基于GIS的森林资源调查空间平衡抽样方法研究

我国现存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。随着新的森林调查技术规程的出台与林区社会经济条件的变化,急剧上升的调查成本与有限的调查经费的矛盾日益突出。与此同时,森林资源调查过程中抽样框变化、无反应样本单元的现象日益突出。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发生。紫金山国家森林公园风景林美学调查空间平衡抽样案例研究表明,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。作为一种具有严格统计学基础的、高效低成本的、适应性强的抽样方法,空间平衡抽样在森林资源调查中具有较大的应用潜力。

Spatial Balanced Sampling for Forest Resources Inventory Based on GIS

There are some weaknesses existing in current forest survey designs,such as spatial auto-correlation,non-flexibility.With the implementation of new regulations on forest resources inventory and rapid change of social and economical conditions in forest areas,the contradiction between the rising costs and limited financial budget becomes more acute,coupled with the problems of sampling frame change and frequent occurrence of non-response sampling units.Spatially balanced sampling(SPB) can generate a probability-based spatially balanced sampling points and filter the points with low inclusion probability,thus greatly reducing the problem of non-response units.The results of case study of recreational forest scenic beauty survey based on SPB in Zijin Mountain National Forest Park showed that SPB outperformed simple random sampling(SRS) both in reducing survey costs and decreasing spatial auto-correlation.However,only when the sampling unit numbers were equal or above theoretical sampling capacity,could SPB demonstrate superiority over SRS at some extent.As mentioned in the paper,being a cost efficient and flexible sampling design with sound statistics basis,SPB has a great application potential in forest resources inventory.