农田大尺度表层土壤电导率的序贯高斯模拟

Sequential Gaussian Conditional Simulation of Soil Surface Specific Conductivity at Large Field-scale

以内蒙古河套灌区部分农田为例,将克立格插值法和序贯高斯条件模拟法分别应用于农田大尺度下表层土壤电导率的空间分布估值和模拟,并将1次实现和多次实现取平均的条件模拟结果与克立格插值结果和实测数据进行了对比分析。结果表明：克立格法具有明显的平滑效应,它改变了电导率数据的空间结构,而条件模拟则能够在尊重实测数据情况下,再现变量的空间变化特点,而且模拟次数在2-3次时可以更好的与原始数据符合。在进行变量的空间分布研究时,可用克立格法来估计,用条件模拟来重现波动性,两者结合才能充分发挥各自的优势,为生产实践提供全面的科学根据。

An example is given for the farm field of Hetao Irrigation Area of Inner Mongolia. Kriging Interpolation Method and Sequential Gaussian Conditional Simulation Method is used to evaluate and simulate spatial distribution of specific conductivity of surface soil at large field-scale. The mean value after simulating 1 time, the mean value after simulating many times, the estimated value by Kriging Method and the measured value are compared. The results indicate that the estimated value by Kriging Method show obviously smoothing effect and the original data configuration have been changed, while the simulated values had the same data configuration as the measured values. When studying spatial distribution of variables, we use Kriging to estimate the variable and use conditional simulation to reconstruct wave characters, and 2-3 times simulation agree with original data better. When Kriging and conditional simulation are combined, they can supply overall scientific basis for production and experience better.