I strongly believe in the importance of long-term datasets for identifying long-term trends and relationships in natural populations. During my tenure at UC Davis, I amassed a large mark-recapture dataset on two neighboring populations of California tiger salamanders at the Jepson Prairie Preserve. The California tiger salamander is an endangered species endemic to California. Due to its large habitat requirements and high land prices in California, conservation of this species is controversial. I now have eight years of data on over 30,000 capture events. I also marked a subset of the salamanders so that I could follow the survivorship and movement of particular individuals.
My dissertation research (with Brad Shaffer and Tom Schoener) focused primarily on the landscape ecology of California tiger salamanders. The underlying questions were how much land is required for the conservation of each salamander population and whether certain portions of this habitat are preferred. These are important questions because the primary threat to this species has always been habitat loss. I used the first five years of data from the Jepson Prairie dataset to determine how far the salamanders migrate (up to ~2 km), how they partition this through time (150-200 meters/night), and which other factors affect their habitat preferences (e.g. microelevational differences). I also used these results to illustrate a new method for calculating mitigation credits such that habitat is valued relative to its importance in ensuring persistence of the threatened/endangered species in question, rather than based on area alone.
I have also used my long-term dataset on California tiger salamanders to look at population dynamics and how this species is regulated by climatic factors and interactions with other community members. To accurately address these questions, I supplemented my mark-recapture data with field surveys of natural vernal pools and mesocosm experiments of vernal pool community dynamics. I have used these data to demonstrate that metamorphosing California tiger salamanders face an evolutionary tradeoff between immediate selection for earlier metamorphosis over their first summer followed by later selection for larger size. This tradeoff is important to consider when modeling population dynamics because delayed life history effects can lead to fixed cycles in population size. These data have also been used to make recommendations on whether hybrid tiger salamanders in California should be conserved and more generally how taxon substitutes should be evaluated to ensure maintenance of ecosystem functions.
Niche models are a powerful tool since they take advantage of two large pools of data – georeferenced museum localities and interpolated climate layers for GIS – to inform us about species’ ecology. However, there needs to be more work ground-truthing how these models are made and which of their outputs are informative based on comparisons to independently derived field data. I recently completed a project that evaluates the utility of different methods of building ecological niche models with Maxent for the purpose of uncovering new localities for endangered species. I also completed an evaluation of Maxent’s metrics for measuring variable importance to determine whether or not they accurately identify the variables that are most important in determining a species’ niche. Encouragingly, one of Maxent’s metrics (permutation importance) seems to perform very well, which indicates that it could be used for generating more accurate predictions or how climate change will affect species range shifts.
My research in the McCauley lab at the University of Toronto focused on a set of 36 experimental ponds that we constructed at the Koffler Scientific Reserve. We built these ponds in a geographically structured array so that they are different distances from existing ponds on the landscape that serve as the source of colonizers. Since their construction, we have monitored the colonization process to determine the relative contributions of determinacy and contingency in community assembly and whether this varies based on landscape context. Preliminary data on one of the colonizing species (green frogs) shows that larger frogs are more likely to disperse farther from the source ponds, indicating that dispersers are not random draws from the population as is often assumed.