Ecologists are increasingly interested in making accurate predictions of plant response to climate change. Many studies have attempted to document plant response to warming by grouping species into functional groups. Within functional groups, however, species often display divergent responses. Determining how foliar functional traits might be used to predict plant responses to warming could reduce analytical complexity while maintaining generalizations across systems.
We conducted a meta-analysis on 18 studies (consisting of 38 species) of plant biomass response to experimental or natural warming. We determined whether plant trait estimates associated with the leaf economics spectrum [leaf life span (LL), leaf mass per area (LMA), leaf nitrogen (Nmass), leaf phosphorus (Pmass), photosynthetic capacity (Amax) and stomatal conductance (Gs)] from a global plant database of experimentally unmanipulated plants, GloPNet, could be used to predict biomass response to experimental warming.
We found that three single leaf traits (LL, Nmass and Amax) were significant predictors for the response of plant biomass to warming treatments, perhaps due to their association with plant growth rates, adaptation rate and ability, each explaining between 21–46% of the variation in plant biomass responses. The magnitude of response to warming decreased with increasing LL, but increased with increasing Nmass and Amax. We found no linear combination of any of these traits that predicted warming response.
These results show that foliar traits can aid in understanding the mechanisms by which plants respond to temperature across species. Because each trait only explained a portion of variation in how plant growth responded to warming, however, future studies that examine how plant communities respond to warming should simultaneously measure multiple leaf traits, especially those most sensitive to warming, across plant species, to determine whether the predictive ability of functional traits changes between different ecosystems or plant taxonomic groups.