[MITgcm-devel] vertical axis origin

[Jean-Michel Campin]

Hi,

It's about setting vertical axis and pressure origin:

Following the fluid isomorphism convention, I added recently a new parameter "phi0Ref"
to be used to set phiRef at the origin, i.e. at r=rF(1),
 with phiRef= the reference pressure/rhoConst (if Z-Coords) 
               or reference geopotential (if P-Coords)
while Ro_SeaLevel (a bad name !) set the value of rF(1)

But I have the impression it would be easier to set vertical axis and pressure origin
independently of isomorphism rule, with 2 (new) parameters (to replace phi0Ref 
& Ro_SeaLevel):
 1) top_Pres = pressure (if P-Coords) or reference pressure (if Z-Coords) at the top.
 2) seaLev_Z = the reference height of sea-level (usually zero) 
    assuming, by default, that the fluid is above sea-level if ATM (fluidIsAir) 
    and below sea-level if OCN (fluidIsWater).

This will be used to set vertical axis and pressure scale:
OCN in Z: top_Pres(Ref) (<-> rhoConst*PhiRef(1)), seaLev_Z (=rF(1), @ the top)
ATM in Z: top_Pres(Ref) (<-> rhoConst*PhiRef(1)), seaLev_Z (=rF(Nr+1), @ the bottom)
OCN in P: top_Pres (=rF(Nr+1)), seaLev_Z (<-> PhiRef(Nr+1)/g, @ the top)
ATM in P: top_Pres (=rF(Nr+1)), seaLev_Z (<-> PhiRef(1)/g, @ the bottom)

In P-Coords and especially OCN in P case, it's more natural to set top_Pres rather 
than the bottom Pressure (currently Ro_SeaLevel).
And in this case (OCN in P), setting the top reference height (i.e., seaLev_Z)
would be better than the default value of phi0Ref (that I kept from earlier code,
 = -Ro_SeaLevel/rhoConst).
It will help also for ATM in Z case, not having to set the altitude of the model
top (currently Ro_seaLevel) but instead the altitude of the bottom: seaLev_Z

For backward compatibility, can keep "Ro_SeaLevel" in namelist for now (probably
with a warning given the bad name), but would remove completely phi0Ref (very recent
addition).
Note that the buoyancyRelation option for ATM in Z is not currently available, but 
I was planning to add this in the near future.

Any comment/suggestion ?

Cheers,
Jean-Michel