兴安落叶松立木材积和树皮材积可加性模型研建

Study on the additivity of individual tree volume and bark volume model of Larix gmelinii Rupr.

【目的】确保立木材积和树皮材积预测的一致性并提高预测精度。【方法】以大兴安岭兴安落叶松为研究对象,分别采用控制法和分解法研建了可加性模型系统。利用SAS统计软件模型模块proc model中的NSUR法进行拟合及参数估计。拟合结果采用确定系数(R2)和均方根误差(RMSE)进行评价;检验结果则通过确定系数(R2)、均方根误差(RMSE)、平均相对误差(MRE)、平均误差绝对值(MAB)和相对误差绝对值(MPB)进行评价。【结果】从模型的整体评价结果来看,两种方法的拟合和检验效果均很好,基于分解法构建的模型略优于基于控制法构建的模型;不同径阶的检验表明,对于中等径阶的树木(20≤D<36 cm),基于控制法的模型相对较好,而对于小径阶(5≤D<20 cm)和大径阶的树木(D≤36 cm),基于分解法的带皮、去皮、树皮材积模型的预测精度要比基于控制法的各立木材积模型要稍好。【结论】总的来说,两种可加性模型系统均能很好地预测单木带皮材积、去皮材积和树皮材积,并确保得到满足一致性的预测结果,在具体应用时可根据实际情况选择适合的可加性材积模型系统。

【Objective】In order to ensure the consistency of the tree volume and bark volume estimation and improve the prediction precision.【Method】Two additive volume model systems(total volume,volume inside bark,and bark volume)were established from total control and total decomposition methods based on Larix gmelinii Rupr.in Daxinganling.All models were fitted using NSUR in SAS proc model.The fitting results were evaluated using the determination coefficient(R2)and the root mean square error(RMSE);the test results were evaluated by determining coefficient(R2),root mean square error(RMSE),mean relative error(MRE),mean error absolute value(MAB)and relative error absolute value(MPB).【Result】The results show that from the overall evaluation results of the model,the fitting and testing results of the two methods are very good.The model based on decomposition method is slightly better than that based on control method;From evaluation results of different diameter classes testing,for medium diameter classes(20≤D<36 cm),the method based on control method is better.For the small diameter classes(5≤D<20 cm)and large diameter classes(D≤36 cm),the prediction precision of volume models with total,inside bark,and bark based on decomposition method is slightly better than that of volume models based on control method.【Conclusion】In general,both additive model systems show good prediction for total volume,volume inside bark,and bark volume,and ensure consistent predictions.The model system based on the decomposition method significantly improves the prediction precision of the bark volume.In the specific application,the appropriate additive volume model system can be selected according to the actual situation.