ORDINATION ANALYSIS OF TAXUS CHINENSIS VAR. MAIREI FORESTS BASED ON ARTIFICIAL NEURAL NETWORK THEORY

Abstract

The artificial neural network is attractive for ecological studies for its power in analyzing and solving complicated and nonlinear matters. The Self-Organizing Feature Map (SOFM) ordination were described and applied to the analysis of Taxus chinensis var. mairei forests in Shanxi province of China in this paper. The data matrix is the important values of 128 species in 95 quadrats. The results showed that SOFM ordination displayed forest communities in species space which reflected ecological gradients successfully. Its first axis is comprehensive gradient of topographical factors and its second axis is a comprehensive gradient of soil variables. SOFM ordination also distinguished 95 quadrats into eight forest types. SOFM ordination made the interpretation of relationships between communities, species, and environments easier.