Mapping local and global variability in plant trait distributions

Abstract

Our ability to understand and predict the response of ecosystemsto a changing environment depends on quantifying vegetationfunctional diversity. However, representing this diversity atthe global scale is challenging. Typically, in Earth system models,characterization of plant diversity has been limited to groupingrelated species into plant functional types (PFTs), with all trait variationin a PFT collapsed into a single mean value that is appliedglobally. Using the largest global plant trait database and state ofthe art Bayesian modeling, we created fine-grained global mapsof plant trait distributions that can be applied to Earth systemmodels. Focusing on a set of plant traits closely coupled to photosynthesisand foliar respiration?specific leaf area (SLA) and drymass-based concentrations of leaf nitrogen (Nm) and phosphorus(Pm), we characterize how traits vary within and among over50,000 ∼50×50-km cells across the entire vegetated land surface.We do this in several ways?without defining the PFT of eachgrid cell and using 4 or 14 PFTs; each model?s predictions are evaluatedagainst out-of-sample data. This endeavor advances priortrait mapping by generating global maps that preserve variabilityacross scales by using modern Bayesian spatial statistical modelingin combination with a database over three times larger thanthat in previous analyses. Our maps reveal that the most diversegrid cells possess trait variability close to the range of globalPFT means.