Climate-driven convergent evolution in riparian ecosystems on sky islands.

Climate-induced evolution will determine population persistence in a changing world. However, finding natural systems in which to study these responses has been a barrier to estimating the impact of global change on a broad scale. We propose that isolated sky islands (SI) and adjacent mountain chains (MC) are natural laboratories for studying long-term and contemporary climatic pressures on natural populations. We used greenhouse common garden trees to test whether populations on SI exposed to hot and dry climates since the end of the Pleistocene have phenotypically diverged from populations on MC, and if SI populations have converged in these traits. We show: (1) populations of Populus angustifolia from SI have diverged from MC, and converged across SI, in reproductive and productivity traits, (2) these traits (cloning and aboveground biomass, respectively) are significantly correlated, suggesting a genetic linkage between them, and (3) the trait variation is driven by both natural selection and genetic drift. These shifts represent potentially beneficial phenotypes for population persistence in a changing world. These results suggest that the SI-MC comparison is a natural laboratory, as well as a predictive framework, for studying long-term responses to climate change across the globe.