森林空间数据的统计与仿真

Statistics and Simulation of Forest Spatial Data

该文介绍一种简单有效的森林空间数据的仿真方法．假设森林中树高为正态分布，单株林木的高度与邻近木相关，与距离远的林木相关极小．利用一个协方差估计公式来描述这种相关性，并以一人工落叶松林样地为例，计算出估计公式中的特征参数ａ，在此基础上用ＣｈｏｌｅｓｋｙＤｅｃｏｍｐｏｓｉｔｉｏｎ方法建立树高空间分布的仿真模型，在与原样地相同的空间位置上产生出一组与样地数据无偏的虚拟树高数据．通过设定仿真程序的各种输入值，如平均值、方差、株距等。

The height of tree in a forest is assumed as a random process with a normal distribution. It is supposed that a single tree's height is closely related with its adjacent trees and almost not related with far trees.One of covariation estimation formulas of spatial data is used to describe such relation.The special parameter a in this formula is estimated accounting to a set of height spatial data in an artificial larch sampling plot. A set of height spatial data in a virtual artificial larch forest is generated by using a simulation model based on Cholesky Decomposition method.The virtual data and plot data are similar in the values of mean and variation.Also different types of virtual set can be simulated by giving different locations of tree,mean and variation.