store canopy turbulent fluxes in cache

[juliasloan25]

Is your feature request related to a problem? Please describe. Currently, when we update the canopy boundary fluxes, we allocate a field of tuples to store the result of canopy_turbulent_fluxes, then unpack each of the tuple entries to store them individually in p (here).

It would be simpler to instead store the whole field of tuples directly in the cache. We will still be able to access the individual values, but can skip some intermediate copying and allocations.

See original suggestion here https://github.com/CliMA/ClimaLand.jl/pull/675/files#r1731979646