Key points are not available for this paper at this time.
BACKGROUND: Heat has become a leading cause of preventable deaths during summer. Understanding the link between high temperatures and excess mortality is crucial for designing effective prevention and adaptation plans. Yet, data analyses are challenging due to often fragmented data archives over different agglomeration levels. METHOD: Using Germany as a case study, we develop a multi-scale machine learning model to estimate heat-related mortality with variable temporal and spatial resolution. This approach allows us to estimate heat-related mortality at different scales, such as regional heat risk during a specific heatwave, annual and nationwide heat risk, or future heat risk under climate change scenarios. RESULTS: We estimate a total of 48,000 heat-related deaths in Germany during the last decade (2014-2023), and the majority of heat-related deaths occur during specific heatwave events. Aggregating our results over larger regions, we reach good agreement with previously published reports from Robert Koch Institute (RKI). In 2023, the heatwave of July 7-14 contributes approximately 1100 cases (28%) to a total of approximately 3900 heat-related deaths for the whole year. Combining our model with shared socio-economic pathways (SSPs) of future climate change provides evidence that heat-related mortality in Germany could further increase by a factor of 2.5 (SSP245) to 9 (SSP370) without adaptation to extreme heat under static sociodemographic developments assumptions. CONCLUSIONS: Our approach is a valuable tool for climate-driven public health strategies, aiding in the identification of local risks during heatwaves and long-term resilience planning.
Wang et al. (Mon,) studied this question.