Self-organizing feature map (SOFM) ordination is a new ordination method based on natural neural networks theory in analysis of community-environment relation. However, it is indirect ordination method and cannot combine environmental data. We combine species data with environmental data to construct a bi-plot of SOFM by use of correlation analysis. This was applied to community-environment analysis in the Taihang Mountain. The results showed that the bi-plot of SOFM described the relationships of samples, communities and species with environmental variables clearly. It provided consistent results with CCA. We can conclude that the bi-plot of SOFM is perfectly useful in vegetation-environment analysis.

direct ordination, bi-plot, self-organizing feature map, plant community.