桉树人工林生长模型研究进展

A Review of Eucalyptus Plantation Growth Model

桉树人工林生长模型包括传统统计模型和现代动态模型。传统统计模型方法包括单因素回归模型、多元回归模型、灰色系统模型、单一分布拟合法、多模型优选法、逐步回归法及其他改进方法。这些统计模型虽然简单、拟合精度高,但只是静态的经验或半经验模型,并未阐明机理,普适性很差,对于动态反馈也无能为力。现代动态模型主要包括3-PG过程模型、人工神经网络模型和系统动力学模型。与传统统计模型相比,动态模型相对较复杂、精度较低,但能够满足多因素交互、非线性预测、动态模拟、多目标决策的研究要求。文中分析了桉树人工林生长模型的特点、各研究方法的比较及系统动力学在林业中的应用,认为系统动力学模型更适合涉及多因素反馈的桉树人工林生长模型的研究。

Eucalyptus plantation growth models could be classified into traditional statistical models and modern dynamic models. Traditional statistical models include univariate regression analysis, multiple regression analysis, gray system model, single distribution fitting, multi - model optimization, stepwise regression and other refined methods. Though the traditional statistical models are simple and accurate in fitting, they are only empirical or semi - empirical models, which cannot present the mechanisms and the dynamic feedbacks, and consequently cannot be applied widely. Modern dynamic models include 3 - PG process model, artificial neural network model and system dynamic model. Modern dynamic models are complex but low in accuracy, while they can meet the requirements in the studies of multi - factor interaction, nonlinear prediction, dynamic simulation, and multi - objective decision. It was recommended that the system dynamic model be more suitable to apply in the study of Eucalyptus plantation growth models.