Plants and climate change: complexities and surprises.

BACKGROUND: Anthropogenic climate change (ACC) will influence all aspects of plant biology over coming decades. Many changes in wild species have already been well-documented as a result of increased atmospheric CO2 concentrations, warming climate and changing precipitation regimes. A wealth of available data has allowed the use of meta-analyses to examine plant-climate interactions on more sophisticated levels than before. These analyses have revealed major differences in plant response among groups, e.g. with respect to functional traits, taxonomy, life-history and provenance. Interestingly, these meta-analyses have also exposed unexpected mismatches between theory, experimental, and observational studies. SCOPE: We reviewed the literature on species' responses to ACC, finding ∼42 % of 4000 species studied globally are plants (primarily terrestrial). We review impacts on phenology, distributions, ecophysiology, regeneration biology, plant-plant and plant-herbivore interactions, and the roles of plasticity and evolution. We focused on apparent deviations from expectation, and highlighted cases where more sophisticated analyses revealed that unexpected changes were, in fact, responses to ACC. CONCLUSIONS: We found that conventionally expected responses are generally well-understood, and that it is the aberrant responses that are now yielding greater insight into current and possible future impacts of ACC. We argue that inconclusive, unexpected, or counter-intuitive results should be embraced in order to understand apparent disconnects between theory, prediction, and observation. We highlight prime examples from the collection of papers in this Special Issue, as well as general literature. We found use of plant functional groupings/traits had mixed success, but that some underutilized approaches, such as Grime's C/S/R strategies, when incorporated, have improved understanding of observed responses. Despite inherent difficulties, we highlight the need for ecologists to conduct community-level experiments in systems that replicate multiple aspects of ACC. Specifically, we call for development of coordinating experiments across networks of field sites, both natural and man-made