4.1. Effects of slope gradient and vegetation cover on runoff
In a field-based runoff scouring experiment, we found that variation of total runoff under different vegetation covers are divided into three stages (slow at first in 0 to 5 mins, then increased rapidly in 5 to 10 min, and then leveled off in10 to 30 mins). The complicated phenomenon is due to the overall effects of several factors including scouring intensity, surface conditions (roughness, sealing layer on topsoil), vegetation covers, overland flow continuity, soil particles and infiltration capacity etc. (Lavee and Poesen, 1991; Cerdà, 2001; Chen et al., 2013). These factors have different effects on the runoff. For example, the scouring intensity affected sealing layer formation (surface conditions) which can significantly reduce infiltration rates (Assouline and Ben-Hur, 2006). Meanwhile, soil infiltration rates is constantly changing. It is also affected by many factors including soil moisture, soil texture, surface roughness, and scouring characteristics. During the process of runoff generation, all these factors are interactive and the overall effects of them are far from perfectly understood (Zhang et al., 2018). In this study, runoff is unstable and changes slowly at first in 0 to 5 mins due to the effect of surface roughness and overland flow continuity. With scouring time increasing, runoff increased rapidly in 5 to 10 min is affected by sealing layer formation. Then the runoff became leveling off in the later scouring period when spatial heterogeneity of infiltration is basically stable.
In addition, when slope gradient was less than 15°, the difference of mean runoff and total runoff was obvious under different vegetation covers, while when slope was steeper than 15°, the difference was less clear. Variation in runoff that occurred on slopes more than 15° was always lower than that on slopes less than 15°, indicating there is a critical slope of runoff change. This result also suggested that the accumulative infiltration on slopes less than 15° was higher than that on slopes more than 15°. This is not exactly the same as the results by some study that the critical slope gradient for overland flow. For example, A studies have found that infiltration decreased greatly with increasing slope when the slope was less than 18°, while infiltration was less influenced by slope beyond 18° (Cheng et al., 2008). Another study conducted by Jin (1996) indicated that the turning point of slope was 15°. Although the results are different, they are basically between 15° and 20° slope (Liu et al., 2017), which supports the results of the present study.
Besides slope gradient, the difference in runoff may be also attributed to vegetation, and interactions between slope gradient and vegetation. Similarly, Sadeghi et al. (2013) emphasized the complex circumstances which govern the interaction between runoff and soil loss. Some research on the effects of vegetation cover on slope runoff erosion indicated that when rainfall intensity reached a steady condition, runoff was negatively correlated with vegetation cover (Joshi and Tambe, 2010; Mohammad and Adam, 2010; Nunes et al., 2011; Zhu et al., 2010; Wang et al., 2016), these studies only considered the effects of vegetation cover and rainfall. However, we found that under different vegetation covers, the changes in runoff tended to be consistent on slopes >15°, and the runoff date curves were basically coincident on the 20° slope, indicating the effect of vegetation cover on runoff was more significant at medium slope gradients (5° and 10°), and slope was more significant at steep slope gradients (15° and 20°). According to the results of this study, the 15° slopes may be a critical threshold for the effect of slope gradient and vegetation cover to overland flow and soil loss.