fix the q2 diagnostics in NoahMP

[CharlesZheZhang]

This PR is to fix the high q2 issue in urban grids in NoahMP, posted here in this thread:
https://forum.mmm.ucar.edu/threads/high-urban-moisture-with-noah-mp-in-mpas-atmosphere-8-2-3.22160/#post-53574
from a closed PR: #1341

Essentially, the fix is to correct the q2 diagnostics in NoahMP in hotfix-v8.3.1:
in the Noah-MP source file: EnergyVarOutTransferMod.F90
Line 139: MPAS-Model/src/core_atmosphere/physics/physics_noahmp/drivers/mpas/EnergyVarOutTransferMod.F90 at ac3866c1e5b05f6d4f5bd41aeab7d3882bace514 · MPAS-Dev/MPAS-Model
change it to the following:
NoahmpIO%Q2MXY(I) = NoahmpIO%Q2MBXY(I) * ( 1 - NoahmpIO%FVEGXY(I) ) + NoahmpIO%Q2MVXY(I) * NoahmpIO%FVEGXY(I)

I have done a short test with the hotfix-8.3.1 version and with this q2_diag fix, and show this fix will help reduce the high q2 over urban points. please see the figure below:

@cenlinhe @barlage

[cenlinhe]

Yes, this bug fix looks good to me and should be included as soon as possible.

[barlage]

Any sense as to why there is such a large difference over urban areas for this code change? Why is this a "bug" fix? This seems to be an order of operations difference. What makes that special for urban area?

[cenlinhe]

I think it is because Q2 (mixing ratio) is very low in urban, so the grid-level mean Q2 computed either from Q2MV and Q2MB or from specific humidity depends on the order of operations since the calculation from specific humidity to Q2 is nonlinear, which has the most impact when it is very small.

[tslin2]

Note that in WRF, Q2 for urban grids is Q2MB because urban assumes no vegetation (for urban bulk scheme urban option is 0), urban vegetation is not considered for bulk urban scheme, so here should follow WRF considering NoahmpIO%FVEGXY = 0.0 or NoahmpIO%Q2MXY(I) = NoahmpIO%Q2MBXY(I) for urban grids

https://github.com/wrf-model/WRF/blob/f52c197ed39d12e087d02c50f412d90d418f6186/phys/module_surface_driver.F#L3358-L3366