Consider the removal of the PModel.estimate_productivity method

[davidorme]

At the moment, PModel has a separate method for scaling light use efficiency calculations up to GPP:

model = PModel(....)
model.estimate_productivity(fapar=..., PPFD=...)

That saves a few calculations if only LUE is required. But, the same can't be applied to the calculation of SubdailyPModel and it would be better to have the two models have closer signatures. So, I'm consulting, but I think we lose this method and align the signatures so that both use:

model = PModel(..., fapar=..., PPFD=...)

We can then add getters for some of the properties so that they are calculated on demand to keep the memory footprint and calculation speed lower.