Genetic consequences of landscape features in two rear edge, highly fragmented metapopulations of a mediterranean conifer.

Context: Habitat fragmentation is expected to erode genetic diversity, which instead needs to be preserved for promoting species adaptation to a changing climate. As this expectation has found mixed support in forest trees, consistent results on the genetic consequences of fragmentation requires adequately replicated experimental designs, as well as an explicit assessment of which landscape features, if any, could mitigate its detrimental effects. Objective: Evaluating the role of several landscape attributes in buffering the detrimental effects of fragmentation in two metapopulations of silver fir. Methods: We genotyped 904 silver fir (Abies alba Mill.) trees from 18 local populations forming two metapopulations comparable for size and extension in the Apennines, a Mediterranean mountain range. We identified the signatures left by the fragmentation process on the genetic features of silver fir local populations. After removing potentially confounding effects due to different evolutionary histories, we used a multivariate approach for testing the relative effect of demographic, geographic, environmental and topographic factors on genetic features of both metapopulations. Results: We found comparable signals of the habitat fragmentation impact on the genetic diversity and structure of both investigated metapopulations. Fragmentation effects were less pronounced in the largest local populations (but not the least isolated), located on gentler slopes with higher soil water availability and lower heat exposure. Conclusions: Our results suggest the existence of a set of demographic and environmental factors that could have coherently buffered the detrimental genetic effects of fragmentation in both metapopulations. These findings could be useful to plan landscape restoration for the evolutionary rescue of mixed forests that once characterized Mediterranean mountain ecosystems. [ABSTRACT FROM AUTHOR]