International Journal of Sediment Research 36 (2021) 668e677

Contents lists available at ScienceDirect

International Journal of Sediment Research

Original Research

Erosion-control mechanism of sediment check dams on the Loess

a r t i c l e i n f o a b s t r a c t

Article history: Severe soil erosion occurs on the Loess Plateau in China, which makes the Yellow River the most
Received 15 November 2020 sediment-laden river in the world. Construction of about 60,000 sediment check dams has remarkably
Received in revised form controlled soil erosion on the Loess Plateau and reduced the sediment load of the middle and lower
31 January 2021
Yellow River. Nonetheless, little is known about the mechanism of erosion control and vegetation
Accepted 5 February 2021
Available online 24 February 2021
development of sediment check dams. The function of a single check dam mainly is trapping sediment,
while the function of a train of check dams comprising dozens of or over hundreds of check dams in a
gully encompasses controlling bed incision and reducing erosion energy. A formula was proposed to
Keywords:
Loess plateau
calculate the potential energy of bank failure and slope failure in a gully, which essentially constitutes the
Check dam erosion energy. The erosion energy increases when gully incision occurs, which is induced by the incision
Energy of soil erosion of the Yellow River and its tributaries on the Loess Plateau. Sediment deposition in many gullies due to
Vegetation-erosion dynamics construction of check dams reduces the erosion energy to almost zero, which in turn greatly reduces soil
Gully incision erosion and sediment yield. Construction of check dams promotes vegetation development. The
vegetation-erosion dynamics model was used to study the effect of check dams on vegetation devel-
opment. Simulation results show that reforestation without check dam construction might result in an
increase of vegetation cover in the first ten years and then a drop of vegetation cover to less than 10% in
the later years. The check dams provide a foundation for vegetation development.
© 2021 International Research and Training Centre on Erosion and Sedimentation/the World Association
for Sedimentation and Erosion Research. Published by Elsevier B.V. All rights reserved.

1. Introduction erosion was severe. One reason for the severe soil erosion was the
high erodibility of loess, which consists mainly of sand and silt with a
The Yellow River carried 1.6 billion tons of sediment annually to small portion of cohesive particles (Tian et al., 2021). Control of soil
the North China Plain and the Yellow River delta in 1919e1960 (Chen erosion on the plateau is the key to flood defense of the lower Yellow
& Zhou,1965). The annual sediment load reduced to 1.3 billion tons in River.
1955e1986, still ranking No.1 among all large rivers in the world The Loess Plateau, which is located between the North China
(Wang & Jiao,1996). Almost every year hyperconcentrated flow (with Plain and the Qinghai-Tibet Plateau at an elevation between
sediment concentration higher than 100 g/L) occurred in the flood 1,000 m and 3,000 m, has an area of about 640,000 km2. There is a
season in the past century (Wan & Wang, 1994). The riverbed 50e250 m thick loess layer covering two thirds of the Loess Plateau.
aggraded at a rate higher than 1 cm/yr during the past millennium, Loess was carried by wind from the Central Asia and the Mongolian
and the lower Yellow River became a perched river. Levee breaches Plateau and deposited in this area 2.6e3.6 million years ago (Liu,
and flood disasters occurred twice every three years (Wang & Liu, 1965). The severe erosion of the Loess Plateau is the result of the
2019). Most sediment came from the Loess Plateau, where soil erosion of the Yellow River and its tributaries (Jiang et al., 2007).
The plateau has been dissected by a network of headward-growing
tributaries (Howard, 1967; Schumm, 1956), envisioned from the
tips of the first-order streams. Degradation of the Yellow River over
* Corresponding author.
the past 150,000 years has been causing incision of tributaries,
E-mail address: [email protected] (S. Yu).

https://doi.org/10.1016/j.ijsrc.2021.02.002
1001-6279/© 2021 International Research and Training Centre on Erosion and Sedimentation/the World Association for Sedimentation and Erosion Research. Published by
Elsevier B.V. All rights reserved.
 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677 669

which has propagated to millions of gullies and resulted in gully system (Shi et al., 2019). Wang et al. (2004) developed the vegeta-
incision and slope erosion. tion-erosion dynamics, which models the relations among the rate
About half of the Loess Plateau was being severely eroded at a of soil erosion, vegetation cover, and human activities with differ-
rate higher than 5,000 t/(km2$yr) in the 20th century (Fu et al., ential equations. The model has been used to simulate and predict
2011). The highest rate of soil erosion occurred in the north part vegetation development in erosion areas including the Loess
of the plateau at an average rate of 30,000 t/(km2$yr). Because of Plateau, the Yungui Plateau, and several deserts (Li et al., 2017; Wang
the extremely high sediment yield from the plateau, the maximum et al., 2015b). Results show that only if erosion is controlled can
sediment concentration reached 1,500e1,700 g/L in the tributaries vegetation on the Loess Plateau be developed by reforestation pro-
and 931 g/L in the main stem (Wan & Wang, 1994). Sediment check jects. It is the reduction of erosion energy by check dams that en-
dams have been constructed to create farmland by trapping sedi- ables vegetation to quickly develop on the slopes and in the gullies.
ment. These dams have become the most effective measure of The aim of this paper is to study the mechanism of erosion energy
erosion control on the plateau, because the dams not only trap reduction and vegetation improvement by check dams.
sediment, but also control gully incision and reduce potential
erosion energy (Wei et al., 2009). A sediment check dam normally is 2. Study area and methods
composed of an earth dam and a culvert for discharging flood
water. Most of the sediment check dams have no spillway (Yu et al., 2.1. Loess Plateau
2019). If the flood discharge is higher than the flow capacity of the
culvert or if the culvert is destroyed, a part of the earth dam may Fig. 1 shows the locations of the Loess Plateau, the Yellow River
fail, a dam-break flood may occur, and a part of sediment deposited and its tributaries, the sampling sites, the investigated small wa-
in the reservoir may be flushed out (Chen et al., 2015, 2019). tersheds, and important cities and towns. To the northwest of the
However, most trapped sediment will remain in the reservoir after Loess Plateau there are 4 deserts: Mu-Us Desert, Kubuqi Desert,
the dam failure and will still reduce the potential erosion energy for Ulan Buh Desert, and Tengger Desert. Wind erosion occurs and sand
many years. dunes develop in these deserts. Nevertheless, there are very few
Large sediment check dams are check dams with a capacity of rivers in the deserts that transport sediment into the Yellow River.
0.5e5 million m3, medium sediment check dams are check dams Therefore, only a small portion of sediment in the Yellow River is
with a capacity of 0.1e0.5 million m3, and small sediment check from the deserts. The red curve in Fig. 1 marks the boundary of loess
dams are check dams with a capacity of 0.01e0.1 million m3 (Li et deposit, which covers an area of about 400,000 km2. The thickness
al., 2019). The productivity of farmland created by check dams is of loess deposit varies between 50 m and 250 m. Many rivers and
8e10 times higher than that of slope farmland (Fu et al., 2017; numerous gullies that develop in the area of loess deposit transport
Wang et al., 2015a). Local farmers have used limited funds from the sediment into the Yellow River.
government to construct sediment check dams mainly to create
farmland. 2.2. Study methods
Scientists have reported that a total of 110,000 sediment check
dams have been constructed in thousands of gullies on the plateau From 2002 to 2017, 6 field investigations have been done on
in the past 60 years, which have trapped 21 billion tons of sediment the Loess Plateau taking samples and doing measurements.
and created 313,000 ha of fertile farmland (Li et al., 2008; Wang Samples were taken from the riverbed of 11 tributaries on the
et al., 2011). Other reports give different figures. The Yellow River plateau and from the riverbed of the Yellow River at Huayuankou
Commission reported that by 2011 a total of 92,758 sediment check after the flood season. The sampling sites are shown in Fig. 1 with
dams had been constructed on the Loess Plateau, including 6,057 an “X” symbol. Sediment in the rivers experiences sorting, so the
large dams, 11,881 medium dams, and 74,820 small dams (Wang, median diameters of sediment at the various sampling sites are
2011). The Ministry of Water Resources of China reported that by different due to different conveyance distances. The mineral
2014 a total of 58,446 check dams with a storage capacity larger composition of the samples which were analyzed with a X-ray
than 10,000 m3 had been constructed, which had trapped 5.7 diffractometer and mineral differential thermal analyzer (DTA)
billion m3 of sediment (MWRC, 2014). Although the total storage reflect the compositions of the original loess and the environ-
capacity of small and mini check dams is less than that of large ments before erosion. Non-clay minerals include quartz, feldspar,
dams, the total volume of sediment trapped by the small and mini calcite, hornblende, and tiny amounts of other minerals. Clay
check dams is not much less than that trapped by the large check minerals mainly comprise illite, smectite, kaolinite, and chlorite.
dams because the small and mini check dams mostly are fully filled. Fig. 2 shows the compositions of clay minerals and non-clay
Construction of sediment check dams has greatly reduced soil minerals of the samples from the 12 rivers (Tian et al., 2021).
erosion and sediment production. Abrupt reduction of sediment The points in Fig. 2(a) are very close to each other, meaning that
load in the streams on the Loess Plateau began in the 1970s, when the compositions of the non-clay minerals from the 12 rivers are
large scale construction of check dams started. Similar to natural almost the same. The scattering of points in Fig. 2(b) indicates
dams the sediment check dams control incision of the stream bed, different compositions of clay minerals from the 12 rivers. Only a
thus the potential energy for bank erosion is greatly reduced (Wang portion of the clay minerals sampled from the riverbeds is iden-
& Zhang, 2019). Statistics show that the volume of sediment trap- tical to the clay minerals in the original loess deposit. Other clay
ped by check dams increased from the 1950se1970s and reached a minerals are results of weathering of loess on the surface of the
peak in the 1980s (Liu et al., 2018). Although the rate of sediment plateau. Different weathering environments result in different
trapping has slowed down since the 1980s, the sediment load compositions of clay minerals.
transported from the Loess Plateau into the Yellow River has kept Samples of the original loess were taken from 7 places on the
on reducing because the check dams continued reducing erosion Loess Plateau. To avoid disturbances of weathering, farming, and
energy after they were filled with sediment. erosion, the samples were taken from 50 cm below the ground
Vegetation development is seriously affected by soil erosion surface. The sampling sites were numbered from north to south,
(Kosmas et al., 2000). Check dams promote revegetation by stabi- and are shown in Fig. 1 with square symbols with numbers 1 to 7
lizing gully slopes and controlling soil erosion. Revegetation and inside. Fig. 3(a) shows the median diameters of the samples. The
check dams were combined to form a soil and water conservation Loess Plateau was formed by aeolian deposition of sediment from
670 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677

Fig. 1. Map of the Loess Plateau showing the boundary of loess deposit, the locations of the Yellow River and its tributaries, the sampling sites, and the six small watersheds
investigated.

the Central Asia and the Mongolian Plateau, which were trans- the north part of the plateau is gentler than in the south part of the
ported by sand and dust storms in the Quaternary Period. The plateau. The landscape is different from north to south.
coarsest particles transported by sand and dust storms deposited Field investigations were done from north to south on the
before reaching the Loess Plateau and formed the deserts. The plateau. Geographical and geometrical characteristics were
median diameter of sand in the deserts varies between 0.05 and measured in the field using GPS receivers and laser range finders,
0.15 mm. Coarse particles deposited in the north part of the plateau. both of which have an accuracy of 0.1 m. The digital map Bigemap
Fine particles were carried over longer distances and deposited in which has a resolution of 0.5 m was used to develop longitudinal
the south part of the plateau. profiles of rivers and gullies.
On the Loess Plateau, the finer the loess is, the more clay min- The rate of soil erosion was measured. There are hydrological
erals it contains. Fig. 3(b) shows the percentage of clay minerals in stations on the rivers, and gauging stations on the selected gullies.
the loess as a function of the median diameter. Loess in the south Regular measurements of water stage, velocity, discharge, and
part of the plateau is finer and more cohesive than loess in the suspended sediment concentration were taken daily at the hydro-
north part of the plateau. Therefore, the erodibility of loess reduces logical stations. Channel bed deformation was measured 1e2 times
from north to south on the plateau. The bank slope of the gullies in per year at the hydrological stations and at selected cross sections

Fig. 2. Compositions of sediment samples from 11 tributaries on the Loess Plateau and the Yellow River: (a) Non-clay minerals; and (b) clay minerals.
 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677 671

Fig. 3. Particle size grading of sediment samples: (a) Median diameter (D50) of original loess from north (No. 1) to south (No. 7); and (b) percentage of clay minerals as a function of
the median diameter.

in the rivers and gullies. In the selected cross sections, bank slope, Loess Plateau by 100e400 m. The incision of the Yellow River
bed gradient, and incision rate were measured using a total station. caused incision in all its tributaries. The incision wave propagated
The vegetation cover was calculated by dividing the area into thousands of gullies and caused intensive erosion. Gully inci-
covered by forest and shrubs by the total area of the watershed. sion resulted in very steep bank slopes. Banks failed and brought a
Each county on the Loess Plateau has its Bureau of Forestry and lot of sediment into flowing water during the flood season. Ac-
Bureau of Soil and Water Conservation. The bureaus measured the cording to the Erosion Distribution Atlas of the Yellow River Basin
rate of soil erosion, vegetation cover, area of revegetation, and published by the Ministry of Water Resources of China, the annual
sediment volume in the reservoirs formed by check dams. The data amount of soil eroded from the Loess Plateau averaged at 1.9e2.1
from local bureaus were collected and analyzed. billion tons/yr in the 20th century. Gully erosion under the action of
flowing water and gravity erosion is the main form of erosion. This
3. Potential energy of soil erosion and mechanism of erosion erosion caused stream incision and bank failures. It was measured
control by check dams that bank failures result in 51%e67% of the total sediment yield
(Yang et al., 2017). Sheet erosion and rill erosion are less important.
Soil erosion on the Loess Plateau can be ultimately attributed to The calculation of erosion energy in gullies is illustrated in Fig. 4.
the incision of the Yellow River. The Yellow River cut through the The angle of repose for loess denoted by a is about 30
in most
Sanmenxia Gorge 860,000e150,000 years ago (Jiang et al., 2007; cases. 30
was used in the calculation. The straight lines in the
Pan et al., 2005), after which the Yellow River has incised into the figure represent slip surfaces. Slope failure may occur if the bank

Fig. 4. Erosion energy of a gully section increases with incision and reduces with aggradation. Bdepo is the final width of the slide block.
672 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677

slope is steeper than the angle of repose. The potential energy of In the south part of the Loess Plateau a stream channel network
the loess body above the slip surface is the effective energy for is developing, and deep-cut gullies are being formed. The fingertip
gravity erosion, which is given by: gullies are developing toward the undissected flat plateau. This
process is already finished in the north part of the Loess Plateau so
ðH that no undissected flat plateau surface remains. In the south part
Ep ¼ gs BðzÞzdz (1) of the plateau the headward growth of deep gullies is continuously
eroding the existing undissected flat plateau surface and is
0
exposing more and more loess to gully erosion. Thus, the total
in which gs is the specific weight of loess; B is the width of the slide erosion energy increases.
block, which varies with z; and z is the height from the bed. If the Fig. 6 shows the longitudinal profiles of the main stem and the
bank slope is smaller than a the energy is 0. branch gullies of Jiuyuangou Gully (a) and Caijiamiao Gully (b). The
If the gully bed is incised down, the value of z at any point and Caijiamiao Gully is a tributary of the Malian River, and Jiuyuangou
the integration range H will increase, and the width of the slide Gully is a tributary of the Wuding River, as shown in Fig. 1. The
block will increase from B to Bincision (Fig. 4). The erosion energy (Ep) gradients of the branch gullies are much greater than that of the
will be much greater than that before incision. Fig. 5 shows a gully main gullies, and the longitudinal profiles of the branch gullies are
named Duigou (Its location is shown as ⑤ in Fig. 1) as an example of mostly parallel to each other. The steps in the profiles of the gullies
gully incision. The gully bed was incised down by 2.2 m. The erosion are sediment check dams constructed mainly from the
energy of loess on the left bank was calculated with Eq. (1). The 1970s to 1990s.
energy was about 80 m,kN per unit length before incision and The Jiuyuangou Gully is in the north part of the plateau and
320 m,kN after incision. the Caijiamiao Gully is in the south part of the plateau, as shown
On the other hand, if the gully bed aggrades due to siltation in Fig. 1. Loess in the Jiuyuangou Gully watershed is noncohesive.
behind check dams, loess on the banks will be below the slip sur- Gully banks fail as the gully incises and the erosion energy in-
face, and the erosion energy (Ep) will become zero, as shown in creases. Therefore, the gully tends to be relatively broad. As a
Fig. 4. The erosion energy after gully incision Ep,incision and the comparison, loess in the Caijiamiao watershed is cohesive and
erosion energy after siltation behind check dams Ep,dam can be the gully wall may remain stable at a slope steeper than 30
. The
computed using Eq. (2) and Eq. (3), respectively. depth of the Caijiamiao Gully is much greater than that of the
Jiuyuangou Gully, exceeding 150 m. Because both the bed and the
slopes of the gullies are more resistant to erosion in the south
ðH
part of the plateau, the gullies develop more slowly in the south
Ep; incision ¼ gs B0 ðzÞðz þ DHÞdz [ Ep (2) part than in the north part of the plateau. The undissected flat
DH plateau surfaces are called “Yuan” by local people. There is no or
very little Yuan in the north part of the plateau spared from gully
erosion.
ðH
Incision caused by incision of the Yellow River propagates from
Ep; dam ¼ gs BD ðzÞðz  HD Þdz z 0 (3)
the main gullies to the branch gullies in the form of head cut
HD erosion. Fig. 7(a) shows a part of the Malian River watershed, in
which numerous branch gullies are creeping into the undissected
in which DH is the depth of incision and HD is the depth of siltation Yuan. Fig. 7(b) shows the gullies incising and widening in the
due to dam construction. B0 and BD is the width of the slide block manner of bank collapse. The speed of headward erosion was
due to gully incision and dam construction respectively. measured at 0.1e0.3 m/yr for large gullies and 0.6e1.1 m/yr for
The total potential energy of erosion in a gully (ET) is the sum of small gullies on average (Shi, 1987; Zhang et al., 2009).
all sectional erosion energy along the gully, which can be calculated Check dams were constructed both in the north and south parts
using the following formula: of the plateau. These dams store water and trap sediment. Fig. 8
shows check dams in the Suncha Gully in Yan'an Prefecture. The
ðL ðH location of the Suncha Gully is shown as ⑥ in Fig. 1. Forty-three
ET ¼ gs BðzÞzdz dx (4) check dams have been constructed in this small watershed,
00 which has a drainage area of 54 km2. These dams are earth dams
and mostly have no spillway (Fig. 8(b)). In the rainy season, the
in which L is the length of the gully, and x is the distance from the dams store water and form small lakes (Fig. 8(a)). Sediment de-
gully mouth. posits in the reservoirs, and, thus, the bank slopes of the gullies

Fig. 5. Example of gully incision in the Duigou Gully in Suide County.

Fig. 6. Longitudinal profiles of the (a) Caijiamiao Gully, (b) Jiuyuangou Gully and their branch gullies.

Fig. 7. Gully development due to erosion. (a) Numerous gullies are creeping into the undissected “Yuan” of the Loess Plateau as a result of headward erosion in the Malian River
basin; and (b) incision and widening of a small gully in the south part of the Loess Plateau.

Fig. 8. Check dams in the Suncha Gully. (a) Water is stored by check dams forming lakes in the Suncha Gully; and (b) a check dam in the Suncha Gully constructed with loess (no
spillway).
674 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677

Fig. 9. Comparison of (a) total sediment storage capacity of constructed check dams and total volume of trapped sediment, and (b) annual sediment load and annual water runoff of
the Yellow River at Tongguan Station.

become gentler. As a result, not only gully erosion is controlled but industrial and agricultural diversion projects, and soil and water
also slope erosion is reduced. conservation programs. The reduction in the sediment load of
Fig. 9(a) shows the total sediment storage capacity of con- the Yellow River is attributed to various driving factors, and
structed check dams and the total volume of trapped sediment different researchers have different conclusions on the relative
from 1950 to 2015, which are obtained by summing up statistics importance (Liu et al., 2020). However, in this period, the
of 58,446 large, medium, and small check dams from the reports reduction of the erosion potential energy played the most
of the counties on the Loess Plateau (Li et al., 2018). As a com- important role in the contribution of check dam construction on
parison the annual sediment load and annual water runoff of the the reduction of sediment yield, and the interception of sediment
Yellow River immediately downstreanm of the confluence of the play a subsidiary role.
Yellow River and the Weihe River at Tongguan are shown in
Fig. 9(b) (Shi et al., 2017; Wang et al., 2009; Wang & Liu, 2019). 4. Effect of check dams on reforestation
The curve of the total storage capacity of the check dams and the
curve of the sediment load in the Yellow River are closely related The following differential equations are the core of the
like the mirror images of each other. Reduction of sediment load vegetation-erosion dynamics model (Wang et al., 2005):
in the Yellow River immediately follows the increase of storage 9
dV
 aV þ cE ¼ Vt >
capacity of the check dams. There are two phases of quick in- >
=
crease of the total sediment storage capacity of the check dams, dt
(5)
namely 1965e1980 and 1990e2010. Concurrently, the annual dE >
>
sediment load in the Yellow River sharply reduced during these  bE þ fV ¼ Et ;
dt
two periods. The increasing rate of the total volume of sediment
trapped by the dams reduced sharply after 1980. The difference where V is the vegetation cover, which is the proportion of land
between the total capacity of the check dams and the total area covered by trees and shrubs in a watershed; E is the amount of
volume of trapped sediment has become larger and larger since erosion per unit area per year; Vt is the net increase of vegetation
1990, but the annual sediment load has been steadily reducing. cover per unit area per year, which equals revegetation minus
The main drivers of this reduction include dam construction, logging and other stresses that cause reduction of vegetation; and
 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677 675

Et equals erosion induced by human activities minus erosion Fig. 10(a) shows the actual process of vegetation cover devel-
reduced by check dams and terrace farmland per unit area per year. opment (black points), calculated vegetation cover using the data of
The parameters a, c, b, and f can be determined by applying a trial revegetation and sediment trapped by the check dams (solid line),
and error method for a given watershed where data on vegetation and calculated vegetation cover assuming no check dam con-
cover, erosion rate, Vt, and Et are available. struction (dashed line).
The parameters a, c, b, and f depend mainly on the local climate, Another example is the Wangjiagou watershed, which is
soil, and topography, and are independent of present vegetation located in the east part of the Loess Plateau (Fig. 1). This small
cover, erosion rate, and various stresses applied on the vegetation. watershed has an area of 9.1 km2, of which 44% is gullies and
Once the parameters are determined for a watershed or an area, 56% is slope land. The annual precipitation is 495 mm, and the
they can be directly applied to the sub-watersheds or neighboring yearly average temperature is 8.8
C. The frost-free period is
areas with the same climate, soil, and topography. The model has 160e180 days. There are 19 gullies in the small watershed. In the
been successfully applied to the Loess Plateau, Yungui Plateau, and 1940s and 1950s, the average erosion rate was 15,850 t/(km2$yr),
many other soil erosion areas for simulation and prediction of 80% of which was from gully erosion, and 20% of which was from
vegetation development (Wang & Wang, 2007; Wang et al., 2015a). slope erosion (Zhang et al., 2001). To control soil erosion, the
Eq. (5) can be applied to study the effect of check dams on the local government launched a comprehensive project in 1955. The
development of vegetation on the Loess Plateau. In this case, Vt main strategies were reforestation and construction of sediment
roughly equals the increase of vegetation cover as a result of check dams.
reforestation (positive); and Et equals the sediment stored by check Using the data of measured vegetation cover, rate of erosion,
dams and reduced erosion due to check dams and terrace farmland area of reforestation, and sediment deposition in the check dams in
(negative). In the 1950s, the Loess Plateau underwent intense the Wangjiagou watershed, the four parameters a, c, b, and f were
reforestation but very few check dams were constructed. The determined to be 0.001/yr, 0.00000137 km2/yr, 0.02/yr, and 400 t/
vegetation cover increased in the first years and then decreased (km2$yr2), respectively.
again because the effective energy of erosion was still large and the Fig. 10(b) shows the actual process of vegetation development
planted trees and shrubs were destroyed due to erosion. (points), calculated vegetation cover (solid line) with the data of Vt
The Anjiagou Gully is in the watershed of the Zuli River on the and Et, and calculated vegetation cover assuming no check dam
Loess Plateau (Fig. 1). The Anjiagou Gully watershed has a drainage construction (dashed line).
area of 9.1 km2 with elevations in the range of 1,900e2,250 m As shown in Fig. 10 the calculated vegetation curves agree very
above the sea level. There are many gullies in the small watershed well with the actual development of vegetation in the Anjiagou and
with a density of 3.14 km/km2. The gullies are 30e50 m deep. The Wangjiagou watersheds. Both gullies experienced two periods of
annual precipitation is only 427 mm, of which 60% occurs in July, intensive revegetation and erosion control. The vegetation cover
August, and September. The annual pan evaporation is 1,526 mm increased from less than 5% to 38%e48% due to the revegetation
and the yearly average temperature is 6.3
C. The soil is composed projects and construction of check dams. Nevertheless, if no check
of gray desert soil, spodosol, and loamy soil. The gully bed and the dams had been constructed on the two gullies, the vegetation in the
gully slope are covered with halogenic soil. two watersheds would first increase following the reforestation
The rate of soil erosion was as high as 10,000 t/(km2$yr) and projects, and then linearly reduce to nearly zero after the revege-
the vegetation cover was only 5.7% before the 1950s. Revegeta- tation projects ended. Then the vegetation cover would increase
tion projects were launched at the mid of 1950s (Li, 1986; Ye, again following the second period of reforestation, but it would
1986; Zhang et al., 1986). Local people planted trees and reduce sharply again after the second period of reforestation pro-
shrubs on the bare hills and slopes. Since the late 1970s, people jects ended. The final vegetation cover without check dams would
have constructed many check dams and further intensified be 30% less than the actual value.
revegetation. The process can be modeled with Eq. (5), in which
Vt and Et were obtained using the data of revegetation and the 5. Conclusions
volume of sediment trapped by the check dams. The four pa-
rameters a, c, b, and f were determined by applying a trial and Severe erosion occurs on the Loess Plateau because gully inci-
error method, which were 0.001/yr, 0.0000018 km2/yr, 0.01/yr, sion increases erosion energy. A formula is proposed to calculate
and 400 t/(km2$yr2), respectively. the erosion energy. Loess in the south part of the plateau consists of

Fig. 10. Comparison of vegetation cover development (black points) with calculated vegetation cover (solid line) and calculated vegetation cover assuming no check dam con-
struction (dashed line). (a) Anjiagou Gully Watershed; and (b) Wangjiagou Watershed.
676 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677

more clay minerals than loess in the north part of the plateau. Ecological Complexity, 8(4), 284e293, https://doi.org/10.1016/
j.ecocom.2011.07.003
Consequently, the gully banks are more stable in the south part of
Fu, B. J., Wang, S., Liu, Y., Liu, J. B., Liang, W., & Miao, C. Y. (2017). Hydrogeomorphic
the plateau than in the north part of the plateau, although the ecosystem responses to natural and anthropogenic changes on the Loess
erosion energy in the south part of the plateau might be higher. The Plateau of China. Annual Review of Earth and Planetary Sciences, 45(1), 223e243,
erosion energy was increased fourfold due to gully bed incision. The https://doi.org/10.1146/annurev-earth-063016-020552
Howard, A. D. (1967). Drainage analysis in geological interpretation. Bulletin of the
main contribution of a single check dam is to trap sediment and American Association of Petroleum Geologists, 51(11), 2246e2259, https://
construct land, and the check dam group can control gully under- doi.org/10.1306/5d25c26d-16c1-11d7-8645000102c1865d
cutting and reduce the erosion energy. The construction of check Jiang, F., Fu, J., & Wang, S. (2007). Formation of the Yellow River, inferred from loess-
palaeosol sequence in mangshan and lucustraine sediments in sanmen Gorge,
dams transformed the undercut gullies into areas experiencing China. Quaternary International, 175(1), 62e70, https://doi.org/10.1016/
sedimentation and uplift, which greatly reduced the erosion en- j.quaint.2007.03.022
ergy, and in some places the erosion energy is reduced to zero. Kosmas, C., Danalatos, N. G., & Gerontidis, S. (2000). The effect of land parameters
on vegetation performance and degree of erosion under Mediterranean con-
Thus, soil erosion and sediment entering the Yellow River have ditions. Catena, 40(1), 3e17, https://doi.org/10.1016/s0341-8162(99)00061-2
been greatly reduced in the Loess Plateau. Li, B. R. (1986). The integrated management of the Anjiagou watershed. Collection of
The check dams promoted development of vegetation because reports of experimental results on soil and water conservation (pp. 20e35). Gansu
Province: Research Institute of Dingxi Soil and Water Conservation. (In Chinese)
the slopes were stabilized due to aggradation of the gullies and Li, P., Xu, G. C., Lu, K. X., Zhang, X. M., Shi, P., Bai, L. L., Ren, Z. P., Pang, G. W., Xiao, L.,
reduction of erosion energy. The vegetation-erosion dynamics Gao, H. D., & Pan, M. H. (2019). Runoff change and sediment source during
model was used to study the effect of check dams on vegetation rainstorms in an ecologically constructed watershed on the Loess Plateau,
China. The Science of the Total Environment, 664, 968e974, https://doi.org/
development. The vegetation cover of two small watersheds
10.1016/j.scitotenv.2019.01.378
increased from 1%e5% to 40%e50% during 1955e1990 as a result of Li, P., Xu, G. C., Ran, Q. H., & Ren, Z. P. (2018). Effects on water and sediment variation
check dam construction and reforestation projects. The results and contribution for erosion control of gully projects on the Loess Plateau. Report
show that if no check dams had been constructed, the vegetation of National Key Research Project. (In Chinese)
Li, Y. F., Li, Z., Wang, Z. Y., Wang, W. L., Tian, S. M., & Jia, Y. H. (2017). Impact of
cover might increase following the revegetation projects but would artificially planted vegetation on the ecological restoration of movable sand
soon quickly reduce. The final vegetation cover without check dams dunes in the Mugetan Desert, northeastern Qinghai-Tibet plateau. International
would be 30% less than the actual value. The check dams provide a Journal of Sediment Research, 32(2), 277e287, https://doi.org/10.1016/
j.ijsrc.2017.02.003
foundation for the development of vegetation. Check dams not only Li, Z. B., Zhu, B. B., & Li, P. (2008). Advancement in study on soil erosion and soil and
control gully erosion, but also reduce slope erosion by promoting water conservation. Acta Pedologica Sinica, 45(5), 802e809. (In Chinese)
vegetation development. Liu, D. S. (1965). Loess deposition in China. Beijing China: Science Press. (In Chinese)
Liu, X. Y., Gao, Y. F., Ma, S. B., & Dong, G. T. (2018). Sediment reduction of warping
dams and its timeliness on the Loess Plateau. Chinese Journal of Hydraulic En-
Declaration of competing interest gineering, 49(1), 145e155. (In Chinese)
Liu, C., He, Y., Li, Z. W., Chen, J., & Li, Z. J. (2020). Key drivers of changes in the
sediment loads of Chinese rivers discharging to the oceans. International Journal
The authors declare that they have no known competing of Sediment Research. , https://doi.org/10.1016/j.ijsrc.2020.05.005
financial interests or personal relationships that could have MWRC (Ministry of Water Resources of China). (2014). Bulletin of the first national
water census for soil and water conservation (pp. 6e8). Ministry of Water Re-
appeared to influence the work reported in this paper.
sources of China. (In Chinese)
Pan, B., Wang, J., & Gao, H. (2005). Terrace dating as an archive of the run-through of
Acknowledgements the Sanmen Gorge. Program in Natural Science, 15(12), 1096e1103, https://
doi.org/10.1080/10020070512331343191
Schumm, S. A. (1956). Evolution of drainage systems and slopes in badlands at Perth
This study was supported by the National Natural Sciences Amboy (Vol. 67, pp. 597e646). New Jersey: Geological Society of American
Foundation of China Fund Project (Grants No. 41731289 and No. Bulletin, https://doi.org/10.1130/0016-7606(1956)67[597:eodsas]2.0.co;2
Shi, N. (1987). Historical evolution of gullies in the Loess Plateau. Forum of Chinese
51879285). We further acknowledge support from the Consultation Geographical History, (2), 3e54. (In Chinese)
and Evaluation Project of the Chinese Academy of Sciences (No. Shi, H. L., Hu, C. H., Wang, Y. G., Liu, C., & Li, H. M. (2017). Analyses of trends and
2018-Z02-A-008), the National Key R&D Program of China (No. causes for variations in runoff and sediment load of the Yellow River. Interna-
tional Journal of Sediment Research, 32(2), 171e179, https://doi.org/10.1016/
2018YFC0407103), and the Basic Research Project of the China j.ijsrc.2016.09.002
Institute of Water Resources and Hydropower Research (No. Shi, P., Zhang, Y., Ren, Z., Yu, Y., Li, P., & Gong, J. (2019). Land-use changes and check
GE0145B0112017). The authors thank Dr. Ruiyu Wang for providing dams reducing runoff and sediment yield on the Loess Plateau. The Science of
the Total Environment, 664, 984e994, https://doi.org/10.1016/
valuable comments and polishing the English. Data for the sedi-
j.scitotenv.2019.01.430
ment samples from the Loess Plateau and the Yellow River were Tian, S. M., Li, Z. W., Wang, Z. Y., Jiang, E. H., Wang, W. L., & Sun, M. (2021). Mineral
originally presented in Tian et al. (2021), and data on the total composition and particle size distribution of river sediment and loess in the
middle and lower Yellow River. International Journal of Sediment Research, 36(3),
storage capacity and siltation storage capacity of the check dam
392e400, https://doi.org/10.1016/j.ijsrc.2020.07.008
were originally presented in Liu et al. (2018), Annual average Wang, Y. S. (2011). Fulfil responsibility and take measures to ensure security of
sediment volume data of the Yellow River were originally pre- sediment check dams on the Loess Plateau in flood season. Yellow River
sented in Wang and Liu (2019), and the vegetation coverage data Conservancy Commission of the Ministry of Water Resources (In Chinese)
www.yrcc.gov.cn
were originally presented in Li et al. (2019), Longitudinal profiles of Wang, Y. F., Fu, B. J., Chen, L. D., Lv, Y. H., & Gao, Y. (2011). Check dam in the Loess
gully come from the Bigemap (http://www.bigemap.com/). Plateau of China: Engineering for environmental services and food security.
Environmental Science Technology, 45, 10298e10299, https://doi.org/10.1021/
es2038992
References Wang, Z. Y., Huang, G. H., Wang, G. Q., & Gao, J. (2004). Modeling of vegetation-
erosion dynamics in watershed systems. Journal of Environmental Engineer-
Chen, Z. Y., Ma, L. Q., Yu, S., Chen, S. J., Zhou, X. B., Sun, P., & Li, X. (2015). Back ing, 130(7), 792e800, https://doi.org/10.1061/(asce)0733-9372(2004)130:
analysis of the draining process of the Tangjiashan barrier lake. Journal of Hy- 7(792)
draulic Engineering, 141(4), 05014011, https://doi.org/10.1061/(asce)hy.1943- Wang, W. Z., & Jiao, J. Y. (1996). Rainfall erosion on the Loess Plateau and the sediment
7900.0000965 transportation in the Yellow River (pp. 1e289). Beijing, China: Science Press. (In
Chen, Z. Y., Ping, Z. Y., Wang, N. X., Yu, S., & Chen, S. J. (2019). An approach to quick Chinese)
and easy evaluation of the dam breach flood. Science China Technological Sci- Wang, Z. Y., Lee, J. H., & Melching, C. S. (2015b). River dynamics and integrated river
ences, 62(10), 1773e1782, https://doi.org/10.1007/s11431-018-9367-4 management (pp. 1e800). Berlin/Beijing: Springer Verlag and Tsinghua Press.
Chen, N., & Zhou, W. H. (1965). Fluvial processes of the lower Yellow River (pp. 978-3-642-25651-6.
1e160). Beijing China: Science Press. (In Chinese) Wang, Z. Y., & Liu, C. (2019). Controlling the Yellow River: 2000 years of debate on
Fu, B. J., Liu, Y., Lu, Y. H., He, C. S., Zeng, Y., & Wu, B. F. (2011). Assessing the soil control strategies (pp. 1e103). Paries, France: UNESCO.
erosion control service of ecosystems change in the Loess Plateau of China.
 Z. Wang et al. / International Journal of Sediment Research 36 (2021) 668e677 677

Wang, X. K., Li, Z. B., & Xu, Y. Y. (2015a). Effects of mulching and nitrogen on soil Yang, J., Yao, W., Zheng, M., & Li, L. (2017). Analysis on gravitational sediment yield
temperature, water content, nitrate-N content and maize yield on the Loess in the check dam controlled basins of Chabagou watershed. Chinese Journal of
Plateau of China. Agricultural Water Management, 161, 53e64, https://doi.org/ Hydraulic Engineering, 48(2), 241e245. (In Chinese)
10.1016/j.agwat.2015.07.019 Ye, Z. O. (1986). Investigation and analysis of erosion control in the Anjiagou water-
Wang, Z. Y., Qi, P., & Melching, C. S. (2009). Fluvial hydraulics of hyperconcentrated shed. Collection of reports of experimental results on soil and water conser-
floods in Chinese rivers. Earth Surface Processes and Landforms, 34, 981e993, vation (pp. 8e16). Gansu Province: Research Institute of Dingxi Soil and Water
https://doi.org/10.1002/esp.1789 Conservation. (In Chinese)
Wang, F. X., & Wang, Z. Y. (2007). Acceleration of vegetation succession on eroded Yu, S., Chen, Z. Y., Yang, X. C., Su, A. S., Li, Y. L., & Zhou, J. W. (2019). Flow discharge
land by reforestation in a subtropical zone. Ecological Engineering, 31(4), model test study of soft spillway on check dam. Chinese journal of hydraulic
232e241, https://doi.org/10.1016/j.ecoleng.2007.07.004 engineering, 50(5), 612e620. (In Chinese)
Wang, Z. Y., Wang, G. Q., & Li, C. Z. (2005). A preliminary study on vegetation- Zhang, F., Li, D. G., & Wan, T. C. (1986). Analysis on the experiments integrated
erosion dynamics and its applications. Science in China - Series D: Earth Sci- management of the Anjiagou Watershed and the effects on soil erosion reduction.
ences, 48(5), 689e700, https://doi.org/10.1360/03yd0074 Collection of Reports of Experimental Results on Soil and Water Conservation
Wang, Z. Y., & Zhang, C. D. (2019). Mechanism of energy dissipation and hazard (pp. 44e53). Gansu Province: Research Institute of Dingxi Soil and Water
mitigation of riverbed structures in southwestern China. Chinese Journal of Conservation. (In Chinese)
Hydraulic Engineering, (1), 124e134. (In Chinese) Zhang, Z. G., Wei, Z. X., & Wang, X. P. (2001). Analysis of runoff yield from terrace
Wan, Z. H., & Wang, Z. Y. (1994). Hyperconcentrated flow. Monograph series of IAHR. system of Wangjiagou Watershed in western Shanxi. Soil and Water Conserva-
Amsterdam: A.A. Balkema Publishers. tion Science and Technology in Shanxi, China, 9, 11e12. (In Chinese)
Wei, X., Li, X. G., Li, Z. B., Zou, B. H., & Lu, K. X. (2009). A study of the gravity erosion Zhang, K., Xu, M. Z., & Wang, Z. Y. (2009). Study on reforestation with seabuckthorn
control effects of check dams on the Loess Plateau. Journal of Xi'an University of in the Pisha Sandstone area. International Journal of Hydro-Environment
Architecture and Technology, 28(6), 856e861. (In Chinese) Research, 3(2), 77e84, https://doi.org/10.1016/j.jher.2009.06.001