Fix Canopy Wind Speed Discontinuity with Massman/MOST Approach

[drnimbusrain]

@zmoon pointed out the unphysical jump in the modeled wind profile from above the canopy to the canopy top (panel on far left below):

The main issue is that the canopy-wind codes were originally only designed to go from canopy-top downward, and then it got a bit fragmented when combining this approach with the MOST and unified RSL approaches I adopted from Noah-MP and Rosenzweig et al., (2021) https://doi.org/10.1029/2021MS002665.

This could be rectified by making the wind profile smooth below the reference height wind (e.g., @ 10 m winds from model). Reference height wind is needed in WAF calculation, so winds can be constant above the reference height wind.

Ultimately, we should have a smoothly varying wind speed below the reference height wind (such as in Bonan's plot below):