Demographic and spatially explicit landscape genomic analyses in a tropical oak reveal the impacts of late Quaternary climate change on Andean montane forests

Abstract

The tropical Andes are one of the most important biodiversity hotspots on Earth, yet our understanding of how their biotas have responded to Quaternary climatic oscillations is extraordinarily limited and the alternative models proposed to explain their demographic dynamics have been seldom formally evaluated. Here, we test the hypothesis that the interplay between the spatial configuration of geographical barriers to dispersal and elevational displacements driven by Quaternary cooling–warming cycles has shaped the demographic trajectories of montane oak forests (Quercus humboldtii) from the Colombian Andes. Specifically, we integrate genomic data and environmental niche modelling at fine temporal resolution to test competing spatially explicit demographic and coalescent models, including scenarios considering (i) isotropic gene flow through the landscape, (ii) the hypothetical impact of contemporary barriers to dispersal (i.e., inter‐Andean valleys), and (iii) distributional shifts of montane oak forests from the Last Glacial Maximum to the present. Although our data revealed a marked genetic fragmentation of montane oak forests, statistical support for isolation‐with‐migration models indicates that geographically separated populations from the different Andean Cordilleras regularly exchange gene flow. Accordingly, spatiotemporally explicit demographic analyses supported a model of flickering connectivity, with scenarios considering isotropic gene flow or currently unsuitable habitats as persistent barriers to dispersal providing a comparatively worse fit to empirical genomic data. Overall, these results emphasize the role of landscape heterogeneity on shaping spatial patterns of genomic variation in montane oak forests, rejecting the hypothesis of genetic continuity and supporting a significant impact of Quaternary climatic oscillations on their demographic trajectories.