Sediment transport characteristics in a highly erosive catchment based on flood events in the northern earth-rocky mountainous areas, China

Abstract

The comprehensive investigation of sediment transport during flood events offers valuable insights into the hydrological and erosion processes of watersheds. It also plays a crucial role in flood disaster prevention and control. In this study, we focused on the northern earth-rocky mountainous areas of China and employed K-medoids clustering to classify a total of 261 flood events spanning from 1959 to 2021 into four distinct types. By comparing the sediment transport characteristics of different flood types and periods, as well as analyzing sediment source distribution using SSC-Q hysteresis loops, we obtained the following results. 1) The study period witnessed a notable decrease in the annual number of flood events and sediment yield, with reductions of 48.09% and 34.01%, respectively. The suspended sediment concentration during flood events exhibited a substantial decline of 76.54% compared to the baseline period. 2) In the Mihe River Basin, the majority of sediment yield could be attributed to flood events classified as Types A and B. These flood types were characterized by short duration, high peak flow, and substantial runoff depth. Among them, Types A were significantly greater than Types B in terms of runoff and sediment transport. 3) The hysteresis loops observed in the Mihe River Basin predominantly displayed a figure-eight and clockwise pattern, indicating potential sediment sources within the river channel and banks. Addressing sediment challenges and ensuring sustainable watershed management practices require connecting these loop characteristics to future management initiatives, given abundant sediment sources near the Huangshan hydrological station.

Graphical abstract

Keywords

flood events / sediment yield / hysteresis loops / northern earth-rocky mountainous areas

1 Introduction

Rivers play a vital role in the natural water cycle by facilitating material and energy transfer within ecological systems (Walling and Fang, 2003; Walling, 2006; Kondolf et al., 2014; Tian et al., 2016). Streamflow serves as the primary pathway for sediment transport (Hu et al., 2019). Changes in runoff-sediment transport correspond to the impact of environmental changes on soil erosion processes in watersheds (Tian et al., 2016; Zhao et al., 2017). Studying the relationship between runoff, sediment transport, and rainfall during a single flood event can help identify the factors influencing sediment transport and the spatial distribution of sediment sources at the catchment scale (Zhao et al., 2017). This deep understanding of the response of soil erosion events to floods is crucial for reducing the risk of regional flood disasters and maintaining water resource security.

Approximately 24% of rivers worldwide have experienced significant changes in their hydrological regimes, which are attributed to human activities and climate change (Li et al., 2020; Zhang et al., 2021). Numerous studies have assessed the characteristics and potential drivers of variability in runoff-sediment transport at global and regional scales, analyzing sediment dynamics and their responses to climate change and land surface modifications across different time scales, such as annual, seasonal, and temporal scales (Horowitz, 2003; López-Tarazón et al., 2009; Gao et al., 2018; Hu et al., 2019). Besides climate, other factors, including changes in land use and land cover, as well as the implementation of soil and water conservation measures like reservoirs and check dams, can significantly affect both runoff and sediment load. It is important to note that the annual sediment transport in the watershed is primarily influenced by a few flood events (Estrany et al., 2009; Lana-Renault and Regüés, 2009; Fang et al., 2012).

“Large floods with high sediment loads” are a defining feature of sediment transport in rivers. Sediment transport in rivers peaks during the rainy season, with suspended sediment transport greatly exceeding bedload (Webb and Walling, 1986; Oeurng et al., 2010). During flood events, the peak sediment transport can happen before, during, or after the peak discharge (Heidel, 1956). Analyzing the temporal variation of suspended sediment concentration (SSC) and discharge (Q) within a single flood event is a commonly employed method for estimating suspended sediment load in rivers. The shape of hysteresis loops can differ depending on the type of rainfall, which indicates potential variations in sediment sources (Fang et al., 2011). Oeurng et al. (2010) previously reported that suspended sediment transport is influenced by various factors, including peak flood discharge, total water yield, and sediment availability. Additionally, Costa et al. (2018) investigated the factors affecting suspended sediment transport and developed a novel rating curve to assess the responses of sediment concentration and discharge to erosive rainfall, ice melt, and snowmelt in an Alpine catchment. Hu et al. (2019) also studied the dynamics of runoff and sediment under different flood patterns in the Loess Plateau catchment in China. These studies have provided valuable insights into the relationship between SSC and Q, as well as the changes in sediment regimes resulting from climate change, land use/cover change, wildfires, and reservoirs. However, previous rarely involved rainstorm, flood and sediment yield in the temperate monsoon region, which is densely populated and unique to China. In recent years, the changes in runoff and sediment yield of rivers caused by human activities and climate change need further elaboration.

The northern earth-rocky mountainous areas of China face the most severe hydraulic erosion, with soil erosion posing a significant constraint on regional development. The Mihe River Basin serves as a representative example of this region and exhibits unique soil and geographical characteristics, such as uneven and intense rainfall distribution, complex and diverse geological structures, and unsustainable human activities. These factors act as both internal and external driving forces contributing to soil erosion. Extensive soil loss due to shallow soil layers has led to severe soil and water loss. The detrimental impacts of soil erosion include reduced availability of usable land resources, declining soil fertility, reservoir siltation, and riverbed elevation. Furthermore, global climate change has increased the frequency and intensity of extreme rainstorms, resulting in more frequent occurrences of extreme floods in certain areas (Jongman et al., 2015). The combined effect of erosion and flood disasters has inflicted severe damage on the local ecological environment and caused degradation of ecosystem functions (Han et al., 2020). Despite the prevalence of heavy rainstorms in many regions, there remains a notable gap in our understanding of the dynamics of runoff and sediment transport. Expanding on this section emphasizes the urgent need for research focusing on runoff and sediment changes in the context of the northern rocky mountainous areas. Addressing this gap through comprehensive research efforts in The Mihe River Basin will be essential for developing effective mitigation and management strategies to safeguard both the environment and local communities from the adverse impacts of soil erosion and flooding. Investigating the dynamic processes of runoff and sediment during flood events in earth-rocky mountainous areas is crucial for understanding the mechanisms underlying soil erosion.

In this study, we aimed to achieve the following objectives: (i) assess the variations in the frequency and sediment transport of flood events during different periods and determine the contributions of human activities and climate change; (ii) investigate the effects of different flood types on sediment transport; (iii) identify the primary types of hysteresis loops (SSC-Q) and analyze the characteristics of sediment sources.

2 Material and methods

2.1 Study area

The Mihe River, situated in the western part of the Shandong Peninsula, extends over a length of 206 km and covers an area of 3847.5 km². For this study, we selected the hydrology station of Huangshan, located in the upper reaches of the Mihe River in Linqu County, which encompasses an approximate area of 375 km² (Fig. 1). In Linqu County, a series of soil and water conservation projects have been implemented since the 1980s. As a result, the area of land experiencing mild to severe soil erosion has witnessed a significant reduction of 50.11% compared to 1985. The predominant soil type in the area is brown loam, and the average annual rainfall measures 691 mm. The primary land use types consist of grassland and cropland. Most of the rainfall, characterized by short duration rainstorms, typically occurs between May and August, contributing to substantial soil erosion and flood disasters. Over the period between 1959 and 2021, the annual average runoff and sediment load at the hydrological gauging station amounted to 8.56 ×  107 m³ and 1.98  ×  105 t, respectively.

2.2 Data sources

Data on runoff and sediment were collected from 1959 to 2021 at the Huangshan hydrologic station located in the Mihe River Basin in Weifang City, Shandong Province, while precipitation data were gathered from 28 rainfall stations distributed throughout the study area (Guo et al., 2021). The collection of data followed international standards and methods for hydrological survey, sampling, and experimental determination. The published data underwent strict verification procedures to ensure its accuracy and reliability.

2.3 Data analysis

2.3.1 Flood event classification

We implemented a set of restrictions to effectively mitigate flood events with low sediment transport. These restrictions included specific thresholds, such as a minimum runoff depth of 0.10 mm, a flood peak exceeding 7.50 m³·s⁻¹, and a flood duration longer than 250 min (Hu et al., 2019). To classify flood events accurately, we employed the widely used clustering analysis method which groups similar objects together (Peng and Wang, 2012). Among the various clustering methods available, we employed the k-medoids algorithm, which employs the squared Euclidean distance measure and the k-means + + algorithm to determine initial cluster medoid positions. The use of k-medoids was essential for determining the necessary number of groups before classification. We then employed discriminant analysis, specifically Fisher’s discriminant function, to identify the most appropriate clusters. This method is based on actual hydrological variables such as H and SY to classify different floods that have occurred in the same basin for years. The operation is convenient and the results are objective. In this study, the clustering analysis was performed using SPSS 26 software, and flood types were classified based on the criteria of runoff depth, flood peak discharge, and flood duration.

2.3.2 Break analysis

Pettitt’s test (Mu et al., 2010) is a non-parametric method used to detect break points in hydro-climatic variables For a given time series X (x₁, x₂, …, x_n), divided into two samples x₁, x₂, ..., x_t and x_{t + 1}, x_{t + 2}, …, x_n, Pettitt’s test uses a version of Mann–Whitney statistic U_t,n, which is calculated as follows:

(1)U_{t,n}=U_{t-1,n}+\sum _{i=1}^n\mathrm{s}\mathrm{g}\mathrm{n}(x_t-x_i)\mathrm{f}\mathrm{o}\mathrm{r}t=2,\dots ,N,

(2)

where U_t,n counts the instances for which members of the first sample are greater than those of the second. In Pettitt’s test, the null hypothesis is the lack of a trend change point. The test statistic K_t,n and the associated probability (p) are derived as follows:

(3)K_{t,n}=max\left|U_{t,n}\right|\begin{array}{cc}& \end{array}(1⩽t⩽n),

(4)p\cong 2\exp \left\{-6(K_{t,n}{)}^2/\left(n^3+n^2\right)\right\}.

2.3.3 Flood variables

To characterize each flood event, multiple hydrological variables were utilized to describe the corresponding flood hydrographs and sediment delivery characteristics. Table 1 lists the relevant variables for runoff and sediment.

For a single flood event, Flood runoff depth (H, mm), Event sediment yield (SY, t·km⁻²), and Flow variability (FV) are calculated as follows:

(5)H\left(t_1,t_2\right)={\displaystyle \frac{{\int }_{t_1}^{t_2}{Q}_t\mathrm{d}t}{A}}={\displaystyle \frac{{\displaystyle \sum Q_t\mathrm{\Delta }t}}{A}},

(6)SY\left(t_1,t_2\right)=\frac{{\displaystyle {\int }_{t_1}^{t_2}{Q}_t·SS{C}_t\text{d}t}}{A}=\frac{{\displaystyle \sum Q_t}·SS{C}_t\mathrm{\Delta }t}{A},

(7)FV={\displaystyle \frac{Q_p}{Q_m}},

where Q_t and SSC_t describe instantaneous discharge and suspended sediment concentration, respectively; the time interval for the hydrological observations is Δt, and A is the control area of Huangshan Hydrology Station.

2.3.4 Hysteretic loops

The sediment transport characteristics of flood events are often studied by the sediment content-discharge (SSC-Q) hysteresis loops of flood processes. Flood event hysteresis loops comprise four types: complex, figure-eight, clockwise, and anticlockwise (Ren et al., 2020). The hysteretic loops type of a flood event is affected by several factors such as runoff, rainfall intensity, rainfall duration, rainfall spatial distribution, flood transport distance, and sediment source. (Hu et al., 2019). The hysteresis loops are related to the dynamic changes of sediment supply and sediment transport processes at the event scale, hence the sediment sources and sediment deposition locations in the watershed of different flood events can be studied by analyzing the relationship between sediment content and discharge during flood processes (Hu et al., 2020).

3 Results

3.1 Interannual variations in runoff and sediment loading

Pettitt’s test was employed to identify breakpoints in annual runoff and sediment. The results indicated no break in streamflow during 1959–2021 (Fig. 2(a)), but sediment discharge experienced a break in the year 1998 (Fig. 2(b)). Based on this, the periods of 1959–1998 and 1999–2021 were identified as the baseline and changing periods, respectively. Based on the break analysis, the double accumulation curves of rainfall runoff for the reference period and the change period were drawn and a linear regression equation established (Fig. 2(c)), as well as the double accumulation curves of rainfall sediment transport for the reference period and the change period (Fig. 2(d)). The cumulative runoff and sediment for the entire period can be simulated without human interference conditions using the reference period regression equation. The calculation results shown the runoff and sediment transport in the Mihe River Basin show a decreasing trend, with a decrease of 9897×10⁴ m³ in runoff and a decrease rate of 1.9%, and a decrease of 392.3×10⁴ t in sediment transport and a decrease rate of 24.9%.

Statistical analyses of annual streamflow and sediment discharge indicators for different periods were conducted (Table 2). The mean annual runoff during the changing period (1999–2021) decreased by 15.28% compared with the baseline period (1959–1999), while the mean sediment yield decreased by 71.16%.

To account for the effects of human interventions on sediment, double mass curves (DMCs) of cumulative sediment and cumulative precipitation were constructed (Fig. 2(c)). Some breakpoints in the DMCs were evident. The identified breakpoints in the DMCs were consistent with those determined using Pettitt’s test, and the detected results are both reasonable and significant. Linear regression was used to assess the relative contributions of precipitation and human activity on sediment changes. The contribution of human interventions to sediment reduction was 83.19% between 1999 and 2021, much higher than that of precipitation (16.81%). These results suggest that human interventions dominated the decline of sediment discharge. On the basis of these factors, the influence of human activities will further regulate the characteristics of sediment runoff during flood events. For example, in highly urbanized areas, flood events can cause more urban runoff and soil erosion to be washed into rivers, increasing the concentration and sediment content of sediment runoff.

3.2 Runoff and sediment characteristics of among flood events

The analysis of runoff-sediment relationships used a data set of 261 flood events from 1959 to 2021, which were categorized into five levels based on their sediment yield, namely relative small flood events (0–10 t·km⁻²), small flood events (10–100 t·km⁻²), medium flood events (100–500 t·km⁻²), large flood events (500–1000 t·km⁻²), and extreme flood events (> 1000 t·km⁻²).

Among the 261 flood events, small flood events had the highest proportion, accounting for 42.53%, followed by relatively small and medium flood events, accounting for 34.87% and 19.16%, respectively. The frequency of large flood events and extremely large flood events was relatively low, with only 7 and 2 times, accounting for 2.68% and 0.77%, respectively. The characteristics of flood event runoff and sediment in the study area are listed in Table 3.

3.3 Sediment transport characteristics of different flood types

The flood event types were classified based on the runoff depth, flood peak discharge, and flood duration (Hu et al., 2019). The 261 flood events were divided into four types (p < 0.01). The scatters of the discrimination functions for each flood type were well-clustered in four distinct regions, indicating a relatively reasonable clustering outcome (Fig. 3). Type C showed relatively decentralized scatters, while Types B and D showed more compact scatters. Additionally, the boundaries between Types B–D were unclear, while Type A was significantly different from the other three patterns. The scatter distribution implied that the flood characteristics of Types B–D shared some similarities, while Type A was markedly distinct.

Hydrological variables of different flood types are depicted in Fig. 4. Among all flood events, only two fell into Type A, characterized by the longest flood duration, the highest runoff depth, the largest flood peak, and the maximum flood variability. Type B had 25 flood events, characterized by a relatively larger runoff depth and flood peak, and a flood duration similar to Type A. Type C had 165 flood events, and was the most common type of flood event, with medium runoff depth, flow variability, duration, and flood peak. Type D had 69 flood events, with the least flood duration, the least runoff depth, the least flood peak, and the maximum flood variability.

3.4 Runoff and sediment relationship under different flood types

Streamflow is the primary means of sediment transport (Hu et al., 2019). Sediment transport in flood events is affected by flood duration, runoff depth, and peak discharge, among others. The runoff and sediment relationships in different flood events can be obtained using the stepwise multiple regression method. Of the four types of flood events, there were only two Type A flood events, the regression equations were not sufficiently accurate to meet the requirements of natural science research. Considering the similarity between Type A and B floods, their data were combined for the regression analysis.

Types A and B,

(8)SY=e^{121.102}{T}^{-2.199}{H}^{-0.024}{Q}_p^{1.869}{R}^2=\mathrm{0.973.}

Type C,

(9)SY=e^{-0.914}{T}^{0.063}{H}^{-0.202}{Q}_p^{1.004}{R}^2=\mathrm{0.725.}

Type D,

(10)SY=e^{-0.568}{T}^{0.033}{H}^{-0.235}{Q}_p^{1.168}{R}^2=0.853.

Regression equation coefficients for different types of floods revealed that sediment transport is negatively correlated with runoff depth and positively correlated with peak discharge for all types of floods. Types A and B floods are negatively correlated with flood duration, while Types C and D floods are positively correlated with flood duration. Flood events of Types A and B are characterized by short flood duration and large flood peaks and total volume (Kvočka et al., 2016).

3.5 Hysteretic loops

The hysteresis loops (SSC-Q) can reflect the source and transportation process of runoff sediment during a flood. The four types of hysteresis loops in the study area are listed in Table 4. Among the 261 flood events, the figure-eight hysteretic loop had the highest proportion, accounting for 42.52%, followed by the clockwise and complex loops, which accounted for 27.96% and 22.98%, respectively. The counter-clockwise curve had the lowest proportion, accounting for only 6.51%.

In Type D floods, the figure-eight curve loop had the highest proportion, but in Type C and B floods, the proportion of the eight-shaped curve decreased, while that of composite and clockwise curves increased. Moreover, the larger the flood scale, the greater the proportion of complex loops. The two Type A flood events were complex loops.

4 Discussion

4.1 Impacts of climate change and human activity on sediment

Statistical analysis of runoff and sediment transport indicators for different flood events at different times was conducted. A total of 201 floods were recorded during the baseline period from 1959 to 1998, while 60 floods were recorded during the changed period from 1999 to 2021. The number of annual flood events during the change period decreased by 48.09% compared with the baseline period, and the sediment yield during the changed period decreased by 34.01%.

Changes in sediment production are mainly affected by human activities, and the variability of sediment input is primarily controlled by the runoff discharge. Compared with rainfall, the sediment transport of rivers is highly sensitive to human activities, including deforestation, dam construction, and implementation of soil and water conservation measures. The soil and water conservation area of the low hills and hilly platforms in the Yimeng Mountains covers an area of 2.11 million km², with a population of about 11.12 million, a population density of 630 people/km², and a per capita cultivated land of 0.067 hm², of which 35% is sloping cultivated land. After years of governance, now the area of soil erosion above mild is 9300 km², still accounting for 44% of the land area. Changes in the suspended sediment discharge to runoff discharge (SSC) represent the impact of erosion environment and rainfall changes on sediment transport.

Based on the comparison of the average suspended sediment concentration of flood events (SSC_mean) during different periods, the impacts of human activities and climate change on sediment transport during floods can be determined. The SSC_mean in the change period decreased by 76.54% compared with the baseline period (Fig. 5), which was most significant in relatively small flood events, with a reduction of 80.39% in SSC_mean. In extremely large flood events, the reduction in SSC_mean was the smallest, at only 34.78%. These findings indicate significant impacts of human activities and climate change on sediment transport during flood events.

The impact of natural environmental factors and climate change on soil erosion and sediment transport within a watershed can be assessed by examining alterations in sediment transport at the watershed outlet (Aneseyee et al., 2020). These changes are influenced by human activities such as modifications in land use, vegetation cover, and the implementation of soil and water conservation measures (Zhang et al., 2020). For example, Feng et al. (2018) demonstrated that terracing reduced upstream soil erosion by 90%, whereas Li et al. (2017) reported an increase in sediment yield trapped by check dams from 27.7% during 1990−1999 to 78.3% during 2000−2012 in the Huangfuchuan watershed located in the northern region of the Loess Plateau, attributed to an increase in the number of check dams.

In addition, since the end of the 20th century, China has widely implemented the policy of returning farmland to forests, which aims to restore the natural environment in mountainous areas, especially upstream rivers, through natural enclosure and artificial planting (Liu et al., 2023). Enhancements in vegetation cover, for instance, result in increased rainfall interception and infiltration, decreased runoff and soil particle displacement, and reduced sediment transport capacity owing to heightened surface roughness (Cochard, 2013). Comparative analysis of rainfall events before and after surface modifications has revealed significant disparities in sediment yield. Additionally, Liu et al. (2013) explored the effects of vegetation restoration on sediment load reduction and found an impressive 81% decrease in sediment load. These findings strongly suggest that the alterations observed in the relationship between runoff and sediment, as well as changes in the magnitude and frequency of flood events, can be ascribed to the effective implementation of soil and water conservation measures.

4.2 Influence of flood types on sediment yield

Types A–D floods contributed 27.22%, 33.64%, 29.94%, and 9.20% respectively to the total sediment yield. For single flood events, the mean sediment yield of Type A floods is much higher than that of other types. Meanwhile, the mean sediment yield of Types C and D floods is only 13.48% and 9.91% of that of Type B floods, respectively.

Types A (0.77%) and B (9.58%) floods contribute 60.86% of the total sediment yield and are characterized by short flood duration, high peak discharge, and large runoff depth, consistent with previous findings indicating that sediment yield in a watershed is mainly concentrated in a few flood events (Estrany et al., 2009; Lana-Renault and Regüés, 2009; Fang et al., 2012). Previous studies have also shown that the runoff-sediment relationship during large floods is relatively stable and does not significantly change with vegetation restoration measures (Zheng et al., 2008; Zheng et al., 2013). For larger flood events, the reduction in sediment concentration at the outlet of the watershed can be attributed to a decrease in runoff volume, i.e., the sediment reduction rate is roughly equal to the runoff reduction rate. Additionally, it is important to consider that changes in sediment yield may also be influenced by factors such as rainfall characteristics. However, in this study, Types A and B flood events showed an 18.07% increase in runoff depth and a 40.03% decrease in sediment yield during the change period, indicating that the vegetation restoration measures have also changed the runoff-sediment relationship during large floods, contrary to previous findings. This difference may be due to the study area; previous research has mainly focused on the loess plateau region with a high sediment content (Zheng et al., 2008; Zheng et al., 2013). Therefore, for Types A and B floods, watershed sediment control should focus on a combination of slope and channel measures, and rational allocation of soil and water conservation measures to reduce sediment yield.

Conversely, for Types C and D floods, there exist differences in the relationship between runoff sediment transport. The variability in runoff sediment transport refers to the fluctuation or inconsistency in the transport of sediment by runoff, which can be influenced by factors such as precipitation intensity, land use changes, and soil erosion patterns. As vegetation recovery measures are continuously implemented, sediment transport decreases. The sediment yield of Types C and D floods decreased by 75.17% and 90.57%, respectively, during the changed period. Type C floods have a very high frequency (63.22%) and contribute the second highest sediment yield among all types of floods, following Type B floods. Type D floods have a relatively high frequency (26.44%); their low sediment yield per event indicates that small floods contribute very little to the total sediment yield at the outlet (9.20%). However, the role of small flood events in the geomorphic processes of the watershed and river channels should be considered (Zheng et al., 2008). Especially for rivers prone to high sediment flows, the sediment trapped by small flood events may temporarily deposit in the channel, becoming the main sediment source for large flood events. For this type of flood event, controlling sediment in the watershed is achieved primarily by changing the runoff sediment relationship on slopes.

4.3 Hysteresis analysis

The flood event-based sediment dynamics can be studied through the analysis of SSC-Q hysteresis loops. The application of SSC-Q hysteresis loops have been widely used to estimate the distance of sediment availability (Williams, 1989; Smith and Dragovich, 2009; Gentile et al., 2010). The sediment yield in the watershed is not only controlled by the sediment transported through the channel network by streamflow but also by the sediment yield in the headwater areas and the area close to the outlet, as indicated by different hysteresis loops (Hu et al., 2019). The variability in sediment yield within the watershed could also be attributed to significant differences in sediment sources and land use distribution across various regions of the watershed. The distribution of sediment sources, such as eroding hillslopes, agricultural fields, urban areas, and natural landscapes, varies across the watershed, influencing the magnitude and composition of sediment transported through the channel network. Additionally, the land use distribution within the watershed, including agricultural practices, deforestation, urban development, and conservation efforts, plays a crucial role in determining erosion rates and sediment delivery processes. Variations in land use practices result in differential sediment production and transport mechanisms, leading to diverse sediment yield patterns observed in different parts of the watershed (Zhou et al., 2019). The complex hysteresis loop indicates that the flood event may have a prolonged duration and could be a result of multiple flood events and peaks occurring in succession (Figs. 6(a) and 6(b)). This complexity can be attributed to the heterogeneous distribution of rainfall across the watershed, leading to varied hydrological processes and sediment sources (Tian et al., 2022).

In the clockwise hysteresis loop (Fig. 6(d)), the peak of sediment discharge arrives before the flood peak (Fig. 6(c)), indicating that the sediment source is relatively close to the measurement station, and the sediment supply is abundant and rapid compared with the rising flood rate (Klein, 1984; Rovira and Batalla, 2006). The predominance of clockwise loops in the Mihe River Basin suggests the presence of rich sediment sources near the measurement stations. Compared with counterclockwise loops, clockwise loops are more common in humid areas (Buendia et al., 2016; Ren et al., 2020). In contrast, in the counterclockwise hysteresis loop (Fig. 6(f)), the flood peak arrives before the sediment peak (Fig. 6(e)), and sediment is lacking during rising water while abundant during falling floods, indicating a widespread distribution of sediment sources in the upstream area (Heidel, 1956). The counter-clockwise loops are more frequently observed in the sediment-rich region.

In this study, most of the occurred flood events during the research period exhibited figure-eight hysteresis loops (Fig. 6(h)), a combination of clockwise and counter-clockwise loops, and intersect at a point where the ratio of sediment concentration to discharge is the same at two different times during the flood event. During figure-eight events, the sediment supply is complicated (Gentile et al., 2010). One possible explanation is that, at the onset of the flood, the sediment source may be located in close proximity to the measuring station, enabling a rapid supply of sediment (Hu et al., 2019). The rising rate of floods is lower than the sediment transport rate. However, as the flood duration increases, the sediment is depleted, and hence sediment from upstream is required to replenish it. The rising rate of floods starts to exceed the sediment transport rate. In the later stages of the flood, the upstream sediment arrives, and the rate of decrease in sediment concentration is lower than the falling rate of water, which leads to the figure-eight hysteresis loop.

The hysteresis loops of the Mihe River Basin are mainly figure-eight and clockwise, indicating high sediment content in the river channel and bank. Intense rainfall events may mobilize sediment from the upstream slopes of the basin. Under such conditions, sediment from riverbanks and within the river channel may increase the sediment load at the beginning of a flood event. In the later stages of a flood event, sediment supply is mainly provided by erosion and sediment yield on the upstream hillslopes. The findings suggest abundant sediment sources in the river channels or banks near the Huangshan hydrological station, which will require significant attention for future runoff and sediment regulation and management.

5 Conclusions

This study examines changes in the runoff-suspended sediment relationship and characteristics of suspended sediment during flood events in the Mihe River Basin from 1959 to 2021. A total of 261 flood events were analyzed. The key findings are as follows.

1) Compared to the baseline period, the number of annual flood events during the change period decreased by 48.09%, and sediment yield decreased by 34.01%. Mean suspended sediment concentration during flood events decreased by 76.54%. Human activities accounted for 83.19% of the reduction in sediment transport, while climate factors contributed 16.81%.

2) Types A and B floods, characterized by short duration, high peak discharge, and large runoff depth, primarily contribute to the total sediment yield in the Mihe River Basin. Therefore, managing these two types of flood events should be a priority in controlling sediment yield.

3) Hysteresis loops in the Mihe River Basin are primarily figure-eight and clockwise in shape. Intense rainfall events can mobilize sediment from the basin’s upstream slopes, contributing to increased sediment load at the onset of flood events. Abundant sediment sources near the Huangshan hydrological station, warranting careful consideration in future runoff and sediment regulation management strategies. Explicitly connecting these hysteresis loop characteristics to future management initiatives will be crucial for addressing sediment-related challenges and ensuring sustainable watershed management practices in the Mihe River Basin.

References

[1] Aneseyee A B, Elias E, Soromessa T, Feyisa G L (2020). Land use/land cover change effect on soil erosion and sediment delivery in the Winike watershed, Omo Gibe Basin, Ethiopia.Sci Total Environ, 728: 138776

[2] Buendia C, Vericat D, Batalla R J, Gibbins C N (2016). Temporal dynamics of sediment transport and transient in‐channel storage in a highly erodible catchment.Land Degrad Dev, 27(4): 1045–1063

[3] Cochard R (2013). Natural hazards mitigation services of carbon-rich ecosystems.,

[4] Costa A, Anghileri D, Molnar P (2018). Hydroclimatic control on suspended sediment dynamics of a regulated Alpine catchment: a conceptual approach.Hydrol Earth Syst Sci, 22(6): 3421–3434

[5] Estrany J, Garcia C, Batalla R J (2009). Suspended sediment transport in a small Mediterranean agricultural catchment.Earth Surf Process Landf, 34(7): 929–940

[6] Fang N F, Shi Z H, Li L, Guo Z L, Liu Q J, Ai L (2012). The effects of rainfall regimes and land use changes on runoff and soil loss in a small mountainous watershed.Catena, 99: 1–8

[7] Fang N F, Shi Z, Li L, Jiang C (2011). Rainfall, runoff, and suspended sediment delivery relationships in a small agricultural watershed of the Three Gorges area, China.Geomorphology, 135(1−2): 158–166

[8] Feng T J, Wei W, Chen L D, Rodrigo Comino J, Die C, Feng X R, Ren K, Brevik E C, Yu Y (2018). Assessment of the impact of different vegetation patterns on soil erosion processes on semiarid loess slopes.Earth Surf Process Landf, 43(9): 1860–1870

[9] Gao G Y, Fu B J, Zhang J J, Ma Y, Sivapalan M (2018). Multiscale temporal variability of flow-sediment relationships during the 1950s–2014 in the Loess Plateau, China.J Hydrol (Amst), 563: 609–619

[10] Gentile F, Bisantino T, Corbino R, Milillo F, Romano G, Liuzzi G T (2010). Monitoring and analysis of suspended sediment transport dynamics in the Carapelle torrent (southern Italy).Catena, 80(1): 1–8

[11] Guo C, Liu W, Wang Q (2021). Statistical analysis of historical rainfall in Mihe River Basin, Shandong Province.Water Resourc Plan Design, 11: 52–55+109

[12] Han J Q, Ge W Y, Hei Z, Cong C Y, Ma L, Xie M X, Liu B Y, Feng W, Wang F, Jiao J Y (2020). Agricultural land use and management weaken the soil erosion induced by extreme rainstorms.Agriculture, Ecosystems & Environment, 301(1):

[13] Heidel S G (1956). The progressive lag of sediment concentration with flood waves.Eos (Wash DC), 37: 56–66

[14] Horowitz A J (2003). An evaluation of sediment rating curves for estimating suspended sediment concentrations for subsequent flux calculations.Hydrol Processes, 17(17): 3387–3409

[15] Hu J, Gao P, Mu X, Zhao G, Sun W, Li P, Zhang L (2019). Runoff-sediment dynamics under different flood patterns in a Loess Plateau catchment, China.Catena, 173: 234–245

[16] Hu J, Zhao G, Mu X, Hörmann G, Tian P, Gao P, Sun W (2020). Effect of soil and water conservation measures on regime-based suspended sediment load during floods.Sustain Cities Soc, 55: 102044

[17] Jongman B, Winsemius H C, Aerts J C, Coughlan de Perez E, van Aalst M K, Kron W, Ward P J (2015). Declining vulnerability to river floods and the global benefits of adaptation.Proc Natl Acad Sci, 112: E2271–E2280

[18] Klein M (1984). Anti clockwise hysteresis in suspended sediment concentration during individual storms: Holbeck catchment; Yorkshire, England.Catena, 11(2−3): 251–257

[19] Kondolf G M, Gao Y, Annandale G W, Morris G L, Jiang E, Zhang J, Cao Y, Carling P, Fu K, Guo Q, Hotchkiss R, Peteuil C, Sumi T, Wang H W, Wang Z, Wei Z, Wu B, Wu C, Yang C T (2014). Sustainable sediment management in reservoirs and regulated rivers: experiences from five continents.Earths Futur, 2(5): 256–280

[20] Kvočka D, Falconer R A, Bray M (2016). Flood hazard assessment for extreme flood events.Nat Hazards, 84(3): 1569–1599

[21] Lana-Renault N, Regüés D (2009). Seasonal patterns of suspended sediment transport in an abandoned farmland catchment in the Central Spanish Pyrenees.Earth Surf Process Landf, 34(9): 1291–1301

[22] Li E H, Mu X M, Zhao G J, Gao P, Sun W Y (2017). Effects of check dams on runoff and sediment load in a semi-arid river basin of the Yellow River.Stochastic Environ Res Risk Assess, 31(7): 1791–1803

[23] Li L, Ni J R, Chang F, Yue Y, Frolova N, Magritsky D, Borthwick A G L, Ciais P, Wang Y, Zheng C, Walling D E (2020). Global trends in water and sediment fluxes of the world’s large rivers.Sci Bull (Beijing), 65(1): 62–69

[24] Liu E J, Zhang X P, Xie M L, Chen N (2013). Change in runoff and sediment in the upper catchment of Beiluo River under the vegetation restoration.Sci Soil Water Conserv, 11: 39–45

[25] Liu S, Dong Y, McConkey K R, Tran L S P, Wang F, Liu H, Wu G (2023). Scientific concept and practices of life community of mountains, rivers, forests, farmlands, lakes, grasslands, and deserts in China.Ambio, 52(12): 1939–1951

[26] López-Tarazón J A, Batalla R J, Vericat D, Francke T (2009). Suspended sediment transport in a highly erodible catchment: the River Isábena (Southern Pyrenees).Geomorphology, 109(3−4): 210–221

[27] Mu X, Zhang L, McVicar T R, Chille B, Gau P (2010). Analysis of the impact of conservation measures on stream flow regime in catchments of the loess plateau, China.Hydrol Processes, 21(16): 2124–2134

[28] Oeurng C, Sauvage S, Sánchez-Pérez J (2010). Dynamics of suspended sediment transport and yield in a large agricultural catchment, southwest France.Earth Surf Process Landf, 35(11): 1289–1301

[29] Peng T, Wang S J (2012). Effects of land use, land cover and rainfall regimes on the surface runoff and soil loss on karst slopes in southwest China.Catena, 90: 53–62

[30] Ren J Q, Zhao M D, Zhang W, Xu Q X, Yuan J, Dong B (2020). Impact of the construction of cascade reservoirs on suspended sediment peak transport variation during flood events in the Three Gorges Reservoir.Catena, 188: 104409

[31] Rovira A, Batalla R J (2006). Temporal distribution of suspended sediment transport in a Mediterranean basin: the Lower Tordera (NE SPAIN).Geomorphology, 79(1−2): 58–71

[32] Smith H G, Dragovich D (2009). Interpreting sediment delivery processes using suspended sediment-discharge hysteresis patterns from nested upland catchments, south-eastern Australia.Hydrol Processes, 23(17): 2415–2426

[33] Tian P, Feng J, Zhao G, Gao P, Sun W, Hörmann G, Mu X (2022). Rainfall, runoff, and suspended sediment dynamics at the flood event scale in a Loess Plateau watershed, China.Hydrol Processes, 36(2): e14486

[34] Tian P, Zhai J Q, Zhao G J, Mu X M (2016). Dynamics of runoff and suspended sediment transport in a highly erodible catchment on the Chinese Loess Plateau.Land Degrad Dev, 27(3): 839–850

[35] Walling D E (2006). Human impact on land–ocean sediment transfer by the world’s rivers.Geomorphology, 79(3−4): 192–216

[36] Walling D E, Fang D (2003). Recent trends in the suspended sediment loads of the world’s rivers.Global Planet Change, 39(1−2): 111–126

[37] Webb B W, Walling D E (1986). Spatial variation of water temperature characteristics and behavior in a Devon river system.Freshw Biol, 16(5): 585–608

[38] Williams G P (1989). Sediment concentration versus water discharge during single hydrologic events in rivers.J Hydrol (Amst), 111(1−4): 89–106

[39] Zhang H, Meng C, Wang Y, Wang Y, Li M (2020). Comprehensive evaluation of the effects of climate change and land use and land cover change variables on runoff and sediment discharge.Sci Total Environ, 702: 134401

[40] Zhang S Y, Zhao G J, Mu X M, Tian P, Gao P, Sun W (2021). Changes in streamflow regimes and their responses to different soil and water conservation measures in the Loess Plateau watersheds, China.Hydrol Processes, 35(10): e14401

[41] Zhao G J, Yue X L, Tian P, Mu X M, Xu W L, Wang F, Gao P, Sun W (2017). Comparison of the suspended sediment dynamics in two Loess Plateau catchments, China.Land Degrad Dev, 28(4): 1398–1411

[42] Zheng M G, Qin F, Yang J S, Cai Q G (2013). The spatio-temporal invariability of sediment concentration and the flow–sediment relationship for hilly areas of the Chinese Loess Plateau.Catena, 109: 164–176

[43] Zheng M, Cai Q, Cheng Q (2008). Modelling the runoff-sediment yield relationship using a proportional function in hilly areas of the Loess Plateau, North China.Geomorphology, 93(3−4): 288–301

[44] Zhou M, Deng J, Lin Y, Belete M, Wang K, Comber A, Huang L, Gan M (2019). Identifying the effects of land use change on sediment export: integrating sediment source and sediment delivery in the Qiantang River Basin, China.Sci Total Environ, 686: 38–49

RIGHTS & PERMISSIONS

Higher Education Press