The effect of soil crust on the generation of runoff on small plots in an arid environment

In arid zones, runoff is frequently generated as a result of the crust development on the soil surface. This crust is a thin layer of greater density, high shear strength, finer pores and has a lower saturated hydraulic conductivity than the underlying soil. The objective of the research reported herein was to study the factors that influence the generation of runoff in small plots under natural rainfall conditions. Factors studied were crust permeability, roughness, soil salt content and time gaps between the rain showers. The field trial was carried out in the Mashash experimental runoff farm in Israel's Negev desert. Runoff was measured on eight plots using a tipping-bucket system (resolution 0.01 mm s^{- 1}). Rainfall intensity was recorded on-site with a rainfall gauge (resolution 0.25 mm s^{- 1}). Two treatments were studied: long-term rainfall-induced crusts (LTC) that had developed over a period of years (three plots), and complete destruction of the crust (ICU) by cultivation to a depth of 0.2 m with a rotary tiller before the beginning of the trial (five plots). Surface roughness was characterized by the surface RMS height obtained from laser micro-relief measurements before and during the season. Prior to the onset of rain, roughness was similar for all the ICU plots. One month thereafter, roughness had decreased sharply, but exhibited no further change until the end of the season. Roughness of LTC plots did not change during the season and was lower than that of the ICU plots. After ~ 21 mm of cumulative rain, the average runoff yield was similar for both ICU and LTC plots, even though roughness in the former did not reach the low values of the LTC plots. Although the variability in roughness among individual LTC plots was very small, large differences were observed in the collected runoff. The same phenomenon was observed for the ICU plots. Moreover, the runoff yields in two ICU plots were consistently higher than those in two LTC plots while three other ICU plots produced much less runoff. Apparent saturated hydraulic conductivity (AHC) was measured on mounds and depressions. In the ICU plots the average AHC of the mounds was markedly higher than that of the depressions while the AHC values of mounds were very similar. No mounds or depressions were observed in the LTC plots, and their AHC was similar to that of depressions in ICU plots. The results indicate that the apparent saturated hydraulic conductivity of the upper soil layer was not directly linked to the runoff generation. Treatment effect was significant only for the first two rainfall events, but the presence of salts in the upper soil layer significantly affected runoff generation during the last four rain events of the season. Analysis of runoff and time gaps between the runoff-producing rain showers showed a clear relationship between runoff yields and average rainfall intensity, the degree of correlation between them improving with a decrease in the length of the gap.