Conclusions
Both phylogenetic and environmental effects are fundamental to understanding the drivers and distribution of plant traits. Combining both in a single model is challenging due to data limitations but is possible by leveraging large scale datasets. This approach allows for improved traits predictions compared to models that rely on either species average or the environment in isolation and allows for robust predictions for species and regions not included in the training data. Across Eastern US both interspecies trait variation (driven by shifts in species’ abundance) and intraspecific variation are key for predicting joint trait distributions, with effects on the LES. The influence of these components varies by species, trait, ecoregion and scale.
Our approach overcomes previous data limitations by integrating multiple sources of biological and environmental information to create a single integrated model. As new traits, phylogenetic, and species inventory data is released globally, the combined approach can be extended to new regions and unlock the potential to study patterns of intraspecific variation for hundreds of traits-species-environment combinations. For example, this is already possible by leveraging national forest inventories in some European countries, Canada, New Zealand (Schelhaas et al, 2006, Rati et al., 2018, Gills et al., 2005, Paul et al., 2021) along with ever-growing plant traits datasets stored in TRY (https://www.try-db.org/). Expanding this work outside the US could contribute to further understanding the mechanisms driving trait distributions across scales and the link between traits, species distribution, forest assembly and function.