Upper-boundary conditions in CAM seem to have no impact

[DirkOlivie]

What happened?

CAM uses upper boundary conditions (UBC) for H2O, and these conditions (file and species) are prescribed via the namelist. The strange thing is that simulations with and without these UBC give bit-identical results.

It gives the impression that these UBC have no impact on the NorESM simulation. Is it because NorESM has a low top?

The atm_in for a simulation with UBC looks like (e.g., on Betzy in /cluster/projects/nn9560k/olivie/cases-noresm3.0-test/n1850esm.ne16pg3_tn4.noresm3_0_beta10.run7.emifv09cmip6.20260127/CaseDocs/atm_in):

&upper_bc_file_opts
 ubc_file_cycle_yr              = 1850
 ubc_file_input_type            = 'CYCLICAL'
 ubc_file_path          = '/cluster/shared/noresm/inputdata/atm/cam/chem/ubc/b.e21.BWHIST.f09_g17.CMIP6-historical-WACCM.ensAvg123.cam.h0zm.H2O.185001-201412_c230509cdf5.nc'
/
&upper_bc_opts
 ubc_specifier          = 'Q:H2O->UBC_FILE'
/

The atm_in for a simulation without UBC looks like (e.g, on Betzy in /cluster/projects/nn9560k/olivie/cases-noresm3.0-test/n1850esm.ne16pg3_tn4.noresm3_0_beta10.run7.emifv09cmip6_ubcnone.20260127b/CaseDocs/atm_in):

&upper_bc_file_opts
 ubc_file_cycle_yr              = 1850
 ubc_file_input_type            = 'CYCLICAL'
 ubc_file_path          = 'NONE'
/
&upper_bc_opts
 ubc_specifier          = 'NOTSET'
/
``

### What are the steps to reproduce the bug?

To impose that no UBC should be used, one can put in user_ln_cam :

ubc_specifier = 'NOTSET'
ubc_file_path = 'NONE'

### What CAM-Nor tag were you using?

noresm3_0_024_cam6_4_121

### What machine were you running CAM (or NorESM) on?

Betzy

### What compiler were you using?

Intel

### Path to a case directory, if applicable

Standard case with UBC : /cluster/projects/nn9560k/olivie/cases-noresm3.0-test/n1850esm.ne16pg3_tn4.noresm3_0_beta10.run7.emifv09cmip6_ubcnone.20260127b

### Will you be addressing this bug yourself?

Yes

### Extra info

_No response_

[gold2718]

I have spent some time tracing this.

• The ubc_mmr for Q is read in in ubc_get_vals in upper_bc.F90 (which calls upper_bc_file_get in upper_bc_file.F90)
• ubc_get_vals is called from vertical_diffusion_tend (in Oslo Aero's vertical_diffusion.F90)
• ubc_mmr is passed as in input to compute_vdiff
• ubc_mmr is not all zero (I inserted an endrun which is not called)
• compute_vdiff is not returning an error string.
• compute_vdiff is being called.

I have no idea why this does not affect answers.

[oyvindseland]

I did an experiment replacing the vertical diffusion routine in Oslo-Aero with standard CAM physics and the results were bit-identical so in principle that routine can actually be removed from the aerosol code.