Resources Science ›› 2016, Vol. 38 ›› Issue (12): 2375-2382.doi: 10.18402/resci.2016.12.16

• Orginal Article • Previous Articles     Next Articles

Effect of sampling on spatial variability in soil salinity in the Yellow River Delta Area

ZHANG Xiaoguang1,2(), WANG Zhigang3, SONG Xiangyun1, LIU Peiru1, LI Shimei1(), YANG Xia4   

  1. 1. College of Resource and Environment,Qingdao Agricultural University,Qingdao 266109,China
    2. State Key Laboratory of Soil and Sustainable Agriculture,Institute of Soil Science,Chinese Academy of Sciences,Nanjing 210008,China
    3. Changjiang River Scientific Research Institute of Changjiang Water Resources Commission,Wuhan 430010,China
    4. Agricultural Bureau of Dongying City,Dongying 257091,China
  • Received:2016-06-23 Revised:2016-09-22 Online:2016-12-20 Published:2016-12-20


The monitoring of spatial variation in soil salinization forms the premise of the rational use of saline soil. The number of sampling points will affect the expression of spatial variation of soil salinity. Here,geostatistical methods were used to extract 12 soil sample subsets for Kenli County in the Yellow River Delta,China. The effect of different sampling numbers on the expression of soil salinity spatial variation was explored. The results showed that a great degree of variability in soil salinity. Effective ranges of most sample sets were between 13.47km and 16.77km. When the sampling number was less than 150 sample points,effective range was decreased (< 6.10km),small samples were better in indicating the variation characteristics at small scales (< 6.10km). With a reduction in sampling numbers,the capacity for describing details was weakened. Spatial variation structure was not expressed well,especially when the number was small(<150). With increasing sampling numbers,the root mean square error (RMSE)was reduced,indicating that the accuracy of the kriging interpolation method was reduced. With increasing sampling numbers,the mean error(ME)had no obvious reduced trend. In summary,the samples number to meet the spatial variability of soil salinity should not be less than 150. We conclude that,in evaluating kriging prediction accuracy the single evaluation index may not be able to completely evaluate predicted results. Combined with the actual fact,a variety of comprehensive indexes may be needed to evaluate predicted results. These findings are applicable to other areas with similar environmental conditions to Kenli County.

Key words: sampling numbers, soil salinity, spatial variability, county scale, Kenli County of the Yellow River Delta Area