FV3-CAM7 problem for cesm3_beta02

[lamaliang]

Hello

In CAM7, there is a new process regarding gravity wave drag. When we use FV3 to build the case, an error message appears stating that FRONTGF can't be undefined.
Therefore, we modified the code in dyn_comp.F90 of FV3 as follows:

subroutine dyn_register()

use physics_buffer, only: pbuf_add_field, dtype_r8
use ppgrid, only: pcols, pver

! These fields are computed by the dycore and passed to the physics via the
! physics buffer.

if (use_gw_front .or. use_gw_front_igw) then
call pbuf_add_field("FRONTGF", "global", dtype_r8, (/pcols,pver/), &
frontgf_idx)
call pbuf_add_field("FRONTGA", "global", dtype_r8, (/pcols,pver/), &
frontga_idx)
end if

end subroutine dyn_register

With this modification, the build completes successfully.
We are curious about whether the gravity wave drag process is turned on when we register FRONTGF and FRONTGA?

Thanks

 

[peverley]

Hi,

You did the right thing by adding the pbuf fields. We don't test gravity wave drag with the FV3 dycore, so you're in uncharted territory. The best I can tell you is that:

• Gravity wave drag is turned on and configured via namelist settings
  • You can see which sources are being used in the "&phys_ctl_nl" section of the run/atm_in file of your case. See "use_gw_*"
  • You can see further configurations in the "&gw_drag_nl" section
• We do test gravity wave drag with the SE dycore, so looking at that code may help you if you run into any issues as you go.

I hope that helps. Good luck!
Courtney

 

[lamaliang]

Hello

Thanks for your reply.

I have run simulations using CESM3-FV3, but I encountered an issue with FV3 for CAM7 and CAM5/CAM6.

When I use CAM5 physics with FV3, the results appear reasonable. However, when using CAM7 physics, there are significant biases in energy balance and temperature.

I suspect that FV3 might not be utilizing some processes specific to CAM7.

Could you help me?

Thanks

 

[jet]

Hi

Currently CAM7 physics are under development and have not been validated with all supported dycores. I am curious what biases you found when running FV3 with CAM7 physics.

 

[lamaliang]

Hello

We found that the temperature is 3-4°C lower than the relative observation. The FNT is -3 W/m².

The experiment is B1850