Windstorm Return Level Model based on WISC/WWIEUR Footprints as described in Priestley et al. (2023, NHESS)

This dataset and code runs the windstorm return level model as documented in Priestley et al. (2023; https://doi.org/10.5194/nhess-23-3845-2023). 

The code uses the NAO co-variate and the two options either estimate return levels and uncertainty for one location, with the other script plotting a map for all of Europe (with upper and lower bounds). The code is constructed using python3.7. All model calculations have already been made, and so the data provided is the parameters calculated on a gridpoint basis. 

Data has been provided for both the WISC footprints (as in the paper), and also the newer Winter Windstorm Indicators for Europe footprints (WWIEUR, https://cds.climate.copernicus.eu/cdsapp#!/dataset/sis-european-wind-storm-indicators?tab=overview), and the model can be run for either. The WISC footprints span 1950-2014, and WWIEUR from 1979-2021.

In the code, you must specify the return period(s) of interest, as well as the NAO state. It is recommended that the NAO state does not exceed -2 or +2, as the data used to build the model does not go beyond these bounds. Therefore, any value outside these bounds has no data trained on it. For return periods, we do not recommend going beyond 500 years. Confidence estimates provided are the 95% intervals.

In order for the code to run, the data must remain in its current file structure/hierarchy.

This model is intended as a tool to test the sensitivity of return levels to NAO state across Europe. For both the WISC and WWIEUR, different estimates are provided due to the different resolutions of the input data, and also the different methodologies used to create the footprints.

The model can be run with new input data, any users wishing to do so should please get in touch and this code can be made available.

Any queries or questions, please contact via email: [e-mail]