The influence of landscape, climate and history on spatial genetic patterns in keystone plants (Azorella) on sub-Antarctic islands.

Abstract

The distribution of genetic variation in species is governed by factors that act differently across spatial scales. To tease apart the contribution of different processes, especially at intermediate spatial scales, it is useful to study simple ecosystems such as those on sub-Antarctic oceanic islands. In this study, we characterize spatial genetic patterns of two keystone plant species, Azorella selago on sub-Antarctic Marion Island and Azorella macquariensis on sub-Antarctic Macquarie Island. Although both islands experience a similar climate and have a similar vegetation structure, they differ significantly in topography and geological history. We genotyped six microsatellites for 1,149 individuals from 123 sites across Marion Island and 372 individuals from 42 sites across Macquarie Island. We tested for spatial patterns in genetic diversity, including correlation with elevation and vegetation type, and clines in different directional bearings. We also examined genetic differentiation within islands, isolation-by-distance with and without accounting for direction, and signals of demographic change. Marion Island was found to have a distinct northwest-southeast divide, with lower genetic diversity and more sites with a signal of population expansion in the northwest. We attribute this to asymmetric seed dispersal by the dominant northwesterly winds, and to population persistence in a southwestern refugium during the Last Glacial Maximum. No apparent spatial pattern, but greater genetic diversity and differentiation between sites, was found on Macquarie Island, which may be due to the narrow length of the island in the direction of the dominant winds and longer population persistence permitted by the lack of extensive glaciation on the island. Together, our results clearly illustrate the implications of island shape and geography, and the importance of direction-dependent drivers, in shaping spatial genetic structure.