A Task-Oriented Perspective on the Role of Hydrogeological Heterogeneity in Transport Modeling

Abstract:

Hydrogeological heterogeneity is often modeled using categorical geostatistics to differentiate between different hydrogeological units, but the method of dividing the hydrogeology into categories varies across the previous literature. Sedimentology classifications, percentile cutoffs of hydraulic conductivity, and geophysical image processing are examples of common techniques. These methods are often applied with the assumption that the field can be divided into only two categories, primarily due to sparse measurements and modeling ease. This poster depicts an analysis of particle paths through a three-dimensional high-resolution aquifer analog in order to provide better insight into what hydrogeological features are important to consider for improved task-oriented field site characterization. The role of each of the analog’s ten hydrofacies in the particle paths with respect to both time and distance are considered, and Markov clustering is applied to guide the combinations of hydrofacies by the particle path transition probabilities into categories in a hierarchical fashion. The analysis is applied for various portions of the particles’ arrival time distribution to highlight the need for a task-oriented perspective, such as when the left tail of the distribution is of interest for risk assessment. By taking into consideration which features of the hydrogeological heterogeneity drive the particular aspect of transport that needs to be predicted, the first step is taken towards capturing the relevant features in categorical geostatistics and thus stochastic modeling.