【作者】 张锐；

【导师】 姜立春；

【作者基本信息】 东北林业大学 ， 森林经理学， 2013， 硕士

【摘要】 本研究以黑龙江省五营林业局丽林林场的11块标准地中293株样木数据分析描述了不同林分条件下落叶松人工林的生长发育规律,根据散点分布趋势,从经验方程中挑选拟合度最优的树高曲线模型为THT=32.0371?(1-e-0.05·DBH)1.044218+1.3;利用逐步回归得到冠长模型为CL=1.6622+0.2922·DBH;从标准地中选取30株人工落叶松得到2190个枝解析数采用逐步回归分析,利用枝条和树木的各个变量来构建枝条生长模型,包括总枝条数、弦长、弓高预估模型；得到最优模型总枝条数模型NBR=1.188955×DBH+2.562554×THT;弦长模型BCL=-0.0728+0.9292BL;弓高模型：BAH=0.243114×BL0.864351。对于基径、枝长、着枝角度本文采用混合模型进行研究,利用逐步回归技术建立了落叶松枝条基径模型为：BD=b1+b2DINC+b3DINC2+b4DBH·DINC2枝条长度模型为：BL=b1+b2DINC+b3DINC2+b4DBH·DINC2,枝条着枝角度模型为：BA=b1+b2DINC+b3DINC2+b4DBH·DINC。然后,利用S-PLUS软件中的LME过程,拟合线性模型。采用AIC、BIC、对数似然值和似然比检验等模型评价统计指标对不同模型的拟合效果进行比较分析。结果表明：当拟合枝条基径模型时,b1、b2、b3同时作为混合参数时模型拟合最好。为了矫正混合模型构建过程中产生的异方差现象,把幂函数和指数函数加入到枝条基径混合模型中。指数函数显著提高了枝条基径混合模型的拟合效果,并且消除了异方差现象。模型模拟表明：对于大小相同树木,枝条基径随着着枝深度DINC的增加而增大,对于大小不同的树木,枝条基径随着胸径(DBH)的增加而增大。林木的胸径变量很好地反映了不同大小树木的枝条基径的变化。在不知道详细林分信息的前提下,可以利用树木变量合理地预测兴安落叶松人工林的枝条基径的变化规律。当拟合枝条长度和角度模型时,b1、b2、b3同时作为混合参数时模型拟合最好。为了描述混合模型构建过程中产生的异方差现象,把幂函数和指数函数加入到枝条长度和角度混合模型中。指数函数显著提高了枝条长度混合模型的拟合效果,幂函数显著提高了角度混合模型的拟合效果,并且消除了异方差现象。模型检验结果表明：混合模型通过校正随机参数值能提高模型的预测精度。因此,混合模型在应用上不但能反映总体枝条长度和角度预测,还能通过方差协方差结构校正随机参数来反映树木之间的差异。更多还原

【Abstract】 In this study, based on the data of293sample trees of11plots from Dahurian larch (Larix gmelini Rupr.) plantations located in Wuying Forest Bureau in Heilongjiang Province, analysis and describes the larch plantation growth and development of different stand conditions. According to the scatter trend, selected fitting degree optimal tree height curve model from the empirical equation THT=32.0371·(1-e-0.05·DBH)1.044218+1.3.the stepwise regression techniques were used to develop a crown length model:CL=1.6622+0.2922·DBH. Used stepwise regression analysis and obtained the data of2190branch diameter samples of30trees from Dahurian larch (Larix gmelini Rupr.), Use of branches and trees of different variables built branches growth models. The model included total number of branches per tree, branch chord length and branch arch height. The optimal total number of branches per tree model is NBR=1.188955×DBH+2.562554×THT. Branch chord length is BCL=-0.0728+0.9292BL. Branch arch height is BAH=0.243114×BL0.864351. For branch diameter,branch length and branch angle, we used linear mixed model to study, the stepwise regression techniques were used to develop a branch diameter model:BD=b1+b2DINC+b3DINC2+b4DBH·DINC2. Branch length model is BL=b1+b2DINC+b3DINC2+b4DBH·DINC2. Branch angle is BA=b1+b2DINC+b3DINC2+b4DBH·DINC. The developed model was fitted using linear mixed-effects modeling approach based on LME procedure of S-PLUS software. Evaluation statistics, such as AIC, BIC, Log Likelihood and likelihood ratio test were used for model comparisons. The results indicated that the branch diameter model with parameters b1, b2, b3as mixed effects showed the best performance. Exponential and power functions were incorporated into the mixed branch diameter model. The addition of the power function significantly improved the mixed-effects model. The plots of residuals indicated that the mixed-effects model with power function showed more homogeneous residual variance than the mixed-effects model. Branch diameters increased with the depth into crown (DINC) increasing for trees with similar DBH. Branch diameters increased with the DBH increasing for different trees. DBH is adequate variable of tree for describing branch diameter variations with different trees. Branch diameters can be predicted from the measurement of some tree-level variables without detailed knowledge of the stand history for Dahurian larch plantations. The results showed that the branch length and branch angle models with parameters b1, b2, b3as mixed effects showed the best performance. Exponential and power functions were incorporated into mixed branch length and branch angle model. The addition of the exponential and power functions significantly improved the mixed-effects model. The plots of standardized residuals indicated that the mixed-effects model with exponential and power functions showed more homogeneous residual variance than the mixed-effects model. Validation confirmed that the mixed model with calibration of random parameters could provide more accurate and precise prediction. Therefore, the application of mixed model not only showed the mean trends of branch length and branch angle, but also showed the individual difference based on variance-covariance structure.更多还原