Statistical extreme value analysis of storm surge levels with Python.

The statistical analysis methods implemented by the code in this repository are documented in the peer-reviewed article:

Markus Reinert, Lucia Pineau-Guillou, Nicolas Raillard, & Bertrand Chapron (2021). Seasonal shift in storm surges at Brest revealed by extreme value analysis. Journal of Geophysical Research: Oceans, 126, e2021JC017794. DOI: 10.1029/2021JC017794

When using the code in this repository, please cite this paper.

Every version (or “Release”) of this code has an individual DOI. Please also cite the DOI of the code version you are using. You find the DOIs for all versions by clicking on the DOI badge above. This badge always redirects to the latest code version.

The data used in the paper cannot be published here for copyright reasons and needs to be obtained separately. The surge levels at Brest can be computed from the tide gauge measurements published on the website https://data.shom.fr/, using the method described by Reinert et al. (2021), or can be obtained from the corresponding author upon request. The GESLA-2 surge dataset of Woodworth et al. (2017) can be obtained from the website https://gesla.org/, as explained in tools_GESLA.py. The climate index data (for example NAO) can be obtained from https://psl.noaa.gov/gcos_wgsp/Timeseries/, as explained in tools_climate.py.

Given the datasets, the main results of the paper can be reproduced with the scripts for Method 1, Method 2, and different stations. The parameter estimates of the full time-dependent GEV model can be calculated with the script Time-dependent_GEV_fit_with_NAO.py.

The functions and methods in this repository can also be used with other datasets or for other studies, see the examples in Publications below. In particular, the script advanced_GEV_analysis.py may be useful. Its main part is an implementation of the methods described in the book “An Introduction to Statistical Modeling of Extreme Values” by Stuart Coles (2001). Example usage of this library is shown for time-independent GEV models, for time-dependent GEV models, and for GEV models of annual maxima. With the surge levels for Brest from the GESLA-2 dataset, the time-independent GEV model of monthly maxima looks like this:

Here is a list of publications using the code of this repository. If any publications are missing, please let me know.

1. Markus Reinert, Lucia Pineau-Guillou, Nicolas Raillard, & Bertrand Chapron (2021). Seasonal shift in storm surges at Brest revealed by extreme value analysis. Journal of Geophysical Research: Oceans. DOI: 10.1029/2021JC017794

2. Jean-Baptiste Roustan, Lucia Pineau-Guillou, Bertrand Chapron, Nicolas Raillard, & Markus Reinert (2022). Shift of the storm surge season in Europe due to climate variability. Scientific Reports. DOI: 10.1038/s41598-022-12356-5

3. Kiesel, J., Lorenz, M., König, M., Gräwe, U., & Vafeidis, A. T. (2023). Regional assessment of extreme sea levels and associated coastal flooding along the German Baltic Sea coast. Natural Hazards and Earth System Sciences. DOI: 10.5194/nhess-23-2961-2023

4. Lorenz, M., & Gräwe, U. (2023). Uncertainties and discrepancies in the representation of recent storm surges in a non-tidal semi-enclosed basin: a hindcast ensemble for the Baltic Sea. Ocean Science. DOI: 10.5194/os-19-1753-2023