Characterizing soil preferential flow using iodine--starch staining experiments and the active region model

Thirteen iodine-starch staining experiments with different boundary conditions and measurement scales were conducted at two sites to study preferential flow processes in natural unsaturated soils. Digital imaging analyses were implemented to obtain the corresponding preferential flow patterns. The test results are used to evaluate a recently proposed active region model in terms of its usefulness and robustness for characterizing unsaturated flow processes at field scale. Test results provide useful insights into flow patterns in unsaturated soils. They show that flow pattern depends on the top boundary condition. As the total infiltrating-water depth increased form 20 mm to 80 mm for the 100 x 100 cm{sup 2} plots, the corresponding flow pattern changed from few preferential flow paths associated with a relatively small degree of stained coverage and a small infiltration depth, to a pattern characterized by a higher stained coverage and a larger infiltration depth, and to (finally) a relatively homogeneous flow pattern with few unstained area and a much larger infiltration depth. Test results also show that the preferential flow pattern became generally more heterogeneous and complex for a larger measurement scale (or size of infiltration plot). These observations support the general idea behind the active region model that preferential flow pattern in unsaturated soils are dynamic and depend on water flow conditions. Further analyses of the test results indicate that the active-region model is able to capture the major features of the observed flow pattern at the scale of interest, and the determined parameter values do not significantly depend on the test conditions (initial water content and total amount of infiltrating water) for a given test site. This supports the validity of the active region model that considers that parameter to be a property of the corresponding unsaturated soil. Results also show that some intrinsic relation seems to exist between active-fracture-model parameter and a random-cascade-model parameter. (The latter model is also developed based on the existence of the fractal flow pattern in unsaturated soils.) Furthermore, our test results demonstrate that the active-region-model parameter is not scale-dependent for a range of scales under consideration. Although further studies are needed to confirm this finding, it seems to be consistent with a consideration that some fractal parameters (e.g., fractal dimension) are universal for a large range of scales