Examples of how to run eddy identification and visualisation with py-eddy-tracker on model data. If you have questions or suggestions, get in touch at stella.berzina@usys.ethz.ch
ICON EERIE runs used in the examples.
These are my personal notebooks to show some examples, I'll try to upload the nice final notebooks on the EERIE github.
Step 1: To identify and track eddies look at the eddy_identification_paralelised_share notebook
Step 2: If you have already identified and teaked eddies, visualisation is done in eddy_viz_share notebook. This also contains figures about how eddies are looking in ICON run.
It is a pain to set up a py-eddy-tracker environment that works but it might be worth it.
What worked for me to create a working environment on levante server:
conda create -n py39 python=3.9.1
conda activate py39
pip install numpy==1.26.4
pip install netcdf4==1.6.5
pip install matplotlib==3.7.1 #REALLY IMPORTANT
pip install scipy opencv-python pyyaml pint polygon3
pip install pyEddyTracker
Then you create a kernel for it
source activate py39
pip install ipykernel
python -m ipykernel install --user --name py39 --display-name "py39"
To run eddy identification with an intake library (as in EERIE project) you need to use the updated py-eddy-tracer version developed by Aaron Wienkers during the first EERIE hackathon (Nov 2023). Link to the py-eddy-tracker updated version and how to use it - https://github.com/eerie-project/EERIE_hackathon_2023/tree/main/RESULTS/pyeddytracker_xarray_dask_parallel.
For this you also need to install stuff for intake to work, so I do
pip install intake
pip install aiohttp
pip install intake-xarray
pip install intake-esm