Weighting Game: Comparing Different Population-Weighted Climate Exposure Methods

Amanda Norton, University of Toronto
Risto Conte Keivabu, Max Planck Institute for Demographic Research (MPIDR)
Monica Alexander, University of Toronto

Aggregate environmental exposure estimates, such as temperature, present a unique issue because populations are not distributed uniformly across space (i.e., urban vs rural) and population level exposure estimates may not adequately capture the spatial variation occurring within a spatial unit. This paper compares several aggregation methods of temperature measures, based on population weighting, to understand differences in aggregates and sensitivity of estimates of excess deaths based on common methods. In preliminary work we use France as a case study. Data on total population by NUTS3 region, gridded population estimates, and gridded temperature measures were sourced from Eurostat, WorldPop, and the Copernicus climate data store, respectively. Three temperature measures (minimum, maximum, mean) were aggregated to a spatial unit with three weighting schemes (population weighting, log of population weighting, and population 65+ weighting). Preliminary findings suggest some regional and seasonal variation in the aggregate measures, while the impact of different measures on resulting excess death estimates may be minimal.

Keywords: Spatial Demography, Geo-referenced/geo-coded data, Population, Environment, and Climate Change, Mortality and Longevity

See extended abstract.