Models of nature are implicitly influenced by the scale of observation of the processes on which they are founded. The continuum model and the hierarchical patch-based model are two alternate approaches for the spatial modelling of fauna distribution. The continuum model aggregates continuous approximations to individual landscape characteristics, whereas the hierarchical patch-based model constructs a hierarchy in which classifications of landscape characteristics describe an interconnected series of patches. We propose the hierarchical patch-based theory for models of population distributions and landscapes in which the spatial patterns can be effectively represented by mosaics at the variety of levels within the set of individual process models. Given that observations are typically made as points or pixels, and that discrete boundaries exist in both natural and human-modified landscapes, we suggest that the hierarchical patch-based method has many applications in conservation and management.