Cross‐Scale Morphology, Geometry, and Dynamics of Ecosystems

This paper tests the proposition that a small set of plant, animal, and abiotic processes structure ecosystems across scales in time and space. Earlier studies have suggested that these key structuring processes establish a small number of dominant temporal frequencies that entrain other processes. These frequencies often differ from each other by at least an order of magnitude. If true, ecosystems therefore will have a few dominant frequencies that are endogenously driven and that are discontinuously distributed. This paper additionally tests the proposition that these structuring processes should also generate a discontinuous distribution of spatial structures coupled with the discontinuous frequencies. If that is the case, animals living in specific landscapes should demonstrate the existence of this lumpy architecture by showing gaps in the distribution of their sizes. This proved to be the case for birds and mammals of the boreal region forest and the short—grass prairie. Alternative hypotheses to explain the body mass clumps include architectural, developmental, historical, and trophic causes. These were all tested by comparing body—mass clump distributions (1) in ecosystems having different spatial structures (forest, grassland, and marine pelagic) and (2) in different animal groups having different body plans (birds and mammals) or feeding habits (carnivore, omnivore, and herbivore). The only hypothesis that could not be rejected is that the body—mass clumps are entrained by discontinuous hierarchical structures and textures of the landscape. There is evidence for at least eight distinct habitat "quanta," each defined by a distinct texture at a specific range of scales. These eight quanta together cover tens of centimetres to hundreds of kilometres in space and at least months to millennia in time. There is a striking similarity, but not identity, between the clump structure of prairie and boreal animals. This indicates that many processes that form qualitative habitat structure are common to both landscapes or ecosystems, but a few are landscape specific, particularly over larger scales. That conclusion is extended to all terrestrial ecosystems by an analysis of the body—mass clump structure of all North American birds. In contrast, there are striking differences in clump structure between landscapes and "waterscapes," indicating that fundamentally different processes shape structure in terrestrial and open ocean systems. The discontinuous body—mass structure provides a bioassay of discontinuous ecosystem structure. Mammalian carnivores, omnivores, and herbivores all show the same number of body—mass clumps, and the gaps in these distributions occur at the same body masses. Mammals and birds show the same number of body—mass clumps, but the mass gaps for mammals occur at larger sizes than those for birds in such a way that the log—transformed body—mass gaps for mammals are correlated linearly with those for birds. Hence there is a simple cross—calibration between the mammal and bird bioassays. I compiled and analyzed published data on home ranges in order to convert body masses into an absolute linear measure of geometric structures in the landscape. A new and general equation was developed relating home—range size to body mass, and was tested by reanalyzing published data for mammalian carnivores, omnivores, and herbivores and for birds. I conclude: (1) Birds and mammals of all trophic levels utilize resources in their foraging areas in the same way by measuring the spatial grain of habitat patches with a resolution defined as a function of their size (i.e., the animal's step length or minimum unit of measurement). The step length is a morphological function of the size of animals and is not significantly affected by trophic status or taxonomy of the groups considered. That explains why all trophic levels and both birds and mammals show the same qualitative body—mass clump structure. (2) Home—range data can convert the body—mass data to a quantitative estimate of texture, i.e., of fractal dimension of the landscape. The landscape forms a hierarchy that contains breaks in object sizes, object proximities, and textures at particular scales. Animals also demonstrate a hierarchy of decisions whose target suddenly shifts at specific scales in space and time. The interaction between these two hierarchies produces the discontinuous body—mass clump structure. The breaks in geometry in the landscape occur because structuring processes exert their influence over defined ranges of scale. The temporal and architectural structure of habitat quanta are in general determined by three classes of processes, each dominating over three different ranges of scale. Vegetative processes that determine plant growth, plant form, and soil structure dominate the formation of texture at fine microscales of centimetres to tens of metres in space and days to decades in time. At the other, macroscale extreme, slow geomorphological processes dominate the formation of a topographic and edaphic structure at large scales of hundreds to thousands of kilometres and centuries to millennia. At the mesoscales in between, contagious disturbance processes such as fire, insect outbreak, plant disease, and water flow dominate the formation of patterns over spatial scales of hundreds of metres to hundreds of kilometres. In addition, the direct impacts of grazing by large herbivores and of human activities, and the indirect effects of large predators and animal disease, further transform spatial patterns over these meso—scales. These processes operate on time scales of years to decades, making them critically important in determining whether present local, regional, and global human influences will trigger a transition in vegetation types, and, if so, how rapidly. The paper provides a direction for the development of programs to evaluate, monitor, and predict ecosystem and community changes across scales. The necessary research elements include (1) models that incorporate a few scale—dependent structuring processes to allow cross—scale analysis; (2) comparative studies of different disturbed and undisturbed landscapes using the animal body—mass bioassay technique to identify critical scales of ecosystem geometry; (3) analysis of remote imagery to identify spatial discontinuities and regions of scale invariance; and (4) behavioral studies of the hierarchy of animal decisions to identify species groups vulnerable to predicted (using models) or observed (using remote imagery) changes in vegetation geometry.