Introduction

Clusters of alpine high mountain ecosystems that are isolated by different environmental conditions in the intervening lowlands can behave as authentic archipelagos. These sky islands contain unique diversity and provide natural laboratories to study evolution, speciation and the effects of habitat fragmentation (Hedberg, 1969; McCormack, Huang, & Knowles, 2009). However, their extent of isolation has changed with climatic fluctuations during the Quaternary, in particular in tropical regions. Colder conditions during glacial periods led to the expansion of high altitude alpine and montane forest ecosystems into lower elevations, thus facilitating population expansion and connectivity between populations previously isolated on mountain tops (Hewitt, 2000). These cycles and fluctuations in gene flow promoted differentiation processes that led to exceptionally high levels of endemism (Williams, Vivero, Spawls, Anteneh, & Ensermu, 2004). On the other hand, these ecosystems are especially sensitive to future climate change, and as a result of climatic warming their area is rapidly diminishing (Moritz & Agudo, 2013), a process which will increase the isolation and extinction risk of their associated flora and fauna. This process is expected to be particularly intense in the scarce alpine environments distributed at high elevations across the tropics. Tropical mountains are predicted to experience the highest levels of extinctions and changes to fauna assemblages under future climate change (Lawler et al., 2009). High altitude species living close to mountain tops may lose their entire range as their suitable climatic regime and its associated ecosystem disappears with increasing temperatures (Pimm, 2009; Williams, Jackson, & Kutzbach, 2007).
This problem is especially acute in under-studied and highly threatened areas like the Ethiopian Highlands, where accelerated land conversion and degradation is placing further pressures on biodiversity (Yalden, Largen, Kock, & Hillman, 1996). The Ethiopian Highlands are a vast extent of high ground rising up to 4620 masl with a low altitude limit of 1500 masl. They cover an area of 519,278 km2 in the otherwise arid zone of the Horn of Africa and include 73% of the Afroalpine biome (areas above 3200 masl) in Sub-Saharan Africa (Williams et al., 2004). The Great Rift Valley crosses the Ethiopian central highlands dividing them into two main blocks, north-east and south-west (Ebinger et al., 2000).
The Afroalpine biome shows one of the highest rate of endemism in the world due to a combination of their relative small areas, high isolation and climatic history (Reyes-Velasco, Manthey, Freilich, & Boissinot, 2018; S. D. Williams et al., 2004). These endemics are either of Palearctic (e.g. the Ethiopian ibex, Capra walie, the Ethiopian long-eared bat, Plecotus balensis ) or Afrotropical origin (e.g. geladas monkeys, Theropithecus gelada, the mountain nyala,Tragelaphus buxtoni ) (Harmsen, Spence, & Mahaney, 1991). Although the source of colonisers of this truly unique environment is probably more complex, including also long-distance vagrants like the endemic plover (Vanellus melanocephala ) (Kingdon, 1990). Rapid human population growth in Ethiopia over the past 150 years, from approximately 6.6 million in 1868 (Nyssen et al., 2014) to more than 100 million (UN, 2019), and the corresponding increase in pressure on natural environments and human encroachment into national parks and other protected areas has resulted in extensive habitat degradation (Kidane, Stahlmann, & Beierkuhnlein, 2012). This pressure has been implicated with biodiversity losses, from the decline of endemic frogs (Gower et al., 2013) to the decline of large flagship mammal species such as the endemic Ethiopian wolf, Canis simensis (Stephens, d’Sa, Sillero-Zubiri, & Leader-Williams, 2001). However, some signs of vegetation recovery have been observed since the early 21st century following the initiation of environmental recovery programs in the 1980s and the establishment of new protected areas closing up to 15% of the country to livestock grazing (Nyssen et al., 2014).
Recently, a new species of the long-eared bat genus Plecotus was described in the Ethiopian Highlands above 3000 m, the Bale long-eared bat, Plecotus balensis , which is endemic to Ethiopia and possibly Eritrea (Juste et al., 2004; Kruskop & Laverenchenko, 2000), and is known by only a few specimens from just seven localities (Benda, Kiefer, Hanák, & Veith, 2004). The genus Plecotus is widely distributed across the Palearctic, and is characterised by high cryptic diversification and specific altitudinal and climatic associations (Spitzenberger, Strelkov, Winkler, & Haring, 2006). Due to their limited long-distance dispersal abilities and habitat specialisations,Plecotus bats are particularly sensitive to habitat loss (Razgour et al., 2014) and the effects of future climate change (Razgour et al., 2013). Bats can be good bio-indicators of environmental change (Jones, Jacobs, Kunz, Willig, & Racey, 2009) and have been used to test biogeographical hypotheses on the role of topographic barriers and past climate change in delimiting species distribution and shaping genetic composition (e.g. Flanders et al., 2009; Razgour et al., 2013).
We use this recently described high altitude bat to study how sky islands and historic versus current patterns of environmental heterogeneity shape the distribution and genetic diversity of their unique fauna. We project ecological niche models (ENMs) across temporal scales to track the effect of past climate changes on range suitability for P. balensis and predict impacts of future changes. We integrate approximate Bayesian computation inference of demographic history with landscape genetics analysis and markers with different mutation rates to disentangle the impact of historic vicariance events versus recent anthropogenic habitat loss and degradation on population size, genetic diversity and genetic connectivity of sky island biodiversity. Our study highlights threats to tropical montane biodiversity due to the combined effects of multiple stressors, being squeezed into higher altitudes due to climate change while losing genetic diversity and suffering population declines due to anthropogenic land-use change.