Dynamically-informed extreme event attribution using circulation imprints

Climate extremes impose an increasing toll on human society, a considerable part of which may be ascribed to anthropogenic climate change. This motivates the study of how anthropogenic climate change may affect specific extreme events, often termed extreme event attribution. Here, we present a novel extreme event attribution approach which incorporates information on the circulation dynamics and quantitatively separates the role of dynamical changes in the atmospheric circulation from thermodynamic changes – such as altered distributions of temperature and specific humidity. Our approach considers multivariate circulation imprints, and allows the use of imprints at multiple timesteps. We further quantify extreme event footprints in terms of impact-relevant hazard indices. This provides a detailed view of the dynamical evolution of the extreme events and the associated hazards. Moreover, the approach is easily reproducible since it builds on historical meteorological data and does not require ad-hoc numerical model simulations. We apply our methodology to three recent high-impact extreme events: the 2025 Los Angeles wildfires, the 2024 Spanish floods and Storm Ciarán which affected Western Europe in 2023. We find a clear contribution of climate change to events like the Los Angeles wildfires, while our results are inconclusive concerning the Spanish floods and show a weakening effect of dynamical changes for Storm Ciarán. These conclusions present both analogies and differences to previous attributions of these events. Some of these discrepancies can be understood by considering the dynamical information that our approach provides, and which was overlooked or combined with thermodynamic information in previous analyses.