Refactor/clean up buildings albedoes across codebase #1101
Description
From my investigation via #851 and more recent analysis, I discovered that at the moment we have different albedoes related to buildings:
1 - properties.land_cover.bldgs.alb: this is a bulk or effective albedo for all building surfaces as a land-cover class, not per facet, not per vertical layer. Single value per site. It feeds into the phenology / surface state (e.g. phenState%alb(2) on the Fortran side). BEERS uses alb_bldg => phenState%alb(2) as the effective building albedo when doing radiative calculations. Is the 1D / BEERS albedo.
2 - vertical_layers.roofs[i].alb, vertical_layers.walls[i].alb: this is per‑layer, per‑facet albedo for SPARTACUS. These are independent of the bulk properties.land_cover.bldgs.alb (i.e. at the moment we do not have any explicit link between them). We might want to add a link/formula between these albedoes and the bulk albedo.
3 - STEBBS case: albedo is encoded via TAR triplet (Transmissivity, Absorbitivity, Reflectivity) for roof, wall, window:
RoofReflectivity, WallReflectivity, WindowReflectivity in BUILDING_ARCHETYPE_PRM. These are per facet (roof, wall, window).
If we want to harmonise/reduce the number of albedoes in the codebase/YAML, we need to link these 3 physics together via some formulas/rules. I can imagine that 1 and 2 can be linked via some weighted function maybe, but for STEBBS it might be more complex.
Happy to discuss more on this to find a good strategy :)