Climate is a key determinant of plant form and function, but a strong mechanistic linkage between these factors remains lacking. Leaf venation networks may provide a solution, because they constrain carbon and water fluxes that determine plant performance and fitness in different climates. Fossil leaves often preserve details of leaf veins - thus a model based on leaf venation network physiology could provide accurate fine-scale estimates of climate. I test two hypotheses: first, that leaf-controlled water supply is matched to climate-controlled water demand, and second, that leaf vein density can be optimally matched to climate independent of evolutionary history or constraint. These hypotheses then yield quantitative predictions via a physiological model that links community-mean leaf vein density to temperature, aridity and carbon dioxide. I test this model in a tropical and a temperate climate gradients spanning a 20 degree C range and then reconstruct 4.6Myr of paleotemperature at points spanning the Cretaceous-Tertiary boundary.
Results/Conclusions
In each gradient, site-mean vein density was a strong positive correlate of temperature, explaining 58 to 79% of the temperature variation, consistent with predictions. Vein density was also significantly more evolutionarily labile than expected under a Brownian model of trait evolution, indicating that it represents an adaptive response to climate. These modern results were also assessed in the Hell Creek fossil flora, to determine if vein networks could be a powerful taxon-independent approach for paleoclimate reconstruction.