文摘
Simulated rainfall of fluctuating intensity was applied to runoff plots on bare dryland soils in order to explore a new method for analysing the non-steady-state responses of infiltration and overland flow. The rainfall events all averaged 10 mm/h but included intensity bursts of up to 70 mm/h and lasting 5–15 min, as well as periods of low intensity and intermittency of up to 25 min. Results were compared with traditional steady-state estimates of infiltrability made under simulated rainfall sustained at a fixed intensity of 10 mm/h. Mean event infiltration rate averaged 13.6% higher under fluctuating intensities, while runoff ratios averaged only 63% of those seen under constant intensity. In order to understand the changing soil infiltrability, up to three affine Horton infiltration equations were fitted to segments of each experiment. All equations had the same final infiltrability fc, but adjusted values for coefficients f0 (initial infiltrability) and Kf (exponential decay constant) were fitted for periods of rainfall that followed significant hiatuses in rainfall, during which subsurface redistribution allowed near-surface soil suction to recover. According to the fitted Horton equations, soil infiltrability recovered by up 10–24 mm/h during intra-event rainfall hiatuses of 15 to 20-min duration, contributing to higher overall event infiltration rates and to reduced runoff ratios. The recovery of infiltrability also reduced the size of runoff peaks following periods of low intensity rainfall, compared with the predictions based on single Horton infiltration equations, and in some cases, no runoff at all was recorded from late intensity peaks. The principal finding of this study is that, using a set of affine equations, the intra-event time variation of soil infiltrability can be tracked through multiple intensity bursts and hiatuses, despite the lack of steady-state conditions. Copyright