- American pikas (Ochotona princeps) are considered an indicator species of climate change. Adaptations for cold climates and active winters make pikas particularly sensitive to increasing temperatures. This, combined with evidence that multiple historically occupied populations have been extirpated within the past century, contributed to American pikas becoming a focal species for climate change research. This dissertation is primarily the product of a collaboration among the National Parks Service and multiple academic institutions whose principal goal was to assess the vulnerability of the American pika within eight national parks representing a variety of habitat types and environmental conditions. While much of the previous climate change-related research on American pikas has focused on how climate affects occupancy, this work sought to establish how landscape characteristics and climate may interact and influence dispersal, population connectivity, and ultimately population vulnerability or resilience. I addressed these questions using population and landscape genetics approaches within nine national parks, two national wildlife refuges, and a national forest.
American pikas represent one of the best described mammalian metapopulation systems. They are restricted to rocky habitats, such as talus and lava, which are frequently patchily distributed on the landscape, both at the local and regional scales. Functional connectivity, the combination of habitat configuration and an organism’s ability to move through the landscape, is essential for maintaining robust genetic populations. This is particularly true within metapopulations which, by definition, rely on dispersal of individuals among habitat patches for overall long-term stability. In Chapters 1 and 2, I used a landscape genetics approach to identify features that either promote or inhibit dispersal of American pikas. There has been significant debate over the past few years about the best statistical approaches within landscape genetics and the appropriateness of certain methods in particular. In Chapter 1, I used Crater Lake National Park as a test case to evaluate the effectiveness of Mantel and partial Mantel tests in a causal modeling framework to correctly identify underlying landscape variables (e.g., topography, aspect, water barriers) and their effect on gene flow. I demonstrated, through simulations, that this approach was able to correctly identify the landscape variables, but not precisely the magnitude of their effect on gene flow. I concluded that while results need to be interpreted with caution, this method was effective for identifying which variables are important in shaping functional connectivity.
In Chapter 2, I applied the causal modeling approach from Chapter 1 to seven additional study sites. I identified a general trend that south-west facing aspects pose greater resistance to dispersal than north and east-facing aspects, suggesting that exposure to high temperatures may limit dispersal in American pikas. I also found that amount and configuration of habitat (i.e., rocky substrate) influenced the degree to which other
landscape variables impact dispersal. I then applied the models of landscape resistance to investigate habitat patch connectivity within each site using a graph theoretic approach. This allowed me to assess overall patch connectivity, as well as identify specific patches and areas within study sites that are particularly important for maintaining functional connectivity, or in contrast, at risk of becoming isolated.
Functional connectivity is an important component when assessing population vulnerability because it describes the ability of genetic material to flow through the landscape. A loss of gene flow can lead to isolation of population segments, reduction of effective population size, loss of genetic diversity and subsequent erosion of evolutionary potential, and ultimately metapopulation collapse. However, other factors including habitat area and quality, affect effective population size and population stability. Genetic diversity is the material upon which evolution acts. In the face of rapid environmental change, species must either shift their range in order to track their ecological niche, or adapt in situ to the new environment. The former strategy requires sufficient available habitat as well as dispersal ability, which is unlikely for American pikas. Genetic diversity, therefore, is an important indicator of long-term population viability. In Chapter 3, I quantified genetic diversity within thirteen study sites and investigated the relationship between genetic diversity and environmental variables related to climate as well as habitat configuration and quality. As expected, habitat area was important at both the local and regional scales. Temperature was also a significant predictor of genetic diversity, with hotter sites having lower genetic diversity. However, and somewhat surprisingly, precipitation was a better predictor of genetic diversity than temperature, with sites receiving moderate levels of precipitation having higher genetic diversity. I
also compared population differentiation among study sites and identified sites that are particularly distinct. In the process of the above analyses, I also identified a previously undescribed contact zone between the northern and southern Rocky Mountain genetic lineages, recognized as separate subspecies, within Rocky Mountain National Park.
Climate change is predicted to affect habitat quality and landscape permeability, both of which have detectable genetic consequences. In Chapter 4, I quantified genetic diversity and structure within two study sites, Yosemite and Lassen Volcanic National Parks, from two sampling periods separated by approximately a century in time. I extracted DNA from historic study skins housed in the University of California Berkeley Museum of Vertebrate Zoology (MVZ) and compared multilocus microsatellite genotypes to those generated from modern fecal and tissue samples. Additionally, the historic survey data, including specimens, field journals, and other artifacts, were used in a series of systematic resurveys and subsequent species distribution modelling by other researchers. The nature of national parks as protected areas affords the opportunity to isolate the effects of climate change to a greater degree than other areas that have also experienced significant land use changes. I found no evidence for a change in genetic diversity within these two sites, consistent with observations from other studies that occupancy has remained relatively stable. I did find some evidence suggesting increasing population differentiation, potentially as a result of eroding landscape permeability. These results provide an important baseline for comparison with other sites, as well as reference points for future genetic monitoring of these populations.
In Chapter 5, I provide general synthesis and conclusions, as well as considerations for future research.