Parameterization of Tree and Shrub Stem Wood Density Adaptions to Multiple Climate and Soil Factor Gradients

Wood density (WD) is an important quality and functional trait of wood. However, despite the relationships between WD and abiotic factors being important to model or predict spatial distributions of functional traits, as well as responses of vegetation to climate changes, in current Earth system models or dynamic global vegetation models (ESMs/DGVMs), WD is often oversimplified, being defined as a globally uniform constant either for all plant functional types (PFTs) or for each individual PFT. Such oversimplifications may lead to simulation biases in the morphology of woody PFTs, as well as ecosystem transition and vegetation–atmosphere interactions. Moreover, existing conclusions about the relationships between WD and abiotic factors drawn from field observations remain mixed, making model parameterization improvements difficult.

This study systematically investigated the influences of climate and soil factors on WD across various PFTs. Optimal fitting models for predicting WD within each PFT were then constructed by utilizing our collated global database of 138 604 observations. For WDs of tree PFTs, climate emerges as a more influential factor than soil characteristics, whereas for shrub PFTs the effects of climate and soil are of equivalent significance. Across all six PFTs, correlation coefficients between predictions by fitting models and observed WD range from 0.49 to 0.93. The predicted and observed WD exhibit good agreement across climate space. It is expected that the incorporation of our research findings into DGVMs will improve the simulation of tree height and forest fractional coverage, particularly in the central forest areas and forest transition zones.