利用分段建模方法建立海南省主要树种立木材积模型

Using segmented modeling method to establish volume models for main tree species in Hainan province

为提高二元立木材积模型的通用性,以海南省桉树、木麻黄、马占相思、橡胶树、加勒比松5种树种为研究对象,采用分段建模的方法改进立木材积估计,并与常规建模方法进行比较分析,结果表明:从模型的整体评价结果来看,常规模型和分段模型的拟合效果均很好,分段模型整体总相对误差和平均系统误差两个指标要优于常规模型;对各个径阶进行检验,分段模型基本保证各径阶下总相对误差和平均系统误差在±3%以内,而常规模型各径阶下总相对误差和平均系统误差在±6%以内。分段模型方法可以有效提高二元立木材积模型的通用性,该方法可以推广应用到林业数表领域。

In this paper, we take the main tree species in Hainan province（Eucalyptus, Casuarina equisetifolia, Acacia mangium, rubber, Caribbean pine） as the objects of study and use segmented modeling method to improve model universal property. Comparing the results of segmented models and conventional models（direct establishing two-way volume model）, the results show that 1） From the overall evaluation results of models, the fitting effects of the conventional models and the segmented models are very good, and the two indexes, TRE and MSE, of segmented model are better than the conventional model. 2） From evaluation results of different diameter classes testing, the segmented model can basically ensure that the TRE and MSE are between-3% and ＋3%. For conventional model, TRE and MSE are between-6% and ＋6%. Therefore, segmented modeling method can effectively improve universal property for two-way volume model and can be applied to the field of other forestry tables.