Estimation of sediment trapping behind check dams using ...

10
Contents lists available at ScienceDirect Journal of Hydrology journal homepage: www.elsevier.com/locate/jhydrol Research papers Estimation of sediment trapping behind check dams using high-density electrical resistivity tomography N.F. Fang a,b,d , Y. Zeng c , L.S. Ni b,d , Z.H. Shi b,c, a State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of Soil and Water Conservation, Northwest A&F University, 26 Xinong Road, Yangling, Shaanxi Province 712100, PR China b Institute of Soil and Water Conservation, Chinese Academy of Sciences and Ministry of Water Resources, 26 Xinong Road, Yangling, Shaanxi Province 712100, PR China c College of Resources and Environment, Huazhong Agricultural University, Wuhan 430070, PR China d University of Chinese Academy of Sciences, Beijing 100049, PR China ARTICLE INFO This manuscript was handled by GeoSyme, Editor-in-Chief, with the assistance of Jesús Mateo-Lázaro, Associate Editor Keywords: Small catchment Loess Plateau Soil erosion Sediment deposition Sediment volume estimation ABSTRACT Check dams are very important structures for torrent control areas throughout the world. Measuring the sedi- ment deposition behind check dams is thought to be an eective method of estimating the sediment yield of small catchments. In this study, we employed a high-density electrical resistivity system to detect sediment at six check dams distributed in the hilly and gully area of the Loess Plateau. A total of 11 lines were arranged, with 60 electrodes along each line. The resultant resistivity data were inverted into subsurface structures using least- squares inversion techniques. The results indicate that the sediment bottom can been identied using an elec- trical resistivity contour method. Using trenches, boreholes and drillings to calibrate the calculated values, the accuracy of the high-density electrical resistivity tomography method was found to be quite high (95.7%). The bulk density increases with the sediment deposit depth and can reach a maximum of 1.6 g cm 3 . Geometric methods considering three shapes (V-shaped, U-shaped and trapezoidal) were used to estimate the prole area of the check dams; the V-shaped geometric method showed high accuracy. This study provides a four-step approach based on high-density electrical resistivity tomography and a geometric method to estimate sediment yield for check dam-controlled catchments. This methodology enables the measurement of the total sediment yield, in- cluding both the suspended and bedload sediment transport, from an ungauged catchment. 1. Introduction Check dams are one of humanitys earliest constructions (Leigh et al., 2013). The building of check dams has been one of the most commonly utilized structural measures for channel stabilization and erosion control in catchments that experience torrents (Garcia and Mario, 2010; Romero-Díaz et al., 2012). In addition, in arid and semi- arid regions, check dams can make land less susceptible to erosion and can help restore eroded land to its former condition (Doolittle, 2013). Measuring sediment deposition behind check dams is thought to be an eective method of calculating sediment yield (Verstraeten and Poesen, 2002). Compared to simultaneous measurement at the plot scale and the creation of sediment-rating curves at the large river scale, mea- suring sediment trapped by check dams provides useful information at the small catchment scale. Plotting the observation data is problematic for large watersheds, and station gauge data provides sediment yield curves, which more closely reect the suspended sediment rather than the total sediment yield. In addition, performing soil erosion observa- tions is a time-consuming and resource-intensive process (Cerdan et al., 2010, Fang et al., 2017). However, using a check dam-based metho- dology allows measurement of the total sediment yield, including both the suspended and bedload sediment transport from an ungauged catchment. Soil erosion leads to widespread land degradation and many other environmental problems (Fang et al., 2017). The Chinese Loess Plateau (CLP) is one of the global hotspots of soil erosion due to its combination of unique soil properties and considerable human pressures (Wang et al., 2007a,b). On the Loess Plateau, an approximately 2.5 × 10 5 km 2 area is subject to extremely severe erosion and has a yearly soil erosion modulus greater than 5000 t km 2 (Fu et al., 2011). This area is mainly distributed in the hilly and gully loess area (Chen et al., 2007). To mitigate soil erosion and improve land productivity, the construction of check dams has been widely implemented since the 1960s. Most of the check dams constructed before the 1980s were without ood discharge https://doi.org/10.1016/j.jhydrol.2018.11.062 Received 26 August 2018; Received in revised form 19 November 2018; Accepted 21 November 2018 Corresponding author at: College of Resources and Environment, Huazhong Agricultural University, Wuhan 430070, PR China. E-mail address: [email protected] (Z.H. Shi). Journal of Hydrology 568 (2019) 1007–1016 Available online 26 November 2018 0022-1694/ © 2018 Elsevier B.V. All rights reserved. T

Transcript of Estimation of sediment trapping behind check dams using ...

Page 1: Estimation of sediment trapping behind check dams using ...

Contents lists available at ScienceDirect

Journal of Hydrology

journal homepage: www.elsevier.com/locate/jhydrol

Research papers

Estimation of sediment trapping behind check dams using high-densityelectrical resistivity tomography

N.F. Fanga,b,d, Y. Zengc, L.S. Nib,d, Z.H. Shib,c,⁎

a State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of Soil and Water Conservation, Northwest A&F University, 26 Xinong Road,Yangling, Shaanxi Province 712100, PR Chinab Institute of Soil and Water Conservation, Chinese Academy of Sciences and Ministry of Water Resources, 26 Xinong Road, Yangling, Shaanxi Province 712100, PR Chinac College of Resources and Environment, Huazhong Agricultural University, Wuhan 430070, PR ChinadUniversity of Chinese Academy of Sciences, Beijing 100049, PR China

A R T I C L E I N F O

This manuscript was handled by Geoff Syme,Editor-in-Chief, with the assistance of JesúsMateo-Lázaro, Associate Editor

Keywords:Small catchmentLoess PlateauSoil erosionSediment depositionSediment volume estimation

A B S T R A C T

Check dams are very important structures for torrent control areas throughout the world. Measuring the sedi-ment deposition behind check dams is thought to be an effective method of estimating the sediment yield ofsmall catchments. In this study, we employed a high-density electrical resistivity system to detect sediment at sixcheck dams distributed in the hilly and gully area of the Loess Plateau. A total of 11 lines were arranged, with 60electrodes along each line. The resultant resistivity data were inverted into subsurface structures using least-squares inversion techniques. The results indicate that the sediment bottom can been identified using an elec-trical resistivity contour method. Using trenches, boreholes and drillings to calibrate the calculated values, theaccuracy of the high-density electrical resistivity tomography method was found to be quite high (95.7%). Thebulk density increases with the sediment deposit depth and can reach a maximum of 1.6 g cm−3. Geometricmethods considering three shapes (V-shaped, U-shaped and trapezoidal) were used to estimate the profile area ofthe check dams; the V-shaped geometric method showed high accuracy. This study provides a four-step approachbased on high-density electrical resistivity tomography and a geometric method to estimate sediment yield forcheck dam-controlled catchments. This methodology enables the measurement of the total sediment yield, in-cluding both the suspended and bedload sediment transport, from an ungauged catchment.

1. Introduction

Check dams are one of humanity’s earliest constructions (Leighet al., 2013). The building of check dams has been one of the mostcommonly utilized structural measures for channel stabilization anderosion control in catchments that experience torrents (Garcia andMario, 2010; Romero-Díaz et al., 2012). In addition, in arid and semi-arid regions, check dams can make land less susceptible to erosion andcan help restore eroded land to its former condition (Doolittle, 2013).Measuring sediment deposition behind check dams is thought to be aneffective method of calculating sediment yield (Verstraeten and Poesen,2002). Compared to simultaneous measurement at the plot scale andthe creation of sediment-rating curves at the large river scale, mea-suring sediment trapped by check dams provides useful information atthe small catchment scale. Plotting the observation data is problematicfor large watersheds, and station gauge data provides sediment yieldcurves, which more closely reflect the suspended sediment rather than

the total sediment yield. In addition, performing soil erosion observa-tions is a time-consuming and resource-intensive process (Cerdan et al.,2010, Fang et al., 2017). However, using a check dam-based metho-dology allows measurement of the total sediment yield, including boththe suspended and bedload sediment transport from an ungaugedcatchment.

Soil erosion leads to widespread land degradation and many otherenvironmental problems (Fang et al., 2017). The Chinese Loess Plateau(CLP) is one of the global hotspots of soil erosion due to its combinationof unique soil properties and considerable human pressures (Wanget al., 2007a,b). On the Loess Plateau, an approximately 2.5×105 km2

area is subject to extremely severe erosion and has a yearly soil erosionmodulus greater than 5000 t km−2 (Fu et al., 2011). This area is mainlydistributed in the hilly and gully loess area (Chen et al., 2007). Tomitigate soil erosion and improve land productivity, the construction ofcheck dams has been widely implemented since the 1960s. Most of thecheck dams constructed before the 1980s were without flood discharge

https://doi.org/10.1016/j.jhydrol.2018.11.062Received 26 August 2018; Received in revised form 19 November 2018; Accepted 21 November 2018

⁎ Corresponding author at: College of Resources and Environment, Huazhong Agricultural University, Wuhan 430070, PR China.E-mail address: [email protected] (Z.H. Shi).

Journal of Hydrology 568 (2019) 1007–1016

Available online 26 November 20180022-1694/ © 2018 Elsevier B.V. All rights reserved.

T

Page 2: Estimation of sediment trapping behind check dams using ...

facilities or spillways. According to the Yellow River ConservancyCommission, there were 31,797 check dams constructed in the cities ofYulin and Yan’an (hilly area of the CLP) during 1950 to 1989. Amongthose check dams, 83% had no spillway. As a result, these dams had lowflood prevention ability and could easily be destroyed by floods. In1986, the Key Dams Project was implemented on the CLP. This projectaimed to construct check dams with large capacity and high quality. By2013, a total of 58,446 check dams, providing more than 10,000m3 ofstorage, had been constructed and remained intact; the resultant sedi-ment silting was estimated to be approximately 5.7 billionm3 (CMWR,2014). However, this number was lower than previous records (CMWR,2003) and most old dams constructed before 1980 were no included.Wang et al. (2011) reported that more than 11× 104 checked dams hadbeen constructed and captured approximately 21× 109m3 sediment onthe CLP as of 2003. It is not clear how many check dams currently existon the CLP. Many studies (e.g. Fu et al., 2011; Wang et al., 2016; Zhao

et al., 2017a,b) have reported that the sediment load of Yellow Riverdecreased from 1.34 Gt yr−1 during 1950–1979 to 0.32 Gt yr−1 during2000–2010 mainly due to construction of check dams and ecologicalrestoration projects on the CLP. However, exact estimations of the se-diment trapped by check dam are scare. Detailed records on the topo-graphy of the channel bed and trapped sediment area available for keycheck dam (with capacity > 5×105m3) funded by the central gov-ernment (see Supplementary information). However, no sediment vo-lume information is available for the considerable number of smallerdams constructed by local farmers.

Due to the semiarid climate of the Loess Plateau, these check damsusually hold no surface water, making experiments more convenient toconduct. The sediment deposited behind check dams exhibits a distinctsedimentary sequence related to each flood, with the thickness of eachcouplet varying from a few centimeters to several decimeters (Chenet al., 2016; Wang et al., 2017), and provides deep insights into the

Fig. 1. Six study catchments on the Loess Plateau Note: SYG, HJW, NYG, LNG, NZG and LJT represent the catchment names of Shayangou, Hujiawan, Nianyangou,Lannigou, Nianzigou and Lijiata, respectively.

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1008

Page 3: Estimation of sediment trapping behind check dams using ...

long-term sediment deposition process in the catchment ecosystems(Chen et al., 2017). The studies conducted on check dams include re-search on the particle size redistributed by soil erosion (Vaez et al.,2017), carbon sequestration opportunities (Lü et al., 2012; Ran et al.,2014), the fingerprint method used to calculate sediment sources (Chenet al., 2016; Liu et al., 2018), and soil erosion modeling (Grum et al.,2017, Zhao et al., 2017a,b). Earlier studies estimated sediment volumeand mass by using one profile or several drillings to represent the wholecheck dam. For example, Wei et al. (2017) used ten drilling holes fromthree check dams to estimate their retained sediment. Li et al. (2016)use one profile per check dam to correlate the sedimentation of threecheck dams with the rainstorm characteristics on the Loess Plateau; thismethod was also used by Zhao et al. (2017a,b) and Wang et al. (2017).Excavating sediment profiles and drilling are laborious, costly and time-consuming tasks; therefore, researchers cannot obtain a complete de-scription of the depth of a large check dam. It is necessary to develop aconvenient method to reveal the true dam depth and accurately cal-culate the sediment volume.

Due to its convenience, low time requirement and high resolution,electrical resistivity tomography (ERT) has been widely used in variousengineering, environmental, and mining studies (Baines et al., 2002;Ritz et al., 1999), for example, for soil layer heterogeneity character-ization with 2-D (Fukue et al. 1999; Besson et al. 2004) and 3-D ima-ging techniques (Bourennane et al., 1998; Chambers et al., 2012). Hsuet al. (2010) used electrical resistivity imaging to accurately detectsediment deposits along the Peikang River. Chambers et al. (2012)showed that ERT provides a quick method to quantify the sedimentfilling volume across large sites with a high degree of precision. In thepresent study, ERT imaging is used to investigate sediment depositsbehind check dams on the Loess Plateau, China. The specific aims ofthis study are (1) to quantitatively assess a simple approach to detectthe depth of check dam fill deposit sediments using the ERT techniqueand (2) to calculate the sediment yield of small catchments based on thesediment volume estimation. To achieve these aims, six check damswith different characteristics were selected in the hilly and gully area ofthe Loess Plateau. Field work included creating deep trenches, bore-holes and manual drillings for calibrating the ERT results.

2. Study area and methods

2.1. Study area and field work

We conducted field experiments at six catchments in the hilly andgully area of the Loess Plateau. Among all the catchments, theHujiawan (HJW), Shayangou (SYG) and Nianyangou (NYG) catchmentswere studied for calibrating the proposed detection method, while theother catchments, the Lijiata (LJT), Lannigou (LNG) and Nianzigou(NZG) catchments, were selected for applying this method (Fig. 1 and

Table 1). The check dams are mostly upper stream check dams. Theclimate of the study area is semiarid continental with an average annualprecipitation that varies between 420 and 510mm. Precipitation occursmainly from June to September, in the form of high-intensity storms.The studied catchment areas vary from 0.13 to 27.02 km2. The loess soildepth is approximately 80–120m in this area (Liu et al., 1994). The soilcan be divided into the Malan Loess, Lishi Loess and Wucheng Loess.The Pleistocene Wucheng Loess, at the bottom of the soil profile, is alsocalled old loess; it has a tight structure and is very resistant to erosion.The new world Q4 secondary loess at the surface ranges from fine silt tosilt and is vulnerable to erosion. Check dams are usually built acrossgullies on bedrock or the base of the Wucheng Loess.

A high-precision Global Positioning System (GPS) and QuickBirdimagery were used to obtain topographic maps (scale 1:10,000) of thecatchments. The ERT method was calibrated using data from trenches,boreholes and drillings in the HJW, SYG and NYG catchments. In boththe HJW and SYG catchments, the trenches were explored using anexcavator. Boreholes were explored in both the SYG and NYG catch-ments. In the other catchments, at least one manual drilling for eachcheck dam was explored (Fig. 2). Bulk soil density (BD, g cm−3) sam-ples were collected from the trenches of HJW and SYG; the sampleswere collected from the top to the bottom of the trenches.

2.2. High-density electrical resistivity tomography

High-density ERT requires many observations. We employed a DUK-2B high-density electrical resistivity system (Chongqing GeologicalInstrument Factory, China). We used 60 electrodes spaced at differentintervals to create survey lines varying from 60 to 295m long (Fig. 3). Atotal of 11 lines were arranged in the six studied catchments (Table 2).The lines at SYG, NYG and NZG were arranged parallel to each otherbut perpendicular to the other nine lines. Wenner configurations(alpha) were adopted according to the experiments of Hauck et al.(2003). Electrodes A and B are power poles with an electric current (I).The electrical resistivity was calculated through the difference in vol-tage value (ΔU) between electrodes M and N, as described by Eq. (1):

=ρ k UI

Δ(1)

in which ρ is electrical resistivity and k is a coefficient related to theelectrode configurations.

The resistivity of fluvial sediment is mainly dominated by the por-osity and water saturation of its pore spaces (Hsu et al., 2010). Sedi-ments deposited behind a check dam exhibit sedimentary sequences offlood couplets. The coarse sediment is always below the fine sedimentin a couplet. A couplet can vary from a few centimeters to severaldecimeters or even meters (Chen et al., 2016). In addition, sedimentdelivery triggered by soil erosion exhibits size selectivity (Wang et al.,

Table 1Descriptions of the study catchments and check dams.

Name Location Catchment area (km2) Silt period Trapping area (ha) Elevation (m)

HJW 36°28′08″–36°32′53″N109°45′05″–109°48′49″E

27.02 1987–2003 10.3 946–1296

SYG 37°54′29″–37°55′02″N110°08′59–110°09″38E

0.69 1960–2017 3.0 1023–1126

NYG 37°35′33″–37°35′54″N110°22′4–110°22′28E

0.18 1960–1990 0.6 1027–1118

LNG 37°53′33″–37°52′34″N109°10′43″–109°11′11″

1.51 1967–2017 2.3 1189–1324

LJT 37°21′31″–37°20′59″N109°39′26″–109°39′48″E

0.31 2012–2017 0.2 1110–1238

NZG 37°34′12″–37°35′11″N109°50′23″–110°51′11″E

1.17 1972–2005 4.8 990–1134

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1009

Page 4: Estimation of sediment trapping behind check dams using ...

2014). The resistivity data sets were processed using Geogiga RImager5.02 (Geogiga Technology Corp, Canada) to check the noise(I < 0.01 A), and then the data were inverted to images using thesoftware package RES2DINV (Loke and Barker, 1995). RES2DINV re-built the tomographic image using a smoothness-constrained least-squares method. The deposit sediment is expected to have a differentresistivity than that of the channel bottom. Electrons are discharged inthe deposited sediment and recharged in the bedload, causing abruptchanges in the inverted image data. Hsu et al. (2010) used a Laplacianoperator method to detect an abrupt change corresponding to Laplacianzero lines. In the present study, the rebuilt tomographic image datawere reprocessed using a contour method to differentiate the layeredstructures. The lines of abrupt change corresponding to possible bottomboundaries were then extracted. This step was processed using Origi-nPro 9 software (OriginLab Corporation, Northampton, USA). Irregularsediment volumes behind check dams were reconstructed using ArcMap10.0 (ESRI Inc, USA) and AutoCAD 2014 (Autodesk Inc, USA).

The ERT method was then calibrated using the results of the fieldwork. We calculated the differences between the results of the ERTmethod and field trench, borehole or drilling method at HJW, NYG andSYG, respectively. The percent bias (PBias), which is expressed as Eq.(2), was used to describe the differences.

=∑

×=

PBiasn

100i 1n (x y )

xi i

i

(2)

where yi is the value of the ERT depth and xi is the depth obtained fromthe trench, borehole or drilling method, which provided the real values.

2.3. Sediment estimation

Once the lithological boundaries are determined, several methodsmay be used to estimate sediment volume (V, m3) (Ramos-Diez et al.,2016). Several geometric methods, e.g., using a prism (Castillo et al.,2007), a pyramid (Romero-Díaz et al, 2007) or trapezoids (Bellin et al.,2011), have been used at small check dams to calculate the dam vo-lume. We employed three of the most common dam shapes on the LoessPlateau, V-shaped, U-shaped and trapezoidal, to calculate the volume ofthe sediment trapped by the studied check dams. We used ArcGIS-3Dand AutoCAD to calculate the real volume. For a quick estimationmethod, we consider V-shaped dams as triangular prisms. We dividedU-shaped sediment into two parts including a cuboid above and asemicylinder below, and consider trapezoidal dam land as prismoid. Weevaluated the prediction capability of each geometric method with re-spect to the real volume of sediment estimated using the field surveymethod. We use the PBias, coefficient of determination (R2) andNash–Sutcliffe efficiency (ENS) values to evaluate the performance ofeach geometric method. These metrics are expressed as follows:

= −∑ −

∑ −=

=ENS

GF

1(F )

(F )i

ave

i 1n

i2

i 1n

i2 (3)

=∑

×=

PBiasn

100i 1n (F G )

Fi i

i

(4)

= ⎧⎨⎩

∑ − × −

∑ − × ∑ −⎫⎬⎭

=RF F G G

F F G G

( ) ( )

[ ( ) ] [ ( ) ]i ave i ave

ni ave

ni ave

2 i 1n

12 0.5

12 0.5

2

(5)

where Fi is the volume calculated by the field survey method, Fave is the

Fig. 2. Field work at the check dams: (a) trench profile, (b) boreholes, (c) manual drilling, (d) samples collecting, (e) bulk density sample, and (f) sediment couplets.

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1010

Page 5: Estimation of sediment trapping behind check dams using ...

mean of the survey values, Gi is the volume calculated by the geometricmethods, and Gave is the mean of the survey values.

The sediment yield of the corresponding catchments were de-termined using the following equation:

=∑ ×=SY

V BDTE

100( )i ii 1

n

(6)

where SY is the sediment yield (t), Vi is the volume of sediment at agiven depth, and BDi is the dry bulk density of the sediment deposit atthe given depth (g cm−3). At the HJW and SYG catchments, BDi wasobtained from the trench samples, while for the other four catchments,BDi was derived from the relationship between BD and depth. TE is the

sediment trap efficiency of the check dam (percent), determined by theproportion of fluvial sediment that is deposited in the check dam. In thisstudy, we used the empirical equation of Brown and Jarvis (1943) toestimate TE, which is described by Eq. (4):

= ×⎛

⎝⎜ −

+ ×⎞

⎠⎟TE

D100 1 1

1 0.0021 CA (7)

where C is the storage capacity of the check dam (m3), A is the catch-ment area (km2), and D is the empirical parameter that depends on thecharacteristics of the reservoir. Based on the studies of Verstraeten andPoesen (2000) and Bellin et al. (2011), we used a D value of 1 in thisstudy.

The area-specific sediment yield (SSY, t km−2 a−1) was then cal-culated as follows:

SSY SYA Ye (8)

where A is the catchment area, and Ye is the sediment silt year (seeTable 1).

3. Results

3.1. Electrical resistivity imaging and sediment bottom detection

Following the approach of Loke et al. (2003) and Chambers et al.(2012), a least-squares method was used to simulate a model that isconsistent with the measured data. To reduce the discrepancy betweenthe simulated model and observed data, iterative programs were usedto produce a geologically realistic model of the inverted image. Fig. 4shows the inversion of the electrical resistivity at survey Lines 1 to 11

Fig. 3. Illustration of the electrode arrangement for ERT.

Table 2Electrode arrangement and field surveys of the six check dams.

ERI site/Line #

Length(m)

Average/Range ofthe boundary (m)

Direction of thegully flow

Field survey

HJW/Line 1 295 13.4/9.2–16.1 Parallel Trench× 1Drillins× 3

SYG/Line 2 295 10.2/8.1–15.6 Perpendicular Trench× 1SYG/Line 3 295 12.7/6.9–22.2 Parallel Borehole× 1NYG/Line 4 120 5.2/1.3–13.5 Perpendicular noneNYG/Line 5 90 6.6/5.4–8.1 Parallel Borehole× 1LJT/Line 6 60 2.4/0.8–7.8 Parallel Drilling×1LJT/Line 7 60 2.7/1.6–5.5 Parallel noneNZG/Line 8 295 8.3/2.5–20.9 Parallel Drilling×1NZG/Line 9 295 7.4/2.1–16.1 Parallel noneNZG/Line

10295 3.8/0.5–6.5 Perpendicular none

LNG/Line11

90 27.5/21.8–29.5 Parallel Drilling×1

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1011

Page 6: Estimation of sediment trapping behind check dams using ...

across the six check dams. Due to differences in topography, sedimentcouplets and hydraulic conditions, the electrical resistivities show largevariances among the check dams. Generally, most upper layers ex-hibited resistivities< 50 ohm.m except for Line 4 and Line 5 at NYG,which were between 50 and 100 ohm.m. Notably, Lines 1 and 2 wererecorded in April 2017 before the planting season, while the other ninelines were recorded in November after the harvest. Thus, a comparisonof Line 2 and Line 3 shows seasonal changes. Integrated and continuouslayered structures were found in all 11 inverted images. We anticipatethat the interface between the sediment and channel bottom coincideswith one of the lines of change in the inverted images.

Following Hsu et al. (2010) and Chambers et al. (2012), we sear-ched for abrupt vertical changes in the resistivity data. Fig. 4 shows thatall 11 images have well layered structures. We used field surveys tolocate the electrical boundary at the HJW, SYG and NYG catchments. Atthe HJW check dam, one trench and three drillings were used to cali-brate the boundary of the check dam determined by ERT. Two bore-holes were used to calibrate the sediment bottom at the SYG and NYGcatchments. The results show that abrupt changes in increasing elec-trical resistance coincided with the dam bottom at HJW. At the NYG

and SYG catchments, however, the channel bottom has a lower electricsignal than that of the upper sediment layer. The results from thesethree check dams show that the channel bottom always coincides withan abrupt change in electrical resistivity. However, the change can beexhibited as either a decrease or an increase, meaning that the electricalresistivity of the upper sediment can be either higher or lower than thatof the channel bottom. As schematically represented in Fig. 5, the dif-ference in the results between the ERT and field survey methods variesfrom 1.7% to 7.3% with an average PBias of 4.3%. The electrode ar-rangement and field survey information from all six check dams aregiven in Table 2. Manual drillings were used to check the boundaries ofthe channel bottom depth. However, drilling at the LNG check damfailed to provide information because the channel bottom depth of thedam is deeper than the drilling. Thus, the ERT bottom boundary at theLNG check dam is inferred based on the results of the other five checkdams.

3.2. Sediment volume and mass estimation

Fig. 6 shows all the boundaries determined by the ERT method and

Fig. 4. Inverted images of the electrical resistivity contours from the 11 survey lines.

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1012

Page 7: Estimation of sediment trapping behind check dams using ...

calibrated by the field data (except for Line 11, which was inferred atLNG). The depth of the check dams vary considerably. The LJT checkdam is the shallowest dam; the average depths from Lines 6 and 7 are2.4 and 2.7m, respectively. The LJT check dam controlled a smalldrainage area of 0.31 km2, and the silt years were from 2012 to 2017,when the “Grain-for-Green” project had been implemented for morethan 10 years. Line 11 at LNG shows the deepest depth, with an averagedepth of 27.5 m. Trenches were explored using an excavator machine inthe HJW and SYG catchments. At the HJW dam, we collected BDsamples from the top to the bottom of the trench. The BD of HJW variedfrom 1.36 to 1.60 g cm−3, while the BD of SYG varied from 1.32 to1.60 g cm−3; both BD trends increased from the top to the bottom ofthese check dams. For the SYG dam, we collected BD samples in the top10m because mud prevented the excavation of a deeper trench. We usethe average value of the bottom five samples to represent the BD ofdepths greater than 10m in SYG and of depths greater than 12.7m inHJW; these averages are 1.53 and 1.56 g cm−3 for the SYG and HJW,respectively.

For the depths of the other four check dams, we used the relation-ship between BD and depth (De) at HJW and SYG to generate a pre-dictive equation, which can be expressed as follows:

Fig. 5. ERT and field survey results from the HJW check dam.

Fig. 6. Boundaries determined by ERT using the contour method for the 11 profiles.

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1013

Page 8: Estimation of sediment trapping behind check dams using ...

= × + = <BD De0.0155 1.39 (R 0.57) ( D 12.7 m)2 (9)

When De > 12.7, we consider BD equal to 1.56 g cm−3. The fluc-tuation of BD can be attributed to the grain size, pore space, and watersaturation difference of each sediment couplet (Chen et al., 2017). Inthe trench profile of the HJW check dam, the sediment water contentincreases with depth; this relationship can be expressed as follows:

= × + = =Wc De R0.0002 0.0767 (n 20, 0.85)2 (10)

where Wc is the water content of the sediment samples and De is thedepth.

Fig. 7 shows the cross-sectional shapes of the six check dams. Thedam body profiles can be grouped into three categories: V-shaped,trapezoidal and U-shaped. The sediment volumes were calculated as631421, 327600, 223960, 31043, 2189, and 378754m3, for HJW, NZG,SYG, NYG, LJT, and LNG, respectively. To provide a quick and con-venient estimate of the sediment volume, we used geometric methods(Ramos-Diez et al., 2016) to estimate the sediment volume in each damand then compared the differences between the geometric methods andfield survey results. For the geometric methods, we need only threeparameters, which are top width (sediment surface at the check dambody), bottom width and the vertical height of the dam. The criteria forexamining the accuracy of the geometric estimation are shown in

Table 3. The overall performance of the V-shaped estimation is the best,as evidenced by PBias, R2 and ENS values of 15%, 0.988 and 0.973,respectively. The trapezoidal estimation is acceptable, while the U-shaped estimation is poor.

Based on the BD and sediment volume, the total sediment yieldswere calculated at every depth interval. The total sediment yields wereestimated to be 947131, 484848, 335940, 5633, 3109 and 575706 t forHJW, NZG, SYG, NYG, LJT, and LNG, respectively (Fig. 8). The SSYvalues for the six check dams were then calculated as 2337, 3061, 8542,8451, 3109 and 7625 t km−2 a-1, respectively.

4. Discussion

4.1. Trapped sediment estimation

By applying the ERT technique, we estimate the boundary between

Fig. 7. Profiles of the six check dams.

Table 3Criteria for examining the accuracy of the geometric methods.

V-shaped U-shaped Trapezoidal

PBias 15% 43% 14%R2 0.988 0.708 0.978ENS 0.973 0.641 0.613

Fig. 8. Sediment yield and SSY for the six check dams.

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1014

Page 9: Estimation of sediment trapping behind check dams using ...

the sediment and underlying subface using inverted images. The trench,borehole and drilling data confirm the high accuracy of the ERTmethods (Fig. 4). The boundary determination with this method is morecomplicated than that described in the studies of Hsu et al. (2010) andChambers et al. (2012). The landform of the study area is fragile andcomplicated, causing the channel bottom to have a very complicatedshape. For example, several slopes extended below the sediment atdeeper deposit depths in the NYG catchment. Our results show that thecontour method can identify a surface of change that coincides with thebottom of a check dam in an inverted electrical resistivity image.

The existing studies of sediment volume evaluation were mainlyconducted in Europe (e.g., Castillo et al., 2007). Ramos-Diez et al.(2017a,b) compared several methods of evaluating the volumes of 25check dams and found significant differences among these methods(geometric: prism and pyramid; topographic: digital terrain models,trapezoids and section methods). The abovementioned studies are themost detailed descriptions of sediment volume estimation of checkdams. However, check dams in Europe are very different from those onthe Loess Plateau of China. Due to severe soil erosion, the volume ofsediment deposited behind check dams in this study is three to fourtimes greater than that of the check dams studied by Ramos-Diez et al.(2017a,b). Existing studies usually use one profile or a limited numberof boreholes to estimate sediment volume, and sediment surveying islimited to the surface layer, e.g., 0.2 m (Wang et al., 2007a,b), 0.6 m(Bao et al., 2005), and 1m (Sun and Guo, 2011). The BD values of thesurface sediment vary from 1.32 to 1.45 g cm−3 (Li et al., 2007; Sunet al., 2004; Wei et al., 2006; Wang et al., 2008) in the existing studies.Here, we obtained BD samples from two deep trench profiles. Based onthe relationship between BD and depth, we suggest that a BD of1.5 g cm−3 is more appropriate when the sediment deposit is deeperthan 7m.

The SSY values of 8542, 8451, 7625 t km−2 a-1 were calculated forthe SYG, NYG and LNG catchments, and those of 3061, 2337,3109 t km−2 a-1 were calculated for the NZG, HJW and LJT catchments,respectively. The primary difference was mainly caused by the differentsilt periods of the check dams and intensities of the local agriculturalactivities. During the 1960s to 1990s, nearly all the slopes in SYG, NYGand LNG were used for farming, and the agricultural activity in thesecatchments was very intense. Since 1999, after the implementation ofthe “Grain-for-Green” project in the hilly loess region, soil erosion re-duced greatly because many croplands were converted to grass or forest(Wang and Shi, 2015; Chen et al., 2007, Fu et al., 2011). Farmland withslopes > 25 was restored due to the “Grain-for-Green” project, throughwhich farmers received subsidies from the government and restoredfarmland with those steep slopes (Fang et al., 2016). The LJT catchmentshows low SSY results because the vegetation cover was healthy duringits silt period (2012–2017); therefore, the main sediment source wasgully erosion. The HJW catchment is the largest of the six studiedcatchments and shows the lowest SYY, which we attribute to the largearea of forestland (62.3%).

4.2. Quick estimation method

For a quick estimation of the silt sediment of a check dam, four stepsare recommended: (i) using the ERT method to obtain an invertedelectrical resistivity image, (ii) using at least one drilling to calibrate theline of abrupt change as the boundary of the sediment, (iii) using anappropriate geometric method to estimate the sediment volume, and(iv) determining empirical BDs based on the dam depth to calculate SYand SSY. In this study, the results show V-shaped got good accuracy.However, an appropriate geometric method should consider the realshape of the objective check dam.

Check dams are a widespread soil and water conservation measurein areas where gully development is a problem (Frankl et al., 2011). Inaddition to the CLP, check dams are constructed in Ethiopia (Nyssenet al., 2009), Italy (Bombino et al., 2006), the US (Norman et al., 2016),

Spain (Romero-Díaz et al., 2012), and Iran (Hassanli et al., 2009). Thesignificant difference between check dams on the CLP and those in theother areas is their capacity. Most check dams on the CLP have capa-cities> 10,000m3 and thousands > 500,000m3, while check dams inother areas are typically small barriers constructed in shallow riversand streams for harvesting water and controlling erosion (Balooniet al.,2008). Previous studies have demonstrated the convenience andprecision of applying the ERT method to shallow sediment (Hsu et al.,2010; Chambers et al., 2012). We are confident that our estimationmethod can be applied to other areas where check dams are con-structed.

In this study, the ERT method with 60 electrodes resulted in 16resistivity layers (Fig. 3). The depth of investigation is directly pro-portional to the length of the arranged electrode lines. The triangle-shaped inverted image has data missing from both ends in the hor-izontal direction. Although the ERT method allows the arrangement ofelectrodes to extended the studied check dam area, the steep slope ofthe gullies make the depth adjustment very difficult. Thus, this methodis suitable for deep sediment deposits with large surface areas. Whenthis method is applied to a new check dam, we recommend that thearrange survey lines be as long as practicable (≤295m). ERT with 3-Dimages has been used to successfully detect sediment deposits (e.g.,Chambers et al., 2012); however, we do not recommend this 3-Dtechnique for check dams on the Loess Plateau. The large surface areasand deep depths of the deposits will require many more electrodes thanare used to create a 2-D image; therefore, the 3-D method is more la-borious and time-consuming.

5. Conclusions

In this study, the ERT method was used to detect the boundary ofsediment deposits and channel bottoms behind six check dams on theLoess Plateau, China. The resistivity data were inverted into subsurfaceelectrical structures using least-squares inversion techniques. Layeredstructures were observed in all the resistivity images. Lithologicalboundaries can be identified in the inverted resistivity image using acontour method. The field surveys showed that the ERT method canaccurately detect the depth of the sediment. However, at least onedrilling (or trench or borehole) is recommended for calibration. Overall,this study provides a quick and inexpensive method to estimate thesediment volume of check dams in the Loess Plateau with high accu-racy. While check dams are widespread on the Loess Plateau, retentioncheck dams have been built in gullies, badlands, creeks, streams andrivers in several other parts of the world. Thus, the ERT methods de-scribed in this study can also be used as a novel way to calculate thesediment yield of other ungauged catchments.

Declaration of Interest statement

We declare that we do not have any commercial or associative in-terest that represents a conflict of interest in connection with the worksubmitted.

Author Contribution Statement

Conceived and designed the experiments: NFF and ZHS. Performedthe experiments: NFF, YZ and LSN. Analyzed the data: NFF, YZ andLSN.

Wrote the paper: NFF and ZHS.

Acknowledgements

Financial support for this research was provided by the NationalNatural Science Foundation of China (41525003 and 41671282) andthe National Key Research and Development Program of China(2016YFC0402401).

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1015

Page 10: Estimation of sediment trapping behind check dams using ...

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.jhydrol.2018.11.062.

References

Baines, D., Smith, D.G., Froese, D.G., Bauman, P., Nimeck, G., 2002. Electrical resistivityground imaging (ERGI): a new tool for mapping the lithology and geometry ofchannel-belts and valley-fills. Sedimentology 49, 441–449.

Balooni, K., Kalro, A.H., Kamalamma, A.G., 2008. Community initiatives in building andmanaging temporary check-dams across seasonal streams for water harvesting inSouth India. Agric. Water Manage. 95 (12), 1314–1322.

Bao, Y.X., Wu, F.Q., Tan, H.C., 2005. Distribution characteristics of soil nutrients in DamLand. B. Soil Water Conserv. 25 (2), 12–15 (In Chinese with English Abstract).

Bellin, N., Vanacker, V., van Wesemael, B., Solé-Benet, A., Bakker, M., 2011. Natural andanthropogenic controls on soil erosion in the Internal Betic Cordillera (southeastSpain). Catena 87, 190–200.

Besson, A., Cousin, I., Samouëlian, A., Boizard, H., Richard, G., 2004. Structural het-erogeneity of the soil tilled layer as characterized by 2D electrical resistivity sur-veying. Soil Tillage Res. 79, 239–249.

Bombino, G., Tamburino, V., Zimbone, S.M., 2006. Assessment of the effects of check-dams on riparian vegetation in the mediterranean environment: A methodologicalapproach and example application. Ecol. Eng. 27, 134–144.

Bourennane, H., King, D., Le Parco, R., Isambert, M., Tabbagh, A., 1998. Three-dimen-sional analysis of soils and surface materials by electrical resistivity survey. Euro. JEnviron. Eng. Geophys. 3, 5–23.

Brown, C.B., Jarvis C.S., 1943. Discussion of Sedimentation in reservoirs, by Witzig B.J.,Proceedings of the American Society of Civil Engineers, pp. 1493–1500.

Castillo, V., Mosch, W., García, C.C., Barberá, G., Cano, J.N., López-Bermúdez, F., 2007.Effectiveness and geomorphological impacts of check dams for soil erosion control ina semiarid Mediterranean catchment: El Cárcavo (Murcia, Spain). Catena 70,416–427.

Cerdan, O., Govers, G., Le Bissonnais, Y., Van Oost, K., Poesen, J., Saby, N., Gobin, A.,Vacca, A., Quinton, J., Auerswald, K., 2010. Rates and spatial variations of soilerosion in Europe: a study based on erosion plot data. Geomorphology 122, 167–177.

Chambers, J., Wilkinson, P., Wardrop, D., Hameed, A., Hill, I., Jeffrey, C., Loke, M.,Meldrum, P., Kuras, O., Cave, M., 2012. Bedrock detection beneath river terracedeposits using three-dimensional electrical resistivity tomography. Geomorphology177, 17–25.

Chen, F.X., Fang, N.F., Shi, Z.H., 2016. Using biomarkers as fingerprint properties toidentify sediment sources in a small catchment. Sci. Total Environ. 557–558,123–133.

Chen, F.X., Fang, N.F., Wang, Y.X., Tong, L.S., Shi, Z.H., 2017. Biomarkers in sedimentarysequences: indicators to track sediment sources over decadal timescales.Geomorphology 278, 1–11.

Chen, L., Wei, W., Fu, B., Lü, Y., 2007. Soil and water conservation on the Loess Plateau inChina: review and perspective. Prog. Phys. Geogr. 31, 389–403.

CMWR (Ministry of Water Resource of P.R. China), 2003. Programming for check dams inthe Loess Plateau (Technical Report). China Water Power Press, Beijing, pp. 47–48 (InChinese).

CMWR (Ministry of Water Resource of P.R. China). 2014. Bulletin of first national watercensus for soil and water conservation. pp 6–8. (In Chinese)

Doolittle, W.E., 2013. Traditional uses of check dams: global and historical introduction.In: Check Dams, Morphological Adjustments and Erosion Control in TorrentialStreams. Nova Science Publishers, Inc., pp. 1–10.

Fang, N.F., Chen, F.X., Zhang, H.Y., Wang, Y.X., Shi, Z.H., 2016. Effects of cultivation andreforestation on suspended sediment concentrations: a case study in a mountainouscatchment in China. Hydrol. Earth Syst. Sc. 20, 13–25.

Fang, N.F., Wang, L., Shi, Z.H., 2017. Runoff and soil erosion of field plots in a subtropicalmountainous region of China. J. Hydrol. 552, 387–395.

Frankl, A., Nyssen, J., De Dapper, M., Haile, M., Billi, P., Munro, R.N., Deckers, J., Poesen,J., 2011. Linking long-term gully and river channel dynamics to environmentalchange using repeat photography (Northern Ethiopia). Geomorphology 129 (3–4),238–251.

Fu, B.J., Liu, Y., Lü, Y.H., He, C.S., Zeng, Y., Wu, B.F., 2011. Assessing the soil erosioncontrol service of ecosystems change in the Loess Plateau of China. Ecol. Complexity8, 284–293.

Fukue, M., Minato, T., Horibe, H., Taya, N., 1999. The micro-structures of clay given byresistivity measurements. Eng. Geol. 54, 43–53.

Garcia, C., Mario, M., 2010. Check Dams, Morphological Adjustments and ErosionControl in Torrential Streams. Nova Science Publishers, New York, pp. 8–15 ISBN978-1-61761-749-2.

Hassanli, A.M., Ebrahimizadeh, M.A., Beecham, S., 2009. The effects of irrigationmethods with effluent and irrigation scheduling on water use efficiency and cornyields in an arid region. Agric. Water Manage. 96 (1), 93–99.

Hauck, C., Vonder Mühll, D., Maurer, H., 2003. Using DC resistivity tomography to detectand characterize mountain permafrost. Geophys. Prospect. 51, 273–284.

Hsu, H.L., Yanites, B.J., Chen, C.C., Chen, Y.G., 2010. Bedrock detection using 2D elec-trical resistivity imaging along the Peikang River, central Taiwan. Geomorphology114, 406–414.

Leigh, D.S., Kowalewski, S.A., Holdridge, G., 2013. 3400 years of agricultural engineeringin Mesoamerica: lama-bordos of the Mixteca Alta, Oaxaca, Mexico. J. Archaeol. Sci.40, 4107–4111.

Li, X.G., Li, Z.B., Wei, X., 2007. Two key physical characteristics indexes of farmlandsediment for check dams in Loess Plateau. Res. Soil Water Conserv. 14, 218–220 (inChinese with English Abstract).

Li, X., Wei, X., Wei, N., 2016. Correlating check dam sedimentation and rainstormcharacteristics on the Loess Plateau, China. Geomorphology 265, 84–97.

Liu, C., Li, Z., Chang, X., He, J., Nie, X., Liu, L., Xiao, H., Wang, D., Peng, H., Zeng, G.,2018. Soil carbon and nitrogen sources and redistribution as affected by erosion anddeposition processes: a case study in a loess hilly-gully catchment, China. Agric.Ecosyst. Environ. 253, 11–22.

Liu, B., Nearing, M.A., Risse, L.M., 1994. Slope gradient effects on soil loss for steepslopes. Trans. ASAE 37, 1835–1840.

Loke, M.H., Acworth, I., Dahlin, T., 2003. A comparison of smooth and blocky inversionmethods in 2D electrical imaging surveys. Explor. Geophys. 34, 182–187.

Loke, M., Barker, R., 1995. Least-squares deconvolution of apparent resistivity pseudo-sections. Geophysics 60, 1682–1690.

Lü, Y., Sun, R., Fu, B., Wang, Y., 2012. Carbon retention by check dams: regional scaleestimation. Ecol. Eng. 44, 139–146.

Norman, L.M., Brinkerhoff, F., Gwilliam, E., Guertin, D.P., Callegary, J., Goodrich, D.C.,Nagler, P.L., Gray, F., 2016. Hydrologic response of streams restored with check damsin the Chiricahua Mountains, Arizona. River Res. Appl. 32 (4), 519–527.

Nyssen, J., Clymans, W., Poesen, J., Vandecasteele, I., De Baets, S., Haregeweyn, N.,Naudts, J., Hadera, A., Moeyersons, J., Haile, M., Deckers, J., 2009. How soil con-servation affects the catchment sediment budget –a comprehensive study in the northEthiopian highlands. Earth Surf Proc Land 34 (9) 1216-1233.3.

Ramos‐Diez, I., Navarro‐Hevia, J., Fernández San Martín, R., Mongil‐Manso, J., 2017a.Final analysis of the accuracy and precision of methods to calculate the sedimentretained by check dams. Land Degrad. Dev. 28, 2446–2456.

Ramos-Diez, I., Navarro-Hevia, J., Fernández San Martín, Díaz-Gutiérrez, R.V., Mongil-Manso, J., 2016. Analysis of methods to determine the sediment retained by checkdams and to estimate erosion rates in badlands. Environ. Monit. Assess. 188, 405.

Ramos-Diez, I., Navarro-Hevia, J., Fernández, R.S.M., Díaz-Gutiérrez, V., Mongil-Manso,J., 2017b. Evaluating methods to quantify sediment volumes trapped behind checkdams, Saldana badlands (Spain). Int. J. Sediment Res. 32, 1–11.

Ran, L., Lu, X., Xin, Z., 2014. Erosion-induced massive organic carbon burial and carbonemission in the Yellow River basin, China. Biogeosciences 11, 945–959.

Ritz, M., Parisot, J.C., Diouf, S., Beauvais, A., Dione, F., Niang, M., 1999. Electricalimaging of lateritic weathering mantles over granitic and metamorphic basement ofeastern Senegal, West Africa. J. Appl. Geophys. 41, 335–344.

Romero-Díaz, A., Alonso-Sarriá, F., Martínez-Lloris, M., 2007. Erosion rates obtained fromcheck-dam sedimentation (SE Spain). A multi-method comparison. Catena 71,172–178.

Romero-Díaz, A., Marín-Sanleandro, P., Ortiz-Silla, R., 2012. Loss of soil fertility esti-mated from sediment trapped in check dams. South-eastern Spain. Catena 99, 42–53.

Sun, W.Y., Guo, S.L., 2011. The spatial distribution of soil organic carbon and it’s influ-encing factors in hilly region of the Loess Plateau. Acta Ecol. Sin. 31, 1604–1616 (inChinese with English Abstract).

Sun, Q.L., Wang, H.X., Ma, J.Q., 2004. Estimating soil erosion status on hill slopes basedon the survey of sediment deposition behind check dams. Soil Water Conserv. Sci.Techn. Shanxi 30 (1), 28–30 (in Chinese with English Abstract).

Vaezi, A.R., Abbasi, M., Keesstra, S., Cerdà, A., 2017. Assessment of soil particle erod-ibility and sediment trapping using check dams in small semi-arid catchments. Catena157, 227–240.

Verstraeten, G., Poesen, J., 2000. Estimating trap efficiency of small reservoirs and ponds:methods and implications for the assessment of sediment yield. Prog. Phys. Geogr. 24,219–251.

Verstraeten, G., Poesen, J., 2002. Using sediment deposits in small ponds to quantifysediment yield from small catchments: possibilities and limitations. Earth Surf. Proc.Land. 27, 1425–1439.

Wang, Y.F., Fu, B.J., Chen, L.D., Lü, Y.H., Gao, Y., 2011. Check dam in the Loess Plateauof China: engineering for environmental services and food security. Environ. Sci.Technol. 45, 10298–10299.

Wang, Y.F., Chen, L.D., Gao, Y., Wang, S., Lü, Y.H., Fu, B.J., 2014. Carbon sequestrationfunction of check-dams: a case study of the Loess Plateau in China. Ambio 43,926–931.

Wang, Y.X., Fang, N.F., Tong, L.S., Shi, Z.H., 2017. Source identification and budgetevaluation of eroded organic carbon in an intensive agricultural catchment. Agric.Ecosyst. Environ. 247, 290–297.

Wang, S., Fu, B., Piao, S., Lü, Y., Ciais, P., Feng, X., Wang, Y., 2016. Reduced sedimenttransport in the Yellow River due to anthropogenic changes. Nat. Geosci. 9, 38.

Wang, X., Guo, S., Ma, Y., Huang, D., Wu, J., 2007b. Effects of land use type on soilorganic C and total N in a small watershed in loess hilly-gully region. J. Appl. Ecol.18, 1281–1285 (In Chiense).

Wang, L., Shi, Z., 2015. Size selectivity of eroded sediment associated with soil texture onsteep slopes. Soil Sci. Soc. Am. J. 79 (3), 917–929.

Wang, H., Yang, Z., Saito, Y., Liu, J.P., Sun, X., Wang, Y., 2007a. Stepwise decreases of theHuanghe (Yellow River) sediment load (1950–2005): impacts of climate change andhuman activities. Global Planet. Change 57, 331–354.

Wang, Y.Q., Zhang, X.C., Han, F.P., 2008. Profile variability of soil properties in checkdam on the Loess Plateau and its functions. Environ. Sci. 29, 1020–1026 (in Chinesewith English Abstract).

Wei, Y., He, Z., Li, Y., Jiao, J., Zhao, G., Mu, X., 2017. Sediment yield deduction fromcheck–dams deposition in the weathered sandstone watershed on the North LoessPlateau, China. Land Degrad. Dev. 28, 217–231.

Wei, X., Li, Z.B., Li, X.G., Lu, K.X., 2006. Distribution law of deposits’ dry bulk density andits application in sediment restoration of check-dam. J. Northwest Sci. Tech. Univ.Agri. Forestry (Nat. Sci. Ed.) 34 (10), 192–196 (in Chinese with English Abstract).

Zhao, G., Kondolf, G.M., Mu, X., Han, M., He, Z., Rubin, Z., Wang, F., Gao, P., Sun, W.,2017a. Sediment yield reduction associated with land use changes and check dams ina catchment of the Loess Plateau, China. Catena 148, 126–137.

Zhao, T.Y., Yang, M.Y., Walling, D.E., Zhang, F.Y., Zhang, J.Y., 2017b. Using check damdeposits to investigate recent changes in sediment yield in the Loess Plateau, China.Global Planet. Change 152, 88–98.

N.F. Fang et al. Journal of Hydrology 568 (2019) 1007–1016

1016