4 DISCUSSION

4.1 Genetic diversity in different parts of Eurasia

In this study, we reported both nuclear and mitochondrial genetic diversity of golden eagles for the entire Eurasia with newly sampled regions, such as Russia and Central Asia. Our analyses of nuclear microsatellites and mitochondrial sequences revealed a relatively high level of genetic variation in the Eurasian golden eagle population, being the highest in Central Asia and Caucasus, and the lowest in Northern Europe.
Mitochondrial genetic diversity in Central Asia and Caucasus (h = 0.82, π = 0.018) was similar to the former findings on golden eagles from mainland Asia (h = 0.79–0.93, π = 0.009–0.012; Nebel et al. , 2015, 2023). Among the previous studies, nuclear genetic diversity was reported only for the Mongol-Altai region, where a slightly higher observed heterozygosity (HO = 0.58) but lower expected heterozygosity (HE = 0.59) and allelic richness (AR = 4.07) were found compared to Central Asia and Caucasus (HO = 0.51, HE = 0.66, AR = 4.98). Notably, our samples from this region dated from 1898 to 1950 (N = 14), with four individuals of unknown year, while Nebel et al. (2023) analyzed the contemporary population. Therefore, despite the small sample size, the observed differences in nuclear genetic diversity may suggest either temporal changes or small-scale genetic variations within mainland Asia. The high genetic diversity in Central Asia detected in our study and in the previous studies, aligns with expectations for areas near past glacial refugia (Hewitt, 2000), the central-marginal hypothesis (CMH, Eckert et al. , 2008), and the latitudinal genetic diversity gradient hypothesis (Smith et al. , 2017; Fonseca et al. , 2023).
On the other hand, mitochondrial diversity was the lowest and nuclear diversity also low in Northern Europe. While peripheral populations commonly have lower genetic diversity compared to populations at the core of the distribution (Eckert et al. , 2008), the observed heterozygosity in Northern Europe (HO = 0.50) was even lower than in previously studied north European continental populations, including Finnish (HO = 0.57; Kylmänen et al. , 2023), Norwegian (HO = 0.56; Nebel et al. , 2023), and Finnish-Estonian (HO = 0.62; Nebel et al. , 2023) populations, with the exception of Scotland (HO = 0.46; Ogden et al. , 2015). Since most of our samples from Northern Europe were from the Northwestern federal district in Russia, our results imply that this area has especially low genetic diversity compared to other northern European regions.
Central and Eastern Europe exhibited similar levels of observed heterozygosity but higher expected heterozygosity (HO = 0.47, HE = 0.61) compared to previously studied golden eagles in the Slovakian population (HO = 0.44, HE = 0.49; Bielikova et al. , 2010) and in the Alps and Mediterranean region (HO = 0.51, HE = 0.55; Nebel et al. , 2023). Additionally, we noted higher mitochondrial genetic diversity in this group compared to the Alpine and Mediterranean region (h = 0.69, π = 0.008; Nebel et al. , 2023), regardless of whether the analyses included only our samples (h = 0.75, π = 0.014) or also GenBank sequences (h = 0.76, π = 0.016). Central and Eastern Europe group contained samples mainly originating from European Russia; thus, the observed high genetic diversity may point to that European Russia harbors a significant reservoir of genetic diversity among European golden eagles.

4.2 Mediterranean and Holarctic groups: genetic diversity and demographic history

By including samples from previously unexplored regions of the golden eagle distribution, we were able to better visualize the spatial distribution of the two mitochondrial lineages and compare it with the findings from nuclear markers. The Holarctic lineage was more widespread, with nearly twice as many golden eagles carrying these haplotypes (N = 284) compared to the Mediterranean lineage (N = 150). Previous studies of large raptors found that range size and historical population size were strong determinants of current genetic diversity (Väli et al. , 2019). Here we discovered that the Holarctic group, occupying a larger geographical range, also had higher mitochondrial genetic diversity than the Mediterranean group. However, the latter exhibited higher allelic and private allelic richness, while other estimates of nuclear genetic diversity were comparable. The Mediterranean group also showed consistent signs of demographic expansion, which could have contributed to an increase in nuclear genetic diversity, whilst almost all demographic analyses pointed to a stable population size of the Holarctic group. The inclusion of samples from previously unstudied areas and temporal periods resulted in slight shifts of the estimated mitochondrial genetic diversity compared to the earlier reports by Nebel et al. (2015); we observed a slight reduction in haplotype and nucleotide diversities in the Holarctic group (h = 0.61 and 0.75, π = 0.0039 and 0.0041, in this study and in Nebelet al. (2015), respectively), and a slight increase in these parameters in the Mediterranean group (h = 0.60 and 0.58, π = 0.0028 and 0.0020).
Our finding of Mediterranean haplotypes in both Central and Eastern Europe and Central Asia and Caucasus suggests that the Mediterranean lineage is spread more eastwards than thought before, where it now coexists with the Holarctic lineage, resulting in high genetic diversity in this region. The Mediterranean lineage likely survived in a glacial refugium around the Mediterranean region, as previously suggested by Nebel et al. (2015), but the location of a refugium for the Holarctic lineage remains uncertain; perhaps it was somewhere in central-eastern Asia. The Mongolian Plateau and the Altai-Saiyan Mountains have been suggested as glacial refugia for several plant and mammal species (e.g., Hais et al. , 2015; Lv et al. , 2016; McLean et al. , 2018), making them plausible candidates also for the golden eagle, especially in the light of the recently discovered golden eagle’s genetic diversity hotspot in the Mongol-Altai region (Nebel et al. , 2023).

4.3 The north-south genetic gradient in Eurasian golden eagles

We discovered both a north-south genetic gradient and genetic differentiation among the geographical groups in Eurasian golden eagles with both mitochondrial and microsatellite analyses. While the two mitochondrial lineages were identified and comprehensively discussed by Nebel et al. (2015), no evidence of their division was associated with nuclear markers. In addition, we highlighted the genetic uniqueness of the Central Asia and Caucasus and the Northern Europe groups in Eurasian golden eagles.
The latitudinal genetic gradient can originate due to several factors, including climatic and environmental instability in the north (Eckertet al. , 2008; Smith et al. , 2017), differences in migratory flyways (Monti et al. , 2018), postglacial colonization history (Thörn et al. , 2021), and a combined influence of the Quaternary climatic changes (Fonseca et al. , 2023). For example, a large-scale study of ospreys (Pandion haliaeetus ), identified two genetic clusters in Eurasia: Mediterranean and Eurasian, which were attributed to different migratory flyways (Monti et al. , 2018). Interestingly, the genetic clustering of ospreys was geographically similar to the genetic clustering observed in Eurasian golden eagles. However, the migratory flyway theory is not applicable to golden eagles because Eurasian golden eagles are generally non-migratory (Watson, 2010). On the other hand, the genetic differentiation between the north and the south as a result of post-glacial colonization history is supported by our findings of genetic diversity and demographic history (see above). Furthermore, golden eagles occupy a variety of habitats, implying diverse dietary and nesting adaptations. Their distribution in southern Eurasia covers Mediterranean-rich habitats, mountains, and steppes, while in northern Eurasia, they are predominantly found in mixed forests and taiga (Watson, 2010). This distinct ecological variation may be a potential underlying reason for the observed genetic differentiation.
However, although this gradient resulting from mixing of two divergent lineages seems to exist, we did not find evidence for isolation by distance (IBD) among the Eurasian golden eagles. Lack of an IBD pattern may be explained by the high dispersal capacity of golden eagles, especially of adolescent birds, documented in multiple studies (e.g., Nygård et al. , 2016; Poessel et al. , 2022). High dispersal potential leading to high gene flow was also mentioned as a reason for a lack of the IBD pattern in British golden eagles (Bourke et al. , 2010) and in some large-scale studies of other philopatric raptors, such as the Eurasian kestrel (Falco tinnunculus ; Alcaide et al. , 2009) and the greater spotted eagle (Clanga clanga ; Väliet al. , 2019).

4.4 Temporal genetic variation and demographic history

The population bottleneck of the 19th–20th centuries has left genetic signs in Palearctic golden eagles. First, we noticed a decrease in the number of haplotypes. While it is possible that the absence of 14 haplotypes in the post-bottleneck period compared to the bottleneck period was due to incomplete sampling, the rarefaction-extrapolation analyses indicated that the Bottleneck group had a significantly higher number of haplotypes compared to the Post-bottleneck group. Second, we observed a decrease in allelic and private allelic richness; a signal of a population bottleneck, as rare alleles are lost at a faster rate than heterozygosity is decreased (Allendorf, 1986). Third, the Bayesian Skyline Plot (BSP) analyses showed a reduction in the effective female population size (Nef) in the Eurasian population starting around 1975, aligning with known population declines in many populations across Eurasia (e.g., Bielikova et al. , 2010; Nebelet al. , 2015; Ollila, 2019; Starikov, 2020). Finally, when comparing temporal variation between the Mediterranean and Holarctic lineages, we found that in both lineages, genetic diversity was higher prior to the recent population bottleneck.
Despite the population bottleneck, golden eagles have nevertheless retained relatively high levels of genetic diversity. Factors such as long generation time, admixed origin of populations, and large distribution range contribute to high genetic diversity (Avise, 2000; Hailer et al. , 2006; Väli et al. , 2019). Therefore, as golden eagles are long-lived (Watson, 2010), occupy vast geographical areas (BirdLife International & Handbook of the Birds of the World, 2022), and likely originate from several glacial refugia (Nebel et al. , 2015; see above), these factors have undoubtedly played an important role in maintaining their genetic diversity. Similarly, high genetic diversity has been observed in other Eurasian raptors, such as the white-tailed eagle (Hailer et al. , 2007), the cinerous vulture (Aegypius monachus , Poulakakis et al. 2008), the crowned solitary eagle (Buteogallus coronatus ; Canal et al. , 2017), and the greater spotted eagle (Väli et al. , 2019).
While the long generation time may have buffered the effects of the 19th–20th-century bottleneck, it is also possible that the recent population decline was not severe enough to cause significant reductions in golden eagle’s genetic diversity (Bourke et al. , 2010). Noteworthy, the bottleneck of the 19th–20th centuries might have occurred at various time points in different parts of Eurasia, distorting the detection of temporal genetic diversity changes on a large-scale. Unfortunately, no studies on long term population trends in golden eagles have been conducted in Russia or elsewhere in continental Asia. For example, the only available records in central Yakutia state that the species was commonly nesting until the mid-1950s but became rare and disappeared from some areas in 1970–1980s, and only during the last 15–20 years the population has started to grow (Isaev et al. , 2021). Similarly, no golden eagle nests were found in Dauriya (east of the Lake Baikal, Zabaikalskiy krai) for the period from 1950s until 1990s (Karyakin & Nikolenko, 2012). On the other hand, nesting in the upper parts of the Don River basin (European Russia) was questioned already in the beginning of the 20th century, but the encounters became more frequent since the mid-1960s (Semago, 2006). Similarly, in Kazakh uplands (Kazakhstan), golden eagles were only reported since 1960s (Starikov, 2020).
When interpreting the population history of golden eagles, it is, too, essential to consider the significant impact of major glaciations on the distribution and dynamics of species (Hewitt, 2004). The glacial periods have had a huge impact on species’ ranges through dispersal, contractions, and even extinctions (Hewitt, 2000). In our study, we detected signals of population expansion of golden eagles in Northern Europe, a region that was long covered by the Scandinavian Ice Sheet. During the last glacial period, the Scandinavian Ice Sheet was the largest component of the Eurasian ice sheet complex, and it covered Fennoscandia and North-Western Russia repeatedly (Hughes et al. , 2016). Upon the end of the Last Glacial Maximum (26.5–19 ka; Clarket al. , 2009), species began to re-colonize new regions as the ice retreated (e.g., Ersmark et al. , 2019; Behzadi et al. , 2022). Although the precise routes of recolonization of Northern Europe remain uncertain, one possible direction could have been from the east of Eurasia, due to the prevalence of Holarctic haplotypes over the Mediterranean ones in this region.