Adjust regular CORDEX grids

[jesusff]

We need to adjust these tables to the new naming -50i, -25i, -12i.

There are also mismatches at the borders:

domain-tables/regular.csv

Lines 14 to 15 in d8d54ec

	13,MNA-44i,Middle East and North Africa,206,106,-26.75,75.75,-7.25,45.25,0.5,0.5
	13,MNA-22i,Middle East and North Africa high res.,410,209,-26.625,75.625,-6.875,45.125,0.25,0.25

To be able to do an upscaling without interpolation, each cell in -50i should be full of cells of the -25i and -12i counterparts.

If there is no specific library or tool interpreting the "ll_lon" naming, it seems more intuitive to keep the "west" header in the old specs.

[larsbuntemeyer]

Good catch! These numbers are actually taken from the archive specs for my convenience to have a machine readable table. I don't know if these numbers have ever been used, the CORDEX-CORE data was published on the AFR-22 grid as far as i know...

Actually, i don't use this file anymore, since i have merged all domains in one table for use in my codes. However, i have to think about restructuring this anyway since some domains seem to change in CMIP6 downscaling.

[jesusff]

OK, in the merged file I understand the "ll_lon" notation. Let's keep it.
I think it is useful to have them stored here, so they can be updated and tracked. The CSV renderer in github can also be used to filter easily e.g. the "-22i" ones, or those for "AFR". We could even generate automatically from them some geojson files, which are also rendered on a map or be used to display the domain borders on a global map.

[gnikulin]

Yes, we need to synchronise the regular grids at the 3 resolution, not so complex task.