[MITgcm-devel] adding ice dyn run time switches

[Martin Losch]

Hi Jean-Michel,

On Feb 12, 2012, at 7:00 PM, Jean-Michel Campin wrote:

> I have a question (to Martin) regarding seaice_summary.F :
> I was tempted (and may have started) to not report parameters
> that are irrevelant (e.g., tAlpha & sBeta when eosType <> LINEAR).
> This would make the summary shorter and, may be, more useful
> (assuming that all the relevant params are reported, which is 
> far to be the case right now, ice/snow/air density, air CP ...).
> The first thhing I had in mind was to skip all the seaice-growth params
> when we use thsice (or usePW79thermodynamics=F).
> But I saw that you went a different path by adding a kind of
> warning before printing the param value, like this one:
> (PID.TID 0000.0001) MAX_HEFF has no effect because SEAICE_CAP_HEFF is undefined
> (PID.TID 0000.0001) MAX_HEFF          = /* maximum ice thickness */
> (PID.TID 0000.0001)                 1.000000000000000E+01
> (PID.TID 0000.0001)     ;
> So, which way should we follow ?
I think that reporting only the relevant flags/values in a summary is the better way (has the potential to be shorter). In pkg/seaice there are (or maybe were) so many cpp-options so I guess there should be some kind of config-check that stops or at least issues warnings, if flags are set (different from the default), that have no effect, but it's only a convenience features for new users and aging (=suffering from dementia) people like me.

Martin