Understanding the relationship between biodiversity and the dynamics, structure, and function of communities at broad spatial scales is central to ecology. Trait-based methods are expanding this understanding by mechanistically linking environmental variables to organismal growth through individual phenotypes. One recently developed approach, termed “Trait Driver Theory” (TDT), combines methods borrowed from quantitative genetics with metabolic scaling theory to investigate how environmental variables ‘drive’ shifts in the distribution of traits within communities which, in turn, ultimately influence higher-order community- and ecosystem-level processes. Focusing specifically on the shapes of trait distributions, we are able to predict ecosystem functions based on individual-level traits and generate specific hypotheses about community responses to environmental change. We apply TDT to a large dataset of tree foliar traits containing 424 field sites as well as remotely sensed LiDAR and hyperspectral data, altogether spanning elevational and latitudinal gradients throughout the world and representing a wide range of environmental and geographic variation. We describe the shapes of community-level trait distributions using central moments (mean, variance, skewness, and kurtosis), identify correlations between these moments and a suite of environmental variables, and finally, scale up from individual-level traits to estimate forest net primary productivity (NPP).
Results/Conclusions
Here we report significant trends in the distributions of several foliar traits in response to shifts in environmental conditions along both elevational and latitudinal gradients. For many traits, trait means exhibit convergent trends with latitude and elevation (either positive or negative), likely due to convergent correlations between latitude/elevation and several important environmental variables (e.g., mean annual temperature, vapor pressure deficit, etc.). Trait variance (almost) always declines with both latitude and elevation. Patterns in skewness and kurtosis reveal that trait distributions are asymmetric on average, generally tending toward beta distributions with low values of kurtosis relative to skewness, potentially indicating the maintenance of phenotypic evenness within communities. Local patterns in trait skewness are consistent with expectations about community shifts toward novel environmental optima, providing evidence of whole-community responses to recent environmental warming. Finally, traits that are tightly linked to growth (e.g., specific leaf area and leaf nitrogen and phosphorus composition) exhibit complex relationships with environmental variables, resulting in both positive and negative influences on NPP along latitudinal and elevational gradients. Together, these results expose broad patterns in trait variation across space and highlight the value of using trait-based approaches to mechanistically link environmental variation to community/ecosystem processes.