PARENT SESSION
Poster Session: Qualitative Relationships Between Landscape Processes and Patterns

Developing Habitat Models for Aquatic-Dependent Wildlife. ^*COPELAND, JANE L. ¹, ^*KUHN, ANNE ² and ^*BRENNAN, MARK ³, ¹ Computer Sciences Corp/ US EPA Atlantic Ecology Division, Narragansett, RI, USA² US EPA/Atlantic Ecology Division, Narragansett, RI, USA³ NH Loon Preservation Committee, Moultonborough, NH, USA

ABSTRACT- Minimizing the risks of chemical contaminants to wildlife populations remains an important goal for the EPA. However, approaches are also needed for evaluating risks from non-chemical stressors on populations of aquatic-dependent wildlife. Habitat modeling provides an empirical approach to summarize effects of multiple stressors on the basis of spatial relationships. Results from these models provide insight into environmental factors and conditions that may be causally related to changes in abundance and distribution of wildlife populations. We are using geospatial modeling approaches to identify environmental factors that have the highest correlations with presence of the Common Loon Gavia Immer. Using loon nest location data collected by the NH Loon Preservation committee, we have developed a set of landscape metrics for over 600 lakes between 1980 and 2000. Generalized linear models (GLM) including logistic regression analysis are used to describe and evaluate the ecological relationship between loon presence and these landscape metrics. These statistical analyses are also used to assess the relative risks of multiple stressors such as dietary methylmercury, lake acidification, habitat alteration and human disturbance to Common Loon populations. This research supports the EPA′s Wildlife and Aquatics Stressors Research Strategy Program and will lead to the development of protective criteria for wildlife and advance the ecological risk assessment process.

KEY WORDS: gavia immer, habitat models, aquatic stressors, generalized linear models