Experimental studies have shown that the modified Green-Ampt (GA) model, which accounts for a velocity-dependent capillary pressure, can describe water infiltration into dry sand columns better than the classical GA model. In this study, we performed a series of downward water infiltration experiments in prewetted sand columns for four different initial water contents (0, 3.3, 6.5 and 13.8%) and three different ponding heights (10, 20 and 40 cm). As expected, an increase in ponding height resulted in a monotonic increase in the infiltration rate. The infiltration rate decreased monotonically as the initial water content increased for initial water contents of 3.3, 6.5 and 13.8%. However, the infiltration rate in the dry sand columns was significantly smaller than in the prewetted sand columns. We linked this anomalous behavior to the strong dynamic effect of the capillary pressure on water infiltration into dry sand columns. The series of downward infiltration experiments were modeled by both the classical and modified GA models. The modified GA model can describe better the experimental results than the classical GA model for infiltration into both dry and prewetted sand columns. The modeling and parameter fitting results show that the magnitude of the dynamic effect on the capillary pressure depends on the initial water content and decreases exponentially as the initial water content increases. Consequently, the classical GA model significantly underestimates the equilibrium suction heads for the infiltration experiments in the dry sand columns.
All Science Journal Classification (ASJC) codes
- Water Science and Technology
- Initial water content
- Modified Green-Ampt model
- Unsaturated flow