文献类型：期刊文献

中文题名：基于贝叶斯模型平均法和逐步回归法构建杉木单木胸径生长模型

英文题名：Individual Tree Diameter Growth Model of Chinese Fir Plantations Using Bayesian Model Averaging and Stepwise Regression Approaches

作者：鲁乐乐[1,2] 王震[1] 张雄清[1,2] 张建国[1]

第一作者：鲁乐乐

机构：[1]中国林业科学研究院林业研究所国家林业和草原局林木培育重点实验室,北京100091;[2]南京林业大学南方现代林业协同创新中心,南京210037

年份：2021

卷号：57

期号：9

起止页码：87-97

中文期刊名：林业科学

外文期刊名：Scientia Silvae Sinicae

收录：CSTPCD;;EI(收录号：20214811256403);Scopus;北大核心:【北大核心2020】；CSCD:【CSCD2021_2022】；

基金：国家自然科学基金项目(31670634)。

语种：中文

中文关键词：单木胸径生长量;气候变量;贝叶斯模型平均法;逐步回归法;林分变量因子

外文关键词：individual tree diameter growth;climate variables;Bayesian model averaging(BMA);stepwise regression(SR);stand variables

分类号：S758.5

摘要：【目的】探索杉木人工林单木胸径生长量变化的驱动因子,比较不同驱动因子的重要性,构建不确定性单木胸径生长模型,为杉木经营管理者科学经营管理杉木人工林提供参考。【方法】以福建省邵武市卫闽林场杉木密度试验林为研究对象,采用贝叶斯模型平均法(BMA)和逐步回归法(SR)分析杉木单木胸径生长量与内部因子(林分变量因子)和气候因子的关系,构建杉木单木胸径生长模型。【结果】杉木单木胸径年均生长量受气候因子影响较小,主要受竞争因子和单木大小因子影响。单木胸径生长量随林分密度、林分平方平均胸径、大于对象木的断面积和、年龄、冬季平均最低温度增加而减小,随期初胸径、胸高断面积、优势木平均高、最冷月平均温度、最热月平均温度、年均降雨量增加而增加。基于SR获得模型的后验概率小于BMA获得最佳模型(最高后验概率)或SR模型不在BMA模型空间前几个后验概率高的模型中。【结论】杉木单木胸径生长量随竞争增加而减小,随温度和降雨增加而增加。贝叶斯模型平均法考虑所有可能变量的组合,能够反映出模型的不确定性。
【Object】Individual tree diameter growth model is one of forest growth and yield basic models.Stepwise regression is widely used in the selection of model variables.However,this method ignores the model uncertainty caused by the variable selection process.So the driving factors of individual tree diameter grouth of Chinese fir(Cunninghamia lanceolata)plantation were explored,the importance of different driving factors was compared,and the uncertain individual tree diameter growth model was constructed,in order to provide reference for Chinese fir managers to manage Chinese fir plantation scientifically.【Method】Data for this study was sampled from Chinese fir stands in Weimin,Shaowu city,Fujian Province.Bayesian model averaging(BMA)and stepwise regression(SR)were used to analyze the effects of endogenous and climatic factors on the individual tree diameter growth of Chinese fir.【Result】Competition and individual tree size were the main factors affecting annual diameter growth comparing with climate factors.Diameter growth decreased with the increase of tree number per hectare,quadratic mean diameter,sum of basal areas of trees larger than the subject tree(BAL),age and winter mean minimum temperature,whereas increased with the increase of diameter at the beginning of growth period,stand basal area,dominant height,mean coldest month temperature,mean warmest month temperature and mean annual precipitation.For most of the treatments of the four models,the posterior probability of the model obtained by SR was smaller than that of the best model obtained by BMA(which exhibited the highest posterior probability).In some cases,SR models did not belong to the top several models with higher posterior probability in the BMA model space.【Conclusion】The diameter increment decreased with increasing competition and increased with increasing temperature and precipitation.BMA considered the combination of all possible variables and reflected the uncertainty of the model.