Analysis of Grouped Growth Patterns in Even-Aged Sugi Forest Stand within the Framework of Mixture Model

• Yanagihara Hirokazu

  Graduate School of Mathematics Hiroshima University, Japan

• Ninomiya Yoshiyuki

  Graduate School of Mathematics Kyushu University, Japan

• Yoshimoto Atsushi

  Graduate School of Environmental Studies Tohoku University, Japan

Abstract

Even in the same even-aged forest stand, trees grow differently due to their own growth capacity, relative spatial conditions, and other environmental factors. In this paper, we use a Gaussian mixture model (GMM) to investigate growth patterns in an even-aged forest to identify groups of trees with the same growth characteristics. Given the Richards growth function for the growth process, the estimated coefficients are used as new response variables in a multivariate GMM to identify growth patterns. The optimal number of grouped growth patterns is searched by minimizing the cross-validation (CV) criterion. We demonstrate our proposed method using growth data from a sugi (Cryptomeria japonica) sample plot in Hoshino Village, Fukuoka Prefecture, Japan. The model identified three separate growth patterns in our sample plot.

Journal

• FORMATH

  FORMATH 7 (0), 39-59, 2008

  FORMATH Research Society

Keywords

• growth analysis
• information criterion
• k-means clustering
• multivariate Gaussian mixture model
• model selection

Details

• CRID: 1390848250112367616
• NII Article ID: 130007845546
• DOI: 10.15684/formath.07.004
• ISSN: 21885729
• Web Site: https://www.jstage.jst.go.jp/article/formath/7/0/7_07.004/_pdf
• Text Lang: en
• Data Source:
  • JaLC
  • Crossref
  • CiNii Articles
• Abstract License Flag: Disallowed