[MITgcm-devel] imrprovement in periodicExternalForcing (?) et al

[Jean-Michel Campin]

Hi Jeff,

I am going to cc to mitgcm-devel so that we can keep a reccord of it more easily.

Also, I opened an issue (#182) in 2018 regarding ocean surface forcing not beeing well 
documented in the manual. This type of information is not exactly what I had in mind but
would fit also into a section dedicated to surface forcing.

Otherwise:
 1) it seems that S/R GET_PERIODIC_INTERVAL would work if we set externForcingCycle=0 ;
  this would need to be checked, but if it does work then we could remove the STOP
  in ini_params.F that prevent to use this (this stop was added a long time agoi, in June 
  1998, by Alistair).
 2) given the stops mentionned above, periodicExternalForcing has the right name,
  since the model prevent to use non periodic forcing.
 3) might be some inconsistency between pkgs, I would not be very surprised by this.

Cheers,
Jean-Michel
 
On Mon, Jul 20, 2020 at 03:44:55PM +0000, Jeffery R Scott wrote:
> 
> Hi Jean-Michel,
> 
> I struggled with this over the weekend, combination of names being sort of confusing and/not clear in the manual.
> 
> What I am trying to do is have time-varying forcing (one month chunks) without it repeating.
> 
> SO I assumed the first thing would be to turn periodicExternalForcing to .FALSE. as one would naturally presume from the (three) collective variable names used here.
> NOPE - you need this set to true for any time-varying forcing, repeating or not (i.e. this variable name was not chosen well  IMO). However this is so
> established as a parm name, the best thing at this point would be to simply clarify in the manual.
> 
> However, then I would assume you could leave externForcingCycle unset then and it would just keep reading in data.
> Again nope: For various reasons I can see why this might be problematic (and, there is an error check to not allow such).
> 
> Finally, this was a pkg cfc run and the behavior for fice and wspeed files was different than the files in data (I don???t think it checks periodicExternalForcing).
> Not sure if this was intended or simply the implementer not really being careful about it.
> 
> Not sure your thoughts on this ??? it would be nice to have a smoother way to implement a time-varying, non-repeating solution (data of some forcing interval), but perhaps this is
> such a rare thing to want to do it has not come up (or people use EXF etc).
> 
> But at the very least, a short comment added to the manual where these variables are described would be helpful.
> Or a could make a issue relating my findings here.
> 
> Best,
> Jeff
> 
> 
> 
>