Abstract
The total land-mass of urban areas is rapidly increasing, forcing wildlife species to respond either by reductions in their distribution and abundance or acclimation to urban and otherwise modified habitats. Coyotes (Canis latrans) are found in the heart of some major cities within their range but relegated to the fringes of others. I examined two urban coyote populations. One population, San Francisco, is relatively new, having become established in patches of habitat within the city only within the last several years, providing a window into the process of colonization. The other, Sacramento, is linear and provides a natural connection between the valley and the foothills. The goal of my research is to investigate genetic connectivity between these two urban coyote populations and surrounding rural populations and to assess the possibility of a single founder pair origin for the San Francisco population. Recent research in provides evidence for natal habitat preference in California coyotes. Using genetic assignment methods, at least four genetic subdivisions, corresponding to bioregions, were found: Great Valley, Cascades/Sierra Nevada/Modoc (hereafter referred to as Sierra Nevada), Northwestern, and Central Western. Coyote DNA from scat and some opportunistically collected blood and tissue samples from our two study sites were genotyped at 14 microsatellite loci. Those genotypes were ran in STRUCTURE along with reference genotypes from previously establish clusters in order to determine assignment. The American River Parkway results indicate both a high level of assignment to the Sierra Nevada cluster and no assignment to the Great Valley cluster. The elevational change along the American River Parkway did not appear to have an effect on the results. When the data were analyzed in STUCTURE with K = 3, most coyotes assigned to a cluster distinct from the Great Valley or Sierra Nevada, which indicates high genetic similarity among the American River Parkway coyotes relative to the surrounding populations. The San Francisco population, having allele counts of greater than four for half of our loci, does not support the single founder pair hypothesis but these counts could have been skewed due to genotyping error. The significant difference between the number of loci with allele counts greater than four in the sample population and the Northwestern reference population provides support for a small founding population. We did not see a strong clustering of the San Francisco genotypes with either the Northwestern or Central Western cluster. The single individual that did assign to the Central Western cluster was from the most southern park sampled, indicating that there is some immigration from the Central Western population. The high relatedness among San Francisco individuals does not support a hypothesis of high immigration from either/or both the Northwestern or Central Western populations. The Central Western displayed high levels of within cluster relatedness, similar to within San Francisco, but Northwestern displayed low levels of relatedness within its cluster. When assessing relatedness among populations we found some error due to probability. This error is evident in relationship levels between the Northwestern cluster and the Central Western cluster, which are geographically isolated from each other. When assessing relationships by park, high relatedness was seen within the Presidio. Among the different parks, the pairs of parks that showed the three highest number of relationships were between Glen Park and each of the other three parks. This shows both evidence of the willingness of coyotes to traverse urban landscapes and the inability of small parks to sustain many individuals. I used noninvasive genetic sampling, which has inherent limitations to for genetic analysis because it yields low quality and quantity DNA. For this reason, genotypic error due to allelic dropout is common. Lack of detection can produce null alleles that may result in false homozygotes. In addition, season/weather, age of feces at collection, and time between preservation and DNA extraction all greatly affect PCR amplification success. Both the American River Parkway and the San Francisco study sites offer insight from a wildlife management perspective. This information can be used, either for the development of corridors and wildlife areas to promote urban wildlife and/or for strategies in managing the population size of current urban coyote populations. Both strategies meet the purpose of limiting the human-wildlife interface within urban areas and thus limiting the opportunity for conflict.