deltamin now defined in ice_in

[JFLemieux73]

For detailed information about submitting Pull Requests (PRs) to the CICE-Consortium,
please refer to: https://github.com/CICE-Consortium/About-Us/wiki/Resource-Index#information-for-developers

PR checklist

• Short (1 sentence) summary of your PR:
  deltamin now defined in ice_in.
• Developer(s):
  @JFLemieux73
• Suggest PR reviewers from list in the column to the right.
• Please copy the PR test results link or provide a summary of testing completed below.
  gx3 test (1 day). BFB for kdyn=1 (deltamin=1e-11) and kdyn=3 (deltamin=2e-9)
• How much do the PR code changes differ from the unmodified code?
  • bit for bit
  • different at roundoff level
  • more substantial
• Does this PR create or have dependencies on Icepack or any other models?
  • Yes
  • No
• Does this PR add any new test cases?
  • Yes
  • No
• Is the documentation being updated? ("Documentation" includes information on the wiki or in the .rst files from doc/source/, which are used to create the online technical docs at https://readthedocs.org/projects/cice-consortium-cice/. A test build of the technical docs will be performed as part of the PR testing.)
  • Yes
  • No, does the documentation need to be updated at a later time?
    • Yes
    • No
• Please provide any additional information or relevant details below:
  I need to do some work about the doc. I have noted in issue Improve the documentation for the dynamics CICE-Consortium/CICE#651 what needs to be done.

[JFLemieux73]

By the way tinyarea has been replaced by deltaminTarea...I am open to suggestions if you have a better variable name.
@dabail10 I have commented tinyarea in the CESM driver.

[apcraig]

Can we add least add the new namelist to the namelist table in the documentation? I think it's useful to keep that up to date even if there is additional documentation needed elsewhere to be added later.

Could we find a shorter string than delatminTarea for the variable. That string really messes up some of the indentation and means we need to adjust a bunch of other lines too. We can do that, but nice if we don't have to.

[JFLemieux73]

Ok I will do it. I think finally I will have to introduce two deltamin (VP and EVP).

[JFLemieux73]

@apcraig I am done. Please let me know if this is ok.
I ran gx3 for 1 day. It is BFB for kdyn=1 and for kdyn=3.
I also added the new variables to the doc.

[apcraig]

Interesting that DminTarea is an array computed initially and stored but DminUarea is computed on the fly each timestep and gridpoint.

If you wanted to use one deltamin namelist, you could do the following (psuedocode). Not suggesting it has to be implemented as one namelist, but it does make things more general and flexible in some ways.

real deltamin
real, parameter :: deltamin_spval = -1.2345e30_dbl_kind
real, parameter :: deltaminEVP = 1.0e-11_dbl_kind
real, parameter :: deltaminVP = 2.0e-9_dbl_kind

deltamin = deltamin_spval
...
read namelist
...
if (deltamin == deltamin_spval) then
  if (kdyn == 1) deltamin = deltaminEVP
  if (kdyn == 3) deltamin = deltaminVP
endif

[apcraig]

It seems like here, you might want to have a G or E formatted output for deltamin since it's very small (e-9, e-11). Or multiple deltamin by 1e6 or 1e9 and then write out with F formatting?

[JFLemieux73]

Good point for the format. I just changed it.
As discussed I use an array Dmintarea to replace the array tinyarea so that it is BFB.
I think I would prefer to keep the code as it is for deltaminVP and EVP...if it's ok?

[apcraig]

The new namelist also need to be added to doc/source/user_guide/ug_case_settings.rst in the correct group. These are alphabetical. thanks.

[apcraig]

Thanks for updating documentation and moving evp_algorithm. The line after that also has to move with evp_algorithm that's "shared_mem_1d".

[JFLemieux73]

Thanks Tony. It should be ok now.

[apcraig]

Just FYI that a followup PR set the capping value to 0. for vp cases for backwards compatibility. The default was 1. for all dynamics options after this change but prior results had it set to 0. for vp cases. If we want to setup capping to 1. for vp and change answers, that's fine, lets just be clear about that.

[JFLemieux73]

Thanks Tony. Sorry I didn't think about backward compatibility...
The implicit solver works fine with both capping = 0 or 1. For our C and CD grid testing (and comparison with EVP solutions) it would be better to have capping=1 as the default. It's ok if it changes the answer of the previous test case. What do you prefer?

[apcraig]

Sounds good @JFLemieux73. I have changed the capping value to 1.0 for VP in #64 and documented there.