依据多种抽样技术对二类调查蓄积量精度控制及比较  

Accuracy Control and Comparison of Forest Management Investigation Stock Volume Using Multiple Sampling Techniques

作　　者：王博一 刘兆刚[1] 颜俊峰 彭湘玲 Wang Boyi;Liu Zhaogang;Yan Junfeng;Peng Xiangling(Northeast Forestry University,Harbin 150040,P.R.China)

机构地区：[1]东北林业大学林学院,哈尔滨150040 [2]东北林业大学,哈尔滨150040

出　　处：《东北林业大学学报》2022年第2期23-28,共6页Journal of Northeast Forestry University

基　　金：国家重点研发计划课题(2017YFC0504103);黑龙江头雁创新团队计划项目(森林资源高效培育技术研发团队)。

摘　　要：森林资源总蓄积量和平均每公顷蓄积量估计是森林资源二类调查的主要内容。为了准确反映森林资源二类小班调查的精度,通常采用不同抽样方式布设固定样地,样地数量要满足抽样控制精度要求,通过总体抽样调查方法获取的蓄积量与小班调查计算的蓄积量进行比较以实现二类小班调查蓄积量精度控制。以东北林业大学帽儿山实验林场为研究区域,2016年帽儿山实验林场森林资源二类小班调查数据和布设在实验林场的256块固定样地数据,采用系统抽样、事后分层抽样和自助法抽样技术进行小班调查蓄积量总体精度控制,并比较几种不同抽样方法的精度和特点。结果表明:有林地小班总蓄积为3263235.87 m^(3),系统抽样估测有林地总蓄积为3991105.77 m^(3)。事后分层抽样按照龄组、林型、郁闭度估测有林地总蓄积分别为3899872.16、4036672.73、3988058.20 m^(3)。自助法抽样在N=20时有林地总蓄积估计值为3948893.46 m^(3)。结果均超出3倍标准误,初次小班调查不符合调查标准要求,应当重新调查小班,直到符合标准为止。在研究区范围内,事后分层抽样按龄组、林型、郁闭度划分理论精度分别为95.9%、94.7%、94.7%,系统抽样理论精度94.5%,两种方法抽样精度接近,对事后分层抽样而言,按龄组林型划分更合理,精度更高。自助法抽样自助样本为20时理论精度是86.5%,在抽样精度要求不高时,可以先抽样再调查,减少调查样地数,节省人力物力。Forest stock volume and stock volume per hectare are the main content of Forest Management Investigation.In order to accurately reflect the accuracy of the sub-compartment survey of Forest Management Investigation,different sampling methods are usually used to set up fixed sample plots.The number of sample plots must meet the requirements of sampling control accuracy.The stock volume obtained through the sampling survey is compared with the stock volume calculated by the sub-compartment survey.In order to realize the precision of the survey accumulation of the sub-compartment stock of Forest Management Inventory,with Maoershan Text Forest Farm of Northeast Forestry University as the research area,and with the second-class investigation data in 2016 and 256 fixed plots in the text forest farm for sample plot data,systems sampling,post stratification sample and bootstrap method sampling are used to control the overall accuracy of the stock volume of the sub-compartment survey.The accuracy and characteristics of several different sampling methods are compared.The total volume of the sub-compartment with forest land is 3263235.87 m^(3).The total volume of forest land estimated by system sampling is 3991105.77 m^(3).Divided by the forest type,the age group,and canopy closure,the post stratification sampling is estimated in the volume of forest land of 3899872.16,4036672.73 and 3988058.20 m^(3),respectively.By Bootstrap sampling when N=20,the estimated total volume of forest land is 3948893.46 m^(3).All of the above exceeds three times the standard error,and the initial sub-compartment investigation does not meet the requirements of the investigation standard.So sub-compartment investigation should be re-investigated until the standard is met.Within the study area,the theoretical accuracies of post-stratified sampling divided by the age group,forest type and canopy closure were 95.9%,94.7%and 94.7%respectively.The theoretical accuracy of systematic sampling was 94.5%.The processing of systems sampled data is simple and

关 键 词：二类调查 系统抽样 事后分层抽样 自助法抽样 精度 

分 类 号：S757.4[农业科学—森林经理学]