长白落叶松等几个树种冠幅预测模型的研究

以20块落叶松云冷杉林为对象，采用多元逐步回归方法，研究其组成树种的单株木冠幅预测模型.因变量为单株木冠幅，自变量包括胸径、树高、枝下高、树冠比、竞争因子和林分密度.共测定了3 099株树木，全部参加了建模.最终建立了长白落叶松、冷杉、红松、云杉、枫桦、水曲柳、色木、白桦和椴树9个树种的冠幅预测模型.结果表明:胸径和林分密度是所有树种中影响冠幅的重要因子；模型的调整决定系数在0.34～0.75之间；绝对误差在-0.000 114～0.054 m之间；相对误差在-0.054 43%～3.440%之间；均方根误差在0.360～0.510 m之间，相对均方根误差在27.4%～37.4%之间. 

Crown-width prediction models for several tree species including Larix olgensis in northeastern China

The mean crown-width of stand-grown trees was modeled as a function of stem diameter at breast height, tree height, crown length, crown ratio, competition index and stand density for several tree species in 20 mixed stands of Larix olgensis, Abies nephrolepis and Picea jazoertsis in northeastern China. Multivariate stepwise regression was performed with the data of 3 099 crown-width observations for nine tree species including L. olgensis, P. jazoensis, A. nephrolepis, Pinus koraiensis, Betula costata, B. platyphylla, Acer mono, Tilla amurensis and Fraxinus mandshurica. Results showed that tree diameter at breast height and stand density were significant variables in crown-width prediction models for all tree species. The adjusted R2 value was between 0.34 and 0.75. Absolute and relative bias values were between - 0.000 114 and 0.054 m, - 0.054 43% and 3.440%, respectively. RMSE and relative RMSE values were from 0.360 to 0.510 m, and 27.4% to 37.4%, respectively.