholds, including analyses stratified by age, sex, socioeconomic status, and chronic health conditions. ResultsThe relativistic H+L+H indicator was associated with significant risk of morbidity and mortality, with similar exposure-response functions for all three weather stations. In contrast, absolute thresholds showed considerable heterogeneity. The relativistic metric also showed different risk functions for some susceptible groups, including those with schizophrenia, Parkinsons disease, and receiving income assistance. ConclusionThe relativistic H+L+H temperature metric integrates cumulative multi-day exposure and addresses the challenge of intra-urban variability in absolute temperatures. This approach provides a more flexible alternative to heat alerts based on absolute temperatures, and highlights differences in risk among susceptible populations.