Experimental and numerical study of water flow in soil under irrigation in northern Senegal: evidence of air entrapment

Irrigation by surge flooding does not always wet the soils thoroughly, and we have investigated the reasons for this on an irrigated plot in northern Senegal by monitoring the water budget during a rice cropping season (100 days). The amount of water added during each irrigation event was measured, and evapotranspiration and infiltration were estimated with lysimeters and Muntz infiltration rings, respectively. At the same time, piezometric levels, neutron probe values and water tension data were recorded at two stations in the plot. These measurements showed unusual results: infiltration rate was less than 1 x 10(-6) mm s(-1) (less than 0.1 mm a day), there was a constant water deficit during the entire irrigation period, around 50 cm deep, and tensiometers at 40 cm reacted very slowly to water infiltration. The water fluxes in the vadose zone derived from these data showed clearly a discrepancy between fluxes calculated from hydraulic gradients and fluxes calculated from mass conservation. The hydraulic gradients suggested a zero flux plane at 40 cm below the surface, but the calculated values of the fluxes overestimated by several orders of magnitude the infiltration rates determined on the plot, whereas fluxes determined from mass conservation matched far better. These results show that air was entrapped between the shallow water table and the wetting front, and this inhibited water infiltration. Modelling water flow down the soil profile with a computer program for simulating one-dimensional water movement (Hydrus) confirmed that single-phase models cannot describe imbibition in this situation. Simple infiltration models based on a modified Green-Ampt equation accounting for air compression and air counter-flow, however, fit experimental infiltration data much better. We demonstrated that where surge flooding is associated with a shallow water table, as in many large irrigation schemes, one must take into account the presence of air to quantify the flow of water into the soil.