资源科学 ›› 2019, Vol. 41 ›› Issue (2): 257-267.doi: 10.18402/resci.2019.02.05

• • 上一篇    下一篇

基于空间聚类分层抽样的黄骅市县域耕地质量等别监测样点布设

杨江燕1(), 殷守强2, 张利1, 门明新1(), 陈影1   

  1. 1. 河北农业大学国土资源学院,保定 071000
    2. 中国矿业大学地球科学与测绘工程学院,北京 100083
  • 收稿日期:2018-07-14 修回日期:2018-11-08 出版日期:2019-02-25 发布日期:2019-02-25
  • 作者简介:

    作者简介: 杨江燕,女,安徽安庆人,硕士生,研究方向为土地资源管理。E-mail:15931895996@163.com

  • 基金资助:
    国土资源部两套指标体系试点项目(20170411);河北省典型样带土地资源与生态环境监测评价项目(20170428)

The layout of county area cultivated land quality monitoring samples in Huanghua City based on spatial clustering stratified sampling

Jiangyan YANG1(), Shouqiang YIN2, Li ZHANG1, Mingxin MEN1(), Ying CHEN1   

  1. 1. Institute of Land and Resources, Agricultural University of Hebei, Baoding 071000, China
    2. China University of Mining, Beijing 100083, China
  • Received:2018-07-14 Revised:2018-11-08 Online:2019-02-25 Published:2019-02-25

摘要:

耕地质量监测是当前国家及时掌握耕地质量和产能变化的重要工作手段。合理布设耕地质量监测样点可以大幅提高耕地质量监测的效率。以河北省黄骅市县域耕地为研究对象,以耕地的质量属性和空间位置属性为耕地质量等别监测区的划定依据,采用空间聚类法生成20种备选监测区方案,通过综合比较方差、抽样误差、抽样效率和抽样弹性系数,选出初始监测区方案,并在此基础上局部优化,最终生成耕地质量等别监测区方案,最后以耕地质量等别监测区为分层依据,运用分层抽样法布设了耕地质量等别监测样点。研究表明:①在抽样误差为1%的要求下,分区数为65的备选耕地质量等别监测区方案的综合样本容量为77,抽样效率相对较高,被选定为初始监测区方案;②随着耕地质量等别监测区数目从5增加到100,相同监测区内耕地单元之间在空间位置和耕地质量等别上的差异程度由快速减少到趋于平稳;③在布设相同数量的监测样点时,本文提出的基于空间聚类分层抽样的样点布设方法在表土质地、剖面构型、盐渍化、有机质含量、排水条件、灌溉条件和国家耕地质量自然等指数等因素方面的抽样误差分别为0.37、1.02、1.39、0.91、0.31、1.53和1.27,均明显低于传统的等别分层抽样、简单随机抽样、网格分层抽样,具有较高的抽样效率。研究成果可为耕地质量等别监测样点布设相关工作和研究提供有效指导。

关键词: 耕地质量等别, 监测样点布设, 抽样误差, 抽样效率, 空间聚类分层抽样, 黄骅市

Abstract:

Cultivated land quality monitoring is an important strategy to grasp the changes in cultivated land quality and productivity for the country. Arranging cultivated land quality monitoring samples in a reasonable way can greatly improve the efficiency of cultivated land quality monitoring. This study took the typical area of the coastal plain, Huanghua City as an example, generated 20 alternative monitoring area scenarios with the method of spatial clustering. The initial monitoring area plan was selected by comprehensive comparison of variance, sampling error, sampling efficiency, and sampling elastic coefficient, and by partial optimization, the project of the cultivated land quality monitoring area was finally generated. Finally, based on the stratified sampling area, the stratified sampling method was used to lay out the monitoring records of cultivated land quality.The results showed that under the requirement of 1% sampling error, the alternative cultivated land quality monitoring area with a partition number of 65 did exhibit a comprehensive sample capacity of 77 and a relatively high sampling efficiency, which was selected as the initial monitoring area. The difference in spatial position and cultivated land quality between cultivated land units in the same monitoring area decreased rapidly first, and finally remained stable, with the number of cultivated land quality monitoring areas increases from 5 to 100. By laying the same number of monitoring samples, the sampling error in terms of topsoil texture, section configuration, salinization, organic matter content, drainage conditions, irrigation conditions, and national natural index were 0.37, 1.02, 1.39, 0.91, 0.31, 1.53, and 1.27 respectively. Based on spatial clustering stratified sampling proposed in this paper, the observations were lower than traditional stratified sampling, simple random sampling, grid stratified sampling, and had higher sampling efficiency. The results are intended to provide effective guidance for the deployment of relevant work and research on the quality monitoring samples of cultivated land.

Key words: cultivated land quality, monitoring sample layout, sampling error, sampling efficiency, spatial clustering stratified sampling, Huanghua City