干旱区剖面土壤盐分空间变异特征及随机模拟

Spatial Variability and Stochastic Simulation of Soil Salinity in the Vertical Profiles

利用便携式土壤盐分计测定干旱区膜下滴灌棉田及裸地剖面土壤电导率,利用GS+软件确定其半方差函数,并进行序贯高斯模拟来研究土壤剖面盐分的空间变异特征。结果显示:干旱区剖面上土壤盐分空间变异性强,膜下滴灌棉田符合球形模型,裸地区符合高斯模型,且棉田变程远小于裸地;棉田埋深50cm以浅土壤盐分变化剧烈,埋深50cm以深变化平缓,在垂向上呈现3个高值区(30,50和75cm深度附近)及3个低值区(埋深40,60和100cm附近);克里格法具有明显的平滑作用,降低了土壤盐分的空间变异性,而序贯高斯模拟数据更离散,更能反映土壤盐分的空间变异性,但多次模拟均值随次数增加趋于稳定。

Based on the measured value of soil electrical conductivity（EC） in different soil profiles in bare soil and cotton field with film-mulched drip irrigation,soil conductivity meter named EC100,and the semi-variance function,Kriging interpolation and sequential Gaussian simulation were carried out.The results indicated that the spatial variability of soil salinity in the profiles was very intense,and the best semi-variance function is spherical model with range of 9.7 cm and Gaussian model with the range of 129 cm to 113.5 cm in cotton field and bare soil respectively.Soil salinity in the cotton field varies rapidly in the shallow area above the 50 cm and varies slowly bellow the depth of 50 cm.It shows three high value areas at the depth of 30 cm,50 cm and 75 cm and three low value areas at the depth of 40 cm,60 cm and 100 cm.The estimated value by Kriging interpolation showed obviously smoothing effect and the original data spatial configuration have been changed.The simulated value by sequential Gaussian simulation was more discrete,which could show the spatial variance better.With the increasing of simulation,the average value trended to be stable and was similar to the value by Kriging interpolation.