Variable naming & metadata

[pbett]

I think we can mostly port over the stuff in the CMIP6 CMOR tables wholesale and then edit.

The table names aren't really very obvious, so I'd like that to be clearer for a start...

Looking at them though, I think there's some tweaks we could do

• e.g. do we want repeated definitions in files for different time-frequencies? Isn't the time frequency independent of the variable definition? (cf monthly data and 3hly data
• near-surface air temperatures ("tas") can explicitly say they're at 1.5 m, rather than "(usually, 2 meter)" (cf. in here
• I don't think we need "modelling-realm"
• Units: The model outputs temperatures in Kelvin of course, but we'll probably be outputting to the user in °C?
  • (and it really makes me squirm to see things like "kg m-2 s-1" rather than "kg m⁻² s⁻¹", but I guess there's some standard lying behind that?)

[agstephens]

These issues are being dealt with in other threads so closing.