Abstract: By using the methods of classic statistics in combining with geo-statistics, this paper analyzed the spatial variability of soil N, P, and K in the main production area of Jingdong chestnut,  Qianxi County of Hebei Province. In the study area, the contents of soil total N, P, and K and soil available N and K were lower, but the available P content in topsoil was higher. The soil N, P, and K had a medium variability of distribution, and soil P had the largest coefficient of variation. In topsoil, the distribution of total N and available N fitted Gaussian model, and their spatial variability was mainly affected by structure factors. The distribution of soil total K and available K fitted spherical model and Gaussian model, respectively. The spatial variability of soil total K was co-affected by structure factors and random factors, and had a medium spatial correlation; while the spatial variability of soil available K was mainly affected by structure factors, and had a strong spatial correlation. At county scale, soil N and K contents were higher in the south and northwest than in the middle and northeast, soil total P content was higher in the north, and soil available P content was higher in the south. Soil total N and K had significant positive correlations with soil available N and K; while soil total P had no significant correlation with soil available P.

Key words: Castanea mollissima, soil nutrient, spatial variability, geo-statistics, irrigation amount, fertigation, water use efficiency, yield, quality, cucumber.