两阶段自适应群团抽样在沙漠边缘植被调查中的比较  

Comparison among Different Designs of Two-Stage Adaptive Cluster Sampling in Vegetation Survey in Desert Edge

作　　者：朱光玉[1,2] 雷渊才[1] 

机构地区：[1]中国林业科学研究院资源信息研究所,北京100091 [2]中南林业科技大学林业遥感信息工程研究中心,长沙410004

出　　处：《林业科学》2010年第6期71-77,共7页Scientia Silvae Sinicae

基　　金：国家自然基金项目(30510103195);科技部社会公益研究专项(2005DIB5J142);国家林业局"948"引进项目(2006-4-34);中南林业科技大学人才引进项目(104-0086)

摘　　要：两阶段自适应群团抽样(two-stageACS)是一种新兴自适应群团抽样(ACS)设计,它能在一定程度上解决自适应群团抽样最终样本量不确定的缺陷。概述两阶段自适应群团抽样方法的技术思路和原理。以实地调查的花棒数据为研究对象,进行4种两阶段自适应群团抽样方法的比较和分析,指出不跨越边界的基于Horvitz-Thompson估计量的两阶段自适应群团抽样的效果最佳。Two-stage ACS （adaptive cluster sampling） is an extended method of ACS,and it can to a certain degree control the final sample size of ACS. The procedure and principle of two-stage ACS are presented in this paper. Based on the survey data of Hedysarum scoparium in the western China,the simulation results of four different two-stage adaptive cluster samplings are compared and analyzed. It is showed that the Two-stage ACS,that is based on the modified Horvitz-Thompson estimator,is most effective,in which the clusters are not allowed to overlap primary unit.

关 键 词：自适应群团抽样 两阶段自适应群团抽样 Horvitz-Thompson估计量 Hansen-Hurwitz估计量 

分 类 号：S759.2[农业科学—森林经理学]