摘要
依据黑龙江省孟家岗林场49株人工落叶松1179个圆盘和轮盘数据,分析了心材半径的纵向变化规律.结果表明:心材半径随树高增高而逐渐减小,与树干外形基本一致,其中去皮半径(XR)、胸径(DBH)及形成层年龄(CA)与心材半径之间关系较显著,利用逐步回归分析建立落叶松心材半径(HR)和面积(HA)模型:HR=b_1+b_2XR^2+b_3CA+b_4XR,HA=b_1+b_2DBH·XR+b_3CA+b_4DBH·XR^2.应用AIC、BIC、对数似然值以及似然比检验等模型评价指标,对利用样地、样木效应拟合的心材半径和面积模型进行比较.当考虑样木效应拟合心材半径和面积模型时,将b_1、b_2、b_3作为混合参数得出的模型最好.混合模型的预测精度高于基本模型.在应用上,总体心材半径和面积可以通过混合模型来预测.采用Beta回归模型模拟了心材比例,模型中各参数均显著,决定系数较高,模型模拟效果较好.
Based on the data of 1179 discs and whorls of 49 trees from larch（Larix olgensis）plantations located in Mengjiagang forest farm in Heilongjiang Province,China,we analyzed the longitudinal variation pattern of heartwood radius.The results showed that the heartwood radius decreased with the increases of tree height,which was basically the same as the trunk shape.The relationship between the xylem radius（XR）,diameter at breast height（DBH）and cambial age（CA）with the heartwood radius was significant.The stepwise regression analysis was used to develop heartwoodradius（HR）and heartwood area（HA）models：HR=b1＋b2XR^2＋b3CA＋b4XR,HA=b1＋b2DBH·XR＋b3CA＋b4DBH·XR^2.We used the evaluation statistics such as AIC,BIC,Log Likelihood and Likelihood ratio test to compare the heartwood radius and heartwood area models which fitted with the plot effect and tree effect.The heartwood radius and heartwood area models with parameters b1,b2,b3as mixed effects performed best when the tree effect was considered.The prediction accuracy of the mixed model was better than that of the basic model.In the application,the total heartwood radius and area could be predicted by the mixed model.Beta regression model was used to simulate the heartwood proportion.In this model,all parameters were significant,and the coefficients of determination were relatively high,with a good simulation effect.
作者
贾炜玮
朱飞燕
李凤日
JIA Wei-wei;ZHU Fei-yan;LI Feng-ri(College of Forestry,Northeast Forestry University,Harbin 150040,China)
出处
《应用生态学报》
CAS
CSCD
北大核心
2018年第7期2277-2285,共9页
Chinese Journal of Applied Ecology
基金
国家重点研发计划项目(2017YFD0600404-2)
中央高校基本科研业务费专项(2572014CA17)
国家自然科学基金项目(31570626)资助~~
