Abundance, physiology, and population structure: Fine-scale landscape genetics of a terrestrial salamander

Background/Question/Methods

Inferring process from pattern can be a challenging undertaking when dealing with ecological complexity. The distribution and abundance of organisms on the landscape is often interpreted through the lens of competition, movement, or physiology, as well as interactions with the abiotic environment. Further, movement, distribution, and abundance often coincide with favorable abiotic environments such as temperature, moisture, or nutrients. At its core, landscape genetics seeks to identify the spatial processes shaping the observed patterns of genetic diversity across the landscape. Unfortunately many landscape genetic studies are predominantly exploratory, seeking correlations with landscape features without well-established hypotheses. To increase understanding of process-driven patterns in landscape genetics, we studied the western slimy salamander (Plethodon albagula) in east-central Missouri with three specific questions: (1) Where are salamanders on the landscape, and what environmental factors influence local abundance? (2) Is there a physiological constraint underlying the observed patterns of distribution and abundance? (3) How is spatial genetic structure shaped by abundance and physiology across the landscape? We utilized a combination of abundance modeling, spatial quantification of water loss using plaster of Paris models, and landscape genetics analyses to assess the factors contributing to salamander population structure across a 1300 ha landscape.

Results/Conclusions

Plethodontid salamanders are highly sensitive to water loss, in part due to their lack of lungs and cutaneous respiration. We found that abundance of salamanders was best predicted by canopy cover, topographic position (ridge, slope, ravine), and the interaction between wetness and solar exposure. The spatial relationships of these parameters are such that abundance is predicted to be highest in forested ravines with lower solar exposure. Plaster models deployed across the landscape served as surrogates for live salamanders to quantify rates of water loss. We found that rates of water loss across the landscape were inversely related to predicted abundance, suggesting that water loss is likely a physiologically-limiting process underlying the distribution of salamanders. Finally, using a multivariate landscape genetics approach, we determined that genetic distances were significantly correlated with both spatial abundance and water loss surfaces, but that the independent landscape features underlying these processes were better predictors. Specifically, topographic position and the interaction between solar exposure and wetness explained more of the spatial genetic variation. Our results highlight the importance of understanding basic ecological and physiological factors as mechanisms for interpreting spatial genetic patterns.