The spatial geological heterogeneity of an aquifer significantly affects groundwater storage, flow and the transport of solutes. In the particular case of coastal aquifers, spatial geological heterogeneity is also a major determining factor of the spatio-temporal patterns of water quality (salinity) due to seawater intrusion. While the hydraulics of coastal hydrogeology can be modeled effectively by various density flow equations, the aquifer geology is highly uncertain. A stochastic solution to the problem is to generate numerical realisations of the geology using sequential stratigraphy, geophysical models or geostatistical approaches. The geostatistical methods (two-point geostatistics, Markov chain models and multiple-point geostatistics) have the advantage of minimal data requirements, e.g., when the only data available are from cores from a few sparsely located boreholes. We provide an extension of sequential indicator simulation by including the uncertainty of the hydrofacies proportions in the simulation approach. We also deal with the problem of variogram estimation from sparse boreholes and we discuss the implicit transition probabilities and the connectivity of simulated realisations of a number of categorical variables. The variogram model used in the simulation of hydrofacies significantly influences the degree of connectivity of the hydrofacies in the simulated model. The choice of model is critical as connectivity determines the amount and extent of seawater intrusion and hence the environmental risk. The methodology is illustrated with a case study of the Andarax river delta, a coastal aquifer in south-eastern Spain. This is a semi-arid Mediterranean region in which the increasing use of, and demand for, groundwater is exacerbated by a transient tourist population that reaches its peak in the summer when the demand for the permanent population is at its highest. The work reported here provides a sound basis for designing flow simulation models for the optimal management of groundwater resources. This paper is an extended version of a presentation given at the 2012 GeoENV Conference held in Valencia, Spain.

含水层的空间地质异质性显着影响地下水的储存，流动和溶质的运输。在沿海含水层的特殊情况下，空间地质异质性也是导致海水入侵的水质（盐度）时空格局的主要决定因素。虽然可以通过各种密度流方程有效地模拟沿海水文地质学的水力学，但是含水层的地质学却高度不确定。该问题的一种随机解决方案是使用顺序地层学，地球物理模型或地统计学方法生成地质的数值实现。地统计方法（两点地统计，马尔可夫链模型和多点地统计）具有数据需求最少的优点，例如，当唯一可用的数据来自几个稀疏钻孔的岩心时。通过在模拟方法中包括水相比例的不确定性，我们提供了顺序指标模拟的扩展。我们还处理了稀疏钻孔的方差图估计问题，并讨论了隐式过渡概率和许多分类变量的模拟实现的连通性。在水相模拟中使用的变异函数模型极大地影响了模拟模型中水相的连通度。模式的选择至关重要，因为连通性决定了海水入侵的数量和程度，进而决定了环境风险。通过对西班牙东南部沿海含水层Andarax河三角洲的案例研究说明了该方法。这是一个半干旱的地中海地区，在夏季，对常住人口的需求最高时，一个短暂的旅游人口达到了顶峰，加剧了对地下水的使用和需求。此处报告的工作为设计用于地下水资源优化管理的水流模拟模型提供了良好的基础。本文是在西班牙巴伦西亚举行的2012 GeoENV会议上所作演讲的扩展版本。此处报告的工作为设计用于地下水资源优化管理的水流模拟模型提供了良好的基础。本文是在西班牙巴伦西亚举行的2012 GeoENV会议上所作演讲的扩展版本。此处报告的工作为设计用于地下水资源优化管理的水流模拟模型提供了良好的基础。本文是在西班牙巴伦西亚举行的2012 GeoENV会议上所作演讲的扩展版本。