Determinism versus chance in canopy gap herbaceous species assemblages in temperate Abies-Betula forests

It is not known if the species composition of herbaceous plant assemblies within tree-fall gaps is determined stochastically or whether it follows species-specific and environmentally-determined patterns. We applied three methods, fitting species abundance distribution models, comparing community similarities, and testing species-environmental variable association, to evaluate the relative importance of chance and deterministic rules in controlling herbaceous composition within canopy gaps in the Abies-Betula forests of Mt. Taibai, Central China. Herbaceous species abundance within canopy gaps was well fit by a neutral model, showing that relative abundance was qualitatively consistent with stochastic processes. Although species composition in gaps significantly differed from closed canopy sites (ANOSIM, R = 0.509, p = 0.001), there were no significant differences among gaps of different age groups (ANOSIM, R = 0.035, p = 0.191) or gaps of different size groups (ANOSIM, R = 0.089, p = 0.057). Similarly, gaps of similar age and size did not show significantly higher similarity (chi(2) = 2.30, df = 3, p = 0.513) in species composition than gaps of distinct age or size. Moreover, there were only eight species, including two light-demanding species, confined to larger gaps among the 69 species, and only an additional seven species found more commonly in larger gaps, whereas most herbaceous species were gap size generalists. Canonical correspondence analysis and random permutation tests suggested that only 27% of species with abundance >= 5 were associated with environmental variables in gaps. In summary, the species composition in gaps was not constrained significantly by gap traits; rather species were distributed stochastically, likely through by random dispersal and recruitment limitation of species from the surrounding available species pool. Measures that introduce gap disturbance, such as selective harvesting, are still recommended when the maintenance of total biodiversity in forests is a concern. (C) 2011 Elsevier B.V. All rights reserved.