Refactor "Prescribed Soil" for canopy and biogeochemistry

[kmdeck]

Purpose

Both the canopy and the biogeochemistry models require time varying soil properties when run in standalone mode. These do not make use of ClimaUtilities TimeVaryingInput, require two "PrescribedSoil" structs, and, when run in integrated mode, also require two "Prognostic Soil" structs.

The idea here is to introduce a "PrescribedSoilDriver" in shared_utilities/drivers.jl which has the fields needed by these standalone models. Then, we can make use of the same "update drivers" technology to update the soil properties as needed, in place.

See kd/soil_bio_consistencykd/soil_bio_consistency for sketch

***Should biogeochemistry be a standalone component with a prescribed soil model, or should this be a new type of soil model?

Cost/benefits/risks

This would have nice parallels between Prescribed/Coupled atmos (Prescribed/Prognostic soil's), and be clearer overall

Producers

Kat

Components

Introduce PrescribedSoil driver with things needed for the biogeochemistry and canopy models.

Introduce a single PrognosticSoil driver which points towards the correct allocated fields in the integrated case

Task breakdown

1. Generalize drivers beyond radiation and atmospheric variables
2. Create PrescribedOrganicCarbon driver using a Spatially Varying Input and use it in biogeochemistry and in all integrated runs. This will also remove an allocation of a 3d field. Eventually, when we add a carbon model, this would be an alternative.
3. Create PrescribedSoil driver and use it in biogeochemistry (temperature and moisture)
4. Use the same PrescribedSoil driver in Canopy
5. Create Prognostic Soil and use in soil canopy model and soil biogeochemistry model

Does (2) encompass LAI? or we will treat Biomass as a different prescribed/prognostic component?

Reviewers

@AlexisRenchon @Sbozzolo