Surface runoff process
The changing trends of runoff in most treatments were similar at both 50 and 100 mm h-1(Figure 6), including two stages: 1) a low starting rate followed by a dramatic increase during the initial runoff-yielding period, and 2) a relatively stable rate that persisted until the end of rainfall experiment. However, the regular changing trends could be interfered with by a ridge rupture in the Hr and Hr+Cm treatments, with runoff rates suddenly rising in the Hr-treated plot at 40 and 25 min under the rainfall intensities of 50 and 100 mm h-1, respectively, and in the Hr+Cm treatment at 60 min under 100 mm h-1 rainfall. In comparison, the average runoff rate of CK was greater than that of Cm, Vr, Hr, and Vr+Cm by 2.9-, 0.8-, 5.0-, and 1.9-fold at 50 mm h-1, respectively, and by 1.8-, 1.0-, 1.7-, and 1.2-fold, respectively, at 100 mm h-1. In addition, the average runoff rate of CK was 3.7-fold greater than that of Hr+Cm at 100 mm h-1.
Compared to CK, the Cm, Hr, and Hr+Cm treatments reduced the runoff loss rates significantly on all points within the entire rainfall experiment (Figure 6). At 50 mm h-1, Hr showed a better capacity for controlling runoff loss rates than Cm. Vr had no notable effects on runoff loss rates at most of the points at 100 mm h-1but could promote the loss rate significantly at 50 mm h-1, including the whole process except for the runoff-yielding point. The runoff loss rates of Vr+Cm were significantly lower than those of CK at 50 mm h-1, with an average runoff rate of 53.6%, while the reduction was very limited at 100 mm h-1.
Figure 6 also illustrates that the stable runoff rates were lower at 50 mm h-1 than at 100 mm h-1 for all treatments. The runoff rates of CK, Cm, Vr, Hr, Vr+Cm, and Hr+Cm stabilized at approximately 91.8, 30.1, 118.7, 20.3, 48.2, and 0 mL s-1 at 50 mm h-1, respectively, and at 198.6, 117.4, 192.5, 122.9, 176.1, and 49.9 mL s-1at 100 mm h-1, respectively.
The results suggested that the mulching treatments, including Cm, Hr+Cm, and Vr+Cm, were able to mitigate rate-changing magnitudes compared to the corresponding tillage measures without mulching, that is CK, Hr, and Vr, indicating that more rainfall was infiltrated or stored under the treatments with mulching compared to those without mulching.
Sediment yielding process
As shown in Figure 7, the sediment loss rates in most treatments varied based on the changing trends of the runoff loss rate (Figure 6), with a relatively low starting level and then varied within a certain range based on rainfall intensity. The four conservation practices could effectively reduce soil loss rate compared to the conventional tillage of CK and Vr, except that the ridges ruptured, and the Vr treatment obviously enhanced the soil loss rate compared to CK. In comparison, the average soil loss rates of CK were 10.04-, 3.68-, and 6.59-fold greater than those of Cm, Hr, and Vr+Cm at 50 mm h-1, respectively, and 13.03-, 3.03-, 16.18-, and 12.60-fold greater than those of Cm, Hr, Vr+Cm, and Hr+Cm at 100 mm h-1, respectively. However, the averaged soil loss rates of Vr were 7.03- and 2.28-fold greater than those of CK at 50- and 100-mm h-1, respectively.
As mentioned earlier in the Materials and Methods section, the regular changing trends were interfered by the ridge rupture in the Hr or Hr+Cm treatments. As Figure 5 shows, Hr could reduce the sediment loss rate throughout the entire rainfall process, averaging 81.95% and 68.40% of CK under the two rainfall intensities, but two of the three ridge rupture time points made the instantaneous rates higher than the earlier rates. The impact of ridge rupture was greater at 100 mm h-1 than at 50 mm h-1, and the subsequent soil loss rates would stay higher thereafter, rather than being at the former level at 100 mm h-1, which dropped to former rates under 50 mm h-1.
During rainfall events, the mean soil loss rates in the three mulching treatments of Cm, Vr+Cm, and Hr+Cm were approximately 0.01, 0.02, and 0 g s-1at 50 mm h-1, and 0.09, 0.07, and 0.09 g s-1 at 100 mm h-1, respectively, being significantly lower than those of CK, which were approximately 0.15 and 1.18 g s-1 at 50 and 100 mm h-1, respectively. The soil loss rates of these mulching treatments were also lower than those of the non-mulching treatments, such as Vr and Hr, which were approximately 1.02 and 0.04 g s-1 at 50 mm h-1 and 2.70 and 0.39 g s-1 at 100 mm h-1, respectively (Figure 7). In addition, mulching also mitigated the changing trends of sediment loss rate, i.e., restricting the rate variation magnitude to a lower scale. Therefore, the mulching treatments were more effective in controlling the sediment yield compared to the non-mulching treatments.