Effects on intra-event-based flood runoff and sediment characteristics
Between the 1960s and 1990s, there was no significant change in rainfall in the Chabagou watershed35. The mean values of runoff and sediment transport in the baseline period and measurement period were calculated. Regardless of rainfall influence, the effect of conservation measures was assessed by the time series contrasting method25.
Table 1 shows the statistics of the characteristics of event-based flood flows and sediment in 1961–1990 (excluding 1970). Compared with those in the baseline period, T and Tr in the measurement period increased by 16.54% and 29.21%, respectively; however, Tp decreased by 55.52% in the measurement period, which showed that the soil and water conservation measures extended the flood duration while reducing the time of increased discharge. Under identical rainfall conditions, long-duration runoff with less time for increased discharge could cause less erosion than short-duration runoff with more time for increased discharge36. Hence, the conservation measures reduced soil erosion by prolonging the flood duration and reducing the time to peak. In addition, the hydrodynamic indices qp, H and qm were 75.2%, 56.0% and 68.0% lower, respectively, in the measurement period than in the baseline period. Moreover, E in the measurement period was only 10.2% that in the baseline period. The results showed that the conservation measures greatly reduced the hydrodynamic energy and thus soil erosion. In addition, the relative erosion indicators SSY, SCE and MSCE, decreased 69.2%, 33.3%, and 11.9%, respectively, in the measurement period compared with the baseline period, which indicated that the conservation measures significantly reduced soil erosion and decreased the mean sediment concentration, although the reduction in the maximum sediment concentration was relatively small. The conservation measures, especially the engineering measures, reduced the runoff velocity, extended the flood duration, and reduced the peak discharge, which sharply reduced the runoff erosion power37,38. As a consequence of the decrease in erosive energy, soil erosion was diminished.
Influence on intra-event-based flood regimes
Classification of flood events and the characteristics of baseline period flood regimes
Figure 2 shows the clustering results of the flood events at the Caoping hydrological station in 1961–1969. The flood events were divided into 4 regimes with a significance level of p < 0.001. The data in the scatter diagrams of different discriminant functions were clustered, which indicated that the classification results were reasonable.
Discriminant analysis of different flood regimes in 1961–1969. F1 and F2 represent the scores of discriminant functions. Regime A: short flood duration and low erosive energy; Regime B: short flood duration and high erosive energy; Regime C: long flood duration and low erosive energy; Regime D: long flood duration and high erosive energy.
The discriminant functions were as follows:
$$F_{1} = 0.00{4}T + 0.00{1}q_{p} – 0.{22}H – {4}.{6}$$
$$F_{2} = 0.00{1}T – 0.00{1}q_{p} + 0.{292}H – {2}.{581}$$
$$F_{3} = 0.00{8}q_{p} – 0.{3}0{5}H – 0.{76}$$
The classification functions of the different regimes were as follows:
$$D_{1} = 0.0{25}T + 0.0{1}q_{p} – 0.{878}H – {23}.{927}$$
$$D_{2} = 0.0{11}T + 0.00{7}q_{p} – 0.{24}H – {6}.{495}$$
$$D_{3} = 0.0{4}T + 0.0{13}q_{p} – {1}.{456}H – {61}.{74}$$
$$D_{4} = 0.0{14}T + {2}.{445}H – {56}.{3}0{2}$$
where F1, F2, and F3 represent the scores of discriminant functions and D1, D2, D3 and D4 represent the classification scores of regimes A, B, C and D, respectively.
Based on the classification of the baseline period (1961–1969), the flood events of the measurement period (1971–1990) were discriminated with a significance level of p < 0.001; Fig. 3 presents the cluster results. The classification results were reasonable considering the scatter diagrams of the different discriminant functions.
Discriminant analysis of different flood regimes in 1961–1990 (excluding 1971). F3 and F4 represent the scores of discriminant functions. Regime A: short flood duration and low erosive energy; Regime B: short flood duration and high erosive energy; Regime C: long flood duration and low erosive energy; Regime D: long flood duration and high erosive energy.
The discriminant functions were as follows:
$$F_{4} = 0.00{3}T + 0.00{1}q_{p} – 0.{233}H – {5}.{2}$$
$$F_{5} = 0.00{1}q_{p} + 0.{288}H – {1}.{621}$$
$$F_{6} = 0.0{11}q_{p} – 0.{447}H – 0.{65}$$
The classification functions of the different regimes were as follows
$$D_{5} = 0.0{21}T + 0.0{14}q_{p} – {1}.{279}H – {19}.{749}$$
$$D_{6} = 0.0{1}T + 0.0{11}q_{p} – 0.{54}H – {6}.{1}$$
$$D_{7} = 0.0{34}T + 0.0{18}q_{p} – {1}.{779}H – {49}.{331}$$
$$D_{8} = 0.00{7}T + 0.0{18}q_{p} + {2}.{695}H – {64}.{322}$$
where F4, F5, and F6 represent the scores of discriminant functions and D5, D6, D7 and D8 represent the classification scores of regimes A, B, C and D, respectively.
Table 2 describes the classification results and the characteristics of different flood regimes. During the baseline period, the flood durations of regimes A and B were short, whereas the flood durations of regimes C and D were long. The qp, H, E, SSY and SCE of regime A, which accounted for 42.86% of all flood events, were small. The T of regime B, which accounted for 44.90% of the flood events, was the shortest, but the qp, H, E, SSY and SCE of regime B were large. Regime C, which accounted for 8.16% of all flood events, had the longest T, but the qp, H, E, SSY and SCE of regime C were small. The qp, H, E, SSY and SCE of regime D, which represented 4.08% of all flood events, were the largest. The runoff erosive energies of regimes A and C were smaller than those of regimes B and D, respectively.
Effect on intra-event-based flood regimes
The average T of the measurement period was 1.17 times longer than the T of the baseline period. In addition, qp decreased by 75.2% in the measurement period. E in the measurement period accounted for only 10.2% of that in the baseline period (Table 1). Consequently, in the measurement period, the flood events transitioned from regimes B and D, which have high erosive energy, to regimes A and C, which have low erosive energy. Compared with those in the baseline period, the proportions of regime A and regime C flood events increased by 33.7% and 94.2%, respectively, during the measurement period; regime B flood events decreased from 44.9% to 26.8%, and regime D flood events did not occur in the measurement period.
Because the conservation measures weakened the erosive energy of runoff, other characteristics within the same regime changed between the measurement period and baseline period. The qp, H, SSY and SCE of regimes A and B were smaller in the measurement period than in the baseline period, and the E of regimes A and B decreased by 79.6% and 87.4%, respectively, in the measurement period. Due to the increase in T and the decrease in erosion in the measurement period, regime D, which is the regime with the maximum erosive ability, transitioned into regime C, which has a long T and low erosive energy. Therefore, the variables of regime C, such as T, qp, H, qm, SSY and SCE, increased in the measurement period compared with the baseline period. In addition, the qp, H, qm, SSY and SCE of regime C were larger than those of regimes A and B in the measurement period and smaller than those of regime D and regimes C/D in the baseline period.
Source: Ecology - nature.com
