You are here:
CONSTRUCTING, PERTURBATION ANALYSIIS AND TESTING OF A MULTI-HABITAT PERIODIC MATRIX POPULATION MODEL
Citation:
Grear, J AND C. E. Burns. CONSTRUCTING, PERTURBATION ANALYSIIS AND TESTING OF A MULTI-HABITAT PERIODIC MATRIX POPULATION MODEL. Presented at The Society of Environmental Toxicology and Chemistry Annual Meeting and 4th Chemistry World Congress, Portland, OR, November 14-18, 2004.
Description:
We present a matrix model that explicitly incorporates spatial habitat structure and seasonality and discuss preliminary results from a landscape level experimental test. Ecological risk to populations is often modeled without explicit treatment of spatially or temporally distributed habitat features, although affected organisms often move through heterogeneous landscapes. For example, temporally distributed processes such as seasonality are frequently subsumed into models with annual time increments. While simple models require fewer data, these spatial and temporal abstractions limit the power to explore impacts of spatially or temporally variable stressors. Often, this exploration provides essential support for management options because exposure to stressors can occur in a subset of an individual's home range or during a discrete phase of a life cycle (e.g., embryonic development). On the other hand, explicit and accurate treatment of time and space requires larger amounts of data and the resulting models sometimes lack heuristic clarity and flexibility. It is therefore critical that the performance of spatially and/or temporally explicit population models be tested and compared to their simpler counterparts so that levels of abstraction provide an appropriate balance between data availability and modeling goals. To this end, we extended demographic models to an intermediate level of realism using periodic (i.e., seasonal) projection matrices that incorporate stage- and habitat-specific classification of survival, reproduction and movement. This presentation focuses on the model's data requirements and estimation procedures and the analytical perturbation analyses that were made possible by this approach (e.g., sensitivity to loss of specific habitats). We present preliminary results of a landscape level experimental test using a density-dependent version of this model for the white-footed mouse (Peromyscus leucopus) and compare these results to those of simpler models.