C3 and C4 refer to the different pathways that plants use to capture carbon dioxide during photosynthesis.
If say, a grid cell is 100% glacier, then PCT_NATVEG should be 0%. I would agree that it seems like the pfts could then be set to any values. However, when I look at a surface dataset, I see that in those cases where glacier is 100%, the bare soil pft (0) seems to be set to 100% and the other pfts are set to 0%. I don't remember why that is, it might have something to do with dynamic landunit requirements (if glacier goes away perhaps). So I would stick to the requirement.that the pfts add up to 100% even in cases where PCT_NATVEG is 0%, in your own dataset.
Why do you say "at most four kinds of pfts"? I'm curious where you got that information?
If say, a grid cell is 100% glacier, then PCT_NATVEG should be 0%. I would agree that it seems like the pfts could then be set to any values. However, when I look at a surface dataset, I see that in those cases where glacier is 100%, the bare soil pft (0) seems to be set to 100% and the other pfts are set to 0%. I don't remember why that is, it might have something to do with dynamic landunit requirements (if glacier goes away perhaps). So I would stick to the requirement.that the pfts add up to 100% even in cases where PCT_NATVEG is 0%, in your own dataset.
Why do you say "at most four kinds of pfts"? I'm curious where you got that information?