OBC in NEMO3.2

[dupontf]

Hi all,

Sorry guys,

I am not sure what is going on with the OBC routines in NEMO3.2. Are they still supposed to be fairly maintained? Some (such as obcini or obcvol) look like they are maintained while others (obcfla) still have the old bugs from NEMO2.3 (and beyond, some tickets on the old system have not been addressed).

I would like to point out that in the new dynspg_ts.F90, ua_e/va_e are now the vertically averaged velocity (versus transport in the older versions). Therefore obcfla should be modified accordingly. There are also the useless vertical loop, etc...

Fred.

[rblod]

The answer is a bit hard to elaborate.

They are supposed to be maintained, but the current state is obviously not satisfying, and I'm really ashamed for it. Since the first application with NEMO/OPA have been originally focused on global ocean, the OBC part have been more or less "running" behind the other developments.

Recently, there was some work done on the filtered free surface case, that's why obcini or obcvol evolved. The current sate is:

• OBC with filtered free surface are ok
  
• BDY with time splitting free surface are ok
  
• OBC with time splitting free surface are not working. (you were kind enough to avoid noticing the obc reading part is this case doesn't even compile)
  

The next step is to revisit this last point , I think we have to do it on a very short time scale, considering the increasing number of users for this kind of applications.

[dupontf]

Thanks Rachid,

We kindly offer to share our OBC routines that we have rewritten for the time-splitting if needed. I note though a small bug on the old system that is still relevant to the filtered mode: line 492 of obctra.F90 there is a missing rtausin factor in the calculation of the restoring coefficient for the southern boundary.

[rblod]

Not totally fixed since OBC could be improved seriously, but major bugs corrected. Major rewriting should append in 2001