When runoff yield was measured over crusted (i.e., surfaces covered by 1–3 mm of cyanobacterial crust) and scalped 2–6 m2 plots at the Hallamish dune field (Negev Desert, Israel) during 5 winter seasons (1990–1995), no runoff was generated from the scalped plots during the first year following scalping while meagre quantities were received during the second year. Nevertheless, during the third season and onward, runoff yield at the scalped plots was consistently and significantly higher than that of the crusted (control) plots and this was so albeit the fact that the crust chlorophyll content of the scalped plots during the third season was as low as half the amount of the crusted plots. These findings appeared to contradict previous findings that showed the existence of a positive linear relationship between the crust chlorophyll content and runoff yield. The apparent contradiction was explained by the differences in the plot microrelief, indicating that minute differences, even of a millimeter-scale, may greatly affect runoff yield and thus may compensate for lower crust chlorophyll content. Surface smoothness, which characterizes many of the arid zones, may thus explain, at least partially, the high runoff yield obtained in deserts.