Local Fractal Dimensions and Multifractal Analysis of the Root System of Legumes

Plant Production Science

Jiro Tatsumi¹  Kalyani Weerasinghe Ketipearachchi² 

[1] Graduate School of Bioagricultural Sciences;Kobe University;

关键词: Fractal dimension;    Legume;    Multifractal;    Root system architecture;    Topology;   

DOI  :  10.1626/pps.3.289

来源: DOAJ

【 摘 要 】

The fractal dimension (D) is the main concept of fractal geometry. D may have the potential to describe morphology in overall terms for root systems quantitatively. However, it is often difficult to discriminate root systems possessing a common D but having a different morphological appearance. We tested the potentiality of multifractal analysis as a tool for characterizing the morphology and architecture of root systems, and applied this method for discrimination of such root systems having similar Ds. We adopted mass fractal measure and determined Ds for small area (local fractal dimension, DLocal) at different locations of root systems. The multifractal spectrum indicated that Dlocal varied with the location on the roots, suggesting that root systems possess a multifractal structure, the complexity is unevenly distributed, and its characteristic can be partly described with the Dlocal spectrum. The skewness and kurtosis of histogram of Dlocal varied among species, even when they possess root systems with a common D which was determined by the box-counting method as an overall morphology of the system (global D). The spectrum of Dlocal was further analyzed in relation to topological parameters. We found significant correlation between topological index and skewness as well as between topological index and kurtosis. These suggest that the variation of Dlocal spectrum can be an indication of variability of architecture in a locally restricted area of root system. The multifractal analysis can be applicable and it has a potential for discriminating and characterizing root systems having a common global D, but with different morphology.

【 授权许可】

Unknown