带度量误差的全林整体模型参数估计研究

Parameter estimate of the whole stand model with measurement error

该文以广西大青山马尾松全林整体模型为基础模型,进行模拟实验设计,并构造株数、直径、优势高、平均高和形高5个观测因子都有度量误差的模拟观测数据,用非线性度量误差联立方程组方法,对5个观测因子都有度量误差的全林整体模型进行估计.结果表明,与最小二乘估计方法一样,优势高、平均高和形高模型没有明显的系统偏差,与通常最小二乘估计方法不同的是断面积、直径和株数模型也没有出现明显的系统偏差.得出结论,在建立有度量误差的全林整体模型时,非线性度量误差联立方程组方法明显优于最小二乘法.

Based on the whole stand model of Autonomous Region, a simulation experiment Pinus massoniana in Daqingshan was devised. The observations of Mountain, Guangxi Zhuang stand density, average diameter, dominant height, average height and form height with measurement error were constructed. Using the nonlinear measurement error simultaneous equation method, the parameters of whole stand model were estimated and results indicated that the dominant height model, average height model and form height model had no clear bias just as the results of the least square method, while basal area model, average diameter model and stand density model had no bias, which was contrary to the results of the least square method. Therefore, the nonlinear measurement error simultaneous equation method was better than the least square method.