基于Cox比例风险函数及混合效应的落叶松云冷杉混交林林木枯损模型研究  

Mortality Model of Larix olgensis-Abies nephrolepis-Picea jazoensis Mixed Stands Based on Cox Proportional Hazard Function and Mixed Effect Model

作　　者：李春明[1] LI Chun-ming(Research Institute of Forest Resource Information Techniques,Chinese Academy of Forestry,Beijing 100091,China)

机构地区：[1]中国林业科学研究院资源信息研究所,北京100091

出　　处：《林业科学研究》2020年第3期92-98,共7页Forest Research

基　　金：国家自然科学基金面上项目“基于混合效应模型的联立方程组及概率分布模型在模拟森林生长中的方法研究”(31570625)。

摘　　要：[目的]将生存分析方法和混合效应模型方法相结合,构建林木枯损模型,提高模型的模拟精度。[方法]以吉林省汪清林业局20块落叶松云冷杉林样地数据为材料,基于生存分析方法中的Cox比例风险函数模型方法,把林分因子和立地因子作为协变量加入到模型中去,构建林木的枯损及生存模型,并在此基础上考虑样地水平的随机效应,最后与不考虑样地水平随机效应的模拟效果进行比较分析。[结果]表明,Cox比例风险函数模型在描述林木枯损时,具有很好的适应性。单木初始胸径与林木的风险函数呈反比,与生存率呈正比;大于对象木断面积与风险函数呈正比,与生存率呈反比;初始林分公顷株数与风险函数呈正比,与生存率呈反比;立地因子对林木的枯损及生存没有显著影响。与固定效应模型相比,Cox比例风险函数模型在考虑了样地水平的随机效应后,模型的模拟精度获得明显的提高,并且达到极显著程度。由于大于对象木断面积和初始林分公顷株数两个变量在考虑了样地水平的随机效应后影响不显著,最后只考虑了单木初始胸径一个变量对枯损的影响,与不考虑随机效应相比,差异也达到显著水平。[结论]林木本身的大小对自身的枯损具有明显的影响,胸径小的林木较胸径大的林木更易枯损。在森林经营中,Cox比例风险函数模型的使用,可为森林经营者在确定合理的经营密度上提供很好的科学依据。[Objective]The survival analysis method and mixed effect model are combined to construct tree mortality model and improve the simulation precision of the model.[Method]Taking 20 plots of Larix olgensis-Abies nephrolepis-Picea jazoensis mixed stand in Wangqing Forestry Bureau of Jilin Province as examples,the tree mortality and survival model were constructed based on Cox proportional hazard function of survival analysis method,and the stand factor and site factor was added into the model as covariates.The plot’s random effect was considered and compared with the simulation effect of the traditional model.[Result]It showed that the Cox proportional hazard function model had fitting goodness in describing tree mortality.The initial DBH of tree was negatively correlated with the hazard function of the tree,and positively correlated with survival rate.The BAL was positively correlated with the hazard function of the tree,and negatively correlated with survival rate.The initial stand density per hectare was positively correlated with the hazard function of the tree,and negatively correlated with survival rate.Compared with the fixed effect model,the accuracy of Cox proportional hazard function model was greatly improved after considering plot’s random effect.Because the BAL and initial stand density per hectare showed no significant effect after considering the plot’srandom effect,only the initial DBH was considered,and the difference also reached a significant level.[Conclusion]In forest management,the Cox proportional hazard function model provides a good basis for determining the reasonable management density for forest operators.

关 键 词：生存分析方法 混合效应模型 枯损 Cox比例风险函数 风险率 

分 类 号：S758.4[农业科学—森林经理学]