iceinfo

environment

I use (mini)conda, and have not had many dependency issues.
The conda packages needed are:

• click-help-colors
• cartopy
• pandas
• scipy
• cmocean
• netCDF4
• eccodes

eccodes is the ECMWF grib tools, which includes the grib_to_netcdf command.

example

$conda config --add channels conda-forge
$conda env create -n iceinfo_env -f iceinfo-env.yml
Note: that some numpy dtype warning appeared when I tested from scratch, but it did not affect the plotting.

workflow

$bash do_ecmwf.sh
-> calls grib_to_netcdf to convert grib2 files to the netCDF format.
-> then, calls "plotECMWF-LHB.py" to do the wind and pressure field fig, "plotECMWFWave-LHB.py" to do the wave field figure.
---> the main python code calls libraries from the iceinfo_libs folder.

update

OpenData file name for wave changed from *wave.nc to *wav.nc. update L105 in do_ecmwf.sh and L42 in plotECMWFWave-LHB.py.

example

$bash do_ecmwf.sh 20231205 00 (new figures are created in the test_fig folder)
Note: right now, the grib files and figure outputs are contained in this directory for convenience (it might not be during JARE).

file managements

This is obviously a preference, so this is just an FYI.
In the Thunderbird app, I used the FiltaQuilla add-on to automatically save attachments to a specific directory (new files).
After I do the new plots on the latest files, I moved the grib2 and converted NC files to an archive folder.
I used to rsync the archive figures folders with the Shirase shared directory, so the forecast figures can be uploaded to the NME intranet page.