Temperature-related mortality risks: effects of different sources of climatic data in the RF regions in 2004–2019

Climate change and increasing thermal stress highlights the need to investigate the temperature-mortality relationship using long-term aggregated temperature data. Globally, two primary sources of temperature data are utilized: ground-based meteorological observations and raster datasets. Ground-based observations from meteorological stations offer precise local temperature measurements but lack comprehensive spatial coverage. In contrast, raster data provide complete spatial coverage but may not accurately represent local microclimatic conditions. This study aims to compare these data sources for analyzing temperature-related mortality across regions of Russia.

To assess the exposure-response relationship, a two-stage modeling approach was applied. At the first stage, region-specific estimates were derived using a distributed lag model. At the second stage, pooled estimates were computed through random-effects meta-regression.
The temperature-mortality relationship in Russia is characterized by a typical J-shaped curve, with cold temperatures posing a higher mortality risk. Heat-related risks were generally higher when estimated using raster data compared to in-situ observations. Minimum mortality risk temperatures typically fall between 15 and 20 °C, with higher thresholds observed in warmer regions.

This study suggests general comparability of raster and point-based temperature data for mortality analysis. However, in certain regions, particularly large and sparsely populated ones, estimates diverged due to multiple factors.