长白山主要树种直径生长的多元回归预测模型:以云杉为例  

Multiple Regression Prediction Model for Diameter Growth of Dominant Tree Species in Changbai Mountains: A Case Study of Picea koraiensis

作　　者：刘洋[1] 亢新刚[1] 郭艳荣[2] 高北延 冯启祥[3] 

机构地区：[1]省部共建森林培育与保护教育部重点实验室(北京林业大学),北京100083 [2]北京林业大学 [3]吉林省汪清林业局

出　　处：《东北林业大学学报》2012年第2期1-4,共4页Journal of Northeast Forestry University

基　　金：林业公益性行业专项(200804027)

摘　　要：选取长白山地区18块皆伐标准地的949株云杉解析木和78株云杉标准木,建立长白山主要树种直径生长的多元回归预测模型。在生长模型中,首先确定林木直径的潜在生长量,建立疏开木的直径潜在生长方程,然后用反映立地质量、林分密度、对象木与周围相邻木最近距离3个竞争指数所表示的修正函数对直径潜在生长函数进行修正。采用SPSS软件对参数进行拟合,依据相关系数最大,剩余平方和最小得到直径的实际生长模型。该模型可以合理预测长白山地区主要树种一定时期内未来直径大小和林分的直径结构,这对把握林分动态、预估林分生长状态有极其重要的意义。另外,用未参加建模的8块皆伐标准地的432株云杉解析木和23株云杉标准木,对直径生长的多元回归预测模型的拟合度进行检验,结果显示拟合效果很好。A total of 949 trees from 18 clear-cut stands and 78 standard trees of Picea koraiensis in Changbai Mountains were selected to establish a multiple regression prediction model for diameter growth of dominant tree species.In the course of model building,the potential increment of tree diameter was determined to build an equation for diameter growth of open-grown trees.Then three competition indexs were applied to express modified function as a quantitative index to measure mean site quality,stand density,and distances to the nearest neighboring trees in order to correct the potential increment of tree diameter.The data were analyzed with SPSS software.The functions with the highest correlation coefficient and lowest surplus sum of squares were chosen as the realistic diameter growth model.The model could reasonably predict the diameter size of the dominant tree species and diameter structure of the future stand,which has a great significance for mastering the stand dynamics and estimating stand growth condition.In addition,the fitting degree was checked with 432 trees of the analytic trees form 8 clear-cut stands and 23 standard trees of P.koraiensis which did not participate model building.It is found that the practical application error of the regression model is small,and the model can achieve the better fitting effect and forecast precision.

关 键 词：直径生长 多元回归预测模型 天然异龄林 长白山 

分 类 号：S758.5[农业科学—森林经理学]