Influence of Check Dams on Flood and Erosion Dynamic Processes Of

water

Article Inﬂuence of Check Dams on Flood and Erosion Dynamic Processes of a Small Watershed in the Loss Plateau

Shuilong Yuan 1 , Zhanbin Li 1,2, Peng Li 1, Guoce Xu 1,*, Haidong Gao 1, Lie Xiao 1, Feichao Wang 1 and Tian Wang 1 1 State Key Laboratory of Eco-Hydraulics in Northwest Arid Region of China, Xi’an University of Technology, Xi’an 710048, China; [email protected] (S.Y.); [email protected] (Z.L.); [email protected] (P.L.); [email protected] (H.G.); [email protected] (L.X.); [email protected] (F.W.); [email protected] (T.W.) 2 State Key Laboratory of Soil Erosion and Dry-Land Farming on the Loess Plateau, Institute of Soil and Water Conservation, Chinese Academy of science and Ministry of Water Resources, Yangling 712100, China * Correspondence: [email protected]; Tel./Fax: +86-29-8231-2651

Received: 21 March 2019; Accepted: 15 April 2019; Published: 19 April 2019

Abstract: As an important soil and water conservation engineering measure, check dams have been constructed on a large scale in the Loess Plateau of China. However, their effects on runoff and sediment processes in the basin are still unclear. In this study, the hydrodynamic processes of the Wangmaogou watershed located in the Loess Plateau were simulated, and the influence of check dams on the flood and erosion dynamic processes in this watershed were also evaluated. The results showed that the check dams obviously reduced the flood peak and flood volume and mitigated the flood process. After the dam system was completed, the flood peak and flood volume were reduced by 65.34% and 58.67%, respectively. The erosion dynamic distribution of the main channel in the small watershed was changed to different extents by the different dam type combinations, and the erosion dynamic parameters of the channel decreased most after the dam system was completed, when the velocity and runoff shear stress of the outlet section were reduced by 10.69% and 31.08%, respectively. Additionally, the benefits of sediment reduction were most obvious after the check dam system was completed, with the sediment discharge in the watershed being reduced by 83.92%. The results of this study would provide specific implications for construction and management of check dams in the Loess plateau.

Keywords: check dam; ﬂood; erosion dynamics; sediment reduction; hydrodynamic model

1. Introduction The Chinese Loess Plateau is an area of about 650 thousand km2 located in the middle and upper reaches of the Yellow River [1]. This region is well known for its severe soil erosion and high sediment yields from its most erosion-prone area of approximately 472 thousand km2, including 91.2 thousand km2 with sediment yields exceeding 8000 t/km2/a [2]. This region has the most serious documented soil erosion rates in the world [3]. Soil erosion on the Loess Plateau has depleted land resources, degraded the environment, and caused severe river bed sedimentation along the lower reaches of the Yellow River, threatening the safety of local inhabitants [4]. Therefore, soil loss management and environmental protection are critical issues in the region [5]. Check dams are the most widely applied engineering structures for soil and water conservation in the erodible regions of the Loess Plateau, and they have been widely constructed to retain ﬂoodwater, intercept soil sediment, improve gully slope stabilities, and increase farmland [6–9]. Indeed, about 110,000 check dams were

Water 2019, 11, 834; doi:10.3390/w11040834 www.mdpi.com/journal/water Water 2019, 11, 834 2 of 16 constructed by 2002, and this number is set to double by 2020 as part of a project launched by the Chinese Ministry of Water Resources [10]. Check dams are also often used in other countries, such as Ethiopia [11], Spain [12,13], Iran [14], Italy [15], and Mexico [16]. The performance of check dams for sediment and runoff control has been investigated in different studies, some of which are reviewed in the following. Check dams are constructed across gullies to reduce the velocity of concentrated water flows, reduce erosion, control sediment, and stabilize gullies [17,18]. The annual runoff was reduced by less than 14.3% by check dams in the Yanhe watershed [19]. Check dams also reduced the sediment load in the Rogativa catchment of Spain by approximately 77% [20]. A series of check dams have been built along waterways, and these have provided dense vegetation, stabilized the channels and considerably decreased the volume of sediments washed from the watershed [21]. Check dams are also found to modify water and sediment transport by impounding storm flow, reducing its velocity and peak rate, decreasing channel slope, and allowing more time for infiltration and sediment settling [22]. The construction of check dams is one of the most effective methods to control gully erosion and reduce sediment transport through field investigations [23]. The effects of check dams on scour dynamics vary greatly depending on initial slope, soil texture, spacing, drop height, and flow depth [24]. The effectiveness of sediment retention by check dams can be associated with different factors such as check dam characteristics, gully characteristics, and water flow conditions [25]. Accordingly, various studies have evaluated the impacts of check dams on controlling runoff, erosion, and sediment transfer [26–28]. Most studies conducted to date have focused on the effects of check dams on runoff and sediment, while few have explored the effects of check dams on floods and their erosion dynamic processes in small watersheds [17–20]. Different types of check dams have different effects on runoff and its erosion dynamic process; accordingly, it is important to identify the roles of different types of check dams. The main goals of the present study were to reveal the mechanisms by which check dams influence floods and their erosion dynamic process, and evaluate the sediment reduction benefits of different dam type combinations. Our specific objectives were to: (1) investigate the influence of different types of check dams and their combinations on the flood process of small watersheds; (2) explore the influence of different types of check dams and their combinations on the erosion dynamic processes in a small watershed; and (3) calculate the sediment reduction benefit of different types of check dam combinations.

2. Data and Methods

2.1. Study Area

The study was conducted in the Wangmaogou watershed (110◦2002600–110◦2204600E, 37◦3401300–37◦3600300N), which is located in the Loess Plateau of China. The Wangmaogou watershed is located in Suide County, Shaanxi Province, which belongs to the third-grade branch ditch of the Wuding River. The watershed covers an area of 5.97 km2 and ranges in altitude from 936 to 1188 m. The main channel length is 3.75 km, and the average slope of the channel is 2.7% [29]. The watershed is characterized by a temperate semiarid continental monsoon climate with distinct seasons and large temperature diﬀerences. According to data from the soil and water conservation test station in Suide, the average annual temperature is 10.2 ◦C and the mean annual precipitation is approximately 513 mm, more than 60% of which occurs between July and September [30]. The study area is underlaid by loessial soil and rainfall during the ﬂood season can result in serious soil erosion. As a well-studied watershed for soil and water conservation, many check dams have been implemented in the Wangmaogou watershed since the 1950s. Indeed, by the end of 2017 there were 23 normal check dams in Wangmaogou watershed, including two key dams, seven medium dams, and 14 small dams (Figure1). The check dams were divided into three types: key dams, medium dams, and small dams. Key dams are composed of the dam body, spillway, and discharge structure; while middle dams Water 2019, 11, 834 3 of 16

are composedWater 2019, 11,of x FOR a dam PEER body REVIEW and a discharge structure; and small check dams consists only3 of 16 one dam body.

Figure 1. Location of the study area (a,b); the Wangmaogou watershed (c); and the layout of check Figure 1. Location of the study area (a,b); the Wangmaogou watershed (c); and the layout of check dam types (d). dam types (d). 2.2. Data Sources 2.2. Data Sources Runoff, sediment, and meteorological data were obtained from the Yellow River Water and soil ConservationRunoff, Suide sediment, Supervision and meteorological Bureau. The data Suide were Supervision obtained from Bureau the established Yellow River a watershedWater and outletsoil Conservation Suide Supervision Bureau. The Suide Supervision Bureau established a watershed hydrologic station in the Wangmaogou watershed. The station monitor runoff and sediment data from outlet hydrologic station in the Wangmaogou watershed. The station monitor runoff and sediment 1962 to 1966, mainly including water level, discharge, sediment concentration, and sediment transport data from 1962 to 1966, mainly including water level, discharge, sediment concentration, and rate. The terrain data is obtained from the State Bureau of Surveying and Mapping. The geodetic sediment transport rate. The terrain data is obtained from the State Bureau of Surveying and reference is the 1980 Xi’an coordinate system, contour distance of which is 5 m. The shape file Mapping. The geodetic reference is the 1980 Xi'an coordinate system, contour distance of which is 5 is generatedm. The shape by paper file is mapsgenerated via spatialby paper registration maps via spatial after scanning,registration splicing, after scanning, and manual splicing, tracking. and Themanual DEM with tracking. a resolution The DEM of 5with m is a thenresolution generated of 5 m by is thethen Hutchinson generated by method the Hutchinson [31]. The method land use [31]. data for theThe watershedland use data was for obtained the watershed from was the landobtained use surveyfrom the of land the use Suide survey Supervision of the Suide Bureau Supervision in 1960s. TheBureau channel in section 1960s. dataThe utilizedchannel forsection the Mike data 11utilized model for (DHI, the Copenhagen,Mike 11 model Denmark) (DHI, Copenhagen, was extracted fromDenmark) the digital was elevation extracted model from (DEM)the digital with elevation 5 m resolution model (DEM) using with the 3D 5 m analyst resolution tool using in ArcGIS the 3D 10.1 (ESRI,analyst California, tool in America),ArcGIS 10.1 then (ESRI, corrected California, based Am onerica), field then measurements. corrected based The on geometric field measurements. characteristics of checkThe geometric dams were characteristics obtained by of field check measurement, dams were obtained including by dam field length, measurement, dam height, including dam dam width, andlength, geometric dam size height, of the dam drainage width, and structure. geometric size of the drainage structure.

2.3.2.3. Model Model Set Set Up Up In thisIn this study, study, the the channel channel hydrodynamic hydrodynamic processesprocesses of the Wangmaogou watershed watershed were were simulatedsimulated using using the the overland overland ﬂow flow modulemodule of of the the distributed distributed hydrological hydrological model model MIKE MIKE SHE and SHE the and the one-dimensional hydrodynamic hydrodynamic model model MIKE MIKE 11. Firs 11.t, First, the watershed the watershed scope was scope defined was deﬁned in the MIKE in the × MIKESHE SHE model. model. The study The study area was area discretized was discretized using a usinggrid of a20 grid 20 of m, 20 with20 a total m, with of 29,700 a total grids. of29,700 The study area DEM, with a 5 m resolution grid, was converted into a shape× point terrain file by ArcGIS

Water 2019, 11, 834 4 of 16 grids. The study area DEM, with a 5 m resolution grid, was converted into a shape point terrain file by ArcGIS 10.1. The processed vector format digital elevation data file was directly imported into the model and interpolated by the triangle interpolation method, which sets the search radius to 20 m depending on the cell size. During the secondary torrential rain and flood, the evapotranspiration of the watershed is very small, so it can be ignored [32]. The precipitation data were transformed into a dfs0 file and imported into the model using the MIKE model software tools. The grid of the model is relatively large, so the land use types were divided into six categories according to the geomorphological characteristics of the study area: cultivated land, grassland, forest land, garden land, residential land, and traffic land. The overland flow module of the MIKE SHE model mainly contains three parameters: Manning number, detention storage, and initial water depth. The model deducts soil infiltration from precipitation data as precipitation input of the model. The amount of soil infiltration, which has an empirical value of 2 to 20 mm/h, mainly depends on the soil types, and the effects of vegetation are taken into account [33]. MIKE SHE and MIKE 11 were used to simulate overland flow and channel confluence processes, respectively. The river network vector file of the watershed was extracted from DEM with 5 m resolution using the ArcGIS 10.1 software and imported into the MIKE 11 model. The river network, which was automatically generated by the model, was manually modified according to the actual situation. The six river channels of the watershed are defined in the MIKE 11 model, and the check dams in the channel are expressed by the hydraulic structures in the river network file. The MIKE 11 model was coupled with the MIKE SHE model, and the storm flood model of the Wangmaogou watershed was established.

2.4. Calibration and Validation of the Model The model of torrential rain and flooding in Wangmaogou watershed was calibrated by trial and error. The runoff process measured at the outlet of the watershed was selected as the calibration parameter. The model was calibrated based on two typical rainstorm flood processes, and was verified by another two rainstorm flood processes during the observation period. Theoretically, the parameters of the distributed watershed hydrological model can be obtained by experimental measurement, but the cell parameters of the model still need to be calibrated because of the difference between observation scale and simulation scale, the errors associated with experimental measurements, etc. [34,35]. At the same time, the distributed hydrological model should select as few parameters as possible when calibrating the model, because too many parameters may not improve the accuracy of the model, and the parameter calibration process will become extremely complex [36,37]. Therefore, fewer parameters were selected to calibrate the model in this study. Specifically, the parameters that need to be calibrated are detention storage, Manning number, leakage coefficient, and soil infiltration rate of different soil types. The performance of stream flow simulations was evaluated by comparing simulated and observed stream flow hydrographs through the following statistics: coefficient of determination (R2) (Equation (1)) and Nash–Sutcliffe efficiency (NSE) (Equation (2)) [38,39].

 2  n   P   (Oi O)(Si S)   i=1 − −  R2 =    s s  (1)  n n   P 2 P 2   (Oi O) (Si S)   i=1 − i=1 − 

n P 2 (Si Oi) i=1 − NSE = 1 ( , 1] (2) − Pn 2 ∈ −∞ (Oi O) i=1 − Water 2019, 11, x FOR PEER REVIEW 5 of 16 Water 2019, 11, 834 5 of 16

3 −1 3 −1 where Oi is the simulated stream flow (m s ), Si is the observed stream flow (m s ), O is the mean 3 1 3 1 where Oi is the simulated stream flow (m s− ), Si is the observed stream flow (m s− ),O is the of the simulated stream flow (m3 s−13), and1 S is the mean of the observed stream flow (m33 s−1). mean of the simulated stream flow (m s− ), and S is the mean of the observed stream flow (m s− ). Additionally, NSE indicates how well the plot of the observed value versus the simulated value fits Additionally, NSE indicates how well the plot of the observed value versus the simulated value fits the the 1:1 line and ranges from 0 to 1, with higher values indicating better agreement. 1:1 line and ranges from 0 to 1, with higher values indicating better agreement.

2.5. Model EEffectffect Evaluation The erosionerosion rainfallrainfall ofof smallsmall watershedswatersheds isis mainlymainly characterizedcharacterized by short durationduration and highhigh intensityintensity inin thethe loessloess plateau.plateau. The four rainfall eventsevents selected in this study were thethe rainrain patternspatterns withwith highhigh frequencyfrequency occurringoccurring in the WangmaogouWangmaogou watershedwatershed from 1962 toto 1966,1966, whichwhich are goodgood representations.representations. WeWe compared compared the the observed observed and and simulated simulated flood flood processes processes during during the calibrationthe calibration and validationand validation periods periods (Figure (Figure2). The 2). results The showedresults sh thatowed there that is good there agreement is good agreement between the between simulated the runosimulatedff and runoff the observed and the runoobservedff in therunoff calibration in the calibration and validation and validation periods. Thisperiods. model This can model be used can tobe simulateused to simulate the dynamic the dynamic changes inchanges the flood in the process, flood whichprocess, can which be used can forbe analysisused for ofanalysis working of conditions.working conditions. The NSE valuesThe NSE of thevalues model of the were model all higher were than all higher 0.8, the thanR2 were 0.8, the 0.90 R and2 were 0.88, 0.90 respectively, and 0.88, andrespectively, the relative and error the relative of peak error flow wasof peak only flow 1.72% was and only 3.33% 1.72% in and the calibration3.33% in the period. calibration The NSEperiod.of theThe modelNSE of was the 0.60 model and was 0.71, 0.60 respectively, and 0.71, respectively, the R2 values the were R2 allvalues 0.72, were and theall relative0.72, and error the ofrelative peak flowerror wasof peak only flow 11.05% was and only 12.05% 11.05% in theand validation 12.05% in period the validation (Table1). period Simplification (Table 1). of Simplification model structure of canmodel cause structure model can error cause [40]. model The accuracy error [40]. of theThe model accuracy in the of validationthe model periodin the validation was lower period than in was the calibrationlower than period,in the calibration which may period, have which been because may have the been model because only couplesthe model the only MIKE couples SHE the overland MIKE flowSHE moduleoverland and flow MIKE module 11. Generalizationand MIKE 11. Generalizatio of the rainfalln infiltration of the rainfall process infiltration is one of process the reasons is one the of accuracythe reasons of the accuracy model in of the the validation model in periodthe validatio is lowern period than that is lower during than the that calibration during the period. calibration period.

Figure 2. Model calibration results (aa,b,b); model validationvalidation resultsresults ((cc,d,d).).

Table 1. Model calibration and validation results of the Wangmaogou watershed.

Water 2019, 11, 834 6 of 16

Table 1. Model calibration and validation results of the Wangmaogou watershed.

Observed Simulated Relative Flood Determination Nash–Sutcliffe Stage Value Value Error/Re 2 Number 3 1 3 1 Coefficient/R Coefficient /NSE (m s− ) (m s− ) (%) 196303 0.58 0.59 1.72 0.90 0.80 calibration 196404 0.90 0.87 3.33 0.88 0.85 196201 1.90 1.69 11.05 0.72 0.60 validation 196304 0.83 0.73 12.05 0.72 0.71

2.6. Working Condition Design In this study, the rainstorm flood process was simulated under different combinations of check dam types in a small watershed, and a total of eight different dam type combinations were designed (Table2). N represents the case without dams; K represents the case with key dams only; M represents the case with medium dams only; S represents the case with small dams only; KM represents the case with key and medium dams; KS represents the case with key and small dams; KMZ represents the case with key, medium, and small dams.

Table 2. Working condition design of diﬀerent dam type combinations.

Diﬀerent Dam Type Working Coding Combinations in Dam Name Conditions Watershed 1 N No dams — — 2 K Only key dams Wangmaogou dam #1, Wangmaogou dam #2 Huangbaigou dam #2, Nianyangou dam #1, Kanghegou dam #2, 3 M Only medium dams Sidizui dam #1, Madizui dam, Guandigou dam #1, Guandigou dam #4 Huangbaigou dam #1, Nianyangou dam #2, Nianyangou dam #3, Nianyangou dam #4, 4 S Only small dams Kanghegou dam #1, Kanghegou dam #3, Sidizui dam #2, Wangtagou dam #1, Wangtagou dam #2, Guandigou dam #2 Key and medium dam 5 KM —— combination Key and small dam 6 KS —— combination Medium and small dam 7 MS —— combination Key, medium, and small 8 KMS —— dam combination

2.7. Erosion Dynamic Parameter Calculation (1) The construction of check dams has changed the connectivity of the channel. In this study, the formula used to calculate channel connectivity was constructed by referring to the formula for calculation of river connectivity [41], which was used to characterize the inﬂuence of the check dam system on channel connectivity. The formula is as follows:

n X S R = i (3) S(n + 1) i=1 i where R is the channel connectivity index, S is the total length of the channel(m), Si is the length of the branch ditch i(m), and ni is the number of check dams on the branch ditch i. The value range of R is 0 to 1, with a value closer to 1 indicating increased channel connectivity. Water 2019, 11, 834 7 of 16

(2) Flow velocity (V). Velocity is an important hydraulic parameter in runoﬀ erosion dynamics and sediment transport. Because of the uncertainty of the measured velocity, the discharge of the section is obtained by solving the Saint–Venant equation, and then the average velocity of the section is obtained by dividing the measured area of the section as follows:

V = Q/A (4)

1 3 1 where V is the average velocity of the section (m s− ), Q is the discharge of the section (m .s− ), and A represents the measured sectional area (m2). (3) Runoff shear stress (τ). Runoff shear stress is a parameter reflecting the amount of soil erosion force exerted by runoff during flow. Foster et al. [42] proposed the formula of runoff shear stress as follows: τ = γRJ (5)

2 3 where τ is the runoff shear stress (N m− ), γ is the bulk density of the runoff (N m− ), R is hydraulic radius (m), and J is the hydraulic energy slope. (4) Runoff erosion power (E). Runoff depth and flood peak flow are two important parameters reflecting the characteristics of storm flood process in the watershed. The product of runoff depth and flood peak flow indirectly reflects the spatial and temporal distribution characteristics of rainfall and the influence of the watershed underlying the surface on runoff confluence processes. The formula for calculating runoff erosion power is defined as [43]:

E = Qm0 H (6)

4 1 2 where E is runoff erosion power (m s− km− ), H is the average runoff depth of the rainstorm(m), 3 1 2 and Qm0 is the peak flow modulus (m s− km− ).

3. Results

3.1. Variation Characteristics of Rainstorm and Flood Process in Small Watershed under Different Working Conditions The flood process of the small watershed changed under the eight working conditions of different dam combinations. Before the channel dam construction, the flood process of the small watershed rose and fell sharply, but this slowed down after dam construction. Comparing different working conditions of the flood process hydrographs of no dams (N), only key dams (K), only medium dams (M), and only small dams (S) revealed that all three kinds of dams reduced the peak discharge and flood volume, but different dam types had different effects on flood regulation (Figure3). To quantitatively explain the influence of different dam type combinations on rainstorm flood processes in the small watershed, the flood characteristic parameters under different working conditions were analyzed. As shown in Table3, construction of the dam system (KMS) will reduce the flood peak by 65.34% and the flood volume by 58.67%, resulting in the strongest flood control ability. When there are only key dams in the watershed, the flood peak and the flood volume will be reduced by 27.28% and 2.18%, resulting in a reduction in the flood peak that is far greater than that of the flood volume, indicating that the role of the key dam is mainly to regulate the flood process. While construction of only medium or small dams in the watershed will reduce the flood peak by 33.39% and 40.13% and the flood volume by 27.08% and 44.89%, respectively, resulting in a reduction in flood peak that is roughly the same as that of the flood volume. This indicates that medium and small dams not only regulate flood processes, but also reduce flood volume. Water 2019, 11, 834 8 of 16 Water 2019, 11, x FOR PEER REVIEW 8 of 16

Figure 3. Simulation results under diﬀerent working conditions. Figure 3. Simulation results under different working conditions. Table 3. Flood characteristic parameters under diﬀerent working conditions. Table 3. Flood characteristic parameters under different working conditions. Working Flood Peak Flood Peak Flood Volume Flood Volume 3 1 3 WorkingConditions FloodDischarge peak discharge (m s− ) ReductionFlood peak (%) Flood(m volume) ReductionFlood volume (%) conditionsN (m 1.263 s−1) reduction - (%) 4853.93(m3) reduction - (%) K 0.92 27.28 4828.87 2.18 N M 1.26 0.84 33.39 - 3556.61 4853.93 27.08 - S 0.76 40.13 2541.04 44.89 KKM 0.92 0.78 38.07 27.28 3532.24 4828.87 27.37 2.18 MKS 0.84 0.51 59.71 33.39 2518.06 3556.61 45.15 27.08 MS 0.50 60.75 1802.36 58.42 SKMS 0.76 0.44 65.34 40.13 1779.58 2541.04 58.67 44.89

KM 0.78 38.07 3532.24 27.37 3.2. Variation of Dynamic Parameters of Channel Erosion under Different Working Conditions KS 0.51 59.71 2518.06 45.15 In this study, two erosion dynamic parameters, velocity and runoff shear stress along the main channelMS section, were 0.50 selected for analysis. 60.75 There are four 1802.36 check dams in the main 58.42 channel, Wangmaogou dam #2, Guandigou dam #1, Guandigou dam #2, and Guandigou dam #4, which have KMS 0.44 65.34 1779.58 58.67 chainages of 2050, 1261, 733, and 529 m, respectively. Velocity is the most basic hydraulic parameter influencing runoff erosion and sediment transport. As shown in Figure4a, the velocity of the cross 3.2. Variation of Dynamic Parameters of Channel Erosion under Different Working Conditions section of the eight working conditions shows an increasing trend from upstream to the downstream portionIn this of the study, main two channel, erosion although dynamic some parameters, sections tended velocity to and decrease. runoff This shear is stress because along the potentialthe main energychannel is section, gradually were converted selected into for kinetic analysis. energy Th duringere are the four downward check movementdams in the of watermain alongchannel, the channel,Wangmaogou and the dam incoming #2, Guandigou water from dam the #1, branch Guandigou ditch dam continuously #2, and Guandigou flows into thedam main #4, which channel have so thatchainages the velocity of 2050, is 1261, generally 733, increasing.and 529 m, Therespectively reason for. Velocity the sharp is the decrease most basic between hydraulic some sectionsparameter is thatinfluencing the construction runoff erosion of check and dams sediment between transport. these sections As shown significantly in Figure reduced4a, the velocity the velocity of the behind cross thesection dam. of Comparison the eight working of the conditions eight working shows conditions, an increasing the velocity trend from along upstream each section to the of downstream the working conditionportion of N the reaches main channel, the maximum, although while some the sections working tended condition to decrease. KMS is the This minimum, is because and the that potential of the otherenergy working is gradually conditions converted were into between kinetic these energy two during conditions, the downward indicating thatmovement the velocity of water along along the mainthe channel, channel and was the reduced incoming and thewater erosion from of the the branch channel ditch by runocontinuouslyff was weakened flows into after the the main dam channel system wasso that completed. the velocity As is shown generally in Figure increasing.4b, the The runo reffasonshear for stress the sharp of the decrease eight working between conditions some sections first increased,is that the then construction decreased, of and check the runodamsff shearbetween stress these of the sections section significantly with the chainage reduced of 1600–3800the velocity m wasbehind greater the dam. than thatComparison of the section of the with eight the working chainage conditions, of 0–1600 the m. Comparisonvelocity along of each the eight section working of the conditions,working condition the runo Nff reachesshear stress the maximum, of each section while alongthe working working condition condition KMS N is was the largest, minimum, while and it wasthat smallestof the other along working working conditions condition were KMS, between and the these other two working conditions, conditions indicating were that between the velocity these twoalong conditions. the main channel These findings was reduced indicate and that the completion erosion of ofthe the channel dam system by runoff reduces was theweakened runoff shear after stressthe dam along system the channel.was completed. Overall, As the shown erosion in dynamicFigure 4b, distribution the runoff shear of the stress main of channel the eight in the working small watershedconditions first was increased, changed to then diff decreased,erent extents and by the di runofffferent shear dam stress type combinations,of the section with and the the chainage erosion of 1600–3800 m was greater than that of the section with the chainage of 0–1600 m. Comparison of the eight working conditions, the runoff shear stress of each section along working condition N was

Water 2019, 11, x FOR PEER REVIEW 9 of 16 largest, while it was smallest along working condition KMS, and the other working conditions were Water 2019, 11, 834 9 of 16 between these two conditions. These findings indicate that completion of the dam system reduces the runoff shear stress along the channel. Overall, the erosion dynamic distribution of the main channel indynamic the small parameters watershed of was the channelchanged decreased to different most extents after by the different dam system dam was type completed combinations, (KMS), and when the erosionthe maximum dynamic velocity parameters and runo ofﬀ theshear channel stress ofdecr theeased outlet most section after were the reduced dam system by 10.69% was andcompleted 31.08%, (KMS),respectively when (Table the maximum4). velocity and runoff shear stress of the outlet section were reduced by 10.69% and 31.08%, respectively (Table 4).

Figure 4.4. DistributionDistribution of of erosion erosion dynamic dynamic parameters parameters along along the channel the channel under different under workingdifferent conditions working. conditions. Table 4. Erosion dynamic parameters of the exit section of the watershed under different working conditions.

TableWorking 4. Erosion dynamic parameters1 of theVelocity exit section of theRuno watershedff Shear underRuno differentff Shear working Stress Velocity (m s− ) 2 conditions.Conditions Reduction (%) Stress (N S− ) Reduction (%) N 1.99 — 15.51 — WorkingK Velocity (m 1.84 Velocity 7.64Runoff shear 13.71 stress Runoff 11.61 shear stress M 1.80 9.70 13.45 13.31 conditions s−1) reduction (%) (N S−2) reduction (%) S 1.78 10.65 13.31 14.20 N KM 1.99 1.79 — 10.30 15.51 13.36 13.89 — KS 1.71 14.22 12.54 19.17 K MS 1.84 1.68 7.64 15.58 13.71 12.27 20.86 11.61 KMS 1.55 22.26 10.69 31.08 M 1.80 9.70 13.45 13.31 S 1.78 10.65 13.31 14.20 3.3. Variation of Runoff Erosion Power along the Channel under Different Working Conditions KM 1.79 10.30 13.36 13.89 Runoff erosion power, as an important index of erosion and sediment yield of floods associated KS 1.71 14.22 12.54 19.17 with a single storm, reflects the erosion dynamics in a watershed controlled by different sections. The runoMSff erosion power 1.68 of different sections 15.58 of the main channel 12.27 of the basin was calculated 20.86 using FormulaKMS (6). As shown 1.55 in Figure5, the 22.26 runo ff erosion power 10.69 of each section along 31.08 the working condition N was the largest, while that along the working condition KMS was smallest, and the other 3.3.working Variation conditions of Runoff fell Erosion between Power these along two the conditions,Channel under indicating Different thatWorking the reductionConditions extent of the runoff erosion power of the watershed was highest after the dam system was completed. Because the Runoff erosion power, as an important index of erosion and sediment yield of floods associated construction of check dams has no effect on the erosion dynamic process in slope, it mainly reduces the with a single storm, reflects the erosion dynamics in a watershed controlled by different sections. The erosion power of the channel. Under eight working conditions, the runoff erosion power fluctuated runoff erosion power of different sections of the main channel of the basin was calculated using greatly in the upper and middle reaches, while it remained basically the same in the lower reaches Formula 6. As shown in Figure 5, the runoff erosion power of each section along the working (less than 2800 m in mileage), and the erosion power of the lower reaches was less than that of the condition N was the largest, while that along the working condition KMS was smallest, and the other upper reaches, demonstrating that the upstream erosion is stronger and the downstream erosion is working conditions fell between these two conditions, indicating that the reduction extent of the weaker, with a basically stable state. According to the Bagnold sediment transport rate, larger velocity runoff erosion power of the watershed was highest after the dam system was completed. Because the and shear stress were associated with a greater sediment transporting capacity of a channel [44]. The construction of check dams has no effect on the erosion dynamic process in slope, it mainly reduces velocity and runoff shear stress in the lower reaches of the Wangmaogou were higher than those in the erosion power of the channel. Under eight working conditions, the runoff erosion power the upper reaches, which explains why the sediment transporting capacity of the downstream of the fluctuated greatly in the upper and middle reaches, while it remained basically the same in the lower main channel is greater than that of the upstream. Overall, the characteristics of sediment yield and reaches (less than 2800 m in mileage), and the erosion power of the lower reaches was less than that transport in the Wangmaogou watershed were as follows: The erosion of upstream branch ditches is of the upper reaches, demonstrating that the upstream erosion is stronger and the downstream intense, which focused on erosion and sediment yield, while the main channel erosion is weak, which erosion is weaker, with a basically stable state. According to the Bagnold sediment transport rate, was primarily a result of sediment transportation in gullies.

Water 2019, 11, x FOR PEER REVIEW 10 of 16

larger velocity and shear stress were associated with a greater sediment transporting capacity of a channel [44]. The velocity and runoff shear stress in the lower reaches of the Wangmaogou were higher than those in the upper reaches, which explains why the sediment transporting capacity of the downstream of the main channel is greater than that of the upstream. Overall, the characteristics of sediment yield and transport in the Wangmaogou watershed were as follows: The erosion of Waterupstream2019, 11 branch, 834 ditches is intense, which focused on erosion and sediment yield, while the10 main of 16 channel erosion is weak, which was primarily a result of sediment transportation in gullies.

Figure 5. Distribution of runoff erosion power along the channel under different working conditions. Figure 5. Distribution of runoff erosion power along the channel under different working 3.4. Impact of Different Dam Type Combinations onconditions. Sediment Discharge in Small Watershed

3.4. ImpactRunoff oferosion Different power Dam Type can betterCombinations reflecterosion on Sediment and sedimentDischarge in yield Small in Watershed watersheds. In the past, many scholars established regression equations between runoff erosion power and sediment transport moduliRunoff using erosion a large power amount can ofbetter measured reflect dataerosion that an cand sediment be used yield to predict in watersheds. sediment In yield the past, in a watershedmany scholars [45,46 ].established Based on theregr dataession describing equations the individualbetween runoff rainfall erosion runoff and power sediment and measuredsediment fromtransport 1961 moduli to 1964 using at the a watershed large amou outletnt of hydrologicmeasured data station that in can Wangaogou, be used to thepredict runo sedimentff depth, floodyield peakin a dischargewatershed modulus, [45,46]. Based and sediment on the transportdata describi modulusng the of individual individual rainfa floodsll in runoff the basin and during sediment the periodmeasured mentioned from 1961 above to were1964 analyzedat the watershed and counted, outlet and hydrologic the corresponding station in runo Wangaogou,ff erosion powerthe runoff was calculated.depth, flood The peak runo dischargeff erosion modulus, power and and sediment sediment transport transport modulus modulus of all of floods individual in Wangaogou floods in from the 1961basin to during 1964 were the period plotted mentioned in a logarithmic above coordinatewere analyzed system, and andcounted, regression and the analysis corresponding was conducted runoff (Figureerosion6 ),power which was resulted calculated. in establishment The runoff of aerosion regression power equation and sediment describing transport the runo ffmoduluserosion power of all andfloods sediment in Wangaogou transport from modulus 1961 of to individual 1964 were rainstorms plotted andin a floodslogarithmic during coordinate the study period:system, and regression analysis was conducted (Figure 6), which resulted in establishment of a regression 0.889 2 equation describing the runoffM serosion= 701157 powerP (andR = sediment0.89, n = 23transport) modulus of individual(7) rainstorms and floods during the study period: 0.889 2 2 where, Ms is the sediment transport modulus of an individual== rainstorm, t km− ; P is the runoff erosion M=s 701157 P ( R 0.89, n 23) (7) 4 1 2 power of an individual rainstorm, m s− km− ; and n is the individual rainstorm flood time. −2 where,The 𝑀 sediment is the sediment transport modulustransport ofmodulus the watershed of an underindividual eight rainstorm, working conditions t km ; 𝑃 was is the calculated runoff accordingerosion power to Formula of an individual (7), and the rainstorm, results are m4 showns−1 km−2 in; and Table 𝑛5i s. Thethe individual sediment transportrainstorm modulus flood time. of 2 2 the watershedThe sediment reached transport 314.99 t.kmmodulus− under of the working watershed condition under N andeight 50.66 working t/km underconditions working was conditioncalculated KMS, according while to the Formula values of 7, other and workingthe results conditions are shown fell in between Table 5. these. The Whensediment compared transport to nomodulus dam construction, of the watershed the sediment reached transport314.99 t.km modulus−2 under was working reduced condition by 83.92% N and after 50.66 the damt/km2 system under wasworking completed, condition indicating KMS, while that the the sediment values of transport other working modulus conditions of the small fell watershed between wasthese. reduced When sharplycompared because to no dam of completion construction, of the the dam sediment system transport completion, modulus and awas large reduced amount by of 83.92% sediment after was the intercepteddam system in was the completed, system, which indicating effectively that reducesthe sediment the sediment transport in modulus the drainage of the ditch. small Comparedwatershed withwas reduced working sharply condition because N, working of completion condition of K, th M,e dam and Ssystem reduced completion, the sediment and transport a large amount modulus of insediment the watershed was intercepted by 24.74%, in 47.11%,the system, and which 64.11%, effect respectively,ively reduces among the sediment which the in sediment the drainage reduction ditch. benefitCompared of small with dams working was condition the most obvious. N, working When condition a number K, ofM, small and S and reduced medium the damssediment were transport present, theremodulus was noin the water watershed release structure, by 24.74%, so 47.11% that the ,and incoming 64.11%, water respectively, and sand inamong the dam which control the areasediment were completelyreduction benefit blocked of andsmall stored. dams was the most obvious. When a number of small and medium dams were present, there was no water release structure, so that the incoming water and sand in the dam control area were completely blocked and stored.

Water 2019, 11, 834 11 of 16 Water 2019, 11, x FOR PEER REVIEW 11 of 16

FigureFigure 6.6. RelationshipRelationship betweenbetween runorunoffﬀ erosionerosion powerpower andand sedimentsediment transporttransport modulusmodulus ofof individualindividual stormsstorms inin thethe WangmaogouWangmaogou watershed. watershed.

Table 5. Calculated sediment transport modulus for diﬀerent dam type combinations. Table 5. Calculated sediment transport modulus for different dam type combinations. Runoﬀ Sediment Sediment Working Flood Peak Flood Volume Erosion Power TransportSediment Modulus Transport Modulus Conditions (m3 s 1) (m3) − Flood (mRunoff4 s 1 km erosion2) (t km 2) SedimentReduction transport (%) Working Flood peak − − transport− 4 4 −1 N 1.26 4853.93volume 1.72power10 (m− s 314.99modulus - reduction conditions (m3 s−1) × 4 modulus (t K 0.92 4828.873 1.25 10−2 237.07 24.74 (m ) ×km 5) (%) M 0.84 3556.61 8.38 10 166.61−2 47.11 × − km ) S 0.76 2541.04 5.42 10 5 113.04 64.11 × − NKM 0.78 1.26 3532.24 4853.93 7.73 1.7210 × 105 −4 155.03314.99 50.78- × − KS 0.51 2518.06 3.60 10 5 78.65 75.03 K 0.92 4828.87 1.25× × −10−4 237.07 24.74 MS 0.50 1802.36 2.53 10 5 57.41 81.78 × − KMS 0.44 1779.58 2.20 10 5 −5 50.66 83.92 M 0.84 3556.61 8.38× × −10 166.61 47.11 S 0.76 2541.04 5.42 × 10−5 113.04 64.11 4. Discussion KM 0.78 3532.24 7.73 × 10−5 155.03 50.78 4.1. ImpactKS of Check Dam 0.51 System on Flood 2518.06 Processes in 3.60 Small × 10 Watershed−5 78.65 75.03

−5 TheMS e ffects of individual 0.50 rainstorm 1802.36 floods in the 2.53 watershed × 10 were 57.41significantly changed 81.78 by check dams throughKMS flood storage 0.44 [13]. Construction 1779.58 of 2.20 small × 10 dams−5 had the 50.66 greatest influence 83.92 on flood processes, followed by medium and then key dams. This was primarily a result of the structural characteristics4. Discussion of these three types of dams [47]. There are no water release structures for small dams. Small check dams can completely block and store the interval and upstream inflow [48]. Therefore, as4.1. long Impact as the of Check dam doesDam System not break, on Flood the flood Processes water in willSmall not Watershed pass, resulting in the greatest reduction in floodThe peak effects and of floodindividual volume. rainstorm Medium floods dams in th aree watershed generally equippedwere significantly with a horizontal changed by tube check or shaftdams and through other flood drainage storage structures [13]. Construction used for flood of discharge. small dams Key had dams the aregreatest not only influence equipped on flood with waterprocesses, release followed structures, by medium but also and with then spillways, key dams which. This can was discharge primarily floodwater a result withof the time. structural As a result,characteristics key dams of leadthese to three a minimal types of reduction dams [47]. in There flood are peak no andwater flood release volume; structures however, for small they dams. play importantSmall check roles dams in protection can completely of medium block and and small store dams the ininterval a watershed and upstream dam system. inflow To further[48]. Therefore, analyze theas long action as mechanismthe dam does of not the break, different the dam flood type wate combinationsr will not pass, on resulting flood processes, in the greatest the connectivity reduction indexin flood (Table peak6) ofand eight flood di volume.fferent dam Medium combinations dams are was generally calculated equipped using Formulawith a horizontal (3). The channel tube or connectivityshaft and other was drainage reduced bystructures 79.0% when used the for dam flood system discharge. was completed. Key dams Figure are not7 shows only equipped the coe ffi cientwith ofwater determination release structures, for the relationship but also with between spillways, the channel which connectivitycan discharge index floodwater and peak with flow time. and totalAs a floodresult, volume. key dams As lead shown to ina minimal the diagram, reduction the di infference flood betweenpeak and the flood channel volume; connectivity however, index they play and theimportant peak flow roles was in as protection high as 0.97, of medium and the correlation and small coedamsfficient in a with watershed the total dam flood system. volume To was further 0.89. Thereanalyze were the close action relationships mechanism between of the thedifferent channel da connectivitym type combinations index and peakon flood flow andprocesses, total flood the volume,connectivity which index indicate (Table that 6) of the eight dam different system layoutdam combinations adjusts the rainstorm was calculated flood using process Formula by changing 3. The thechannel channel connectivity connectivity. was reduced by 79.0% when the dam system was completed. Figure 7 shows the coefficient of determination for the relationship between the channel connectivity index and peak flow and total flood volume. As shown in the diagram, the difference between the channel connectivity index and the peak flow was as high as 0.97, and the correlation coefficient with the total flood volume was 0.89. There were close relationships between the channel connectivity index and

Water 2019, 11, x FOR PEER REVIEW 12 of 16

peak flow and total flood volume, which indicate that the dam system layout adjusts the rainstorm flood process by changing the channel connectivity.

Water 2019, 11, 834 Table 6. Channel connectivity index under different working conditions. 12 of 16

\ Channel connectivity index Flood peak discharge (m3 s−1) Flood volume (m3) Table 6. Channel connectivity index under diﬀerent working conditions. N 1 1.26 4853.93 Channel Connectivity Index Flood Peak Discharge (m3 s 1) Flood Volume (m3) \ K 0.70 0.92 − 4828.87 NM 10.58 1.26 0.84 4853.933556.61 K 0.70 0.92 4828.87 MS 0.580.39 0.84 0.76 3556.612541.04 S 0.39 0.76 2541.04 KM 0.52 0.78 3532.24 KM 0.52 0.78 3532.24 KSKS 0.280.28 0.51 0.51 2518.062518.06 MS 0.26 0.50 1802.36 MS 0.26 0.50 1802.36 KMS 0.21 0.44 1779.58 KMS 0.21 0.44 1779.58

Figure 7. Relationships between channel connectivity index and peak flow (a); relationships between Figure 7. Relationships between channel connectivity index and peak flow (a); relationships channel connectivity index and flood volume (b). between channel connectivity index and flood volume (b). 4.2. Impact of Check Dam System on Erosion Dynamics Process in Small Watershed 4.2. Impact of Check Dam System on Erosion Dynamics Process in Small Watershed Runoff and sediment yield are extremely complex physical processes caused by the interaction betweenRunoff precipitation and sediment and the yield underlying are extremely surface complex of the basin physical [49]. processes Under the caused condition by ofthe individual interaction rainstorms,between precipitation the final result and ofthe their underlying interaction surface is the of the flood basin characteristics [49]. Under the of thecondition watershed of individual outlet, whichrainstorms, can indirectly the final reflect result theof their comprehensive interaction influenceis the flood of precipitationcharacteristics and of the underlying watershed surface outlet, characteristicswhich can indirectly on the erosion reflect and the sedimentcomprehensive yield of influence watershed. of Runoprecipitationff depth andand peakunderlying discharge surface are twocharacteristics important parameters on the erosion that and reflect sediment the characteristics yield of watershed. of the individual Runoff depth rainstorm and peak flood discharge process inare thetwo basin. important Runoff parametersdepth represents that reflect the total the amountcharacteristics of flooding of the produced individual by rainstorm an individual flood rainstorm, process in whichthe basin. reflects Runoff the precipitationdepth represents amount the total and amount the redistribution of flooding produced impact of by the an di individualfferent underlying rainstorm, surfaceswhich reflects on precipitation the precipitation in the basin amount [50]. Theand peakthe re dischargedistribution represents impact of the the flood different intensity, underlying which reflectssurfaces the on characteristics precipitation of in temporal the basin and [50]. spatial The peak distribution discharge of precipitation represents the and flood the influenceintensity, of which the basinreflects underlying the characteristics surface on of runo temporalff confluence and spatial processes distribution [51]. The of runo precipitationff depth and and peak the influence discharge of reflectthe basin some underlying characteristics surface of the on individual runoff confluence rainstorm flood, processes and cannot [51]. The comprehensively runoff depth reflect and peak the effdischargeect on erosion reflect and some sediment characteristics yield. Therefore, of the indivi thedual influence rainstorm of di ffflood,erent and dam cannot type combinations comprehensively on erosionreflect and the sedimenteffect on yield erosion is analyzed and sediment by calculating yield. theTherefore, erosion powerthe influence of individual of different rainstorm dam runo typeff, whichcombinations shows that on the erosion runoff erosionand sediment power inyield the loweris analyzed reaches by of thecalculating basin is lowerthe erosion than that power in the of upperindividual reaches. rainstorm The results runoff, of some which studies shows have that shown the runoff that the erosion runoff erosionpower in power the lower in the reaches Yanhe Basin of the isbasin large inis lower upstream than areas that andin the small upper in downstreamreaches. The reachesresults of [52 some], which studies is consistent have shown with that the resultsthe runoff of theerosion present power study. in The the area Yanhe of theBasin Yanhe is large Basin in is upst 7725ream km2 ,areas whereas and thatsmall of in Wangmaogou downstream watershed reaches [52], is 2 onlywhich 5.97 is km consistent2. Even thoughwith the the results two basins of the have present a huge study. diff erenceThe area in area,of the the Yanhe runo Basinff erosion is 7725 power km , 2 showswhereas the samethat of characteristics, Wangmaogou indicating watershed that is theonly distribution 5.97 km . Even of runo thoughff erosion the powertwo basins depends have not a huge on thedifference watershed in area, area, but the rather runoff on theerosion convergence power shows process the of rainfallsame characteristics, runoff. The erosion indicating characteristics that the of the Wangmaogou watershed were as follows: The erosion of upstream branch ditches is intense, resulting in erosion and sediment yield, while the main channel erosion is weak, mainly resulting in sediment transportation in gullies, which is consistent with the research results reported by Liao [53]. Water 2019, 11, x FOR PEER REVIEW 13 of 16 distribution of runoff erosion power depends not on the watershed area, but rather on the convergence process of rainfall runoff. The erosion characteristics of the Wangmaogou watershed Waterwere2019 as follows:, 11, 834 The erosion of upstream branch ditches is intense, resulting in erosion and sediment13 of 16 yield, while the main channel erosion is weak, mainly resulting in sediment transportation in gullies, whichFigure 8is is consistent the siltation with modulus the research diagram results of Wangmaogou reported by Liao watershed, [53]. Figure from 8 which is the itsiltation can be seenmodulus that diagramthe siltation of Wangmaogou modulus of the watershed, branch ditches from inwhich the upstream it can be areaseen is that obviously the siltation larger modulus than that of of the branchmain ditch ditches in the in the downstream upstream area,area is which obviously also indirectly larger than explains that of thethe conclusionsmain ditch in above. the downstream area, which also indirectly explains the conclusions above.

Figure 8. Siltation modulus diagram of the Wangmaogou watershed. Figure 8. Siltation modulus diagram of the Wangmaogou watershed. 5. Conclusions 5. Conclusions In this study, the hydrodynamic processes of the Wangmaogou watershed located in the Loess PlateauIn this were study, simulated the hydrodynamic and the influence processes of check of dam the Wangmaogou construction on watershed the flood andlocated erosion in the dynamic Loess Plateauprocess inwere this simulated small watershed and the was influence evaluated. of ch Theeck checkdam damsconstruction reduced on the the flood flood peak and and erosion flood dynamicvolume and process mitigated in this the small flood watershed process, was while evalua differentted. The types check of check dams dams reduced played the variousflood peak roles and in floodflood regulation.volume and After mitigated the dam the system flood wasprocess, completed, while different the flood types peak andof check flood dams volume played were reducedvarious rolesby 65.34% in flood and regulation. 58.67%, respectively. After the dam The system erosion was dynamic completed, distribution the flood of thepeak main and channel flood volume in the smallwere watershedreduced by was 65.34% changed, and 58.67%, to different respectively. extents, by The the erosion different dynamic dam type distribution combinations, of the and main the channel erosion indynamic the small parameters watershed of was thechannel changed, decreased to different most extents, after the by damthe different system was dam completed, type combinations, when the andvelocity the anderosion runo dynamicff shear stress parameters of the outlet of the section channel were decreased reduced by most 10.69% after and the 31.08%, dam system respectively. was completed,Under different when dam the typevelocity combinations, and runoff the shear runo streff erosionss of the power outlet fluctuates section were greatly reduced in the upperby 10.69% and andmiddle 31.08%, reaches, respectively. while it remains Under basically different the samedam intype the lowercombinations, reaches, andthe therunoff erosion erosion power power of the fluctuatesdownstream greatly becomes in the lower upper than and that middle of the reaches, upstream while portion. it remains Check dambasically construction the same can in etheffectively lower reaches,reduce the and sediment the erosion transport power capacity of the downstream in the watershed. becomes When lower compared than that with of thethe absenceupstream of portion. dams in Checkthe basin, dam the construction sediment transport can effectively modulus reduce of the key, the medium,sediment and transport small dams capacity was reducedin the watershed. by 24.74%, When47.11%, compared and 64.11%, with respectively, the absence with of dams the largest in the amplitude basin, the reductionsediment beingtransport observed modulus for smallof the dams. key, medium,The benefits and of small sediment dams reduction was reduced are most by 24.74%, obvious 47.11%, after the and check 64.11%, dam systemrespectively, is completed, with the with largest the amplitudesediment discharge reduction in being the basin observed being for reduced small bydams 83.92%.. The benefits of sediment reduction are most obvious after the check dam system is completed, with the sediment discharge in the basin being Authorreduced Contributions: by 83.92%. Conceptualization, S.Y. and Z.L.; methodology, P.L. and G.X.; formal analysis, H.G. and L.X.; software, F.W. and T.W.; S.Y. wrote the manuscript and all authors contributed to improving the paper. Author Contributions: Conceptualization, S.Y. and Z.L.; methodology, P.L. and G.X.; formal analysis, H.G. and Funding: This study was supported by the National Key Research and Development Program of China L.X.;(2016YFC0402404), software, F.W. the and National T.W.; S.Y. Natural wrote Science the manuscrip Foundationt and of all China authors (51779204, contribute 41731289),d to improving the Consultation the paper. and Funding:Evaluation This Project study of thewas Chinese supported Academy by the of National Sciences (2018-Z02-A-008),Key Research and and Development the School FoundationProgram of of China Xi’an University of Technology (310-252071711). (2016YFC0402404), the National Natural Science Foundation of China (51779204, 41731289), the Consultation Acknowledgments:and Evaluation ProjectWe of thank the theChinese reviewers Academy for their of Sc usefuliences comments (2018-Z02-A-008), and suggestions. and the School Foundation of Xi’anConflicts University of Interest: of TechnologyThe authors (310-252071711). declare no conflicts of interest. Acknowledgments: We thank the reviewers for their useful comments and suggestions.

Water 2019, 11, 834 14 of 16

References

1. Pal, D.; Galelli, S.; Tang, H.; Ran, Q. Toward improved design of check dam systems: A case study in the Loess Plateau, China. J. Hydrol. 2018, 559, 762–773. [CrossRef] 2. Zhao, G.; Kondolf, G.M.; Mu, X.; Han, M.; He, Z.; Rubin, Z.; Wang, F.; Gao, P.; Sun, W. Sediment yield reduction associated with land use changes and check dams in a catchment of the Loess Plateau, China. CATENA 2017, 148, 126–137. [CrossRef] 3. Zhou, Z.; Li, J. The correlation analysis on the landscape pattern index and hydrological processes in the Yanhe watershed, China. J. Hydrol. 2015, 524, 417–426. [CrossRef] 4. Li, X.-G.; Wei, X. Soil erosion analysis of human influence on the controlled basin system of check dams in small watersheds of the Loess Plateau, China. Expert Syst. Appl. 2011, 38, 4228–4233. [CrossRef] 5. Shi, H.; Shao, M. Soil and water loss from the Loess Plateau in China. J. Arid Environ. 2000, 45, 9–20. [CrossRef] 6. Ran, D.-C.; Luo, Q.-H.; Zhou, Z.-H.; Wang, G.-Q.; Zhang, X.-H. Sediment retention by check dams in the Hekouzhen-Longmen Section of the Yellow River. Int. J. Sediment Res. 2008, 23, 159–166. [CrossRef] 7. Li, P.; Xu, G.; Lu, K.; Zhang, X.; Shi, P.; Bai, L.; Ren, Z.; Pang, G.; Xiao, L.; Gao, H.; et al. Runoff change and sediment source during rainstorms in an ecologically constructed watershed on the Loess Plateau, China. Sci. Total Environ. 2019, 664, 968–974. [CrossRef] 8. Vaezi, A.R.; Abbasi, M.; Keesstra, S.; Cerdà, A. Assessment of soil particle erodibility and sediment trapping using check dams in small semi-arid catchments. CATENA 2017, 157, 227–240. [CrossRef] 9. Shi, P.; Zhang, Y.; Ren, Z.; Yu, Y.; Li, P.; Gong, J. Land-use changes and check dams reducing runoff and sediment yield on the Loess Plateau of China. Sci. Total Environ. 2019, 664, 984–994. [CrossRef][PubMed] 10. Jin, Z.; Cui, B.; Song, Y.; Shi, W.; Wang, K.; Wang, Y.; Liang, J. How Many Check Dams Do We Need To Build on the Loess Plateau? Environ. Sci. Technol. 2012, 46, 8527–8528. [CrossRef] 11. Guyassa, E.; Frankl, A.; Zenebe, A.; Poesen, J.; Nyssen, J. Effects of check dams on runoff characteristics along gully reaches, the case of Northern Ethiopia. J. Hydrol. 2017, 545, 299–309. [CrossRef] 12. Boix-Fayos, C.; Barberá, G.G.; López-Bermúdez, F.; Castillo, V. Effects of check dams, reforestation and land-use changes on river channel morphology: Case study of the Rogativa catchment (Murcia, Spain). Geomorphology 2007, 91, 103–123. [CrossRef] 13. Castillo, V.; Mosch, W.; García, C.C.; Barberá, G.G.; Cano, J.N.; López-Bermúdez, F.; Cano, J.A.N. Effectiveness and geomorphological impacts of check dams for soil erosion control in a semiarid Mediterranean catchment: El Cárcavo (Murcia, Spain). CATENA 2007, 70, 416–427. [CrossRef] 14. Hassanli, A.M.; Nameghi, A.E.; Beecham, S. Evaluation of the effect of porous check dam location on fine sediment retention (a case study). Environ. Monit. Assess. 2009, 152, 319–326. [CrossRef] 15. Bombino, G.; Gurnell, A.M.; Tamburino, V.; Zema, D.A.; Zimbone, S.M. Adjustments in channel form, sediment calibre and vegetation around check-dams in the headwater reaches of mountain torrents, Calabria, Italy. Earth Surf. Process. Landf. 2010, 34, 1011–1021. [CrossRef] 16. Leigh, D.S.; Kowalewski, S.A.; Holdridge, G. 3400 years of agricultural engineering in Mesoamerica: lama-bordos of the Mixteca Alta, Oaxaca, Mexico. J. Archaeol. Sci. 2013, 40, 4107–4111. [CrossRef] 17. Castillo, C.; Pérez, R.; Gomez, J.A.; Rodríguez, C.C. A conceptual model of check dam hydraulics for gully control: efficiency, optimal spacing and relation with step-pools. Hydrol. Earth Syst. Sci. 2014, 18, 1705–1721. [CrossRef] 18. Mekonnen, M.; Keesstra, S.D.; Stroosnijder, L.; Baartman, J.E.; Maroulis, J. Soil Conservation Through Sediment Trapping: A Review. Land Degrad. Dev. 2015, 26, 544–556. [CrossRef] 19. Xu, Y.D.; Fu, B.J.; He, C.S. Assessing the hydrological effect of the check dams in the Loess Plateau, China, by model simulations. Hydrol. Earth Syst. Sci. 2013, 17, 2185–2193. [CrossRef] 20. Boixfayos, C.; De Vente, J.; Martínez-Mena, M.; Barberá, G.G.; Castillo, V. The impact of land use change and check-dams on catchment sediment yield. Hydrol. Process. 2010, 22, 4922–4935. [CrossRef] 21. Kondolf, G.; Piégay, H.; Landon, N. Channel response to increased and decreased bedload supply from land use change: contrasts between two catchments. Geomorphology 2002, 45, 35–51. [CrossRef] 22. Mishra, A.; Froebrich, J.; Gassman, P.W. Evaluation of the SWAT Model for Assessing Sediment Control Structures in a Small Watershed in India. Trans. ASABE 2007, 50, 469–477. [CrossRef] Water 2019, 11, 834 15 of 16

23. Xu, X.-Z.; Zhang, H.-W.; Zhang, O. Development of check-dam systems in gullies on the Loess Plateau, China. Environ. Sci. Policy 2004, 7, 79–86. 24. Lenzi, M.A.; Comiti, F. Local scouring and morphological adjustments in steep channels with check-dam sequences. Geomorphology 2003, 55, 97–109. [CrossRef] 25. Bracken, L.; Wainwright, J.; Parsons, A.J.; Poeppl, R.E.; Keesstra, S.D. Introduction to special issue on connectivity in water and sediment dynamics. Earth Surf. Process. Landf. 2015, 40, 1275–1277. 26. Romero-Díaz, A.; Marín-Sanleandro, P.; Ortiz-Silla, R. Loss of soil fertility estimated from sediment trapped in check dams. South-eastern Spain. CATENA 2012, 99, 42–53. 27. Quiñonero-Rubio, J.M.; Nadeu, E.; Boix-Fayos, C.; de Vente, J. Evaluation of the Effectiveness of Forest Restoration and Check-Dams to Reduce Catchment Sediment Yield. Land Degrad. Dev. 2016, 27, 1018–1031. [CrossRef] 28. Zema, D.A.; Bombino, G.; Denisi, P.; Lucas-Borja, M.E.; Zimbone, S.M. Evaluating the effects of check dams on channel geometry, bed sediment size and riparian vegetation in Mediterranean mountain torrents. Sci. Total Environ. 2018, 642, 327–340. [CrossRef][PubMed] 29. Xu, G.; Cheng, S.; Li, P.; Li, Z.; Gao, H.; Yu, K.; Lu, K.; Shi, P.; Cheng, Y.; Zhao, B. Soil total nitrogen sources on dammed farmland under the condition of ecological construction in a small watershed on the Loess Plateau, China. Ecol. Eng. 2018, 121, 19–25. [CrossRef] 30. Zhao, B.; Li, Z.; Li, P.; Xu, G.; Gao, H.; Cheng, Y.; Chang, E.; Yuan, S.; Zhang, Y.; Feng, Z. Spatial distribution of soil organic carbon and its influencing factors under the condition of ecological construction in a hilly-gully watershed of the Loess Plateau, China. Geoderma 2017, 296, 10–17. [CrossRef] 31. Qin, C.-Z.; Ai, B.-B.; Zhu, A.-X.; Liu, J.-Z. An efficient method for applying a differential equation to deriving the spatial distribution of specific catchment area from gridded digital elevation models. Comput. Geosci. 2017, 100, 94–102. [CrossRef] 32. Guo, D.; Westra, S.; Maier, H.R. Impact of evapotranspiration process representation on runoff projections from conceptual rainfall-runoff models. Water Resour. Res. 2016, 53, 435–454. [CrossRef] 33. Van Cuyk, S.; Siegrist, R.L. Virus removal within a soil infiltration zone as affected by effluent composition, application rate, and soil type. Water Res. 2007, 41, 699–709. [CrossRef] 34. Madsen, H. Parameter estimation in distributed hydrological catchment modelling using automatic calibration with multiple objectives. Adv. Water Resour. 2003, 26, 205–216. [CrossRef] 35. McMichael, C.E.; Hope, A.S.; Loaiciga, H.A. Distributed hydrological modelling in California semi-arid shrublands: MIKE SHE model calibration and uncertainty estimation. J. Hydrol. 2006, 317, 307–324. [CrossRef] 36. Cibin, R.; Sudheer, K.P.; Chaubey, I. Sensitivity and identifiability of stream flow generation parameters of the SWAT model. Hydrol. Process. 2010, 24, 1133–1148. [CrossRef] 37. Wu, Q.; Liu, S.; Cai, Y.; Li, X.; Jiang, Y. Improvement of hydrological model calibration by selecting multiple parameter ranges. Hydrol. Earth Syst. Sci. Discuss. 2016, 21, 1–24. [CrossRef] 38. Henriksen, H.J.; Troldborg, L.; Nyegaard, P.; Sonnenborg, T.O.; Refsgaard, J.C.; Madsen, B. Methodology for construction, calibration and validation of a national hydrological model for Denmark. J. Hydrol. 2003, 280, 52–71. [CrossRef] 39. Mo, X.; Pappenberger, F.; Beven, K.; Liu, S.; De Roo, A.; Lin, Z. Parameter conditioning and prediction uncertainties of the LISFLOOD-WB distributed hydrological model. Hydrol. Sci. J. 2006, 51, 45–65. [CrossRef] 40. Butts, M.B.; Payne, J.T.; Kristensen, M.; Madsen, H. An evaluation of the impact of model structure on hydrological modelling uncertainty for streamflow simulation. J. Hydrol. 2004, 298, 242–266. [CrossRef] 41. Hooke, J. Coarse sediment connectivity in river channel systems: a conceptual framework and methodology. Geomorphology 2003, 56, 79–94. [CrossRef] 42. Foster, G.R.; Huggins, L.F.; Meyer, L.D. A Laboratory Study of Rill Hydraulics: I. Velocity Relationships. Trans. ASAE 1984, 27, 790–796. [CrossRef] 43. Kexin, L.; Zhanbin, L.; Hua, J. Application of Runoff Erosion Power in the Calculation of Soil Erosion and Sediment Yield on Hillslopes. J. Water Resour. Water Eng. 2009, 20, 70–73. (In Chinese) 44. Suyanto; Rintis, H.; Rian, M.S.P. The Analysis of Sediment Transport Using Yang Method, Engelund-Hansen Method, and Bagnold Method in Bah Bolon River, Simalungun Regency of North Sumatera. Appl. Mech. Mater. 2016, 845, 30–34. [CrossRef] Water 2019, 11, 834 16 of 16

45. Sun, Q.; Li, Z.; Yu, K.; Lu, K.; Tang, S. Parameters of sediment transport model for individual rainstorm in the Loess Hilly and Gully Region under changing environment. Sci. Soil Water Conserv. 2016, 14, 1–9. (In Chinese) 46. Zhang, L.-T.; Li, Z.-B.; Wang, H.; Xiao, J.-B. Influence of intra-event-based flood regime on sediment flow behavior from a typical agro-catchment of the Chinese Loess Plateau. J. Hydrol. 2016, 538, 71–81. [CrossRef] 47. Wei, X.; Li, X.; Wei, N. Fractal features of soil particle size distribution in layered sediments behind two check dams: Implications for the Loess Plateau, China. Geomorphology 2016, 266, 133–145. [CrossRef] 48. Yazdi, J.; Moghaddam, M.S.; Saghafian, B. Optimal Design of Check Dams in Mountainous Watersheds for Flood Mitigation. Water Resour. Manag. 2018, 32, 4793–4811. [CrossRef] 49. Fang, N.; Shi, Z.; Chen, F.; Zhang, H.; Wang, Y. Discharge and suspended sediment patterns in a small mountainous watershed with widely distributed rock fragments. J. Hydrol. 2015, 528, 238–248. [CrossRef] 50. Modrick, T.M.; Georgakakos, K.P. The character and causes of flash flood occurrence changes in mountainous small basins of Southern California under projected climatic change. J. Hydrol. Reg. Stud. 2015, 3, 312–336. [CrossRef] 51. Parsons, A.; Stone, P.; Parsons, A. Effects of intra-storm variations in rainfall intensity on interrill runoff and erosion. CATENA 2006, 67, 68–78. [CrossRef] 52. Gong, J.; Li, Z.; Li, P.; Ren, Z.; Yang, Y.; Han, L.; Tang, S.; Sun, Q. Spatial distribution of runoff erosion power based on SWAT Model in Yanhe River Basin. Trans. Chin. Soc. Agric. Eng. 2017, 33, 120–126. (In Chinese) 53. Liao, Y.S.; Cai, Q.G.; Zhuo, M.N.; Zheng, M.G.; Luo, X. Influence of Channel Networks on the Sediment Yield Under Variant Temporal and Spatial Scales:A Case Study of Chabagou Watershed. Prog. Geogr. 2009, 28, 47–54. (In Chinese)

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).