Characterizing heterogeneous flow patterns in porous media using cluster analysis

The objectives of this study are to develop a quantitative method to describe heterogeneous flow patterns and identify the heterogeneous information characteristics based on flow patterns. Dye tracer experiments were performed on plots with different conditions. The flow patterns were separated into similarly stained horizontal layers based on the width distributions of the stained areas. Then the layers identified in the flow patterns were partitioned into groups with similar patterns by hierarchical clustering. The results show that the dye coverage and width of stained structures are the most indicative factors for the different clusters. The number of clusters affects the degree of layer discrimination. Three and eight clusters seem to be sufficient to capture the main characteristics of the heterogeneous flow patterns. For the experimental plots where the vertical flow is dominated, the clusters contain the layers with the homogeneous flow, preferential flow, and those with flow changes from homogeneous infiltration to preferential flow. For the experimental plots where the horizontal heterogeneous is significant, the soil horizons has a strong influence on the water flow patterns.