渗透系数的空间变异性是影响渗流场不确定性的重要因素。将渗透系数作为各向同性对数正态平稳随机场，通过数值模拟研究渗透系数空间相关性对渗流场统计特征的影响。选取了3种具有代表性的相关函数模型，即指数模型、高斯模型和球状模型，分别以渗透系数的相关距离和变异系数为变化因子对水头方差作敏感性分析。结果表明，渗透系数相关函数模型的选取、相关距离和变异系数的取值对水头方差具有较大影响。

The spatial variability of permeability coefficient is a main cause for the uncertainty of seepage field. With the basis that the permeability coefficient is an isotropic random field which follows lognormal distribution, the influence of the spatial correlation of permeability on the statistics of seepage field is researched by stochastic numerical modeling. Three representative correlation functions, such as exponential, gaussian and spheric models, are adopted. The sensitivity analyses to variance of hydraulic head for above three models are conducted respectively with correlation distance and variation coefficient of permeability as variable factors. The results show that the choice of correlation model and the value of correlation distance and variation coefficient have an important impact on variance of hydraulic head.