Variable gravity in fv3/atmos_cubed_sphere

[WChen-NOAA]

Description

The model uses the shallow atmosphere approximation with constant gravity.
This is problematic for a space weather model (WAM) where Ztop/r0 )~10%
The method to resolve this issue is to Implement a scaling factor wherever gravity is applied to account for variations with height:
g(r0/r)^2 while setting var_grav = .true. in the name list
The new gravity as grav_var(i,j,k)

Issue(s) addressed

• fixes issue 1017

Testing

Tested on Hera with intel compiler for several tests.

Dependencies

• waiting on Variable gravity  NOAA-GFDL/GFDL_atmos_cubed_sphere#395 (comment)

[grantfirl]

@WChen-NOAA Since the graviational acceleration is also used in physics (as a constant), is there any need to make physics changes? Otherwise, wouldn't this introduce an inconsistency between the dycore and physics?

[yangfanglin]

The variable gravity option should only be used for deep-atmosphere modeling. This capability should be controlled by a namelist option. I think the impact is mostly on dycore formation. The impact on physics where grav is used has not been included or even explored. @WChen-NOAA @XiaqiongZhou-NOAA please correct me if I am wrong. In this case, the constant g can still be used in CCPP.