Statistical methods (such as Sequential Gaussian Simulation) are often used to create geological models to serve as the basis for reservoir simulations. Most of these methods rely on the assumption that the spatial distribution of a property is Gaussian and can be described only in terms of its mean and covariance. What if this assumption is false? What if extreme values dominate economic significance? These traditional methods do not allow the representation of asymmetric spatial structures. The impact of asymmetry on connectivity may have particular significance for flow models. This paper discusses the spatial asymmetry of coal properties and investigates the impact of including spatial asymmetry on flow simulations representing coal seam gas production.