Connectivity, metapopulation dynamics, and genetic structure of tiger salamanders in a heterogeneous landscape

Metapopulation biology has been integral for understanding the impact of spatial habitat structure on ecological and evolutionary processes. In fragmented landscapes, theory predicts that species occupancy and turnover dynamics depend on habitat area and isolation, and isolation has historically been an important predictor of gene flow. However, metapopulation theory is neutral with respect to the effects of habitat heterogeneity on population processes. Landscape ecology approaches have begun to account for effects of habitat quality and matrix structure on occupancy and gene flow, but few empirical studies have integrated the area-isolation and habitat paradigms to understand metapopulation dynamics and genetic structure in the same system. Here, I employ both approaches to understand the spatial population dynamics and genetic structure of tiger salamanders (Ambystoma tigrinum tigrinum ) in an agricultural landscape in Illinois.;First, I assessed the degree to which matrix heterogeneity influences A. tigrinum movement behavior. Using a field experiment, I showed that a physiological constraint, desiccation risk, varied significantly among matrix habitats (corn, soybean, forest, prairie). Water loss was greater in corn and prairie than in forest and soybean, indicating that dispersal costs can vary among agricultural crops. To assess whether movement decisions were influenced by desiccation risk, I tracked the movements of individuals released on habitat boundaries for two treatment combinations: soybean-corn, soybean-prairie. I observed that movements were oriented towards soybean in both cases, suggesting that variation in desiccation risk among matrix habitats influenced salamander movement decisions.;Next, I examined the effects of area, isolation, and habitat heterogeneity on metapopulation dynamics of A. tigrinum. Emphasis was placed on understanding the role of connectivity in moderating interactions between A. tigrinum and predatory fish. Occupancy and turnover of A. tigrinum were documented in 90 wetlands for three years. Since desiccation risk influenced A. tigrinum movements, I tested whether a connectivity metric that accounted for desiccation was a better predictor of occupancy and turnover than metrics based on Euclidean distance or expert opinion. Occupancy and colonization probabilities were related positively to connectivity and negatively to fish presence. Extinction probability was related positively to fish presence, but extinction risk was low in connected networks, suggesting a rescue effect. A desiccation-informed connectivity metric was a better predictor of colonization probability than alternative metrics, whereas a Euclidean model was the best predicator of occupancy and extinction probabilities. The results indicated that the effect of desiccation risk on individual movement can scale up to influence metapopulation processes, and that the effects of predatory fish on metapopulation dynamics depended on spatial connectivity.;Finally, I evaluated whether ecological factors underlying occupancy and turnover were also important predictors of metapopulation genetic structure. Newly colonized populations were more genetically differentiated than established populations, indicating that founder effects influenced genetic structure. However, the degree of genetic differentiation varied spatially. Genetic differentiation was related negatively to both wetland area and spatial connectivity. Differentiation was not strongly related to habitat quality, suggesting that metapopulation factors were more effective at reflecting the historical strength of genetic drift and gene flow than current habitat suitability.