Detection of fine scale niche assembly in a tropical forest through analysis of indirect environmental variables

Topographic or forest structural variables have been called indirect environmental variables in forest ecology studies. We tested the hypothesis that relative elevation and tallest tree height in the sampling unit, two of these indirect environmental variables, play a role in fine scale community organization in an Atlantic forest in Southeastern Brazil. One hundred 10 m × 10 m sampling units were used to sample the tree community in a 6.5 ha stand. We modeled space through Moran's eigenvectors maps and selected spatial eigenfuctions through an individual species based approach to reach an effective control of type I error. Redundancy analysis and variation partitioning were applied to correlate community composition and environmental variables while maintaining space constant. A significant global p-value for the Redundancy analysis was found. The pure environmental fraction of variation in community composition was also significant, increasing from 1% (adjusted R2) for the simple environmental data to 3.3% for the polynomial (third degree) environmental data. The pure and spatially structured environmental fractions summed up 4% and 6% for the simple and polynomial environmental data, respectively. The influence of the environmental variables varied between species functional groups: non-pioneer species were predominantly affected by relative elevation while pioneers were almost strictly associated with tallest tree height. Our findings amplify the importance of niche assembly processes in fine scale community organization because of the choice of a grain size (10 m × 10 m) and an extent (6.5 ha) smaller than those considered elsewhere, along with rigorous control of type I error rate.