Initialising with specified SSH

[angus-g]

As mentioned in #17, it might be good to be able to start the model including the snapshot SSH from the forcing, rather than starting with a flat surface. From what I can tell, there might be a method for altering the surface after this.

To initialise temp/salt, we use INIT_LAYERS_FROM_Z_FILE, which then reads these fields and their vertical grid, and remaps onto the model grid. This essentially gives us the initial thickness configuration for the model, which is the reference vertical profile, i.e. the Z* configuration without any stretching.

There's another option which we currently set to false: DEPRESS_INITIAL_SURFACE. With this set to true, two new options become available: SURFACE_HEIGHT_IC_FILE and SURFACE_HEIGHT_IC_VAR. I think that providing this should be enough to do what we need.

[angus-g]

I ran 3 months from cold using this, it's available at /scratch/x77/ahg157/mom6/archive/panan-01-zstar/archive/output000. I can also provide the file with the surface initial condition for a different run. Maybe somebody could give this a look over and then we can see if this is affecting #17 at all?

[adele-morrison]

Awesome, thanks @angus-g ! @MatthisAuger did you want to check the SSH and anomalous u at the boundary in this new run compared to the other one?

[MatthisAuger]

Sure!

Here is the zonal mean of the SSH at the boundary for the obc file, previous run and new run.

Currently looking at the differences in U but from the SSH it seems like it does not change that much the outputs.

[MatthisAuger]

U across the boundary. New run seems to weaken the zonal current at the boundary? Edit - Changed the zonal mean plot to compare with the 01degree global model

At one longitude (70E):

U_ref_new_lon70E.1.mp4

Zonal mean:

U_ref_new_access_zonalmean.1.mp4