Exploring Drivers of Gene Flow in Jaguars and Pumas in Southern Mexico via Molecular Scatology and Eco-Evo Simulations

Thesis (Ph.D.)--University of Washington, 2017-01The profound fragmentation and degradation of Neotropical forests over the past 50-100 years poses a significant threat to the wildlife populations in Mesoamerica. Neotropical large carnivores, jaguars (Panthera onca) and pumas (Puma concolor), are at particular risk from forest conversion due to their large spatial requirements and high vagility, and are key contributors to ecosystem function in their roles as top predators. Diminishing structural connectivity of the landscape is likely to impede gene flow for both species, with potential impacts on population or species persistence. However, the mechanistic drivers behind gene flow are poorly understood. In this dissertation, I explore how landscape patterns and habitat selection interact to influence gene flow of jaguars and pumas in southern Mexico. The first half of this dissertation is dedicated to the quantification of jaguar and puma landscape use, gene flow, and genetic diversity in southern Mexico, where we know little about the remaining populations (Chapters 1 & 2). This work was based on noninvasive genetic samples, collected with the aid of wildlife detector dogs, in the Uxpanapa valley of Veracruz, and northern Quintana Roo. Resource selection analysis suggests less ubiquitous use of the landscape by jaguars due to greater habitat specificity for natural vegetation, rugged terrain, and avoidance of human activity, as compared to use of a broader array of habitats by pumas. However, I did not find evidence of gene flow restriction within Uxpanapa despite low predicted connectivity between forest patches. At the regional scale between study locations, pumas exhibited greater genetic discontinuity than sympatric jaguars. These findings are also echoed in the literature and highlight an apparent disconnect between predicted structural connectivity at fine-scales and gene flow at broader scales, suggesting that behavioral components of movement ecology may differ between resource use within home ranges and juvenile dispersal. In the second half of this dissertation, I turned to computer simulations to explore the possible drivers of gene flow by scaling-up fine-scale processes, such as resource selection, to broader-scale patterns, such as gene flow (Chapters 3 & 4). I explored the utility of an individual-based modeling (IBM) platform, HexSim, for integrating population dynamics, movement ecology and behavior, and evolutionary processes on spatially explicit landscapes. I then employed this modeling platform to build a biologically and spatially realistic eco-evo IBM of large felid gene flow. I used this model to conduct a pilot test of hypothesized drivers of gene flow through Mexico, Guatemala, and Belize. Results suggest that gene flow was decreased by territorial habitat specialization and increased by sensitivity to landscape features during dispersal. My results showcase the model’s ability to investigate how specific components of complex movement behavior drive of gene flow. The large-felid eco-evo IBM offers a powerful tool for future investigations of mechanistic connections between fine-scale resource selection and gene flow at broader scales, as well as for forecasting the effects of habitat preservation versus connectivity conservation. Jaguars appear to have greater forest selectivity and sensitivity to human activity as compared to pumas, highlighting the need to bolster the existing national incentives for forest preservation in order to protect this declining species. Finally, my results stress the need for state or federal protection of the Uxpanapa valley, Veracruz, as a biological hotspot that provides a stepping stone for movement between Central and North American wildlife populations