Climate-driven selection results in powerful geographic framework for predicting phenotype

Our ability to prepare for and mitigate the likely ecological and evolutionary impacts of climate change largely depends upon our ability to predict the phenotypic responses of organisms that allow them to persist, adapt, and migrate along environmental stress gradients. Using fifteen populations of cottonwoods, a dominant riparian forest tree, that are distributed along elevation gradients and represent three genetic provenances, we hypothesized and show that: 1) populations within a provenance demonstrate parallel evolutionary responses to climatic gradients associated with elevation; and 2) the evolutionary effects of elevation on bud-break phenology varied by provenance. Across all populations, we find strong evidence of directional selection on bud-break phenology in response to variation in potential evapotranspiration. Overall, there is a 4-day difference between high and low elevation sites when averaged across the western United States. The difference between high and low populations was nearly 11 days in southern latitudes compared to a single day in northern latitudes; a 90% difference in the evolutionary response in bud-break phenology to climatic gradients associated with elevation. Our results raise questions about the general consequences of limiting study locations to a single clime but demonstrate the broad applicability for using elevation gradients and associated environmental gradients in predicting plant phenotypic responses to climate.