Investigations of community assembly often involve the use of functional diversity indices to make inferences about how ecological selection shapes community composition. A commonly used index is the convex hull volume (CHV), which is a multidimensional analog of the trait range. Blonder et al. (2014, 2016) suggested that CHV can overestimate functional diversity because of holes (missing trait combinations) inside the CHV. Does the degree of functional holeyness vary in any systematic way across the regional landscape? We sampled 4 functional plant traits (SLA, leaf dry matter content, leaf size, and height) in a total of 1868 individuals in 64 m2 nested plots in six habitat types (open canopy vs. forest in dry, mesic, and wet sites). Holeyness was calculated as the proportion of total CHV of the first two principal components of the plant traits that was unoccupied ((CHV – hypervolume)/CHV) using the geometry and hypervolume packages in R.
Results/Conclusions
Mesic sites in both forests and open grasslands had lower functional holeyness (missing trait combinations) than both dry sites and wet sites. The missing trait combinations in dry and wet sites were estimated to be about 50% of the total CHV and so CHV overestimated functional diversity by about 100%. Mesic forest had the lowest holeyness (< 25% of the CHV). In addition, the observed CHV was largest in the dry sites and this was unexpected. These results suggest that ecological selection in stressed habitats may require more specialized trait combinations and that there may be alternative trait combinations that confer success under stress. In addition, the notion of functional holeyness has barely been investigated and it would seem that this may be an important new aspect of functional community assembly.