A key challenge in ecology is to understand the relationships between organismal traits and ecosystem processes. Here, with a novel dataset of leaf length and width for 10,480 woody dicots in China and 2,374 in North America, we show that the variation in community mean leaf size is highly correlated with the variation in climate and ecosystem primary productivity, independent of plant life form. These relationships likely reflect how natural selection modifies leaf size across varying climates in conjunction with how climate influences canopy total leaf area. We find that the leaf size‒primary productivity functions based on the Chinese dataset can predict productivity in North America and vice-versa. In addition to advancing understanding of the relationship between a climate-driven trait and ecosystem functioning, our findings suggest that leaf size can also be a promising tool in paleoecology for scaling from fossil leaves to paleo-primary productivity of woody ecosystems.

This dataset includes two Excel files, one for the data of China and the other for the data of North America from different data sources, and files for related R code. Both Excel files include leaf size measures and environmental variables used in the study entitled as "Leaf size of woody dicots predicts ecosystem primary productivity". Data was compiled from published floras, the online databases and peer-reviewed publications described in the Methods of the paper and Data supplement in the appendices. Leaf size measures were calculated by integrating species distributions with the median leaf length, leaf width and 2/3*length-width products per species separately. Each environmental layer was resampled into the same resolution as those of species distributions in China and North America (see Data supplement for details).

The definitions for all columns (i.e. variables) within each sheet have been listed in the metadata sheet in each Excel file.