Extreme events

Extreme events arise from the chaotic nature of the climate system through complex interplay between its components. These are out-of-the-ordinary events that may have significant ecological, economic, and societal impacts across the globe. Heat waves, for instance, can be devastating, in particular for vulnerable communities without the capacity to cope. During the hot summer of 2023, for example, tens of thousands of deaths all over Europe were attributed to heatwaves. 

Climate models are commonly used to understand how these events develop, since they can provide a large enough sample of current and future climatic conditions to identify extreme events. A combination of atmospheric high-pressure systems, soil moisture, and land interactions has been identified as the main drivers in the development of heat waves. For other types of extremes, such as extreme precipitation events, climate models face greater challenges  mainly because the small-scale processes governing precipitation must be parameterized, and because of the fine spatial and temporal structure of extreme precipitation fields.

Since the importance of the impacts of extreme events is undeniable, the question of how extreme events will change with climate change is a subject of intensive research. Changes in their frequency and intensity will determine the adaptation strategies that we will have to undertake. However, extreme event research poses several challenges. Firstly, due to the rare occurrence of these events, the confidence in future projections of certain variables from climate models is low. Furthermore, climate models only approximate the real world, and trustworthy observations are not available everywhere across the globe to evaluate the models. However, while projections can disagree on local changes, aggregating changes over larger regions can lead to increased consistency of the projections. 

In the climate physics group, we run the advanced CESM2-ETHZ model in-house. This allows for climate model output tailored to specific research questions, but also for the development of “ensemble boosting”, a numerical method of exploring extreme events more efficiently. Ensemble boosting perturbs initial conditions around already extreme events to produce a new set of potentially more extreme events, thereby pushing models to their physical limits and producing model-based storylines for unprecedented (record-shattering) extreme events.