Specific Humidity for get_phi_fv3_run

[danrosen25]

Where is specific humidity calculated for input into get_phi_fv3_run as gq01? This variable isn't referenced anywhere else in the physics suite. Does the ccpp code generator alias all variables with the same standard name? Does that mean that identical variables names with different standard names are actually different variables?
https://github.com/NCAR/ccpp-physics/blob/main/physics/get_prs_fv3.F90#L91

[climbfuji]

If you look at the associated .meta file, you see that gq01 = specific_humidity_of_new_state. That variable gets calculated in GFS_stateout_update from the old state specific_humidity plus the collected tendencies from the physics until this point in the physics group. If you grep for specific_humidity_of_new_state on all scheme metadata file, you should be able to see that.

In general, only one (host model) variable can be associated with a standard name is the same quantity, but it can have different local/Fortran names in each scheme.

[danrosen25]

Where is GFS_stateout_update?

What I'm trying to figure out is why gt0 [air_temperature_of_new_state] and gq01 [specific_humidity_of_new_state] have NaN values 10 hours into a simulation. The NaN values cause the MYNN surface scheme to crash.

[climbfuji]

grep is your best friend in this case ...

ccpp-physics/physics/GFS_suite_interstitial.F90

Line 448 in 4bb6e8d

subroutine GFS_suite_stateout_update_run (im, levs, ntrac, dtp, &

and https://github.com/NOAA-EMC/fv3atm/blob/eb463bccde5d19c15c86b6187bffd47e1bd63f9a/ccpp/suites/suite_FV3_GFS_v16.xml#L68 (as one example suite)