Partitioning temperate plant community structure at different scales

Abstract

Three stem-mapped field plots, each representing a specific forest developmental stage, were established in a temperate forest in Northeastern China: a young secondary conifer and broadleaved mixed forest (YSF), an old secondary conifer and broadleaved mixed forest (OSF), and an old-growth Korean pine and broad-leaf forest (OGF). The focus of this study is to test an environmental control hypothesis. The spatial variations of community structure (species diversity, forest density and size differentiation) were partitioned into pure environment, pure space, and spatially-structured environmental processes in the three research plots. The principal coordinates of neighbor matrices (PCNMs) method was included in the procedure of variation decomposition with respect to spatial and environmental components. The significant PCNM variables could be directly interpreted in terms of spatial scales. The results indicate that the explanatory power of the soil data was much greater in the secondary forests (YSF and/or OSF) than in the old-growth forest regarding species diversity, forest density and size differentiation. Nearly half (48.35% and 44.86%) of the variation of species richness was explained by soil properties in the young secondary forest and the old secondary forest, respectively. However, only 4.87% of that variation was explained by soil properties in the old-growth forest. Over 14% of the variation of the tree size differentiation was explained by soil properties in two secondary forests, and only 4.23% in the old-growth forest. In this study, the spatial variation of species richness and size differentiation was related to environmental variables at multiple scales. Soil variables had a significant effect on species richness and size differentiation at broader scales in the secondary forests, but mainly at medium and fine scales in the old-growth forest. The results challenge the commonly held assumption that tree distributions simply reflect patterns of seed dispersal at local scales.