Associations of climate, soil resources, and species-level plant traits with ecosystem productivity across aridity gradients in China

Net primary productivity (NPP) varies with climate and soil resources, but whether species-level plant traits provide information beyond these environmental controls remains unclear. We linked trait data for 25,210 vascular plant species across China with climatic variables, depth-resolved soil properties, and long-term MODIS NPP. NPP increased from approximately 150 g C m −2 yr −1 in the arid region to 700 g C m −2 yr −1 in the humid region. Soil–NPP relationships varied among soil properties, depths, and aridity classes. Joint soil-profile models improved spatial prediction in the dry sub-humid and humid regions but not in the arid region. Nutritional traits showed stronger marginal relationships with NPP than reproductive traits, although most adjusted relationships were concentrated in the dry sub-humid region and individual traits provided little additional predictive information at geographically separated locations. Climate factors had the largest relative importance in all three regions, followed by soil factors. The relative importance of nutritional and reproductive traits was greater in the arid region but remained small in the two wetter regions, and much of the explained variation was shared among predictor groups. The piecewise structural equation models showed that temperature, precipitation, and soil resources had the clearest direct associations with NPP, whereas direct paths from nutritional and reproductive traits were small and mostly nonsignificant. These results identify how the environmental and species-level trait correlates of productivity differ across aridity gradients while defining the limits of inference from species-level trait records.