[Application of indicator species in predicting forest management effect on forest species diversity and community composition].

By using indicator species analysis (ISA) method, this paper studied the feasibility of using indicator species to reflect the responses of species diversity and community composition of subtropical forests in Huitong of China to forest management. Ninety-four significant indicator species from 357 understory species were identified, and a new indicator species dataset (community level) was constructed to examine the association between indicator species dataset and original community dataset, and to evaluate the predictive potential of indicator species in reflecting forest management effect. There existed a strong association between the two datasets (Mantel r = 0.898). The indicator species dataset could well predict the management effect on species diversity (regression analysis, R2 > 0.74) and community composition (ANOVA, F >16.79). When the two datasets were applied to Nonmetric Multi-Dimensional Scaling (NMDS) ordination and K-mean cluster analysis, the indicator species dataset could well identify the forest types with different management treatments, as the original community dataset did. Also, the indicator species dataset nearly played the same role as the original community dataset in identifying the species diversity, community composition, and forest type. It was suggested that for saving costs in overall investigation of forest ecosystem, indicator species could be used as a surrogate of full community to predict forest management effect.