Plant model needs to return understorey transpiration

[vgro]

Transpiration is only returned for filled canopy layers, but we need the understorey for the energy balance, too.

[davidorme]

OK - so it is written to data but currently in a separate variable:

virtual_ecosystem/virtual_ecosystem/models/plants/subcanopy.py

Lines 464 to 471 in d9a335b

	# Write nutrient uptakes and transpiration - one m3 is 1000 mm/m2
	for name, values in (
	("subcanopy_ammonium_uptake", ammonium_uptake_kg),
	("subcanopy_nitrate_uptake", nitrate_uptake_kg),
	("subcanopy_phosphorus_uptake", phosphorus_uptake_kg),
	("subcanopy_transpiration", subcanopy_volume_m3 / 1000),
	):
	self.data[name] = DataArray(values, coords=coords)

Is there any value to keeping them separate or should it be written to the transpiration layer along with the canopy transpiration?

[vgro]

It can be written along with the canopy, I can then just select the correct layer. Unless you want to keep it separate for any reason, both fine by me.

[davidorme]

At the moment, data["transpiration"] is repopulated at each time step from within PlantsModel.estimate_gpp, so it is implicitly canopy_transpiration. If we want to make it include subcanopy transpiration, then we'll have to declare it in _setup and _update so that it is defined in data for both the Subcanopy and estimate_gpp to write to.

@sallymatson This feels like it makes the code a bit more complex - but it reduces the number of variables and is probably more extensible in the long run (shrub_transpiration? grass_transpiration?). I think this may be the only variable we'd need to aggregate like this?

[sallymatson]

I don't 100% follow - are we just going to add a new "understory_transpiration" or will it be a multidimension array?

In terms of declaring it, I think that solution is fine. There are more and more variables that I think we'll need to declare in _setup which I don't think is inherently an issue. It's generally quite clear and easy to parse so even if _setup gets long I personally don't see it as an issue. Esp if it makes it more streamlined further down in the code.

... but still not 100% sure I understand the exact requirement & proposal 😅

[davidorme]

It's creating the transpiration array in setup before running estimate_gpp and then writing to it from two locations (estimate_gpp for the canopy and Subcanopy.calculate_dynamics for the subcanopy). Those two calls populate different vertical layers in the same array.

[sallymatson]

Okay that was my best guess. In that case yes I think it's totally fine to do so.