Wood density and leaf size jointly predict woody plant growth rates across (but not within) species along a steep precipitation gradient

Abstract

Abstract Functional traits have been proposed to define key dimensions of plant ecological strategies, but we lack consensus on whether traits can accurately predict plant demography. Despite theoretical expectations, it has been challenging to find consistent relationships between functional traits and growth. In this study, we quantified inter‐ and intraspecific trait variation (ITV) and individual growth rates of woody plants across a steep moisture gradient that varies 10‐fold in annual precipitation (350–3700 mm) in southern Chile and used a hierarchical Bayesian model to predict growth as a function of trait values. We show that large‐leaved species with lower stem tissue density exhibited the fastest growth rates, and these two traits exhibited the highest proportion of interspecific variation. Predictions of growth improved considerably (R2 of the best model increased from 0.28 to 0.49) when species‐level multiple traits and their interactions were considered. The inclusion of ITV, however, did not improve models of growth rate. We found that trait–growth rate relationships were not always consistent across levels of biological organization; relationships observed at the interspecific level did not necessarily hold at the intraspecific level. We found that the relationships between wood density or leaf size and growth were consistent in direction across the precipitation gradient, and the relationships between leaf economics traits and growth were weak and site‐specific. Synthesis. Although using more than one functional trait considerably improved growth predictions, wood density and leaf size successfully predicted growth rates across (not within) species, which is consistent with a whole‐plant carbon economy. We assert that these two traits are intimately linked and ultimately describe a continuum of plant architecture and carbon economy that covers multiple trait syndromes.