Vertical structure of herbaceous canopies:: the importance of plant growth-form and species-specific traits

The factors determining herbaceous canopy architecture are poorly understood, especially in natural and semi-natural plant communities. In this study, we tested three main hypotheses: (1) the structure of herbaceous canopies can be explained by the vertical distribution of functional groups defined by leaf width and the presence/absence of leaves on upright stem; (2) the degree of canopy stratification is greater in habitats that experience lower spatial heterogeneity in the supply of light (i.e., grasslands as opposed to forest herb layers); and (3) there is significant variation among species within a growth-form, with respect to their vertical position in the canopy. We used plant foliage height distribution data from 14 grassland and 13 forest herbaceous communities to test these hypotheses. A general linear mixed model was applied to specify the proportions of total variance in the foliage height, accounted for by the fixed effects of plants' basic growth-form properties (growth-form) and community type (forest/grassland), and by the random effects of sampling site, sampling point, and individual species. We were also interested in the correlation of the degree of the stratification with various community characteristics (productivity, other canopy properties, species richness, variation of species' traits) and light availability. There was some evidence of overall canopy stratification according to plant growth-form, since plants with leafy stem were located significantly higher. However, such a pattern of two more or less distinct layers (grasses+upright forbs and rosette forbs) occurred with consistency only in grasslands (greater homogeneity in light). The between-species variation within a growth-form was a highly significant predictor of canopy vertical structure in the 27 communities. The proportion of total observed variance, explainable through species-specific effects, was comparable to that caused by between-site differences. The effect of community horizontal pattern was less obvious, but still significant. The site by site analysis revealed that the degree to which horizontal patchiness explained variation in vertical canopy structure was negatively related to the relative importance of species-specific effects, showing that small between-species differences lead to a more obvious within-community horizontal pattern, and vice versa. The upper bound of the degree of foliage stratification, according to growth-form, was related to the variability of species light requirements and to relative (to community pool size) richness, indicating that certain aspects of canopy architecture might be explained through community species composition and diversity pattern.