PREDICTING PATTERNS OF MAMMALIAN SPECIES DIVERSITY FROM A PROCESS-BASED SIMULATION MODEL

Abstract

Large-scale ecological patterns emerge from multiple biological processes that interact on several scales and are affected by the heterogeneity of the environment. I use a landscape simulation model to explore and analyze how interactions among various ecological processes affect patterns of mammalian species diversity according to body size. The model incorporates ecological realism by using mammalian allometric relationships with realistic values taken from field data. It uses energy as a common currency to model and compare how species of different sizes use their environment. Components of the model include landscape, habitat, patch, community, species, and population. Processes include continuous-time population growth, saturation effect, species-habitat matching, demographic stochasticity (local-scale), dispersal, and catastrophic stochasticity (landscape-scale). I simulated a 16-patch landscape, with each patch having a unique habitat, and a 10-species pool in which species differ only in body size. Summed over the landscape, a log-normal relationship between body size and species abundance emerges. Additionally, geographic range increases nonlinearly with the log of body size. These patterns are qualitatively consistent with patterns of mammalian species diversity observed in natural systems, suggesting that they can emerge from scaling of habitat use with body size. I also simulated a simple 4-patch landscape, with each patch having a unique habitat, and a 26-species pool to draw from in which species differ only in body size. The results show that interspecific competition by itself reduces species diversity in each habitat and in the entire landscape. Stochasticity depresses mean population sizes but opens opportunities for species to avoid competitive exclusion. It also allows different habitats to have different communities that are characterized by the consequences of which large species become locally extinct at random. Demographic and catastrophic stochasticities differ, however, in their characteristic effects on species diversity. These characteristic effects can be used as fingerprints to identify processes acting in nature. Dispersing individuals move between habitats and reestablish the local populations of their species. Overall, the model provides us with predictions of how process-interactions in heterogeneous landscape may affect species composition and community structure.