首页|期刊导航|林业科学|浙江省马尾松天然林枯损与进界模型构建及影响因素分析

浙江省马尾松天然林枯损与进界模型构建及影响因素分析OA

Model Construction and Influencing Factors Analysis of Mortality and Ingrowth for Natural Pinus massoniana Forests in Zhejiang Province

中文摘要英文摘要

[目的]构建浙江省马尾松天然林枯损与进界模型,分析影响林分动态的关键因素,为该地区马尾松林的可持续经营提供理论依据.[方法]基于浙江省第五次(1994 年)、第六次(1999 年)、第七次(2004 年)和第八次(2009 年)森林资源连续清查固定样地数据,筛选出马尾松天然林样地389 块,其中311 块作为模拟样本、78 块作为检验样本.采用 Kendall's tau-b相关系数筛选林分与立地因子,分别构建泊松模型(Poisson)、负二项模型(NB)、零膨胀泊松模型(ZIP)、零膨胀负二项模型(ZINB)、Hurdle-泊松模型(HP)和 Hurdle-负二项模型(HNB)共 6类广义线性模型,对林分枯损与进界过程进行拟合与比较.利用赤池信息准则(AIC)、贝叶斯信息准则(BIC)、-2倍对数似然函数值(-2logL)和 Vuong检验比较各模型间的拟合效果,选取平均绝对误差(MAE)和均方根误差(RMSE)评价其预测效果.[结果]1)ZINB模型在枯损与进界模拟中表现最优;2)负二项系列模型(NB、ZINB、HNB)在拟合效果上显著优于相应的泊松系列模型(Poisson、ZIP、HP),ZIP与 HP模型拟合效果相当,ZINB模型拟合效果略优于HNB模型;3)林分枯损主要受林分密度、林分胸高断面积、阔叶树胸高断面积和海拔的影响;4)林分进界主要受林分密度、林分胸高断面积和阔叶树胸高断面积的影响,立地因子对进界无显著影响.[结论]林分密度是调控枯损与进界过程的核心变量,但其作用方向相反;阔叶树组成增加有助于降低马尾松林枯损风险并促进更新.建议在马尾松林实际经营中合理控制林分密度并适当补植阔叶树种,从而维持马尾松林的长期健康与生产力.

[Objective]This study aims to develop stand-level mortality and ingrowth models for natural Pinus massoniana forests in Zhejiang Province,and identify the key factors influencing stand dynamics,so as to provide a theoretical basis for the sustainable management of P.massoniana forests in this region.[Method]Based on the data from the 5th(1994),6th(1999),7th(2004),and 8th(2009)Continuous Forest Inventories in permanent sample plots of Zhejiang Province,a total of 389 sample plots of natural P.massoniana forests were selected,with 311 used for model simulation and 78 for validation.The Kendall's tau-b correlation coefficient was used to screen stand and site factors.Six types of generalized linear models,including the Poisson model(Poisson),the negative binomial model(NB),the zero-inflated Poisson model(ZIP),the zero-inflated negative binomial model(ZINB),the Hurdle-Poisson model(HP),and the Hurdle-negative binomial model(HNB),were constructed to fit and compare the stand mortality and ingrowth processes.The fitting performance of each model was compared using Akaike information criterion(AIC),Bayesian information criterion(BIC),-2 times the log-likelihood(-2logL),and the Vuong test.The prediction performance was evaluated using mean absolute error(MAE)and root mean square error(RMSE).[Result]1)The ZINB model demonstrated the best overall performance for both mortality and ingrowth simulations.2)The negative binomial series models(NB,ZINB,HNB)showed significantly better fit than their corresponding Poisson series counterparts(Poisson,ZIP,HP),with comparable performance observed between the ZIP and HP models,and a slight advantage of the ZINB model over the HNB model.3)Stand mortality was primarily influenced by stand density,stand basal area at breast height,broad-leaved tree basal area at breast height,and elevation.4)Stand ingrowth was mainly affected by stand density,stand basal area at breast height,and broad-leaved tree basal area at breast height.The site factors showed no significant influence on ingrowth.[Conclusion]Stand density is a core variable that simultaneously regulate both mortality and ingrowth processes,but with opposite effects.The increase in the proportion of broad-leaved trees helps reduce mortality risk in P.massoniana stands and promotes its regeneration.It is recommended that practical forest management for P.massoniana should involve in rational control of stand density and appropriate enrichment planting with broad-leaved species,thereby maintaining the long-term health and productivity of P.massoniana forests.

毛阳骏;马立帮;郭昊天;李同欣;王懿祥

浙江农林大学环境与资源学院/碳中和学院 杭州 311300||浙江农林大学浙江省农林生态系统固碳减排重点实验室 杭州 311300浙江农林大学环境与资源学院/碳中和学院 杭州 311300||浙江农林大学浙江省农林生态系统固碳减排重点实验室 杭州 311300浙江农林大学浙江省农林生态系统固碳减排重点实验室 杭州 311300||浙江农林大学林业与生物技术学院 杭州 311300浙江农林大学环境与资源学院/碳中和学院 杭州 311300||浙江农林大学浙江省农林生态系统固碳减排重点实验室 杭州 311300浙江农林大学环境与资源学院/碳中和学院 杭州 311300||浙江农林大学浙江省农林生态系统固碳减排重点实验室 杭州 311300

农业科技

马尾松枯损模型进界模型广义线性模型零膨胀模型

Pinus massonianamortality modelingrowth modelgeneralized linear modelzero-inflated model

《林业科学》 2026 (10)

82-92,11

"十四五"国家重点研发计划项目"南方典型次生林全周期多功能经营技术研究"(2022YFD2200504).

10.11707/j.1001-7488.LYKX20260068

评论