INTRODUCTION
Population differentiation and speciation is a complex process that depends on species-specific factors as well as extrinsic landscape processes. An ongoing question in evolutionary biology is whether taxa differentiate in communities in response to extrinsic forces—for example, geographic or climatic change—or in species-specific manners based on their intrinsic characteristics such as physiological thermal tolerance or dispersal ability. In speciation scenarios dominated by geologic or climatic change, entire communities may differentiate in pulses synchronous with the changes in the Earth’s template (e.g., Barber & Klicka, 2010; Musher et al., 2019; O’Connell et al., 2018). In contrast, when communities have species that vary in dispersal ability, niche breadth, or population history, diversification may be asynchronous across taxa based on those species’ intrinsic characteristics (e.g., Naka & Brumfield, 2018; Oswald et al., 2017; Papadopoulou et al., 2009).
In well-studied geographic regions, it has become clear that with increasing taxonomic scope, a combination of both synchronous and asynchronous diversification arises that depends on the taxonomic and temporal breadth of study (Shafer et al., 2010; Smith et al., 2014). These results demonstrate the need for sampling multiple taxonomic groups, with a range of intrinsic characteristics and histories, to best understand the drivers of diversification in different geographic contexts. This work is arguably most needed in biodiversity hotspots, the regions of the world that are richest in diversity while simultaneously most at risk due to human activities (Myers et al., 2000; Zachos & Habel, 2011). Not only are these regions some of the richest in biodiversity on our planet, but undescribed biodiversity—especially species and genetic diversity—is often concentrated in these regions (Hamilton et al., 2010; Miraldo et al., 2016; Mora et al., 2011).
The Horn of Africa biodiversity hotspot is rich in species (Friis et al., 2001; Largen & Spawls, 2010; Yalden & Largen, 1992), landscape perturbation (Dessie & Kleman, 2007; Zeleke & Hurni, 2001), and elevational heterogeneity; a majority of the topographic complexity is found in Ethiopia, with elevations ranging from more than a hundred meters below sea level in the Danakil Depression to greater than 4500 meters above sea level in the Simien Mountains (Fig. 1). The Ethiopian Highlands are a largely continuous region of tropical high elevation habitats with major lowland biogeographic barriers including the Great Rift Valley (GRV) and the Blue Nile Valley (Fig. 1). Both the Blue Nile Valley and the GRV are part of the large East African rift system (Frisch, Meschede, & Blakey, 2010). These lowland biogeographic barriers have shaped geographic population structure in a variety of montane species, including mammals, frogs, and plants (Belay & Mori, 2006; Bryja et al., 2018; Evans et al., 2011; Freilich et al., 2016; Gottelli et al., 2004; Kebede et al., 2007; Manthey et al., 2017; Reyes-Velasco et al., 2018; Reyes‐Velasco et al., 2018; Silvestrini et al., 2007; P. J. Taylor et al., 2011).
Birds are often assumed to be good dispersers, making them a good focal taxon to identify whether the GRV is a significant biogeographic barrier for relatively highly dispersive species. Despite the general assumptions of bird dispersal ability due to flight, tropical birds—especially those that are non-migratory—may not always disperse long distances. For example, rivers have been shown to be long-term dispersal barriers in Amazonian birds (Naka & Brumfield, 2018). Additionally, isolated sky islands in other regions of the East African rift system have promoted diversification in some avian taxa, suggesting at least some montane birds have limited dispersal across lowland biogeographic barriers (Habel et al., 2015). These patterns suggest that even in birds, species-specific diversification patterns may exist due to intrinsic characteristics of each species such as dispersal ability.
What remains lacking are studies of comparative population structure in Ethiopian montane birds; to date, there have been no studies investigating phylogeographic or population genetic patterns in this diverse community. To help fill this gap, we used a comparative framework to study the effects of the GRV on population genetic differentiation in montane forest birds of the Ethiopian Highlands. We included six bird species in this study: Cossypha semirufa(Rüppell’s Robin-chat), Crithagra tristriata (Brown-rumped Seedeater), Melaenornis chocolatinus (Abyssinian Slaty Flycatcher), Sylvia galinieri (Abyssinian Catbird), Turdus abyssinicus (Abyssinian Thrush), and Zosterops poliogastrus(Ethiopian White-eye) (Table 1). We chose these species for several reasons, as they are all (1) highland forest species in Ethiopia, (2) relatively common where found locally, and (3) endemic to Ethiopia or the Horn of Africa region. These six bird species are associated with various types of forests and woodlands, including forest edge, and they can often be found in the same communities; indeed, we observed and captured all species for this study in the same general sampling areas (Fig. 1). However, some of these species have different habitat associations and minimum elevational affinities (Fig. 2). For example, the Abyssinian Catbird and Rüppell’s Robin-chat often preferJuniperus and Podocarpus forests, the Abyssinian Thrush and Brown-rumped Seedeater are occasionally found in highland scrub habitat, and the Abyssinian Slaty Flycatcher is often found near woodland streams (Clement, 2020; Collar, 2020; Collar & Robson, 2021; del Hoyo, Collar, & Kirwan, 2020; B. Taylor, 2020; van Balen et al., 2020). In addition to differential habitat preferences, the species studied here have different wing shapes as measured by the hand-wing index (HWI; Table 1) (Claramunt et al., 2011; Kipp, 1959). Because the HWI is positively correlated with dispersal ability in birds, we may expect species with higher HWI, such as the Abyssinian Thrush and Ethiopian White-eye, to have maintained relatively higher population connectivity relative to other species studied here.
We used genome resequencing data to estimate genomic diversity and differentiation, timing of diversification, and demographic histories for these six bird species on either side of the GRV. We then tested whether species-specific characteristics, including dispersal ability and demographic history, could explain the comparative patterns of population genomic diversity and differentiation.