利用非线性混合模型进行杉木林分断面积生长模拟研究  

Simulating basal area growth of fir plantations using a nonlinear mixed modeling approach

作　　者：李春明[1] 

机构地区：[1]中国林业科学研究院资源信息研究所

出　　处：《北京林业大学学报》2009年第1期44-49,共6页Journal of Beijing Forestry University

基　　金：中央级公益性科研院所基本科研业务费专项资金项目(RIFRIGTZGZ2007005);"十一五"国家科技支撑计划课题(2006BAD03A08)

摘　　要：该文描述了传统生长收获模型存在的问题及非线性混合效应模型在预估林分断面积方面的应用,介绍了非线性混合效应模型的表达形式。选择常用的Richards和Schumacher两种基本模型形式,通过变化自变量进行断面积生长模拟,采用非线性最小二乘回归法进行拟合;然后选择复相关系数及预估精度较高的函数作为基础模型,并在此基础上构造非线性混合效应模型;最后对非线性混合效应模型与传统模型拟合的效果进行了比较分析。结果表明,考虑初植密度效应的非线性混合效应模型比传统的回归模型估计精度高。The paper describes the problem of conventional growth and yield model and the use of nonlinear mixedeffect models in modeling basal area. The forms of nonlinear mixed-effect models were introduced. By varying independent variables, two growth models, Richards and Schumacher, were used to simulate the basal area of fir plantations. To design the nonlinear mixed-effect model, the authors used a nonlinear least squares regression approach to find the best correlation coefficient and simulation accuracy. The results show that the precision of nonlinear mixed effect model, which takes account of the random effects of different initial planting density, is better than that of conventional regression models.

关 键 词：非线性混合效应模型 林分断面积 随机效应 杉木 

分 类 号：S758.55[农业科学—森林经理学]