Finding the potential height of tropical cyclone storm surges in a changing climate using Bayesian optimization

EarthArxiv preprint now available https://doi.org/10.31223/X57T5R!

We want to answer the question of what the potential height of a storm surge could be now and in a changing climate. To do this we first calculate the potential intensity and size from CMIP6 (tcpips & w22), and then use a Bayesian optimization loop (adbo) to drive an storm surge model ADCIRC with idealised tropical cyclones (adforce). We then show that knowing the upper bound can be useful in the context of an evt fit (worst).

All (or almost all) of the key experiments are carried out as slurm jobs, so go to slurm/ to see these. data/ contains some of the key data, and img/ most of the key figures. docs/ contains the source for the readthedocs documentation https://worstsurge.readthedocs.io/en/latest/MAIN_README.html.

tcpips

Tropical Cyclone Potential Intensity (PI) and precursors for Potential Size (PS) calculations. Includes pangeo script to download and process data from CMIP6 for calculations. Currently regrids using cdo or xESMf. Uses the tcpypi pypi package to calculate potential intensity.

w22

Used to calculate tropical cyclone potential size as in Wang et al. (2022). Used to use the Chavas et al. 2015 profile calculation in matlab (using octave instead from w22/cle5m.py). But now we have a new python implementation that is much faster and more stable, see w22/cle15.py. The implementation of Potential Size is in w22/ps.py.

# to calculate the example potential size
python -m w22.ps
# to calculate the tests against W22
python -m w22.test

adforce

An ADCIRC wrapper that does ADCIRC forcing and processing. Some generic mesh processing is also included. Assumes gradient wind reduction factor V_reduc=0.8 for 10m wind. Runs using hydra for config management.

python -m adforce.wrap

Needs an executatable directory for ADCIRC that includes a padcirc and adcprep executable stored in adforce/config/files/.

A repo with our compilation settings for ADCIRC, and small edits, is available at https://github.com/sdat2/adcirc-v55.02.

adbo

Bayesian optimization using adforce to force ADCIRC with a trieste Bayesian optimization loop. Runs using argparse for config management, and then calls adbo which uses hydra (it was not possible to have two layers of hydra configs for the two modules).

python -m adbo.exp_1d --test True --exp_name test

worst

Statistical worst-case GEV fit using tensorflow. Varies whether to assume an upper bound ahead of time. Uses hydra for the config management.

Getting started

Developer install:

Use conda to install a python virtual environment:

conda env create -n tcpips -f env.yml
conda activate tcpips

Alternatively use micromamba (faster):

micromamba create -n tcpips -f env.yml
micromamba activate tcpips

Install repository in editable version:

pip install -e .

Citations

If you use this code, please cite the following paper:

@article{potential_height_paper_2025,
  title={Finding the potential height of tropical cyclone storm surges in a changing climate using Bayesian optimization},
  author={Thomas, Simon D. A. and Jones, Dani C. and Mayo, Talea and Taylor, John R. and Moss, Henry B. and Munday, David R. and Haigh, Ivan D. and Gopinathan, Devaraj},
  journal={EarthArXiv (Submitted to Environmental Data Science)},
  year={2025},
  doi={10.31223/X57T5R}
}

And cite the code as:

@software{potential_height_code,
  author = {Thomas, Simon D. A.},
  doi = {10.5281/zenodo.15073504},
  month = {Oct},
  title = {{Finding the potential height of tropical cyclone storm surges in a changing climate using Bayesian Optimization}},
  url = {https://github.com/sdat2/PotentialHeight},
  version = {v0.1.2},
  year = {2025}
}

We used to use the matlab code from Chavas, Lin, and Emanuel (2015), which should be cited as:

@software{cle_2015_code,
  title = {Code for tropical cyclone wind profile model of Chavas et al (2015, JAS)},
  month = {Jun},
  url = {https://purr.purdue.edu/publications/4066/1},
  year = {2022},
  doi = {10.4231/CZ4P-D448},
  author = {Daniel Robert Chavas}
}