Extend farquhar mechanism to spatially varying

[imreddyTeja]

Purpose

Extend FarquharParemeters and OptimalityFarquharParameters to support a mechanism parameter that is
<: Union{FT, ClimaCore.Fields.Field}
Closes #762

To-do

• Add to NEWS.md
• Add to ClimaArtifacts repo

Content

This change is applied to both OptimalityFarquharParameters
and FarquharParameters. Previously, the mechanism parameter
was held as an empty struct of type ::C3 or ::C4 to allow
for dispatch. Because it is not possible to create a
ClimaCore field of different structs, the mechanism param is changed to a
is_c3 param that holds 1s and 0s. This choice was made because
reading a NetCDF file requires regridding, which must be performed over Numbers.
If the input data file must store the mechanism as a float, then it is easiest to
also internally represent the mechanism as a float. The C3/C4 structs
are now removed. The functions that previously used C3/C4 structs for dispatching
are converted to two seperate functions, which are called based on the is_c3 value.

If you have a better name for the parameter than is_c3, please suggest it. I previously considered
c3_mechanism_flag.

The constructor checks that the mechanism value of field is
between 0 and 1, and throws an error otherwise. Regridding a field of
0s and 1s may result in values between 0 and 1, so the constuctor rounds them
to the nearest integer. This implementation allows for future support of
mixed mechanisms in an area represented by proportion c3 without many changes.

All calls to the FarquharParameters constructor are changed to
work with the changes to mechanism representation. Tests are also
changed to work with new constructor and representation of mechanism.

The tests in canopy_model are changed to run twice using a for loop. The first run has all the parameters
as floats, and the second run has all canopy parameters that can be fields as fields. In this
case, the photosynthesis mechanism is tested as a field that alternates between 1.0 and 0.0,
which represents alternating between C3 and C4. The tests were previously all run over a Point
domain. The tests that use FarquharParameters are changed to run on a spherical surface.
Additionally, a test for vcmax as a field is added.

Land.jl and land_region.jl are changed to use the mechanism map from
the clm data.

---

• I have read and checked the items on the review checklist.

[Sbozzolo]

Thank you!

[kmdeck]

since we only allow for C3 right now with the Optimality model, should we @Assert all(is_c3 .== 1) in the constructor?

[imreddyTeja]

I followed the behavior of the previous constructor, which ignores any kwarg for is_c3 and sets is_c3 to FT(1). Should this be changed?

[kmdeck]

I see that now, I think that is fine. I guess someone could set up the parameters using the default constructor with is_c3 = 0, but we note that this only supports C3 for now in the doc string.

[kmdeck]

very nice, thank you! I left some questions in the comments but otherwise this looks good to me.