The Northern Goshawk (Accipiter gentilis; hereafter goshawk), is a large diurnal forest raptor, and is a raptor of management interest across the American West (Reynolds et al. 1992; Kennedy 2003). When considering management objectives for any species, understanding dispersal capabilities is of critical importance. Wiens et al. (2006) found that local natal dispersal distance of goshawks in their study area of the Kaibab Plateau in Arizona was a median of 15km. Distance of breeding dispersal in adults within the northern Great Basin was 2.1 to 5.8km (Bechard et al. 2006). Wiens et al. (2006) recorded juvenile dispersal distances as far as ~400 km beyond their study area.
Our study’s objective was to better understand gene flow and dispersal in goshawks in the Northern Great Basin of western North Amercia, and to place the goshawks in our study area within the biogeographical context set forth by Bayard de Volo et al. (2013). To do this we focused on a small and relatively isolated population of goshawks in south-central Idaho and northern Utah (Miller et al. 2014). We sampled 36 goshawk nests in 2012, 2014, and 2015 to evaluate the genetic variation of mitochondrial DNA sequences (mtDNA haplotypes) and dispersal using microsatellite markers. In addition to our focused study in south-central Idaho, we collected a smaller sample of tissues from goshawks across Idaho for comparison.
We extracted DNA from blood and molted feathers of distinct individuals. If adult females were sampled, their nestlings were not, and vice versa; paternal goshawks were an exception. We amplified domain I of the mtDNA control region following the methods of Baker and Marshall (1997) and Bayard de Volo et al. (2013). Successful PCR amplications were purified using Exo-SAPit (Affymetrix, Cleveland, Ohio) and sequenced by Genewiz (Genewiz, Plainfield, New Jersey). Sequences were edited using Phy-DE (Müller et al. 2005), and compared to published haplotypes (Bayard de Volo et al. 2013; Sonsthagen et al. 2004).
In the more widespread sampling, we found the H-haplotype (accession number AY699835.1), which had not been previously described in the Rocky Mountain region. Here we discuss the implications of finding this mitochondrial haplotype in terms of dispersal and gene flow for goshawks in the American West.
Southwest Refugium Specific Haplotype: H-Haplotype
Previous research indicates that following the most recent glacial period, goshawks expanded within western North America from two primary refugia (Bayard de Volo et al. 2013). One refugium was in the “Pacific” (Cascade and Sierra Nevada Mountains) and the other in the “Southwest” (Colorado Plateau and the Jemez Mountains, NM; Bayard de Volo et al. 2013). Bayard de Volo et al. (2013) found several haplotypes specific to the Southwest refugium. One of these haplotypes, the H-haplotype, was found in low densities and was restricted to the Colorado Plateau and Arizona Sky Island populations. Interestingly, we detected a single occurance of the H-haplotype in central Idaho, near the town of Obsidian.
There are a few possible explanations for the presence of the H-haplotype in the Idaho population. The first involves natal dispersal of individuals from surrounding populations. Bayard de Volo et al. (2013) suggested female mediated gene flow from the Arizona Sky Islands (with a relatively high frequency of the H-haplotype) to the Rocky Mountains may occur at a rate of zero individuals per generation (95% CI= 0-2). Goshawks in the Colorado Plateau region also have the H-haplotype, but at a much lower frequency than the Arizona Sky Islands (Bayard de Volo et al. 2013). From the Colorado Plateau, the rate of female mediated gene flow to the Rocky Mountains is four birds per generation (95% CI= 2-17; Bayard de Volo et al. 2013). Thus, the H-haplotype is more frequent in the Arizona Sky Islands, but the probability of a goshawk dispersing to Idaho from the Colorado Plateau versus from the Arizona Sky Islands is higher. Finding a goshawk of this haplotype supports the hypothesis that gene flow from the Colorado Plateau and Arizona Sky Islands may sometimes occur over large distances. Wiens et al. (2006) study of juvenile dispersal from the Kaibab Plateau of northern Arizona led them to hypothesize that the primary factors influencing natal dispersal were environmental conditions, competition for breeding opportunities, and avoidance of inbreeding. Such factors likely explain the appearance of the H-haplotype in Idaho.
A second explanation is that this haplotype is not endemic to only the Southwest and Colorado Plateau, but occurs regularly in the Rocky Mountains in such low frequencies that it was not detected by Bayard de Volo et al.’s (2013) sample size of 18 Rocky Mountain goshawks. A third explanation is that the H-Haplotype has two independent mutational histories involving two biogeographic regions. While this explanation is possible, Eo and DeWoody (2010) found that the rate of base pair substitution in birds is 2.56 x 10-9 per site per year. Eo and DeWoody (2010) also calculated the rate of base pair substitution in Accipitridae and found it to be slightly higher, around 4 x 10-9 substitutions per site per year. However, rates of mutation in control regions of birds can be much higher than in the genome as a whole (Ruokonen and Kvist 2010). This suggests that homoplasy could be much higher within the control region.
Investigating these alternative explanations requires additional intensive sampling – with additional molecular markers – focusing on the Rocky Mountain population as a whole. This would include sampling in Northern Utah, Idaho, Montana, the Pacific Northwest, and Canada. Since the Rocky Mountains encompass a large biogeographic area, increasing the sample size across this range will lead to a better understanding of changes in haplotype frequency on smaller scales, and increase the probability of detecting other haplotypes beyond the regions previously identified. Increased sampling of birds between the Rocky Mountains and the Southwest could help clarify current gene flow between the Southwest and Rocky Mountains. Identifying a pattern of gene flow between these areas could lead to a better understanding of current and past dispersal among goshawks throughout these regions.