榆林沙区樟子松人工林胸径-树高生长模型研究

Study on DBH-Height Growth Model of Pinus sylvestris Plantation in Yulin Sandy Area

为研究榆林沙区樟子松胸径-树高生长模型,填补榆林沙区樟子松在胸径-树高生长模型研究的空白,为榆林沙区樟子松林分生长预测、生产经营管理提供参考依据,以榆林沙区樟子松人工林为研究对象,选用15个常用的胸径-树高生长模型,通过5个评价指标(R^(2)、决定系数(RMSE)、平均绝对残差(MAE)、残差平均和(SSE)、Akaike信息量准则(AIC))进行对比分析拟合,从而进一步确定,适宜的榆林沙区樟子松胸径-树高生长模型。结果表明:除了Gompertz(1825)模型(14号模型)无法输出参数,剩余的模型均可以。在剩余的14个模型中,双曲线型(2号模型)、混合型(5号模型)、二次多项式(6号模型)、Korf(10号模型)、修正Veibull(11号模型)这5个模型评价指标较优,拟合效果较好。综合比较分析可知,这5种基础胸径-树高模型中,修正Veibull模型可以更好地拟合樟子松胸径-树高的关系,精度也比较准确,建议选用修正Veibull模型。

Studying the DBH-tree height growth model of Pinus sylvestris in Yulin sandy area can fill the gap in the research on DBH-tree height growth model of Pinus sylvestris in Yulin sandy area.It can also provide a reference for the growth prediction and production management of Pinus sylvestris in the sandy area of Yulin.This study takes the Pinus sylvestris plantation in Yulin sandy area as the research object,selects 15 commonly used DBH-tree height growth models.Through 5 evaluation indicators(R^(2)),coefficient of determination(RMSE),mean absolute residual error(MAE),residual mean sum(SSE),Akaike information content criterion(AIC)for comparative analysis and fitting.Therefore,the most suitable DBH-tree height growth model of Pinus sylvestris in the sandy area of Yulin is further determined.The result shows:except for the Gompertz(1825)model(model number 14),which cannot output parameters,the rest of the models can.Among the remaining 14 models,these five model evaluation indicators are better,and the fitting effect is better,including the hyperbolic model(Model No.2),the hybrid model(Model No.5),the quadratic polynomial model(Model No.6),the Korf model(Model No.10),the modified Veibull model(Model No.11).Finally,a comprehensive comparative analysis shows that the modified Veibull model among the five basic DBH-tree height models can better fit the relationship between the DBH and tree height of Pinus sylvestris,and the accuracy is relatively accurate.It is recommended to use the modified Veibull model.