improve interface of patch counts

[rgknox]

The host land-model is solely in charge of how much space is available for patches. FATES does not really ask the HLM how much space is available, nor does it weigh in on this. We are just making sure that the maximum number of fates patches is smaller than the number the HLM comes up with. See EDTypesMod.F90: integer, parameter :: maxPatchesPerSite = 10

This situation could be improved by either leaving the decision completely up to the host model but at least performing a check to see if the desired number set in EDTypesMod is lower than the space allocated by the HLM, or better yet, by weighing in during clm_varpar when numpft is being decided, and when natpft_lb and natpft_ub are being calculated.

[rgknox]

flagging this to be investigated/implemented