基于非线性混合模型的杉木标准树高曲线  

Generalized Height-Diameter Model for Chinese fir Based on Nonlinear Mixed Effects Model

作　　者：董云飞[1] 孙玉军[1] 许昊[1] 梅光义[1] 

机构地区：[1]北京林业大学,北京100083

出　　处：《东北林业大学学报》2014年第11期72-76,81,共6页Journal of Northeast Forestry University

基　　金：林业公益性行业科研专项(200904003-1);国家林业局重点项目(2012-07);林业科技成果国家级推广项目([2014]26)

摘　　要：树高曲线是研究森林生长和收获的重要基础,以福建省将乐县国有林场29块杉木人工林实测数据为例,首先运用非线性最小二乘法对9个经典常用标准树高曲线方程进行拟合,选出拟合效果最好的模型作为基础模型,利用基础模型及建模数据构建非线性混合效应模型。通过变化混合参数个数,并利用SAS软件进行模拟,选择结果收敛且AIC、BIC值和负二倍对数似然值最小的混合模型作为最优模型,最后利用检验数据与传统非线性最小二乘法进行精度比较。结果表明:方程H=1.3+a0Ha1t(1-e-a2D-a3gD)a4在模拟杉木树高的生长趋势效果较好,因此,选为构建非线性混合模型的基础模型。在选择混合参数时,发现当a1和a3同时作为混合参数时,拟合效果最好。综合分析,非线性混合模型比固定模型的模拟精度高,可更好地描述杉木的树高生长。We used the data of 29 plots of Chinese fir located in national forest farm of Jiangle in Fujian Province to study the best function as the base model with the least square method among nine regular generalized height-diameter equations. The nonlinear mixed model was constructed based on the base model and modeling data.The convergence model, and the mixed model of the values of -2 Log Likelihood, AIC and BIC models were considered as the best model by changing the number of mixed parameters in fitting process.The precision of mixed models was compared with that of conventional non-linear least square method based on validation data.The precision of the model H=1.3＋a0H at1（1-e-a2D-a3g D）a4 was higher than that of the other eight models when simulating the growth of Chinese fir.The fitted effects of and simultaneously as mixed parameters were the best when choosing mixed parameters.Therefore, the precision of mixed model is better than that of fixed model, and the mixed model can be used to describe the height growth of Chinese fir better.

关 键 词：杉木 标准树高曲线 非线性混合模型 

分 类 号：S791.27[农业科学—林木遗传育种]