Species diversity of communities based on different vegetation organizational scales in China’s Malan forest region

Aims Patterns of species diversity are closely related to vegetation organizational scales.Our aims were to ana-lyze the relationship between species diversity and vegetation organizational scales and discuss its internal mechanism,which may help us better understand the general law of species diversity.Methods We investigated 208 10 m × 10 m representative quadrats in the Malan forest region on China’s Loess Plateau.Each plot had five 2 m × 2 m subplots for shrubs and five 1 m × 1 m subplots for herbs.We recorded height and diameter at breast height (DBH) of trees and height and cover of shrubs and herbs.We determined communities and their relationships with Two Way Indicator Species Analysis (TWINSPAN) and Principal Com-ponents Analysis (PCA).Based on species richness,Shannon-Wiener and Simpson diversity indices,the relation-ship between species diversity and vegetation organizational scales was analyzed using the additive partition ap-proach.Important findings For the four components constituting regional diversity (α1,β1,β2 and β3) ― no matter if diversity was measured as species richness,Shannon-Wiener or Simpson indices ― within-association diversity always contributed the most to regional diversity in tree,shrub and herb layers (except species richness in the herb layer),indicating that it was the most appropriate scale for measuring species diversity.The proportion of Shannon-Wiener diversity at the vegetation type,formation or association scales was higher than the proportion measured by Simpson diversity,which resulted from the distribution pattern of rare species on each scale.Differ-ences among the orders of β diversity components were observed in scales of between-vegetation type,be-tween-formation and between-association,which were mainly correlated with tree,shrub and herb layers’ species composition and distribution in the region,as well as environmental heterogeneity among different scales.