Tuesday, August 7, 2007 - 11:00 AM

COS 31-10: Predictions and postdictions from the hierarchical representation of habitat

Jurek Kolasa, McMaster University, Craig R. Allen, University of Nebraska, Craig Stow, NOAA, and Jan Sendzimir, International Institute for Applied Systems Analysis.

Any habitat, aquatic or terrestrial, is a nested mosaic of smaller habitats such that larger fragments contain subcategories, which are smaller, and those in turn are composed of even finer microhabitat subcategories that are even smaller.  A regional pool of species responds to such a universal hierarchical structure in predictable ways in terms of species diversity, species densities, and their variability.  This perspective stimulated research and produced new insights but its quantitative form was scarcely explored.  We present a quantitative version of the habitat-based model accounting for many aspects of community-habitat relationship.  Its core terms include species’ habitat specialization, amount of habitat resources available to each species, habitat fragmentation experiences by a species, and adult body size. We review diverse testable predictions derived from the model.  For example, we show that: (a) habitat specialists should be relatively more variable and prone to extinction than habitat generalists; (b) population decline should be an accelerating function of habitat loss and fragmentation; (c) predators should show a faster decrease in population energy use with body size as compared to omnivores; (d) perturbations to habitat structure should favor habitat generalists over specialists; (e) animal body sizes should form clumps for animals constituting a community; (f) invader species should be most successful at the edges of body size groups. While some of these predictions can be inferred using other considerations and models, we see the value in striving to develop and refine a unifying perspective as the one hinted above.