Simplified leaf sun-shade fraction for two-stream

[rgknox]

Description:

This updates the two-stream calculation of leaf sunlit fraction. The previous implementation attempted to dis-entangle the leaf sunlit fraction from that of the scattering media (leaf+stem+snow). The new implementation simplifies the calculation by assuming the leaf sunlit fraction is comparable to that of the media. The old method may be implemented as an alternative at a later date, but this method is more comparable to how CLM and others (such as Sellers) have calculated sunlit fraction in the past.

The the old and new approach do not yield large differences at moderate to high zenith angles when plants perform the bulk of their photosynthesis. Plots below.

This allows for more straightforward checks on energy conservation.

Collaborators:

Expectation of Answer Changes:

Checklist

If this is your first time contributing, please read the CONTRIBUTING document.

All checklist items must be checked to enable merging this pull request:

Contributor

• The in-code documentation has been updated with descriptive comments
• The documentation has been assessed to determine if updates are necessary

Integrator

• FATES PASS/FAIL regression tests were run
• Evaluation of test results for answer changes was performed and results provided

Documentation

• Technical Note update:
• User's Guide update:

Test Results:

CTSM (or) E3SM (specify which) test hash-tag:

CTSM (or) E3SM (specify which) baseline hash-tag:

FATES baseline hash-tag:

Test Output:

[rgknox]

The following shows results from a sensitivity test conducted at BCI. This test ran the two-stream solver for a generic broad-leaf evergreen canopy using the default parameters, with a a total LAI of 5, and SAI of 0.5. Meteorologic data from the site drove these calculations from 2003-2016. These results show the mean difference between the new-old methods, partitioned by depth in the canopy and zenith angle (denoted by different shaded lines). The right plot normalizes that difference by the mean values of the old method.

Note*: "Sun/shade fraction" means sunlit fraction.

[rgknox]

Technical documentation on the two-stream code is here: https://www.overleaf.com/read/nwmhkyrvszqr#c15e6f

Go to section 5 for derivation of the sunlit fraction calculation.

[rgknox]

Removing the "not ready" label.

[mpaiao]

Thanks for this update @rgknox! I think this is good to go. I only have one question related to scelg%Kb in my review, but that is not something you are changing in this pull request.

[mpaiao]

I agree it makes sense to use vai as opposed to lai in here. From what I see the clumping index is already accounted for in the calculation of scelg%Kb, so I think your edits are all reasonable, thanks for updating it, Ryan!

[mpaiao]

One question, is there any risk for scelg%Kb to be zero or undefined? I am just looking at its definition, and it seems there could be a singularity if both lai and sai are zero. I don't know if this can ever happen, but I wondered about grasses if they are deciduous.

[rgknox]

You make a good point, I'm looking at this line:
https://github.com/NGEET/fates/blob/sci.1.82.4_api.39.0.0/radiation/TwoStreamMLPEMod.F90#L1038

And you're right we do normalize by sai+lai... In the past we always had some stem, even in grasses, but I think we are now allowing to break from that constraint. Either way, it would be better to accommodate this. I'm not sure if a min(x,inf) would return x, but best to avoid the calculation.

[mpaiao]

Maybe something like this would work?

               if ( (scelg%lai+scelg%sai) > nearzero) then
                  scelg%Kb = min(kb_max,(scelg%lai*scelg%Kb_leaf + scelg%sai*Kb_stem)/(scelg%lai+scelg%sai))
               else
                  scelg%Kb = 0.5_r8 * ( scelg%Kb_leaf + Kb_stem)
               end if

(Not sure if the dummy average in the else case is the best way to go).

[mpaiao]

Good catch, this is not needed.

[rgknox]

Here are difference maps from a comparison of two SP runs, base uses the main branch, test uses this branch; simulations are two years, maps are averages of the second year.

https://drive.google.com/file/d/1TPwpCLOzu3gYy9-yJKuVKf8pfUzrAVFE/view?usp=sharing

Differences are subtle and within ranges I was expecting. Thoughts? I added a bunch of water diagnostics because I wanted to rationalize differences in reflected solar radiation (FSR). Changing the sunlit fraction (here) should not change albedo as a direct affect. It can change FSR due to changes in water balance and its subsequent impacts on soil and canopy albedo.