Comment by swiftcoder

All their growth models do involve light, however. So it probably does make sense to have a common light representation - ideally some sort of sparse cone storage though, not a voxel grid the size of a forest.