湖南栎类-马尾松天然混交林单木生长模型研究

A Study on the Growth Model of Individual Trees in Natural Mixed Forests of Oak and Pinus Massoniana in Hunan

建立湖南栎类-马尾松天然混交林单木材积和断面积生长模型,为该林分的提质增量提供科学指导。选取湖南省9个地区共23块栎类-马尾松天然混交林,构建栎类、马尾松单木材积和断面积生长模型。通过分析可知:两种树种的材积和断面积生长模型均以Richards生长模型为最优,其确定系数(R2)最高,残差平方和(SSE)最小;混合效应模型模拟结果显示,栎类单木材积和断面积生长模型的随机效应在参数a1处最好,而马尾松的材积和断面积生长模型的随机参数分别为b1和a1+b1;相比基础模型,混合效应模型的拟合精度(P)、确定系数均有所提高,平均误差(ME)、平均误差绝对值(MAE)、相对平均误差(RME)和相对平均误差绝对值(RMAE)均大幅降低。结果表明,天然林林木的材积和断面积生长受树种与立地类型差异的影响较大,而混合效应模型能够充分考虑立地之间的差异,可以很好地解决这一难题。

In order to provide scientific guidance for the increment of quality of the mixed Oak and Pinus massoniana natural forest,establishing the growth model of individual volume and basal area.Twenty-three mixed Oak and Pinus massoniana natural forests in nine areas of Hunan province were selected to construct the growth models of individual volume and basal area.Through analysis,we can know that Richards model was the best model for volume and basal area growth of the two tree species,the determining coefficients(R2)were the highest and the sum of squares of residuals(SSE)were the smallest.The result of mixed effect models show that the random effects of the Oak single-wood volume and basal area growth model are best at parameter a1,while the volume and basal area growth model of Pinus massoniana are best at parameter b1 and a1+b1.Compared with the basic model,the fitting precision(P)and determination coefficients of the mixed effect model are improved.Mean error(ME),mean absolute error(MAE),mean percent error(RME)and mean absolute percent error(RMAE)are all slumped substantially.The results show that wood volume and basal area growth of natural forests are greatly affected by the differences between tree species and site types.Mixed-effect model can take full account of the differences between sites,and solve this problem well.