Fractal nature of spatial patterns in Japanese evergreen oak forest trees

Abstract

Spatial patterns of canopy tree distribution were analyzed through the Box Counting Method in fractal geometry. A map showing the position of 4102 individual canopy trees, belonging to 55 species, was compiled from an inventory of a fourteen hectare research plot, established in an evergreen oak forest, in Kasugayama Forest Reserve, Nara City, western Japan, in 1991. A newly proposed fractal model allows the saturation of the number of boxes in smaller box sizes, and was successfully used for the calculation of fractal dimensions of sixteen common species, with a sufficient sample size for computation. The calculated dimensions were in a range between 1.19 of Ckamaecyparis obtusa Sieb. et Zucc. and 1.73 of Quercus salicina Blume, and their variations among species were compared with the dimensions of valley (1.47) and ridge (1.43) lines. Larger fractal dimensions tended to be confined to a species with a wider distribution and with habitat less dependent on topography. Smaller fractal dimensions were observed from species with patchy distribution. Species with intermediate fractal dimensions had a distribution pattern fairly dependent on a valley or ridge site. In conclusion, fractal geometry is applicable to the spatial patterns of forest trees and results in fractal dimensions which appear to indicate a degree of topographic dependency in species' distribution.