含海拔哑变量的杉木优势木高-树龄曲线模型研究  

Study on multi-form site index curve model of Cunninghamia lanceolata plantation with elevation dumb variable

作　　者：易烜 黄朗 朱晋梅 朱光玉[2] 王琢玙 黄靓 YI Xuan;HUANG Lang;ZHU Jinmei;ZHU Guangyu;WANG Zhuoyu;HUANG Liang(Hunan Qingyang Lake State-owned Forest Farm,Ningxiang 410627,Hunan,China;Central South University of Forestry and Technology,Changsha 410004,Hunan,China;Hunan Biodiversity Conservation Center,Changsha 410004,Hunan,China)

机构地区：[1]湖南省青羊湖国有林场,湖南宁乡410627 [2]中南林业科技大学,湖南长沙410004 [3]湖南省生物多样性保护中心,湖南长沙410004

出　　处：《湖南林业科技》2023年第2期33-40,共8页Hunan Forestry Science & Technology

基　　金：湖南省林业科技创新专项项目(XLK202001)。

摘　　要：人工纯林立地质量评价通常采用立地指数法,即利用林分平均优势木高和树龄的关系来评价立地质量。本文以湖南青羊湖、高泽源、江华等45个国有林场的518个杉木纯林标准地为研究对象,利用其中388个标准地数据,对林分平均优势木高与树龄的关系进行理查德、单分子式和舒马切尔3种常用的非线性基础模型方程模拟。通过对比分析,选出最优为理查德方程(参数a、b、c分别为23.67023、0.05856、1.59679),其决定系数为0.79489。以最优模型为基础,考虑到不同海拔对优势木高生长过程的影响,将海拔划分等级变量F(以100 m为一个等级,共分为12个等级),纳入含海拔哑变量(F)的优势木高-树龄生长曲线,并以绝对残差均值ARM、决定系数DC、模型精度3个模型精度检验指标对模型进行精度检验。模拟结果表明:加入以海拔等级为哑变量的3个模型精度较高,其中,将哑变量(F)加在参数b上的模型精度最高(绝对残差均值1.20630、决定系数0.8444、模型精度97.77%),并利用130个标准地数据进行再验证,证明该模型可行性强,可为区域性杉木立地质量评价提供建模依据。The quality evaluation of artificial pure forest land usually adopts the site index method,it means,the relationship between high average stand advantage and age is used to evaluate site quality.In this paper,518 Cunninghamia lanceolata pure forest data from 45 state-owned forest farms including Qingyang Lake,Gaozeyuan and Jianghua in Hunan Province were used as the research object,including 388 data for modeling and 130 data for model applicability testing.Three commonly used nonlinear equations(Richard,Single molecule formula and Schumachel)were used to simulate the relationship between high stand average dominance and age.The model results were compared and analyzed,and the optimal Richard's equation was selected with a determination coefficient of 0.79489.After then,based on the optimal model,considering the influence of different altitudes on the dominant high growth process,the altitude was divided into grade variables F(in this paper,100m was a grade,a total of 12 grades),the regression analysis method containing dumb variable(F)was used to construct the dominant high-age growth curve,and the accuracy test of the model was carried out by three model accuracy test indicators:absolute residual mean ARM,coefficient of determination DC,and model accuracy,and the feasibility of the model was analyzed.The results were shown that those three models with altitude grade as the dumb variable have high accuracy,among which the model with the dumb variable(F)added to the parameter b had the highest accuracy(mean absolute residuals:1.20630,determination coefficient:0.8444,model accuracy:97.77%),and had strong applicability.The research results provide a modeling basis for the quality evaluation of regional C.lanceolata sites.

关 键 词：杉木纯林 立地指数 曲线模型 哑变量 立地质量评价 

分 类 号：S791.27[农业科学—林木遗传育种]