Sean A. McKenna, Katherine A. Klise, David B. Hart and Vince Tidwell, Sandia National Laboratories
For the past 25 years, numerical applications of stochastic subsurface hydrology have relied on geostatistically-based models of aquifer heterogeneity. These approaches employ reproduction of a two-point model of spatial covariance, or Markov transition probability, as the basis for the generation of the heterogeneous fields. This reliance on measures of variability between two points separated at some distance along a straight line precludes the simulation of complex spatial structures often observed in sedimentary and fractured rock aquifers. More recent developments in the area of simulation of stochastic spatial fields allow for reproduction of spatial structure beyond that afforded with two-point statistics. These approaches include Boolean and random token models, facies-based simulation and simulation techniques derived from spatial filtering of existing images and, in general, are referred to as “multipoint” statistical tools. Inference of the spatial model for these approaches comes from training images, not a variogram, and is well suited to incorporating conceptual model information, outcrop analog data and subsurface data from other locations. To date, these tools have seen relatively limited use in ground water applications.
We examine several of these simulation approaches for the creation of stochastic permeability models and compare the results to more traditional stochastic fields created with a multiGaussian simulation algorithm using a two-point spatial covariance model. The performance measures for the comparison are based on the simulation of steady-state ground water flow and advective transport through high resolution stochastic permeability models created with various simulation algorithms. These performance measures include detailed analysis of the early and late time tails of the travel time distribution and measures of flowpath spreading/focusing with the flow domain.
Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.
The 2007 Ground Water Summit
