A New Statistical Approach for Rapid Attribution of Extreme Weather Events

Extreme weather events, including heatwaves and heavy rainfall, have become increasingly frequent over Japan in recent decades. Quantifying the relative contributions of anthropogenic climate change and natural internal variability to individual events through event attribution (EA) is therefore an important scientific challenge. Here, we develop a statistical method for rapid EA based on existing long-term large-ensemble climate simulations and observational datasets. The method constructs regression models linking large-scale sea surface temperature patterns and associated atmospheric variability to the probability distributions of surface air temperature and precipitation. This approach enables attribution analyses without performing event-specific numerical simulations. We have already applied this method to recent heatwave events over Japan, and the results of the rapid EA analyses have been made publicly available through the “Weather Attribution Center (WAC Japan)”. Furthermore, the method can also be applied to heavy rainfall events, yielding reliable estimates. In this presentation, we present an overview of the newly developed statistical approach, its applications, and our ongoing efforts.