Clarify the coding and usage of frac_nobio

[jpolcher]

An analysis of the usage of frac_nobio in the model (​https://forge.ipsl.jussieu.fr/orchidee/wiki/Documentation/Frac_Nobio) has revealed some issues which need to be addressed. This will clarify the code and allow new evolutions which are planed.

frac_nobio and totfrac_nobio contain redundant information as totfrac_nobio=SUM(frac_nobio) over nnobio. But both are passed to most subroutines and nowhere is it verified that the information matches.

The code would be much more easier to follow if only frac_nobio would be passed. When totfrac_nobio is needed the sum would be done locally.

totfrac_nobio is for instance the only variable used for STOMATE and this is logical. So at least slowproc.f90 will need to compute this variable to pass it to STOMATE.

The other issue is that there is no guarantee that the model conserves all stocks if frac_nobio changes over time. So the model should check that frac_nobio does not change. Especially in slowproc.f90 where frac_nobio can change with the reading and interpolation of the PFT maps.

Furthermore the introduction of nnobio has not been done fully. Many other variables need to have this dimension as well, for instance the parameters to be used on the surfaces other than ice.