More general function for delineating basins

[angus-g]

See discussion in #73: it would be nice to do the description of basins by a list of rectangles, instead of hand-coding the whole routine.

[navidcy]

What about regionmask package?

[angus-g]

That looks pretty cool!

[navidcy]

cc @rmholmes (in relation to a chat in slack...)

[rmholmes]

FYI: For the OMIP run post-processing for CMIP6 I have to create a basin mask like the one we have for the 1-degree model for the 1/4-degree model. I plan to do this early next week by using the boundaries from the 1-degree mask and updating for the 1/4-degree land-sea mask differences. I'll post here with what I end up doing.

[Thomas-Moore-Creative]

Aloha, I don't have at hand a clean piece of code to share here but I have started to use regionmask, xesmf, & xgcm with the "marine area" shapefile linked at Natural Earth as this link

I don't think there is anything particularly "oceanographic" about these NE definitions but it's a start. I also note there are apparently other efforts that I have yet to trial that seem more promising?

[rmholmes]

Thanks @Thomas-Moore-Creative.

I have made a basin mask a very non-general way for the 1/4-degree, based on the same file that was lying around in the input folders for the 1-degree, some old masks I had lying around for a paper and matlab. The two files are at: /scratch/public/rmh561/basin_mask.nc % 1-10 labelled basin mask for 1-degree model /scratch/public/rmh561/basin_mask_025deg.nc % 1-10 labelled basin mask for 1/4-degree model

These files may be useful to others. However, a more general approach would most likely be better (e.g. to make the same file for the 1/10-degree).