Problem with hybrid runs

[adamrher]

Hybrid runs 64(b,c) have very large global SSTs that suggest they are not being initialized to the year 43 SST state. @cecilehannay relayed to me that @gustavo-marques verified that these runs are reading in the year 43 restart file. If so, I still suspect it's not being used. Runs 64(e,f) are branch runs but w/ small perturbation to clubb gamma, and those look like there are initialized correctly.

We need to be able to do hybrid runs, and so this issue needs to be resolved.

[gustavo-marques]

I was looking at the wrong log file previously.
Cecile and I looked again at 64b this afternoon and we indeed see it was not initialized from the desired restart file.
We will look into fixing this issue asap.

[adamrher]

great thx Gustavo.

[gustavo-marques]

Here is a short-term hack for fixing this:

1. Add the following into user_nl_mom (MOM6 restart filename must be changed accordingly)

INPUTDIR = "INPUT"
INIT_LAYERS_FROM_Z_FILE = False
THICKNESS_CONFIG = "thickness_file"
THICKNESS_FILE = "b.e23_alpha16g.BLT1850.ne30_t232.064.mom6.r.0043-01-01-00000.nc"
TS_CONFIG = "file"
TS_FILE = "b.e23_alpha16g.BLT1850.ne30_t232.064.mom6.r.0043-01-01-00000.nc"
TEMP_IC_VAR = "Temp"
SALT_IC_VAR = "Salt"
VELOCITY_CONFIG = "file"
VELOCITY_FILE = "b.e23_alpha16g.BLT1850.ne30_t232.064.mom6.r.0043-01-01-00000.nc"

2. Go to the run directory and do the following

mkdir INPUT
cd INPUT
ln -s /glade/campaign/cesm/cesmdata/inputdata/ocn/mom/tx2_3v2/* .
ln -s /glade/campaign/cesm/cesmdata/inputdata/cesm2_init/b.e23_alpha16g.BLT1850.ne30_t232.064/0043-01-01/b.e23_alpha16g.BLT1850.ne30_t232.064.mom6.r.0043-01-01-00000.nc .  # or the MOM6 restart file that must be used

For reference, here is a version of run 064b with this hack applied:

/glade/derecho/scratch/gmarques/b.e23_alpha16g.BLT1850.ne30_t232.064b.clone

[jedwards4b]

Would you like to brainstorm on how to fix this or do you have a plan?

[gustavo-marques]

I would like to brainstorm. Thanks, Jim! I will send you a meeting invite.