bioRxiv preprint doi: https://doi.org/10.1101/2021.01.12.426363; this version posted January 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1 Glacial connectivity and current population fragmentation in 2 sky-islands explain the contemporary distribution of genomic 3 variation in two narrow-endemic montane grasshoppers from a 4 biodiversity hotspot 1 1 5 Vanina Tonzo and Joaquín Ortego 6 7 1 Department of Integrative Ecology, Estación Biológica de Doñana (EBD-CSIC); Avda. 8 Américo Vespucio 26 – 41092; Seville, Spain 9 10 Author for correspondence: 11 Vanina Tonzo 12 Estación Biológica de Doñana, EBD-CSIC, 13 Avda. Américo Vespucio 26, E-41092 Seville, Spain 14 E-mail: [email protected] 15 Phone: +34 954 232 340 16 17 18 19 20 21 22 23 24 bioRxiv preprint doi: https://doi.org/10.1101/2021.01.12.426363; this version posted January 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 25 Abstract 26 27 Aim: Cold-adapted biotas from mid-latitudes often show small population sizes, harbor 28 low levels of local genetic diversity, and are highly vulnerable to extinction due to 29 ongoing climate warming and the progressive shrink of montane and alpine 30 ecosystems. In this study, we use a suite of analytical approaches to infer the 31 demographic processes that have shaped contemporary patterns of genomic variation 32 in Omocestus bolivari and O. femoralis, two narrow-endemic and red-listed Iberian 33 grasshoppers forming highly fragmented populations in the sky island archipelago of 34 the Baetic System. 35 Location: Southeastern Iberia. 36 Methods: We quantified genomic variation in the two focal taxa and coupled 37 ecological niche models and a spatiotemporally explicit simulation approach based on 38 coalescent theory to determine the relative statistical support of a suite of competing 39 demographic scenarios representing contemporary population isolation (i.e., a 40 predominant role of genetic drift) vs. historical connectivity and post-glacial 41 colonization of sky islands (i.e., pulses of gene flow and genetic drift linked to 42 Pleistocene glacial cycles). 43 Results: Inference of spatial patterns of genetic structure, environmental niche 44 modelling, and statistical evaluation of alternative species-specific demographic 45 models within an Approximate Bayesian Computation framework collectively 46 supported genetic admixture during glacial periods and postglacial colonization of sky 47 islands, rather than long-term population isolation, as the scenario best explaining the 48 current distribution of genomic variation in the two focal taxa. Moreover, our analyses 49 revealed that isolation in sky islands have also led to extraordinary genetic 50 fragmentation and contributed to reduce local levels of genetic diversity. 51 Main conclusions: This study exemplifies the potential of integrating genomic and 52 environmental niche modelling data across biological and spatial replicates to 53 determine whether organisms with similar habitat requirements have experienced 54 concerted/idiosyncratic responses to Quaternary climatic oscillations, which can 55 ultimately help to reach more general conclusions about the vulnerability of mountain 56 biodiversity hotspots to ongoing climate warming. bioRxiv preprint doi: https://doi.org/10.1101/2021.01.12.426363; this version posted January 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 57 58 KEYWORDS: Approximate Bayesian Computation, ddRAD-seq, demographic history, 59 environmental niche modelling, landscape genetics, Pleistocene glaciations 60 bioRxiv preprint doi: https://doi.org/10.1101/2021.01.12.426363; this version posted January 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 61 1. INTRODUCTION 62 63 Distributional shifts in response to Quaternary climatic oscillations had a dramatic 64 impact on biogeographical patterns of species diversity, abundance and local 65 endemism (Hewitt, 1996; Sandel et al., 2011). These cycles, in particular the high- 66 amplitude climatic changes characterizing the Middle-Late Pleistocene (Jouzel et al., 67 2007), have also shaped the distribution and spatial patterns of genetic variation in 68 many organisms (Hewitt, 2000). However, multiple studies have documented 69 considerable heterogeneity across regions and taxa in the demographic consequences 70 of Pleistocene glacial cycles (Hewitt, 2000). On the one hand, the impact of Pleistocene 71 glaciations strongly depended on latitude and regional topography, with extinction- 72 recolonization dynamics at higher latitudes (i.e., “southern richness to northern purity” 73 paradigm; Hewitt, 1996) and elevational shifts and more complex processes of 74 population fragmentation and connectivity at lower latitudes such as the tropics or 75 temperate regions (e.g., “refugia within refugia” concept; Gómez & Lund, 2006). On 76 the other hand, the way organisms respond to climate changes strongly depend on 77 species-specific niche requirements and life-history traits, which define 78 favorable/unfavorable climatic periods (glacials or interglacials; Bennett & Provan, 79 2008) and their capacity to deal with population fragmentation (e.g., micro-habitat 80 preferences, dispersal capacity, etc.; e.g., Massatti & Knowles, 2016; Papadopoulou & 81 Knowles, 2016). These aspects determined the location and extension of Pleistocene 82 refugia, which have played a predominant role on species´ persistence during 83 unfavorable climatic periods and acted as source populations from which species 84 expanded their ranges at the onset of more favorable conditions (Bennett & Provan, 85 2008). Temperate species currently inhabiting low elevation areas generally restricted 86 their distributions to southern refugia during glacial phases (i.e., glacial refugia) and 87 expanded during interglacial periods whereas cold-adapted species, nowadays 88 presenting fragmented populations at high elevations/latitudes (i.e., interglacial 89 refugia), had much more widespread distributions in glacial stages (Hewitt, 2000). 90 Mid- and low-latitude alpine and montane taxa represent small-scale replicates 91 of cold-adapted species living at high latitudes (Hewitt, 1996; 2000). Sky islands are a 92 paradigmatic example of interglacial refugia in which alpine and montane organisms bioRxiv preprint doi: https://doi.org/10.1101/2021.01.12.426363; this version posted January 13, 2021. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 93 currently form highly isolated populations in mountain ranges separated from each 94 other by intervening valleys with unsuitable environmental conditions (e.g., DeChaine 95 & Martin, 2005; Knowles & Massatti, 2017; McCormack et al., 2008; Mouret et al., 96 2011). Changing climatic conditions cause suitable habitats to expand or shrink along 97 elevational gradients, leading to alternating periods of population isolation and 98 connectivity (DeChaine & Martin, 2005; Knowles, 2000). Periods of isolation on 99 mountain tops are expected to lead to demographic bottlenecks, genetic drift and 100 divergence among populations, while periods of connectivity allow for dispersal and 101 gene flow among formerly isolated populations (DeChaine & Martin, 2005). Species 102 restricted to sky islands often show unique patterns of population genetic structure 103 resulted from climate-induced distributional shifts (Mouret et al., 2011), processes 104 that in some cases have led to lineage diversification and speciation (Knowles, 2000; 105 Knowles & Massatti, 2017). Therefore, montane habitats are often important hotspots 106 of intraspecific genetic diversity, species richness and local endemism, providing an 107 excellent study system for understanding how climatic changes associated with 108 Pleistocene glacial cycles impacted species distributions, demography and genetic 109 diversification (Mairal et al., 2017). Their study takes particular relevance if we 110 consider that species restricted to sky-islands often show small population sizes, 111 harbor reduced levels of local genetic diversity and are highly vulnerable to extinction 112 due to ongoing climate warming and the progressive shrink of alpine ecosystems 113 (Rubidge et al., 2012). 114 Southeastern Iberia constitutes an important biodiversity hotspot due to the 115 interplay among a vibrating geological history (Meulenkamp & Sissingh, 2003), an 116 extraordinarily complex topography (Braga et al., 2003), and a limited impact of 117 Quaternary glaciations (Hughes & Woodward, 2017). The presence of vast areas free 118 of permanent ice during the coldest stages of the Pleistocene made this region an 119 important glacial refugium for warm-temperate taxa (Gómez & Lunt, 2007; Hewitt, 120 2011). At the same time, the mountain ranges of the region (Baetic system; > 3,000 m) 121 constitute important interglacial refugia for several cold-adapted organisms that 122 currently persist in severely fragmented populations at high elevation patches of 123 suitable habitat and contribute