You are here:
EFFECTS OF HABITAT LOSS ON POPULATIONS OF WHITE-FOOTED MICE: MATRIX MODEL PREDICTIONS WITH LANDSCAPE-SCALE PERTURBATION EXPERIMENTS
Citation:
BURNS, C. E. AND J. S. GREAR. EFFECTS OF HABITAT LOSS ON POPULATIONS OF WHITE-FOOTED MICE: MATRIX MODEL PREDICTIONS WITH LANDSCAPE-SCALE PERTURBATION EXPERIMENTS. LANDSCAPE ECOLOGY. Springer, New York, NY, 23:817-831, (2008).
Impact/Purpose:
EPA’s ecological risk assessors have identified the need for wildlife population models incorporating spatial dynamics and density dependence and for testing of wildlife risk assessment models in general. This publication extends our existing field-parameterized spatial matrix models (Grear and Burns, Landscape Ecology, 2007) that were designed to capture the roles of specific habitats in landscape level population dynamics. Specifically, density dependent dynamics are incorporated so that landscape level effects of spatially discrete stressors (habitat loss in this case) could be predicted and tested in field experiments. Qualitative agreement between model predictions and experiment results are reported and ecological mechanisms underlying areas of disagreement are discussed. Because population models used in wildlife risk assessment are so rarely tested, this manuscript is expected to be an important contribution to EPA’s risk assessment methodology as well as the general ecological literature.
Description:
Habitat loss is the leading cause of decline in wildlife diversity and abundance throughout the world, and understanding its impacts on animal populations is a critical challenge facing conservation biologists. Population viability analysis (PVA) is a commonly used tool for predicting the effects of habitat loss. In spite of the growing number of studies using this type of modeling approach, and its wide use in conservation practice, PVA model predictions are rarely tested. This is particularly true for studies of animal populations in landscapes with multiple habitat types (mosaics) linked by movement of individual organisms between habitats. We constructed and tested spatially structured demographic matrix models for populations of white-footed mice, Peromyscus leucopus, in northeast Connecticut, USA. Our models were parameterized with four years of field data on demographic and movement rates at three mosaic sites. The models predicted that while loss of low quality habitat may not have a large impact on the population dynamics of animals in the surrounding landscape, loss of intermediate or high quality habitat could trigger rapid population declines extending throughout the mosaic. We then conducted landscape manipulations designed to test these model predictions. We carried out clear-cut logging or prescribed burning of the highest quality habitat at two sites (the third site remained unmanipulated), and monitored subsequent Peromyscus movement, demography and population dynamics. These experiments indicated that the models were qualitatively good predictors of Peromyscus response to habitat loss—population declines were observed across the mosaic sites following habitat manipulation—however, natural deviations from model assumptions did lead to discrepancies between observed and predicted responses. We show that PVA-based modeling can be a useful tool for understanding how populations respond to habitat loss in mosaic landscapes. Further, we show that perturbation of a single habitat can impact dynamics throughout the mosaic landscape. We believe that experimental testing of model predictions is a crucial and often overlooked component of PVA-based studies, and that the iterative cycling of modeling and experimental validation stands to greatly enhance our confidence in the predictive capabilities of PVAs and their utility in wildlife conservation.