PARENT SESSION
Contributed Oral Session 38: Aquatic Faunal Assemblage Responses to Stressors
Tuesday, August 9, 8:00 AM - 11:30 AM, Meeting Room 513 C, Level 5, Palais des congrès de Montréal

Using landscape level processes to predict areas of concern for aquatic freshwater fishes.

Poos, Mark ^*,1, Jackson, Donald¹, ¹ Department of Zoology, Toronto, Ontario, Canada

ABSTRACT- There is a paradox with the conservation of freshwater species in the Great Lakes. Although there are seemingly endless freshwater habitats as 1/5 of the world's freshwater is found in the Great Lakes, the Great Lakes is heavily industrialized and urbanized marginalizing these habitats. In particular, influences related to urbanization have impacted many rare aquatic species, such as the fish redside dace (Clintostomus elongatus), which relies on high aquatic vegetation and low turbidity. To date, the ability to formulate useful predictive models for this species (and others) has been limited because of the difficulties with the confounding temporal and spatial influences within this region. We evaluated the applicability of two modeling approaches, logistic regression and artificial neural networks, in linking species occurrence with environmental features. Species data were combined with landscape level factors such as geology, type of landcover, porosity and position. Areas with high conservation value were produced from the resultant habitat variables and overlaid using geographic information systems. We demonstrate the applicability of such modeling approaches in evaluating important habitat features, identifying areas of concern for species conservation, and predicting the response in species occurrence with changes in habitat features.

Key words: Freshwater fish ecology, Urbanization, Predictive modelling, Biodiversity and imperilment