A framework for the automated detection and attribution of heatwaves

Event attribution has become an established field in climate science, with robust methodologies routinely used to assess the influence of anthropogenic climate change on different types of extreme weather events. At the same time, societal and scientific demand for timely attribution statements has increased, particularly in the immediate aftermath of extreme events. While rapid attribution approaches have made substantial progress and allow for the dissemination of attribution statements within weeks or even days, the overall workflow remains largely ad hoc and manual. In particular, event selection is often triggered by reported impacts or media coverage, which leads to uneven spatial coverage and systematic underrepresentation of events in regions with limited reporting, such as the Global South. Additionally, the manual nature of attribution analyses limits the number of events that can be assessed.  

Among extreme weather events, heatwaves are particularly interesting for systematic attribution, as they often exhibit robust signals of anthropogenic climate change and are associated with substantial impacts on human health, ecosystems and socio-economic systems. Here we present an automated, hazard-based framework for the detection and attribution of heatwaves that enables continuous and systematic analysis with global coverage and flexible spatio-temporal scales. Heatwaves are identified directly from temperature data and attributed using established probabilistic methods. The framework is based on reanalysis products, forecast data and CMIP6 model simulations, allowing for both historical assessments and (near-)real time analyses.

The resulting inventory of heatwaves and associated attribution statements enables systematic comparison with existing attribution studies, including rapid attribution efforts such as World Weather Attribution [1], ClimaMeter [2] and Qasmi et al. (2025) [3]. Detected events are further compared with disaster databases, e.g., EM-DAT, to assess reporting biases and the consistency between hazard-based and impact-based heatwave records. Beyond comparisons of individual events, the resulting inventory of historical heatwaves also provides a basis for analysing regional changes in heatwave characteristics over time.

References: 

[1] Philip, S., Kew, S., van Oldenborgh, G. J., Otto, F., Vautard, R., van Der Wiel, K., ... & van Aalst, M. (2020). A protocol for probabilistic extreme event attribution analyses. Advances in Statistical Climatology, Meteorology and Oceanography, 6(2), 177-203.

[2] Faranda, D., Messori, G., Coppola, E., Alberti, T., Vrac, M., Pons, F., ... & Vautard, R. (2024). ClimaMeter: contextualizing extreme weather in a changing climate. Weather and Climate Dynamics, 5(3), 959-983.

[3] Qasmi, S., Ribes, A., Cattiaux, J., Barbaux, O., Robin, Y., & Dulac, W. (2025). An automatic procedure for the attribution of extreme events at the global scale: a proof of concept for heatwaves. Bulletin of the American Meteorological Society, BAMS-D.