关键词: 青冈栎, 林分条件平均高, 林分形高模型, 林分蓄积量, 留一交叉验证

Abstract:

Exploring the construction and evaluation method of the stand form height model could provide a theoretical basis for the precise measurement of stand volume of Quercus glauca secondary forest in Hunan Province.Based on the measured data of 1055 individual trees from 35 plots,we established the univariate form height model and the binary form height model of Quercus glauca secondary forest in Lutou Forestry Farm and Qingyanhu Forestry Farm.The Quercus glauca tree height curve was derived from the Kunze,M.stem curve to calculate the stand condition mean height.Based on the univariate and binary correlations een form height,stand diameter at breast height and stand condition mean height,6 univariate basic models and 8 binary basic models were selected,and the coefficient of determination(R²),root mean square error(RMSE),explained variance(EA),mean absolute error(MAE)and model prediction accuracy(P₁)were used as evaluation indicators.After evaluation and comparison,the univariate form height model was found to be f(D)=2.870 7×D^{0.261 8}、f(H)=2.022 4+0.339 5×H and the binary form height model was f(D,H)=0.590 6×exp(-0.008 3×D)H、f(D,H)=(0.710 2-0.070 1×logD)H and f(D,H)=2.479 9×(D²×H)^{0.109 2}.The models were tested through cross Leave-One-Out Cross-Validation and Paired T-test.The results showed that both univariate and binary stand form high models had high accuracy and strong applicability.It verified that the method of constructing a stand form height model based on the stand mean diameter at breast height and the stand condition mean height is reasonable and feasible.It can also provide demonstration and reference for the construction of other types of stand form height models and the preparation of stand form height tables.

Key words: Quercus glauca, stand condition mean height, stand form height model, stand volume, Leave-One-Out cross-validation