[MITgcm-support] adding uniform background flow to all water grid points smoothly.

[Martin Losch]

> On 13. Nov 2020, at 16:04, Li Jianing <li.jia.ning at outlook.com> wrote:
> 
> Hi Martin,
> 
> Thanks for your advice.
>  Yes, I do want the flow to interact with the topography. Yesterday, after I send this email, I ran the model with 'mypackage' modification for longer time and the result suggested that it's not a proper way. So, I'll follow your suggestion to do it. Since the west and east boundaries are all open, I'll impose the flow at west boundary from still initial condition to force a eastward flow. 
> And there are some little puzzles:
> 1.This is quite similar to set U field at open boundary, except the restoring process, what kind of difference will this resotring make? 
If you have real open boundaries OBCS will also work. with RBCS you have the option of making a smoother transition (becauce you can specify 3D masks at your convenience), also I was not sure if you wanted to use period boundaries (simpler!), but then only RBCS would work.
> 2. Is it necessary to set the outflow inthe east boundary?
Probably yes, same as inflow. You’ll still get reflections of the perturbations. I have had good experience with the Stevens OBCS (see exp4 as an example, and the documenation)
> 3. I'm a little confused about the rbcs parameters 'rbcsForcingPeriod' and 'rbcsForcingCycle'. Since the rbcs forcing cover the whole simulation with no change in this case, will these two setting work the same way if I set both of them to 0 or set them to 'whole simulation time'? And more generally, what is the relation between 'rbcsForcingPeriod', 'rbcsForcingCycle' and 'whole simulation time’?
If your rbcs forcing never changes, that these parameters can be left alone (default = 0). The model will read the mask/forcing once and use it as constant in time. “period” is the interval between time records to read, if you have time varying forcing, “cycle” is the length that the forcing spans in time, before it starts at the first record again. For example, if you have a 10y integration and you have 12 monthly forcing sets, then you’d set period=1month (2592000.) and cycle=1year (31104000.). The model with load the correct forcing until the 12th record is reached, then starts again at the first record and so on until the simulation is over.

Martin
> 
> Many thanks!
> Jianing Li
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support