Forecasting how biodiversity will change in the future due to natural and anthropogenic impacts is a primary focus of both ecology and evolutionary biology. By understanding the historical processes influencing the current geographic distribution of biodiversity, we can determine the relative importance of different factors shaping biodiversity, now and in the future. The main question that drove this dissertation: What historical processes led to the current distribution of diversity across the landscape? One omnipresent influence on the geographic distribution of diversity at multiple levels– e.g., genes and species– is climate. Understanding the spatial distributions of intraspecific genetic diversity and the role of climate and climate refugia in evolutionary and ecological processes is important because it shapes species potential for persistence in the face of future climate change. My dissertation focuses on how populations have responded to past climate change, and how the historical distributions and past areas of climate refugia will influence future climate change responses, using mammals as the study system.
Studying population histories through time, we can uncover how different populations with similar genetic reservoirs respond differently to the same environmental stressor (climate). Determining how the distribution of intraspecific diversity of North American taxa was directly influenced by climate and landscape changes may illuminate broad-scale patterns of species’ responses to other climatic events, or more generally, to barriers impeding or constraining gene flow. My dissertation research utilized an interdisciplinary approach– next generation sequencing; GIS data; ecological modeling; bioinformatics– to understand how historical events have shaped the current distribution of genetic diversity within mammals. My aim was to study how mammal populations respond to climate change through time by determining the range dynamics and potential areas of refugia (Chapter 2 and 3). Understanding the spatial distributions of intraspecific genetic diversity and the role of climate refugia in the evolutionary and ecological processes of populations is important because it may determine their potential for persistence in the face of future climate change. My dissertation examined how two small mammals responded to the glacial cycles of the Pleistocene in North America. First, I determined Neotoma fuscipes has three historical populations in California—two northern and one southern population (Chapter 2). The major split between the northern and southern populations is older than 1.7 million years and occurred in the San Francisco Bay-Delta region, a historically significant region with high lineage diversification in mammals, amphibians, and reptiles. I detected multiple refugia within the species, including several origins of expansions and contractions (particularly with the northern populations). Second, I examined two western lineages within Peromyscus maniculatus and identified three main populations: 1) southern California; 2) a small population nested within the broader Pacific Northwest; 3) the Pacific Northwest through central California and across the Rocky Mountains (Chapter 3). These populations diverged within the last 160,000 years with very little migration, and evidence of recent population expansion. I found evidence for multiple areas of refugia including southeastern Alaska, a known refugia for several mammal species.
In order to connect patterns and processes observed in the past with projections of change for the future, I pair my focus on generating empirical genetic & genomic data with different modeling types. Specifically, ecological niche models (ENM) are powerful tools for approximating the abiotically suitable area of a species by comparing environmental conditions at localities where the species occurs with the overall conditions available in the study region. Many ENM studies struggle with small sample sizes, but modeling widely distributed and well-sampled cosmopolitan species raises computational issues as well (Chapter 1). I thus determined the number of localities needed to model the distribution of a cosmopolitan species (Peromyscus maniculatus) with many occurrence records. I discovered when modeling species with a large number of occurrence records, it may not be necessary to use all localities for ENMs and could potentially affect model performance negatively.
