The variation and attribution analysis of the runoff and sediment in the lower reach of the Yellow River during the past 60 years Open Access

Rivers are an indispensable and important resource for human survival and development. For most rivers, the impacts of climate change and human activities on water and sediment processes are particularly significant. The change of land use, large-scale drainage and overgrazing have reduced the water-holding capacity of basins, resulting in an increase in the frequency and scale of downstream floods in the wet season, and a serious decrease in river water volume in the dry season (Leopold 1968). The construction and storage of reservoirs obviously changes the hydrological process of rivers, which is the human activity that has the largest impact on rivers and the most far-reaching disturbance from rivers (Petts 1984).

The Yellow River is the second longest river in China. The complex and intractable crux of the Yellow River lies in the nature of less water and more sand and the inharmonious relationship between water and sand (Wang 2015). With the superposition of natural factors and human activities, the amount of water and sediment in the basin shows an obvious trend of fluctuation, and the historical water and sediment conditions in the lower reaches of the Yellow River have been completely different from those in modern times. In the new period, the relationship between water and sediment in the Yellow River has changed, and the sediment inflow in the lower reaches of the Yellow River has decreased sharply (Hu & Zhang 2018). The serious shrinkage of the lower reaches of the Yellow River, the reduction of flood discharge and sediment load capacity, etc., seriously threatens the flood control safety of the basin and seriously restricts the sustainable development of regional economy and society.

The evolution of water and sediment in the Yellow River basin has always been a concern of human beings, and it is also a focus and difficult problem in Yellow River management. At present, there have been a large number of studies on the changes of water and sediment in the Yellow River, most of which focus on the changes of main sections, branch flows and sediment as well as the middle and upper reaches of the basin, and the characteristics of the water and sediment process are not comprehensive enough (Ouyang et al. 2016; Zhang et al. 2018) and there are relatively few quantitative analyses on the changes of water and sediment along the main stream of the Yellow River and its influencing factors. Then, under the influence of various coupling factors, what have been the characteristics of the water and sediment change of the Yellow River in the last 60 years? How much did each factor contribute to the change?

Therefore, the purpose of this study is: (a) to statistically detect the trend, change-points and periodic relationship of the annual runoff and sediment discharge of the four main hydrological stations in the lower Yellow River; and (b) to analyze the contribution of climate change and human activities to runoff and sediment in the lower reaches of the Yellow River basin. The study will provide scientific guidance for the management of the Yellow River basin.

The Yellow River is known as the cradle of Chinese civilization and is a world-famous river with high sand content. The Yellow River basin is located between 95°53′ to 119°05′E longitude and 32°10′ to 41°50′N latitude. It has a continental climate with a multi-year average annual precipitation of 446 mm. The basin covers an area of 752,443 km², showing a trend of ‘high in the west and low in the east’, descending gradually from west to east. The western section of the Yellow River is located on the Qinghai-Tibet Plateau with an altitude of more than 4,000 m. The middle section is dominated by the Loess Plateau, with an altitude of 1,000 to 2,000 m. The eastern section is located in the North China Great Plain, with the main landforms being plains. The total area of the Yellow River basin is 795,000 square kilometres, the main stream is 5,464 kilometres in length, and the water drop is 4,480 metres. The main stream has a variable curvature and uneven distribution of tributaries. There are 76 first-level tributaries with a drainage area greater than 1,000 km², among which 13 are first-level tributaries with a drainage area greater than 10,000 km² or with a sediment inflow of more than 50 million tons. According to the characteristics of the river, with Hekou Town and Taohuayu as the dividing point, the Yellow River is divided into three parts: the upper, middle and lower reaches. The section from Mengjin to Gaocun on the lower Yellow River is a typical wandering section. The section has the characteristics of wide and shallow water flow, unsteady mainstream swing, and drastic changes in river regime. During the operation of the Sanmenxia Reservoir for water storage and sediment interception, the amount of sand entering the downstream river channel was greatly reduced, resulting in strong erosion in the lower Yellow River. During this period, the lower reaches of the Yellow River scoured a total of 2.14 billion m³ of sediment, of which the total scouring of the upper reaches of Gaocun (1.52 billion m³) accounted for 71% of the total erosion of the lower channel. Since the operation of the Xiaolangdi Reservoir, the amount of sediment entering the lower reaches of the Yellow River has been greatly reduced, and the riverbed in the lower reaches of the river has been scoured severely.

In order to analyze the temporal changes of runoff and sediment conditions, four major hydrological stations in the lower reaches of the Yellow River, Huayuankou station, Gaocun station, Aishan station, and Lijin station, were selected as the case study locations (Figure 1). Hydrological data (annual runoff and sediment load) from the Yellow River Yearbook and China River Sediment Bulletin, as well as annual rainfall series from the China Meteorological Data Network (1960–2016), were used in this study. Three missing rainfall data were supplemented by the linear interpolation method.

The Huayuankou Hydrological Station, which is the starting point of the suspension of the Yellow River, is about 4,700 km from the source of the Yellow River and 770 km from the mouth of the Yellow River. Its catchment area accounts for 97% of the total area of the Yellow River basin. As an important flood reporting station for major rivers in China, the Gaocun Hydrological Station is also the first important control station for the Yellow River to enter Shandong. Its cross-section is a compound riverbed with banks and troughs, about 579.1 kilometres away from the mouth of the river. Aishan Hydrological Station is used to monitor and forecast downstream floods. It is located in a narrow river with a section 384 kilometres away from the mouth of the river, and its control basin area is 749,000 square kilometres. As the last hydrological station of the Yellow River, Lijin Hydrological Station provides important hydrological data in the lower reaches of the Yellow River.

Note: HYK: Huayuankou station, GC: Gaocun station, AS: Aishan station, LJ: Lijin station, LYX: Longyangxia Water Conservancy Project, LJX: Liujiaxia Water Conservancy Project, SMX: Sanmenxia Water Conservancy Project, XLD: Xiaolangdi Water Conservancy Project.

The Mann–Kendall non-parametric test method (Liu & Wang 2015) is used to statistically test the long-term change trend of hydrological elements, and quantitatively determine the possible mutation points of the hydrological time series. At the same time, the T-test for differences method (Liu et al. 2007) is used to carry out the possible mutation points test. The wavelet analysis method (Wang & Wu 2009) reveals the multiple change cycles of the water and sediment time series under different time-scales. The slope change ratio method (Wang et al. 2012, 2015) was used to quantify the contribution rate of natural factors and human activities to the evolution of water and sediment quantity, and the double cumulation curve method (Mu et al. 2010; Qin et al. 2018; Guo et al. 2019) was used to analyze the stage change characteristics of the water and sediment sequence.

The standardized Mann–Kendall statistic Z follows the standard normal distribution with a mean of 0 and variance of 1. In the test for a trend, if ⁠, where Z is asymptotically normally distributed and is the critical value of the standard normal distribution with a probability ⁠, the trend of the sequence will be significant. A positive value of Z denotes an increasing trend, and the opposite corresponds to a decreasing trend.

The ability of the M-K test to detect trends can be influenced by autocorrelation in the data. Given this, the ‘trend-free pre-whitening’ (TRPW) technique was applied before the M-K test to avoid this issue in this paper. More details of calculation procedures for the TRPW technique can be obtained from Yue et al. (2002) and Shahid et al. (2020). If the statistical value is small enough, it can be directly tested by the traditional Mann–Kendall trend test method without preprocessing.

The Mann–Kendall mutation test calculates the normalized variable UF of the time series data and compares it with a critical variable at a certain confidence level (taken as 0.01). When UF is greater than 0, it indicates an upward trend, and when it is less than 0, it indicates a downward trend; when UF exceeds a critical value, it indicates a significant upward or downward trend. At the same time, the same statistical calculation is performed on the inverse sequence of the original time series, so that UB = −UF. If the two curves appear at the intersection point within the 99% confidence level, it indicates that the mutation occurred at this time point.

Among them: and S₁ are the average and standard deviation of years before the baseline; and are the mean and standard deviation of years after the baseline; and are the sample lengths of the two sequences before and after the benchmark.

In this study, the complex Morlet wavelet function is applied to distinguish temporal runoff and sediment load oscillations. Using the wavelet transform, wavelet coefficients and their variances are calculated. Wavelet power spectra and multiscale periodicity features are obtained with wavelet coefficients.

The double cumulative curve is a commonly used method, which tests the consistency of the relationship between two parameters and their changes. The double cumulative curve is the continuous cumulative value of one variable and the continuous cumulative value of another variable during the same period plotted by a relationship line in a rectangular coordinate system. If the curve has a significant turn at a certain point, that is, the abrupt change in the slope of the curve has changed, it indicates that the relationship between the parameters has changed (Mu et al. 2010). In this study, double cumulative curves of runoff vs sediment are plotted to estimate the relative effects of human activities.

The annual runoff and sediment load of each hydrological station are shown in Figure 2. During this period, generally speaking, the annual runoff and sediment load show a downward trend.

In Table 1, the mean Mann–Kendall values of annual runoff at the HYK, GC, AS and LJ stations were −3.99, −4.17, −4.71, −5.44, respectively, with absolute value great than 1.96, and there were obvious downward trends at the four stations (p < 0.01); the mean Mann–Kendall values of sediment load at the HYK, GC, AS and LJ stations were −6.51, −7.03, −6.68, −6.47, respectively, with absolute value great than 1.96, and there were obvious downward trends at the four stations (p < 0.01).

The Mann–Kendall test was used to analyze the abrupt change of annual runoff and sediment load of the four hydrological stations in the lower Yellow River. The statistical results of Mann–Kendall are shown in Figure 3. In the past 60 years, the changes of multi-year runoff and sediment load at the four hydrological stations are divided into two stages, that is, the trend of first increasing and then decreasing. The annual runoff and sediment load of the four hydrological stations all had obvious mutation, but the occurrence time was different, and the mutation year is shown in Table 2. The abrupt transition points of the annual runoff series of HYK, GC and AS hydrological stations all passed the confidence test of 0.05, while the abrupt transition points of the annual runoff series of LJ Hydrological Station all passed the confidence test of 0.01. The abrupt transition points of the annual sediment load sequence of HYK, AS and LJ hydrological stations all passed the 0.01 confidence test. The intersection point of the detection curve of the annual sediment load sequence of GC Station is outside the critical limit of the 99% confidence level, which indicates that the annual sediment load sequence of GC Station had no high reliability of abrupt change in 1996. The annual runoff sequence of each station changed in the early and mid-1980s, which was mainly related to the implementation of a series of soil and water conservation measures in the middle reaches of the Yellow River. Abrupt change occurred in the sequence of annual sediment load in the middle and late 1990s. Since the 1990s, water conservancy facilities in the Yellow River basin have been built and put into use one after another, resulting in a significant decrease in the sediment load of the Yellow River compared with the previous period, resulting in abrupt change in sediment load. The impoundment of the reservoir has held back a lot of sediment and reduced the sediment load quantity of the Yellow River. After entering the 20th century, with the increase of human activities, the capacity of sediment load further decreased.

The Mann–Kendall non-parametric test may have multiple mutation points or low mutation credibility in the test process, so the T-test of mean difference was used to verify the mutation results. In this paper, the critical value is t = 2.704 and the significance level 0.01, and the statistical results are shown in Table 3. The abrupt transition points of annual runoff of AS station are 1980, 1981 and 1985, among which the fluctuation range of 1980 and 1981 is relatively small and the time series from 1980 to 1985 is short, so the abrupt transition points of annual runoff of AS Station are all attributed to 1985.

The Morlet wavelet analysis of annual runoff and sediment load at the four hydrological stations in the lower Yellow River was used to draw the real-part isolines of the wavelet coefficients (Figure 4) and the variance (Figure 5), which represent the periodic changes of the sequence at different time-scales and their distribution in the time domain. The positive and negative values of the real part of the small wave coefficients in the contour map reflect the water–sand sufficient or deficient regions in the time scale. A series of runoff and sediment sequences in the Yellow River includes several different periods, and the runoff and sediment load sequences of the four stations show obvious interannual and interdecadal changes.

The variation of water and sediment in the lower Yellow River has multi-time-scale characteristics. The periodic behavior and duration are shown in Table 4. During the study period, the annual runoff of the Yellow River is 4-A, 7-A and 11-A, and the dominant runoff cycle is 11-A. The annual sediment load is presented as periodic of 4-A, 5-A, 8-A, 12-A, 13-A, 20-A, 21-A, 22-A and 23-A, and the dominant period of the four stations is 20-A, 21-A, 22-A and 23-A, respectively. The periodic variation of runoff in the Yellow River is the same, but the periodic variation of sediment load is different, and the periodic characteristics of runoff are more obvious than those of sediment load. The main reason for this difference is that the implementation of water conservancy projects and soil and water conservation measures in the Yellow River basin has stopped a large amount of sediment, leading to abrupt changes in the sediment load volume of the Yellow River.

The change of runoff and sediment load relationship is affected by natural disasters, underlying surface conditions, climate change, human activities and other factors (Guo et al. 2020), among which climate change and human activities are the main driving factors leading to the change of water and sediment in the Yellow River basin (Ma et al. 2014; Shi et al. 2014; Yan et al. 2015). The rainfall–sediment correlation model is a commonly used method to analyze the causes of sediment change in rivers. According to the occurrence time of the abrupt transition point, the time series was divided into stages and the results are shown in Table 5. According to the characteristics of hydrological abrupt transitions, the relationship curves of rainfall and runoff, and rainfall and sediment load accumulation before and after abrupt transitions at the four stations were drawn as shown in Figure 6. The determination coefficient R of the fitting relation in each stage is above 0.95, and the fitting degree is good.

The contribution rate of precipitation and human activities to the variation of water and sediment volume of the four hydrological stations is calculated quantitatively by the cumulative slope change rate method (see Tables 6 and 7). The influence of rainfall on the yield of water and sediment mainly lies in the size and amount of rainfall in the region. When the rainfall is too much, it will even lead to the damage of water conservancy and water conservation projects with low standards and cause a change of water and sediment quantity. In recent years, the magnitude and frequency of heavy rain have decreased. According to statistics, since the 1980s, the average number of heavy rain days in the middle and upper reaches of the Yellow River have decreased, and the number of heavy rain days has decreased more, which directly affects the sediment inflow into the Yellow River (Shi & Zhang 2013; Zhai et al. 2020). The contribution of precipitation in the lower Yellow River basin to the reduction of water and sediment volume of the four hydrological stations is different. Compared with the periods of T_b and T_a, the contribution of precipitation to the reduction of runoff volume of the four hydrological stations is 3.89%, 5.52%, 8.71% and 0.15% respectively, and the contribution of precipitation to the reduction of sediment volume is −2.21%, −0.06%, 22.32% and −0.04% respectively. It can be concluded that the change of water and sediment quantity mainly depends on the influence of human activities.

The variation of water and sediment quantity in the lower Yellow River basin is closely related to human activities. With the passage of time, human activities are increasing, and their contribution to the reduction of water and sediment quantity in the lower Yellow River basin is dominant. The response degree of each station to the contribution rate of human activities to the variation of water and sediment volume is different. Compared with the T_d and T_c periods, the contribution rate of human activities to the runoff reduction of the four hydrological stations is 96.11%, 94.48%, 91.29% and 99.85% respectively, and the contribution rate of human activities to the sediment load reduction is 102.21%, 100.06%, 77.68% and 100.04% respectively.

When the slope of the double accumulation curve is inclined to the cumulative runoff axis or the cumulative sediment load axis, it can be expressed as the decrease or increase of hydrological variables respectively. The double accumulation curves of runoff and sediment load in HYK, GC, AS and LJ are shown in Figure 7. The double cumulative slope of the four hydrological stations all deviated towards the cumulative runoff, indicating that the cumulative sediment load tended to decrease. There were two turning points on the double accumulation curve of runoff and sediment load, so the accumulation curve was divided into three stages, and the linear fitting equation of sediment load and runoff accumulation was established. The results of stage sediment reduction and annual sediment reduction are shown in Table 8.

To sum up, since the abrupt change of annual runoff in the lower Yellow River, the sediment load volume of all hydrological stations has decreased significantly and the sediment reduction volume has decreased along the river. Due to the impact of human activities, the sediment load volume of each hydrological station decreased by about 381 million t, 269 million t, 224 million t and 67.2 million t respectively every year, which is closely related to the large-scale water conservancy and water conservation project implemented since the end of 1970s and the project of returning farmland to forest and grassland in the late 1990s.

The implementation of soil and water conservation activities, the construction of water conservancy projects, and the continuous expansion of irrigation area from the Yellow River are the main human factors affecting the variation of water and sediment in the lower reaches of the Yellow River, which not only change the hydrological cycle process and the temporal and spatial distribution law, but also have a significant impact on the runoff and sediment yield conditions of the Yellow River.

Soil and water conservation is one of the main influencing factors. The Loess Plateau began soil erosion control in the 1950s. Researchers have carried out a variety of rapid dam-building technology explorations, such as water pouring in water, water pouring in soil and directional blasting. In 1957, the application of hydraulic filling damming technology promoted the construction of warping dams to a great extent and accelerated the progress of comprehensive treatment of small watersheds. Since 1960, soil and water conservation on the Loess Plateau has been transformed into comprehensive control and comprehensive planning, instead of the former state of disordered control. At the end of 1970s, the key method to reduce sediment flow into the Yellow River was defined as strengthening the construction of dams and reservoirs for water storage in the upper and middle reaches of the Yellow River. All the above are the initial stage of implementing soil and water conservation measures, which cannot play a significant role. In the 1980s, warping dams became the ‘backbone project’ in the comprehensive management mode and the key measures of soil erosion control for small watersheds in the middle reaches of the Yellow River (Zuo et al. 2016). From 1980 to 1985, soil and water conservation measures played an important role in the period, and it was a stage of comprehensive treatment with the small watershed as a unit. Therefore, the annual runoff of the lower Yellow River had a sudden change.

The construction of water conservancy projects is another important reason for water and sand changes (Li 2008). Since 1960, a series of water control projects have been built in the upper main stream of the Yellow River, such as Qingtongxia, Liujixia and Longyangxia water control projects. Since 1986, the combined action of Longyangxia and Liujiaxia reservoirs has blocked most of the sediment above Lanzhou station, and the water volume of the upper reaches of the Yellow River in the flood season is 54% less than that of the Liujiaxia reservoir. Therefore, this is also the reason for the abrupt change of annual runoff in the lower Yellow River. After the completion of Xiaolangdi Water Control Project in 1999, the control of downstream water and sand was further strengthened. In December 1996, the lower sluice of Lijiaxia stored water. Under the joint action of the two, the sediment load volume of the lower Yellow River has hydrological variation.

The irrigation project along the Yellow River main stream is another main aspect that causes the change of water and sediment. Most of the irrigation works along the main stream of the Yellow River are carried water. They are mainly located in the lower reaches of the Yellow River and the Ningxia–Inner Mongolia reaches, accounting for 52.4% and 44.5% of the total irrigation area respectively, among which the lower reaches of the Yellow River have many irrigated areas and a large irrigation area. Since the 1990s, industry and agriculture have developed rapidly, and the water drawn along the Yellow River is still mainly used for agriculture. The agricultural water of the lower Yellow River accounts for 90% of the total water supply of the lower Yellow River, and the water consumption is 115.619 billion m³ (Li 2008). Therefore, the irrigation project along the Yellow River is another main factor that causes the change of water and sediment.

Sediment problem has always been an important problem in the control and development of the Yellow River. In the qualitative description of the causes of water and sediment in the Yellow River, researchers agree that climate conditions, changes in the environment of the Yellow River basin, the construction of water conservancy projects in the main and tributary streams, the excessive exploitation of resources and other factors together promote the change of water and sediment conditions in the Yellow River basin. The contribution rate of climate change and human activities to runoff and sediment change is analyzed synthetically by a hydrologic method in this paper. It should be pointed out that the contribution of climate factors to precipitation factors is mainly quantified. Climate change mainly includes precipitation and temperature in the basin, among which the change of precipitation directly affects the change of yield and discharge, while the change of temperature leads to the change of evapotranspiration, which will also lead to the change of yield and discharge. However, evapotranspiration has little influence on runoff and data acquisition is difficult. Therefore, this study does not consider the influence of air temperature on runoff and sediment in the basin, and climate change only considers the influence of precipitation change on water and sediment change.

The environment for aquatic and sediment production in the Yellow River Basin is extremely complex, and the impact mechanisms of natural factors such as human activities and rainfall on water and sediment changes are different. In this study, only the comprehensive effects of the two factors on water and sediment on a longer time-scale were considered. There is no research on the internal relationship between the variation of rainstorms, extreme climate phenomena, the influence of large-scale human activities on the tributary channel and the change of water and sediment. In addition, it is of greater guiding significance to explore the influence of human activities on future changes of water and sediment in a short time.

Based on the hydrological data of four hydrological stations in the lower reaches of the Yellow River, a variety of statistical methods were used to analyze the variation trend year of abrupt change period, influencing factors and their contribution rates to the variation of runoff and sediment volume. The major findings can be summarized as follows:

• (1)

  With the passage of time, the annual variation of runoff and sediment discharge of the Yellow River from 1960 to 2019 showed an obvious decreasing trend, and the decreasing trend of annual sediment discharge was more obvious than that of annual runoff.

• (2)

  There were abrupt changes in runoff and sediment load at all hydrological stations, among which the abrupt changes in runoff occurred in the 1980s and in the 1990s.

• (3)

  There are multi-time-scale periodic changes in the water–sediment series of hydrological stations, and the cycle changes of runoff are the same, while the variation scales of sediment discharge are slightly different.

• (4)

  Since the 1980s, the abrupt change of annual runoff and sediment has been caused by the combined action of precipitation and human activities, and human activities are the main reason for the change of runoff and sediment in the lower reaches of the Yellow River.

This work was supported by the National Nature Science Foundation of China (Grant No. 51679090, 51609085, 51779094); Water Conservancy Science and Technology Project of Guizhou Province (KT202008); National Natural Science Foundation of Henan, China (Grant No. 162102110015, 152300410113); Fund of Innovative Education Program for Graduate Students at North China University of Water Resources and Electric Power, China (Grant No. YK2020-08).

The authors declare no conflict of interest.

All relevant data are included in the paper or its Supplementary Information.