论文部分内容阅读
基于95株落叶松解析木、843个枝条实测数据,构建了枝条直径和长度的传统非线性模型、含树木变量的非线性模型、非线性混合模型和含有误差结构的非线性混合模型,比较了各模型的拟合精度,分析了枝条直径和长度的变化规律。结果表明:含有误差结构的落叶松人工林枝条直径和长度非线性混合模型拟合精度最优,决定系数R2分别为0.783 7和0.865 7,均方根误差(RMSE)分别为0.377 7和0.286 6,且消除了异方差现象,独立样本检验结果也证实该模型是最优预测模型;落叶松人工林枝条直径和长度与枝条所在位置和树木胸径、冠幅、高径比紧密相关,在树冠顶部分布的一级枝条直径和长度都较小,而在树冠底部分布的则较大,树冠下部枝条的直径和长度受树木形态因子的影响更明显。
Based on the measured data of 95 branches of Larch and 843 branches, the traditional nonlinear model of branch diameter and length, the nonlinear model with tree variables, the nonlinear mixed model and the nonlinear mixed model with error structure were compared The fitting accuracy of each model, the variation of branch diameter and length were analyzed. The results showed that the fitting accuracy of the nonlinear model of branch diameter and length with the error structure was the best, the determination coefficients R2 were 0.783 7 and 0.865 7 respectively, and the root mean square error (RMSE) were 0.377 7 and 0.286 6 , And eliminated the heteroscedasticity. The independent sample test results also confirmed that the model is the optimal prediction model. The diameter and length of the branches of Larix gmelinii plantation are closely related to the location of the branches and the DBH, crown width and height-diameter ratio of the branches, The diameter and length of the primary branches distributed are smaller, but larger at the bottom of the canopy. The diameter and length of the branches under the crown are more affected by the morphological factors of the trees.