基于随机森林模型的东北三省乔木林生物质碳储量预测_田惠玲.pdf

第 58 卷 第 4 期2 0 2 2 年 4 月林业科学SCIENTIASILVAESINICAEVol.58,No.4Apr.,2 0 2 2doi:10.11707/j.1001-7488.20220405收稿日期:2021-11-16;修回日期:2022-03-12。基金项目:“十三五”国家重点研发计划课题“人工林生产力形成的关键生理生态与环境控制机制”(2016YFD0600201)。朱建华为通讯作者。基于随机森林模型的东北三省乔木林生物质碳储量预测田惠玲1 朱建华1,2何潇3陈新云4简尊吉1李宸宇1郭学媛1黄国胜4肖文发1,2(1.中国林业科学研究院森林生态环境与自然保护研究所国家林业和草原局森林生态环境重点实验室北京 100091;2.南京林业大学南方现代林业协同创新中心南京 210037;3.中国林业科学研究院资源信息研究所国家林业和草原局森林经营与生长模拟实验室北京 100091;4.国家林业和草原局调查规划设计院北京 100714)摘要:【目的】利用全国森林资源清查固定样地连续监测数据,通过机器学习算法构建基于多因子的森林生长模型,提高森林生长和固碳量的模拟精度,预测东北三省乔木林未来碳汇潜力,探索乔木林碳汇的潜在分布,为准确定位我国东北森林在增汇减排中的作用以及科学制定国家“碳中和”行动路径和目标管理提供科学指导。【方法】利用 19992018 年 4 次全国森林资源连续清查固定样地监测数据,结合区域气候、土壤、林分和地形因子,采用随机森林模型构建区域主要优势树种(组)的生长-消耗模型,运用未来气候情景与未来乔木林面积扩增情景,预测东北三省 20152060 年间乔木林生物质碳储量变化与碳汇潜力。【结果】东北三省乔木林生物质碳储量 2060年可达 3 393.15 TgC,比 2015 年增加 1 895.23 TgC,20152060 年间年碳汇量为 42.12 TgCa-1,其中天然林是主体。辽宁省、吉林省和黑龙江省乔木林生物质碳储量分别由 2015 年的 139.19、463.58 和 895.15 TgC 增至 2060 年的 328.95、915.83 和 2 148.37 TgC,乔木林平均生物质碳密度分别由 2015 年的 32.71、59.75 和 45.11 MgChm-2增至 2060 年的 75.20、109.32 和 85.24 MgC hm-2。20152060 年间辽宁省、吉林省和黑龙江省乔木林生物质年碳汇量分别为 4.22、10.05 和 27.85 TgC a-1。【结论】本研究构建的随机森林模型表现效果较好,能够用于东北三省未来乔木林碳储量预测。20152060 年东北三省乔木林生物质碳储量将增加 1 895.23 TgC,未来仍具有较大碳汇潜力。黑龙江省的碳汇潜力最大,年碳汇量达 27.85 TgC a-1,是未来重要的碳增汇区域;辽宁省的碳汇潜力较弱,年碳汇量仅为 4.22 TgC a-1。加强中、幼龄林经营管理,适度更新成、过熟林,有助于提升东北三省乔木林碳汇功能,发挥我国东北森林在增汇减排以及实现区域“碳中和”目标中的作用。关键词:生长-消耗模型;随机森林模型;森林资源清查;多变量;生物质碳储量中图分类号:S718.55文献标识码:A文章编号:1001-7488(2022)04-0040-11Projected Biomass Carbon Stock of Arbor Forest of Three Provinces in Northeastern China Based on Random Forest ModelTian Huiling1Zhu Jianhua1,2He Xiao3Chen Xinyun4Jian Zunji1Li Chenyu1Guo Xueyuan1Huang Guosheng4Xiao Wenfa1,2(1.Ecology and Nature Conservation Institute,CAFKey Laboratory of Forest Ecology and Environment,National Forestry and Grassland AdministrationBeijing 100091;2.Co-Innovation Center for Sustainable Forestry in Southern China,Nanjing Forestry UniversityNanjing 210037;3.Institute of Forest Resource Information Techniques,CAFKey Laboratory of Forest Management and Growth Modelling,National Forestry and Grassland AdministrationBeijing 100091;4.Academy of Forest Inventory and Planning,National Forestry and Grassland AdministrationBeijing 100714)Abstract:【Objective】Based on the data of permanent monitoring plots of the successive national forest inventories,a multi-factor forest growth model was constructed through the machine learning algorithm,which improves the simulation accuracy of forest growth and carbon sequestration,predicts the future carbon sink potential of arbor forests of the three provinces in northeast China,and explore the potential distribution of arbor forest carbon sink.It provides a scientific guidance for accurately locating the role of forests in northeast China in increasing sinks and reducing emissions,and 第 4 期田惠玲等:基于随机森林模型的东北三省乔木林生物质碳储量预测scientifically formulating the national“Carbon Neutral”action path and target management.【Method】Based on the data of permanent monitoring plots of the successive national forest inventories from 1999 to 2018,and combined with multiple influence factors such as regional climate,soil,forest stand and topography,etc.,we used the random forest algorithm to construct a growth-loss model of the main dominant tree species(groups)in the region.Combining the future climate and arbor forest area expansion scenario,the biomass carbon stock(BCS)and BCS changes of arbor forests of the three provinces in northeast China from 2015 to 2060 were predicted.【Result】The BCS of arbor forests of three provinces in northeast China reached 3 393.15 TgC in 2060,an increase of 1 895.23 TgC compared with that in 2015,and the annual mean BCS changes was 42.12 TgCa-1 during 2015-2060.Natural forests played the major role in the BCS and BCS changes.The BCS of arbor forests in Liaoning,Jilin,and Heilongjiang will be increased from 139.19,463.58,and 895.15 TgC in 2015 to 328.95,915.83,and 2 148.37 TgC in 2060,respectively.The average biomass Carbon Density(BCD)of arbor forests will be increased from 32.71,59.75,and 45.11 t hm-2 in 2015 to 75.20,109.32,and 85.24 t hm-2 in 2060,respectively,and the average annual biomass carbon sinks of arbor forests in Liaoning,Jilin,and Heilongjiang from 2015 to 2060 are 4.22,10.05,and 27.85 TgCa-1,respectively.【Conclusion】The random forest model constructed in this study performs well and can be used to predict the future BCS of arbor forests of three provinces in northeast China.From 2015 to 2060,the BCS of arbor forests of three provinces in northeast China will be increased by 1 895.23 TgC,which will still have a large carbon sink potential in the future.Heilongjiang Province has the largest carbon sink potential,with an annual increase of carbon storage up to 27.85 TgCa-1,which is an important carbon sink area in the future;while Liaoning Province is weaker,with an annual increase of carbon storage by only 4.22 TgCa-1.Strengthening the management of the young and middle-aged forests,moderately regenerating the mature and over-mature forests,and enhanc