南京林业大学学报(自然科学版) ›› 2018, Vol. 42 ›› Issue (02): 170-176.doi: 10.3969/j.issn.1000-2006.201703103

• 专题报道 • 上一篇    下一篇

基于广义非线性混合效应的华北落叶松天然次生林枝下高模型

段光爽1,2,李学东3,冯 岩4,符利勇1*   

  1. 1. 中国林业科学研究院资源信息研究所,北京 100091; 2. 信阳师范学院数学与统计学院,河南 信阳 464000; 3.中国林业科学研究院华北林业实验中心,北京 102300; 4.中国林业科学研究院,北京 100091
  • 出版日期:2018-04-12 发布日期:2018-04-12
  • 基金资助:
    基金项目:国家林业公益性行业科研专项项目(201404417); 国家自然科学基金项目(31570628,31470641); 河南省科技开放合作项目(172106000071) 第一作者:段光爽(oliverdgs@163.com),博士。*通信作者:符利勇(fuly@caf.ac.cn),副研究员。

Generalized nonlinear mixed-effects crown base height model of Larix principis-rupprechtii natural secondary forests

DUAN Guangshuang1,2,LI Xuedong3, FENG Yan4, FU Liyong1*   

  1. 1. Research Institute of Forest Resource Information Techniques, Chinese Academy of Forestry, Beijing 100091, China; 2. College of Mathematics and Statistics, Xinyang Normal University, Xinyang 464000, China; 3. Experimental Centre of Forestry in North China, Chinese Academy of Forestry, Beijing 102300, China; 4. Chinese Academy of Forestry, Beijing 100091, China
  • Online:2018-04-12 Published:2018-04-12

摘要: 【目的】立木枝下高模型的构建是森林经营的核心内容,利用非线性混合效应模型方法构建华北落叶松枝下高模型,可为森林生长与收获研究提供理论依据。【方法】基于112块华北落叶松天然次生林样地单木数据,从7个备选的枝下高-树高模型中选出一个最优基础模型; 分析9个不同单木或林分因子及其因子之间的组合对枝下高的影响,将影响显著的因子作为模型预测变量以提高模型精度。在此基础上,考虑区组以及嵌套在区组里的样地对枝下高的影响,即构建嵌套两水平的非线性混合效应枝下高模型。【结果】Logistic模型预测精度较高并且模型参数可解释,因此选为基础模型。除树高之外,立木胸径、样地内所有大于对象木胸径的立木断面积总和、平均冠幅和林分密度与枝下高相关显著,故作为模型预测变量。与传统模型相比,所构建的两水平嵌套非线性混合效应模型对应的决定系数提高了53.26%,均方根误差降低了24.73%,因此明显提高了模型预测精度。【结论】区组和嵌套在区组里的样地对立木枝下高随机干扰较大,当考虑这些随机效应对枝下高的影响时能明显提高模型的预测精度。

Abstract: 【Objective】As a core tool for forest management, a model of crown base height,which utilizes the method of nonlinear mixed effects, was established in this study for researching the growth and yield of Larix principis-rupprechtii natural secondary forests.【Method】Based on individual tree data of 112 plots of a Larix principis-rupprechtii natural secondary forests, an optimal basic model was selected from seven candidate crown base height models.The influence on crown height base of nine disparate individual tree or stand characteristics and their combinations were analyzed, and these significant factors were regarded as predictor variables for improving the precision of the model. Taking into account the disturbance from block- and plot-level random effects, a nested two-level nonlinear mixed-effects model of crown height base was constructed.【Result】The Logistic model was selected as the basic model on account of its higher prediction accuracy and interpret ability of model parameters. Tree diameter at breast height, total basal area of all trees with diameter larger than that of the target tree, mean crown length and stand density were used as predictor variables because of their significant correlation to crown height base, and not to tree height. Compared with the conventional model, the prediction accuracy of the nested two-level nonlinear mixed-effects model was enhanced distinctly, and its coefficient of determination and root mean square error were increased by 53.26% and reduced by 24.73% respectively.【Conclusion】The disturbance from block and plot level random effects significantly influenced the prediction of the crown height base.However, the prediction accuracy of the crown height base model with these random effects was still improved obviously.

中图分类号: