Changes in the distribution of cloud forests during the last glacial predict the patterns of genetic diversity and demographic history of the tree fern Alsophila firma (Cyatheaceae)

AimWe investigated changes in distribution of cloud forests during the last 130 kyr, and tested whether these changes explain the spatial patterns of genetic diversity of the tree fern Alsophila firma (Cyatheaceae), a species restricted to this habitat. LocationMexican cloud forests. MethodsWe sampled 204 individuals from 16 localities. Genetic data consisted of DNA sequences for five chloroplast microsatellites and one nuclear gene. We used distribution modelling to predict the historical distribution of cloud forests during the last glacial period, using two palaeoclimate models: the Model for Interdisciplinary Research on Climate (MIROC) and the Community Climate System Model (CCSM). We tested the correlation between temporal cloud forest stability and genetic diversity and used an approximate Bayesian computation (ABC) framework to test two plausible demographic scenarios. ResultsThe range fluctuations observed for cloud forests during the last 130 kyr are key factors affecting the distribution of genetic variation in A. firma. Increased genetic diversity in areas with high temporal environmental suitability is probably the result of increased population sizes and higher interpopulation connectivity. In accordance with the expansion of cloud forests predicted by CCSM, the genetic data supported the scenario of a population expansion occurring c.110 ka, followed by population divergence c.20 ka. However, population dynamics involving expansion of suitable microclimates could reconcile the stability of cloud forests predicted by MIROC and the observed genetic patterns. Main conclusionsThe predicted changes in the distribution of cloud forests were congruent with the population genetics of A. firma. However, the choice of palaeoclimate model has a substantial impact on the inferences drawn from the observed genetic and demographic patterns. The use of alternative palaeoclimate hypotheses and biome modelling can provide a common analytical framework for evaluating the historical cohesiveness of forest communities.