parameter file updates

[rgknox]

Description:

This is a group of changes to the parameter file, not backwards compatible.

Collaborators:

Many, see : #616

Expectation of Answer Changes:

Implementing non-answer changing modificaitons first, will test incrementally

Checklist:

• I have read the CONTRIBUTING document.
• [O] FATES PASS/FAIL regression tests were run
• [O] If answers were expected to change, evaluation was performed and provided

Test Results:

Tested addition of Medlyn stomatal parameters, matches baseline, moving on to next set

/glade/scratch/rgknox/clmed-tests/fates-sci.1.36.0_api.11.1.0-clm5.0.30-v1-Cd525994c-Fed442bb3.fates.cheyenne.gnu

[rgknox]

Added rooting profile updates, per #290
These changes make it so that the same rooting profile is used for both btran and decomposition, but also is a non-b4b change. Results should be nearly identical though for all non-decomposition related output. Test results at BCI:
bci-test-ufnrt_plots.pdf

[rgknox]

Compared commit after adding in the fire-threshold and dbhmin parameters with commit before it, results are b4b:
/glade/scratch/rgknox/clmed-tests/fates-sci.1.36.0_api.11.1.0-clm5.0.30-v3-Cd525994c-F8a1317a9.fates.cheyenne.gnu

[rgknox]

paired with: ESCOMP/CTSM#1027

[ckoven]

do we want to pull in @rosiealice's PFT matrix term into this?rosiealice@f0170fd
and rosiealice@a69ac89

[ckoven]

Looks good to me. Simplifying the rooting profile scheme makes sense at this point, and glad to get these parameters in.

[jkshuman]

few small changes, and a question on if we should suggest the range for fire_threshold

[serbinsh]

@rgknox so for the tests, I assume you will start by confirming the BallBerry model produces similar results as before and then in the future we would want to test both versions or eventually just the default (Medlyn?)

[rgknox]

@serbinsh yup, exactly.

I'm wondering what extra tests, if any, we should perform before switching to Medlyn as the default model in the parameter file. I feel it might be a good idea to run a global simulation with it for a few decades just to see if anything weird happens (like an f10_f10 grid). Since the default parameter set (right now) is more about stress testing the model in different modes, and less about providing scientifically reasonable base set, the bar is low in terms of producing reasonable global benchmarks. Although, arriving at a default parameter set that does the latter is the long-term goal, and adding Medlyn as the default their may help get us to that goal.

[serbinsh]

@rgknox Both @Qianyuxuan and I would be happy to talk through a baseline run. We can even do that run here if helpful. But before we do that we should perhaps think about exactly what we need.

[jkshuman]

@rgknox thanks for your continued work on this. looks good.

[rgknox]

Test updates look nominal:
/glade/scratch/rgknox/clmed-tests/fates-sci.1.36.0_api.11.1.0-clm5.0.30-Cdab82c29-Feaa02cf4.fates.cheyenne.gnu/

I did the cheyenne gnu test suite, which is nice for small syntax updates like these (two tests), and baseline comparisons were b4b.

I propose we freeze changes on this PR. In your hands now @glemieux