Calibration of Green‐Ampt wetting front suction for ponding and non‐ponding infiltration scenarios

Abstract

AbstractMathematical formulations for layered soil profiles and corresponding parameterization to improve performance of the Green‐Ampt model (GA) have not always shown accurate. We evaluated GA maintaining its assumption of saturated flow but calibrating the effective matric suction at the wetting front (hf) by damping the initial matric suction (hi). The proposition was evaluated comparing cumulative infiltration simulated by GA to that simulated by Hydrus‐1D for five soils from three texture classes, three initial matric suctions, and five 24‐h water supply regimes. For each case, GA was run with a decreasing series of 20 values of hf from the initial matric suction (10, 100, or 1000 cm) to values near 1 cm. To minimize the relative error (RE) between cumulative infiltration simulated by GA and Hydrus‐1D for the 24‐h period we found the required damping of hf to be dependent on soil type, water supply regime and initial matric suction, the latter being the most important factor. The calibration of hf was efficient to approach GA predictions to numerical Hydrus‐1D simulations, even for cases with alternating ponding and non‐ponding conditions. Results suggest a large reduction of wetting front matric suction hf may be necessary to approach cumulative infiltration predicted by GA to corresponding Hydrus‐1D predictions, especially during rainfall events that do not imply in continuous ponding. Taking hf somewhere between 1 and 3 cm, values of RE of cumulative infiltration during the first 24 h were below 10% for all cases, independent of soil type, rainfall characteristics or initial water content.