5 m × 0.5 m) indicated that

the average steady infiltration rate decreases with slope gradient in this region (Li et al., 1995). However, the loess soil is very susceptible to soil crust (Luk and Cai, 1990). The development of soil crust can significantly selleck screening library decrease infiltration rates (Römkens et al., 1990a). Luk and Cai (1990) observed that multiple cycles of soil crust development and destruction occur in the rainfall processes. Zhang and Cai (1992) found that soil crustability of loess is varied with slope gradients. The rainfall intensity also affected surface crust development (Römkens et al., 1990b). In addition, rill development is very active on the sloping lands in this region and the threshold of rill formation is varied with slope gradients and rainfall intensity (Wang and Zhang, 1992). Infiltration between inter-rill and rill areas may be different due to the destruction of crusts in rill areas. The combined effect of the above individual factors on runoff generation was highly complicated selleck chemicals llc and difficult to separate. At slope angles of 5°, 10°, 15°, 20°, 25°, and 30°, the mean annual soil loss per unit area was 1633.5, 1941.1, 3278.5, 3896.3, 4663.8, and 6658.2 g/m2 on SSP, in comparison of 2320.3,

2109.2, 2752.4, 3417.4, 3238.1, and 5878.8 g/m2 on LSP. Soil loss per unit area increased with slope steepness in both SSP and LSP (Fig. 6b). Although LSP generated 36.4% less annual runoff per unit area than SSP, ranging from 25.7% at 15° to 46.7% at 30°, they produced an average of only 3.6% less annual soil loss per unit area than SSP. In addition to the difference in rainfall between the two periods, this may also imply that the runoff infiltration and detention on long slope was higher than that on short slope, and that the concentrated flows on long slope had greater flow velocities Metformin and thereby erosion power than runoff generating from short slope (Wischmeier, 1972 and Lal, 1982). The annual runoff and soil loss per unit area showed wide variations

among years of observation on both SSP and LSP (Supplementary Table 2). The coefficient of variation ranged from 0.59 to 0.73 in runoff and 0.56–1.18 in soil loss on SSP, in comparison of 0.91–1.26 in runoff and 0.67–1.83 in soil loss on LSP. This reflected the great variation in precipitation among years. As an extreme, there was no runoff and soil loss on LSP in 1965. The year had the lowest annual precipitation of 243.3 mm, among which 126.9 mm fell in the rainy season but none of it generated runoff. However, annual soil loss did not increase linearly with yearly precipitation either. The greatest yearly precipitation in 1964 did not produce the highest soil loss on LSP. The highest annual soil loss occurred on SSP in 2000. That year had a total of precipitation of 487.2 mm, which was even considerably below the mean annual precipitation of 522 mm over the7-year SSP monitoring period.