ssh change in ana_ssh.h not reflected in velocity results

[tasha]

I would like to set ssh in ana_ssh.h higher in one part of the domain and then observe the effect on ocean velocities in response to the raised ssh. In theory this should work, no?

I defined ANA_SSH and modified ana_ssh.h (and have it in an appropriate directory where the modified file should be read, where other ana_* files are definitely being read). I set ssh to 1 in one part of the domain and 0 everywhere else. I also have set ANA_WINDS (=0), ANA_INITIAL (constant T&S), ANA_HUMIDITY, ANA_TAIR, ANA_PRES, ANA_SRFLUX. When I run the model with and without the ssh set, I see no difference at all between the different runs (for example comparing velocity results). So either it isn't being read in correctly or the model is not designed to work like this.

I did a search on ANA_SSH, and not much information came up, so I decided to post this new thread. Or did I miss the discussion somewhere?

Your help is appreciated!
Thanks, Tasha

[kate]

The model generally computes its own SSH. The purpose of having ANA_SSH would be for setting a climatology of SSH to be used for say nudging. I'm not sure we have examples of it actually being used.

[arango]

Actually, ANA_SSH is an option that I forgot to clean. No manipulations in ROMS free-surface are allowed, except during variational data assimilation. It is incorrect to do any of manipulation in ROMS free-surface because it is an expression of the vertically-integrated continuity equation. You will get in a lot of trouble if such manipulations are done. You will erroneously alter the transport. Any forcing needs to be done instead in the prognostic vertically-integrated momentum equations (ubar, vbar) for consistency. Recall that the free-surface is an expression for the div(ubar,vbar). This also applies for tidal forcing which has been discussed so many times this forum.

This is a fundamental concept that should be clear to all ROMS users.