Behind the paper: modelling extreme precipitation in Europe

We talk to Birthe Steensen and colleagues, authors of the recent PLOS Climate publication “How climate models reproduce the observed increase in extreme precipitation over Europe“

What led you to decide on this research question?

Recent summers have seen several extreme precipitation events in Europe, causing severe infrastructural damage, crop failure, and loss of life. Myhre et al. (2019) showed that over Europe since the 1950s, total precipitation from extreme events nearly doubles per degree of warming, primarily driven by increases in frequency, and that the rate of increase becomes stronger with the rarity of the event. We wanted to assess how this behavior is represented in CMIP6 models. Specifically, this study investigates how well CMIP6 models capture the observed increase in heavy to rare extreme precipitation, given that Myhre et al. (2019) found such events were underrepresented in CMIP5 models. We were especially interested in the change in frequency vs. the intensity of extreme precipitation.

How did you go about designing your study?

The same methods used in Myhre et al. (2019) were applied to observational datasets, reanalysis, the CMIP6 models and CORDEX simulations to evaluate how well they reproduce the observed increase in heavy to extreme precipitation over Europe. A key objective of this study was to assess changes in rare events (e.g., those occurring less than once a year). Such events increase more than events that occur every year, and they increase in areas that otherwise experience general drying and a decrease in the less extreme heavy precipitation events. It was also important to identify an effective way to illustrate model diversity in the projected increases in both frequency and intensity of extreme precipitation.

Did you encounter any challenges in collecting or interpreting your data?

Since we primarily worked with model data and openly available observational datasets, data collection and analysis were straightforward. However, a new version of the E-OBS gridded observational dataset introduced changes in the interpolation method between stations, which noticeably affected the representation of extreme precipitation events and led to inconsistencies compared to earlier results. To address this, we also included the older version of E-OBS for comparison.

What struck you most about your results? What are the key messages and who do you hope might benefit from these new insights?

The results showed that many CMIP6 models were able to reproduce the observed increase in extreme precipitation. The results also showed that including multiple ensemble members, up to 50 for some of the models, revealed that changes in extreme precipitation that could be explained by natural variability in the models was relatively small compared to the individual model diversity. This suggests that the simulated response of extreme precipitation to global warming is more strongly influenced by model dynamics and thermodynamics than by internal variability. The higher-resolution CORDEX models showed a smaller change in frequency but a larger change in intensity compared to observations, driven by a few extremely intense precipitation events that occurred in both study periods. These issues reduce the reliability of the CORDEX simulations for assessing trends in extreme precipitation. We hope the results will improve climate adaptations for flood management and infrastructure planning, water planning in reservoirs in addition to adaptations to avoid crop failure.

What further research would you like to see in this area?

A natural step onwards is to investigate how the models project extreme precipitation over Europe for the different future scenarios for possible mitigation of these extreme events. There is a large diversity between the models over historical periods, and how this evolves in the future is interesting. A more detailed study on why the models have a different response in the hydrological cycle to global warming is also needed, as well as how the models represent the rarest events, considering also aerosols and cloud microphysics parameterizations. How the new generation of CMIP7 models represents the change in extreme precipitation over Europe will also be of interest.

What made you choose PLOS Climate as a venue for your article?

We chose to publish our results in PLOS Climate because it is a fully open access journal with a clear focus on interdisciplinary climate science, including the societal impacts of climate change. Results presented in this study are directly important for future mitigation and adaptation. The journal’s broad audience helps ensure that our findings are accessible.

Ready to submit your own work to PLOS Climate? Follow our step-by-step guide to the submission process!