INFILTRATION PATTERNS FROM SIMULATED RAINFALL ON A SEMIARID RANGELAND SOIL

Currently proposed approaches to the estimation of surface runoff and erosion are based on the Green-Ampt equation. Processes controlling water movement in uncultivated semiarid rangelands, such as macropore flow for example, may be incompatible with assumptions fundamental to the Green-Ampt equation. In this study, patterns of water and tracer (bromide, methylene blue, and alizarin red S) movement under two simulated rainfall intensities were compared with those consistent with the Green-Ampt equation. Macropore flow was an important process at the site. The reasonably uniform and abrupt wetting front indicated that this was not necessarily incompatible with the use of the Green-Ampt equation. Apparently the effective macropores were sufficiently close together and shallow that lateral movement effectively obliterated the effects of individual macropores at depths of about 0.3 m. Two other inconsistencies with the Green-Ampt equation were observed. First, soil cover type strongly affected infiltration. Bare interspace areas in microtopographic lows had relatively low infiltrabilities and were hydraulically interconnected. These areas, therefore, control time to ponding and runoff initiation. Explicit consideration of these differences may improve runoff and infiltration estimation. Second, a contrasting argillic horizon appeared to inhibit water movement. Subsoil properties not usually considered in the Green-Ampt equation may dominate infiltration under near-field-capacity soil moisture conditions and large rainfall events.