摘要
我国现行的森林资源调查方法存在着空间关联性强、适应性差的缺陷.空间平衡抽样(SBS)强调样本点抽取的随机等概和空间上的均衡分布,通过包含概率栅格层的过滤运算,极大地减少了无反应样本单元现象的发生,在森林资源调查中具有较大的潜在应用价值.文章在简要回顾基于模型的抽样、基于设计的抽样两大类森林资源调查方法特点,分析4种传统概率抽样方法(简单随机抽样、系统抽样、分层抽样、整群抽样)弊端的基础上,介绍了基于通用随机方格分层算法(GRTS)的空间平衡抽样的原理和步骤,最后以紫金山国家森林公园为研究对象进行了风景林蓄积量调查空间平衡抽样案例研究.研究表明,空间平衡抽样在降低调查成本、减少空间关联性强方面明显优于简单随机抽样;但在提高抽样精度方面没有表现出明显的优势,只有当样本容量大于或等于理论计算容量时,空间平衡抽样才表现出一定的抽样精度优势.
Current methods for forest resource inventory have a defect of strong spatial connection but poor adaptability. Spatial Balanced Sampling （SBS） highlights random equal probability and evenly distribution in space, so it hosts large potential values in applying to forest resource inventory. The paper briefly reviews the characteristics of sampling two methods for forest resource inventory based on models and designs respectively. It analyzes the malpractice of traditional probability sampling methods and introduces the principle and procedures of spatial balanced sampling based on GRTS. Finally, it borrows a case study on stock volume inventory of scenery forest, taking Zijinshan National Forest Park as an objective. The result shows that spatial balanced sampling presents certain advantage on sampling precision only when sample volume is bigger than or equal to theoretically calculated volume.
出处
《林业调查规划》
2008年第4期1-6,共6页
Forest Inventory and Planning
基金
国家自然科学基金(30571490)
国家留学基金(2005832062)
