Validation of reanalysis products for extreme event attribution at regional and national levels

The WASITUS project was established to build towards an operational event attribution capability for Ireland. The project’s aim is to deep dive into the effect climate change has on extreme weather events at a national level, while also providing additional support to international attribution groups such as project collaborators; World Weather Attribution. 

By focusing on smaller national scales, and investigating data products used in event attribution, attribution studies can become more accurate and offer deeper insight for local responders and policy makers. A main focus of the WASITUS project is to take advantage of the small geographical size of Ireland and work directly with end-users to better understand how event attribution can help them prepare for future changes in extreme weather. These end-users include members of the public, local representatives, and national policy makers. This directly links attribution with real-world planning and damage mitigation measures.

Focusing on the data side of event attribution, most datasets used, whether reanalysis or models, have been tested at large regional or continental scales. However, we have found that the reanalysis data for Ireland, an island nation on the western boundary of most European datasets, is not as accurate as the data over continental Europe. This is quite possibly the case for other nations globally, where a variety of geographical and observational factors may have led to reanalysis products inaccurately representing the weather and climatology. As Ireland sits on the East of the Atlantic ocean, it is prone to weather threats of marine origin. Therefore, it is important to question the data used in creating the reanalysis and model products for Ireland as changing climate trends impact Ireland in different ways to the rest of Europe. 

A particular issue found for reanalysis products is their retrospective extension to earlier decades. To combat this potential issue, we are developing a toolbox to ascertain if reanalysis products reliably characterise the temperatures experienced in a given region for the entirety of the available time-series. The toolbox also aims to identify if shorter subsets of the entire reanalysis timeseries better represent the changing climate than the full dataset. Focusing on ERA5 daily maximum and minimum temperature data over the Republic of Ireland, station observations are being statistically compared to location-specific reanalysis data. While the initial focus will be temperature in Ireland, this toolbox should be readily adaptable for use in different regions globally, as well as on different meteorological parameters, provided sufficient long-term records are available.

In future, it is hoped that other national attribution capabilities, which are being newly formed, can collaborate and aid one another in conducting analysis and report writing. National groups also allow for further research into the methods used for extreme event attribution, where a focus can be placed on improving and expanding the existing attribution capability. In addition, time and focus placed on smaller geographical regions allows for data used in attribution analysis to be thoroughly quality controlled and checked.