PATTERN ANALYSIS AND ASSOCIATION ANALYSIS OF FOREST COMMUNITIES IN MT. DINGHU

The data of density and abundanceof the plant species from three communities in Dinghu Shan forest were used as the ecological information for the pattern analysis and the association analysis. The four methods of pattern analysis, Greig-Smith's(1952) pattern analysis, Kershaw's (1970) intensity analysis, Hill's(1973)two-term local variance analysis, and spectral analysis, were applied to describe and indicate the population pattern of the tree, shrub species. The other three methods of association analysis, joint analysis and covariance analysis(Kershaw, 1900), correlation analysis(Austin, 1968)were utilized to explore the interspecific association of the representative species in three communities. The comparisons between the methods that had been used were made in this paper. The pattern analysis are very useful for separating the intensity, scale and grain of the pattern and indicating the patchiness of horizontal dispersion of plant population. Of the four methods, the Greig-Smith's and the Hill's are appropriate for the utility in the lower subtropical forest, especialy being applied cooperatively. It had been found from the analysis graphs that Kershaw's not only has no suitable statistics-test but also reduces the peak values. The spectral analysis has some advantages of describing the pattern, but no enough evidence to prove its efficiency for a small number of the samples. The association analysis gave some interesting results from which the reasons of interspecific association could be deduced. The positive association is the fruit either of the similar responses of two species to the environmental factors or of the dominants pattern being exerted on another species. The negative association is the consequence either of the variant responses of two species to the environmental factors or of the automatic division of the horizontal spaces occupied by two high density species, The appearances of the association between plant species could not be restricted by the block size and the type of pattern. Among the three methods that had been used the joint analysis is the best one not only for its clear graph but also for its statistical significance. Doubulessly, all the analyzing methods used in this paper, have played the important role on the advance researches of the plant species space distribution, in particular, on showing the patchiness and the mosaic of the plant population individuals.