The spatial organization of plant communities in a deciduous forest: A computational-geometry-based analysis.
The use of Voronoi tessellation and Delaunay triangulation in studying the organization of plant communities is described. Data on the species identity, diameter at breast height and position of over 1500 trees, from three approximately one hectare plots with low, medium and high tree densities, located in the deciduous forest of Mudumalai sanctuary, Tamil Nadu, Southern India, form the basis of this analysis. By defining the local neighbourhood in terms of the Delaunay neighbours, we have devised an objective and unambiguous measure of intra- and inter-specific associations. The distribution of the number of neighbours is seen to be relatively independent of the tree density. The area of the voronoi cell of a tree is seen to be significantly and positively corelated with its diameter, and the strength of the association seems to be density-dependent. Other features such as the distribution of nearest-neighbour distances, the area-perimeter relationships, etc., have been analysed. The potential of this approach for the study of the long-term dynamics of plant communities is discussed.