Catena 149 (2017) 417–425

Contents lists available at ScienceDirect

Catena

j ourna l homepage: www.e lsev ie r .com/ locate /catena

Impact of cornstalk buffer strip on hillslope soil erosion and itshydrodynamic understanding

Ximeng Xu a, Fenli Zheng a,b,⁎, Chao Qin a, Hongyan Wu a, Glenn V. Wilson c

a Institute of Soil and Water Conservation, State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Northwest A & F University, Yangling 712100, Shaanxi, PR Chinab Institute of Soil and Water Conservation, CAS & MWR, Yangling 712100, Shaanxi, PR Chinac USDA-ARS National Sedimentation Laboratory, Oxford 38655, MS, USA

⁎ Corresponding author at: No. 26, Xi'nong Road,Conservation, Yangling, Shaanxi 712100, PR China.

E-mail addresses: xuxm@nwsuaf.edu.cn (X. Xu), flzh@fenlizheng1960@gmail.com (F. Zheng), glqinchao@nwsuawhyyun4511@126.com (H. Wu), glenn.wilsom@ars.usda.

http://dx.doi.org/10.1016/j.catena.2016.10.0160341-8162/© 2016 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 10 June 2016Received in revised form 19 October 2016Accepted 21 October 2016Available online 9 November 2016

Soil erosion is still a serious concern on the Loess Plateau of China. Cornstalk buffer strips are not commonly uti-lized for erosion control on the Loess Plateau, and there is little hydrodynamic understanding of this soil erosioncontrol practice. A simulated rainfall experimentwas designed to investigate howa cornstalk buffer strip affectedsoil erosion and to enhance the hydrodynamic understanding of this method. Large loessial soil beds (10m-long,and 3 m-wide) with slope gradient of 20° were subjected to three successive simulated rainfall events with in-tensities of 100mmh−1 for each experimental run. The rainfall events were conducted by a down sprinkler rain-fall simulator system. Two treatments (with and without a cornstalk buffer strip) were tested in the followingfour runs: 1)without cornstalk buffer strip, 2)with cornstalk buffer strip in the third rain event, 3)with cornstalkbuffer strip in the second rain event, 4) with continuous cornstalk buffer strip in all three successive rainfallevents. In treatments with buffer, a 1 m-width cornstalk buffer strip was applied. The results showed that, com-pared with the run without cornstalk buffer strip, the run with continuous cornstalk buffer strip in three succes-sive rainfall events reduced sediment yield by 29.1% while the other two runs with cornstalk buffer strip in asingle event only reduced sediment yield by 2.0%–9.1%, and early buffer run had a larger reduction in soil erosionthan late buffer run. The runoff-sediment relationship coefficients revealed that cornstalk buffer decreased thesediment concentration and increased the runoff threshold required for soil erosion initiation. Moreover, thebuffer strip increased sheet flow velocity in interrill areas, while it decreased concentrated flow velocity in rills.This promoted a shift of rill flow to subcritical laminar flow which reduced sediment yield. Cornstalk bufferstrip also increased the critical hydrodynamic forces required for the initiation of soil erosion.

© 2016 Elsevier B.V. All rights reserved.

Keywords:Rainfall simulationsRill erosionShear stressStream powerRunoff thresholdThe Loess Plateau

1. Introduction

Mulches, e.g. straw, stalk, leaves, or plastic film, are often used toprotect the soil surface from raindrop (splash) erosion and runoff de-tachment during the critical period of plant establishment (Smets etal., 2008). Many field and laboratory studies, focusing on the impact ofmulches on water erosion, have been conducted in a wide range of en-vironmental and topographic conditions (Kamara, 1986; Brown et al.,1989; Bradford and Huang, 1994; Döring et al., 2005; Mulumba andLal, 2008; Wilson et al., 2008; Jordán et al., 2010; Shi et al., 2013;Montenegro et al., 2013; Prosdocimi et al., 2016a,b). Lal (1976) reportedthat organic mulches reduce soil loss by reducing raindrop impact, in-creasing surface storage and infiltration, decreasing runoff velocity,and improving soil qualities (soil structure, soil porosity and biological

Institute of Soil and Water

ms.iswc.ac.cn,f.edu.cn (C. Qin),gov (G.V. Wilson).

activity). Mulumba and Lal (2008) used a long term field plot to studythe effects of mulching on soil physical properties and determined anoptimum mulching rate for increasing soil porosity, available water ca-pacity, soil moisture retention and aggregate stability whichwasmean-ingful for reducing soil erosion by water. Based on a 3-year experiment,Jordán et al. (2010) found that mulch helped to improve the soil physi-cal properties and to reduce runoff coefficient and sediment yield, withan optimum rate of mulch application under semi-arid conditions insouthern Spain set to 5 Mg ha−1 yr−1. By conducting an intermittentsimulated rainfall, Montenegro et al. (2013) found that residue coverstrongly affected infiltration, soil moisture, runoff and erosion.Prosdocimi et al. (2016a,b) conducted an experiment testing the effectsof barely straw mulching on soil erosion on vineyards in eastern Spainand concluded that strawmulchwas very effective in reducing soil par-ticle detachment and surface runoff, and this benefit was achieved im-mediately after the application of the straw.

Cornstalk, as one kind of typical organic mulch, can greatly reducewater erosion (Gilley et al., 1986; Wen et al., 2014). Cornstalk retentionin fields is also important for promoting physical, chemical, and biolog-ical attributes of healthy soil in agricultural systems (Turmel et al.,

418 X. Xu et al. / Catena 149 (2017) 417–425

2015). The term buffer strip here implies to a strip of vegetation thatacts as a filter for sediment and attached nutrients and pollutants(Barling and Moore, 1994; Hussein et al., 2007). Xu et al. (2015a) ap-plied a cornstalk buffer strip using the whole plant cornstalk after har-vest, which proved to be more efficient in erosion control than clippedshort cornstalks applied in the simulated gully. It also served to shortenthe slope length for runoff convergence and intercepted the sedimentfrom upslope.

Soil erosion bywater is a process of detachment and transport of soilparticles by rainfall and runoff. Flow detachment is often described byenergy-based approaches and linked to the soil surface condition (Guoet al., 2013; Shen et al., 2016). Flow hydrodynamics are related to parti-cle detachment and sediment transport. It is of great significance toevaluate the hydrodynamic characteristics of runoff on hillslopes withorganic mulches. Gilley and Kottwitz (1995) measured the Darcy-Weisbach roughness coefficients for surfaces with different kinds ofcrop residues under controlled laboratory conditions. They found thatsmaller diameter residue materials did influence hydraulic resistancewhen they substantially increased the total volume of resistance ele-ments. Cassol et al. (2004) conducted a laboratory study to test the im-pact of crop residue on flow hydraulic conditions on sandy clay loamsoil. They proposed that, due to the increase in the viscous forces fromthe physical interference of residue on runoff, soil surface residuecover caused an increase in water flow depth and hydraulic roughnesswhich decreased the mean flow velocity, thus contributing to a reduc-tion in interrill soil detachment and transport rate. Xu et al. (2015a,b)examined rill flow characteristics by conducting an indoor experimentwith and without cornstalk buffer strip. Their results showed that rillflow velocity was decreased and rill flow energy was consumed afterit went through the cornstalk buffer strip.

Shear stress, unit stream power, and unit energy of cross section arebasic hydrodynamic parameters used to evaluate soil detachment ratesand characterize critical conditions required to initiate soil erosion(Nearing et al., 1997; An et al., 2012; Reichert and Norton, 2013;Zhang et al., 2015). Although hydrodynamic understanding of soil ero-sion has received more attention lately, the hydrodynamic characteris-tics associated with organic mulches are still unclear and need morequantification and understandings.

The Loess Plateau is well known for its serious soil erosion caused byconcentrated annual precipitation with intensive rainfall, steep slopes,less vegetation cover, and highly erodible silty soils (Cai, 2001). Al-though the Grain for Green Project (conversion steep slope farmlandsto permanent vegetation cover), which is a national ecological project,has greatly increased the vegetation cover in this region, there are stilllarge areas of farmlands needed for the food security on the Loess Pla-teau (Zhao et al., 2013; Chen et al., 2015). Corn (Zea mays L.) is one ofthemost common crops grown on the Loess Plateau in order to producea large amount of grain necessary for local people. As one kind of by-product, cornstalks are often utilized as biofuel or animal feed, whichmay contribute to air pollution (Li et al., 2002). Cornstalks are also pul-verized and left on farmland to increase soil organic matter, but thismethod is not easy to implement on steep slopes of the Loess Plateau.So, it may be environmentally-friendly and sustainable to set up corn-stalk buffer strips on loessial hillslopes for soil erosion control.

Rainfall simulations have been recognized as an important methodfor water erosion research (Cerdà, 1998). Rainfall simulation is consid-ered by several studies as a rapid and efficient method to study erosion,which can be better controlled than natural rainfalls (Cerdà, 1997;Iserloh et al., 2013a; Prosdocimi et al., 2016a,b). It has been widelyused to assess the impact of several factors on soil erosion, such asslope, soil type, soil moisture, aggregate stability, surface structure,and vegetation cover on soil erosion processes (Arnaez et al., 2007;Blavet et al., 2009; Iserloh et al., 2012; Lassu et al., 2015; Marzen et al.,2015; Xiao et al., 2015). Field rainfall simulation experiments are oftencarried out with small portable rainfall simulators (Iserloh et al., 2012,2013b; León et al., 2013; Rodrigo Comino et al., 2015, 2016a,b,c).

While laboratory-based rainfall simulations are often conducted withfixed rainfall simulator systems that can provide special requirementsfor laboratory environments that enable a large range of hydrologic,pedologic, and surface treatment conditions (de Lima and Singh, 2002;de Lima et al., 2003; Zhang et al., 2010; Shen et al., 2015; Li et al., 2016).

The main goals of this research performed under laboratory condi-tions were to: a) quantify the reduction of soil erosion induced by acornstalk buffer strip; b) determine the relationship between runoffand sediment yield on the hillslope; c) enhance the hydrodynamic un-derstanding of cornstalk buffer effects on hillslope soil erosionprocesses.

2. Materials and methods

2.1. Experimental materials

2.1.1. Rainfall simulator systemThe experiments were carried out using a rainfall simulator under

laboratory conditions in the rainfall simulation laboratory of the StateKey Laboratory of Soil Erosion and Dryland Farming on the Loess Pla-teau, Yangling City, Shaanxi Province, China. A down sprinkler rainfallsimulator systemwas used (He et al., 2014). This rainfall simulator sys-tem consists of three sets of nozzles inwhich the rainfall intensity can beset to the range from 30 to 350mmh−1 by adjusting the nozzle size andwater pressure. The nozzle type used in this study is SP (1.9 cm) typedeveloped by the Institute of Soil and Water Conservation, ChineseAcademy of Sciences & Ministry of Water Resources. The nozzles wereinstalled 18 m above the ground as this height was enough for the ma-jority raindrops to reach terminal velocity. Spatial distribution of rainfalland its intensity was measured using 10 rows and 3 columns of equallyspaced rain gauges. The spatial uniformity of the simulated rainfall wascontrolled to above 90%.When rainfall intensitywas set to 100mmh−1,the simulated raindrop diameters were 0.2 to 3.1 mm with 85.7%(±2.4%) of raindrop diameters less than 1.0 mm as calibrated by thestain method (Cerdà et al., 1997;Wang et al., 2015). Prior to the exper-iments, calibration of rainfall intensity was carried out in order that thetested rainfall intensity reached the requirement.

2.1.2. Soil bedA 10-m long, 3-mwide and 0.5-m deep soil panwithmany drainage

holes (2 cm aperture) at the bottom was used in the experiments. Thesoil pan can be inclined to the slope gradient from 0 to 30° with adjust-ment steps of 0.5°. A runoff collector was installed on the soil pan outletand used to collect the sediment and runoff samples during the rainfallsimulation. In this study, the soil panwas set to 20° whichwas the aver-age slope gradient for rill erosion development on farmland of the LoessPlateau.

The soil used in this study was loessial soil (fine-silty and mixed),classified as a Calcic Cambisols (USDA NRCS, 1999). Tested soils werecollected from the top layer (20 cm) in the Ap horizon of a well-drainedfarmland site (tilled by hand) in Ansai County (36°45′N, 109°11′E),Yan'an City, Shaanxi Province, which is located in the hilly-gullied re-gion of the Loess Plateau in northwest China. The soil texture was28.3% sand (N50 μm), 58.1% silt (50–2 μm), 13.6% clay content(b2 μm) determined by pipette method according to USDA soil classifi-cation system. Soil organicmatterwas 5.9 g kg−1 determined by the po-tassium dichromate oxidation-external heating method. The pH inwater was 7.95, measured with a 1:2.5 solid-to-water ratio on a weightbasis. Prior to the experiment, the soil was air-dried at 25 °C, and then,big clods were broken by hand into subangular-blocky clods less than4 cm in size, but was not sieved and ground to keep the in-situ soil ag-gregation fabric.

Soil water contentwas tested to calculate the soil amount needed forpacking the soil pan. The lowest 10 cmof soil panwasfilledwith sand toallow free drainage of excess water. A highly permeable cloth wasspread on the sand surface to separate the sand layer from the soil

419X. Xu et al. / Catena 149 (2017) 417–425

layer. A 10-cm plow pan layer at a soil bulk density of 1.35 g cm−3 waspacked above the sand layer, and a 20-cm tilled layerwith a bulk densityof 1.10 g cm−3 was packed above the plow pan layer. During the pack-ing process, both the plow pan and tilled layer were packed in 5-cm in-crements, and each layer was raked lightly before the next layer waspacked to ensure uniformity and continuity in soil structure. Manualtillage with a shovel (conventional tillage method on the Loess Plateau)was applied on the soil bed to a depth of approximately 20 cm along thecontour line to simulate the natural tillage practice on croplands.

To accelerate the surface channel development rate and simulate thenatural slope condition on the Loess Plateau, a preformed rill was madeon themid-slope area of the soil bed (Fig. 1). According to the rill distri-bution andmorphological characteristics on the loessial hillslope (Shenet al., 2015), a 3-m long, 25-cm wide and 15-cm deep rill was madefrom 5 to 8 m of the slope length of the soil bed, and soil materials inthis rill were removed from the soil pan. After making the rill, the soilbed was self-settled for 48 h.

2.1.3. Cornstalk buffer stripCorn (Zea mays L.) is one of the most popular crops and occupies a

large area of farmland on the Loess Plateau. Whole plant cornstalkswere found to bemore efficient than clipped short cornstalks in control-ling soil erosion (Xu et al., 2015a), so this method was applied in thisstudy. The whole plant cornstalks were collected by using a sickle inYangling (south fringe area of the Loess Plateau, where rainfall experi-ments were conducted) after harvest at August of 2014, and theywere air dried before being applied on the soil bed.

Fig. 1. Different runs applying cornstalk buffer strip mulching on the soil bed surface at differeapplied before second rain, (4) continuous buffer strip before all three rains.

2.2. Experimental design

A total of 24 rainfall events (2 repetitions of 4 different experimentalruns, each run consisting of 3 successive rainfall events) were conduct-ed (Fig. 1). According to the standard of high intensity and short dura-tion rainstorms in summer on the Loess Plateau (I5 =1.52 mm min−1) (Zhou and Wang, 1987), the rainfall intensity of thisstudy was set to 100 mm h−1 (1.67 mm min−1) for each event. Twotreatments, with and without cornstalk buffer strip, were tested infour experimental runs to evaluate the soil erosion control effectivenessand timing effects of corn buffer strips (Fig. 1). All experimental runs in-volved three successive rainfall events and the duration of each rainfallevent was 30 min. As a result, a total of 90 min rainfall duration and150 mm precipitation were carried out on the soil bed for each experi-mental run. 24 h after the former rainfall event, the next rainfall eventwas conducted to allow water redistribution within the soil bed (Liand Shao, 2008).

As shown in Fig. 1, the control run was conducted without cornstalkbuffer strip during any of three rainfall events; the late buffer run wasconducted with cornstalk buffer strip applied prior to the third rainfallevent; the early buffer runwas conductedwith cornstalk buffer strip ap-plied prior to the second and removed before the third event; the con-tinuous buffer experimental run was implemented with cornstalkbuffer strip applied prior to all three successive rainfall events.

Cornstalks were laid on the soil surface as a 3 m long by 1 m widestrip perpendicular to and immediately above the initial rill head atabout 4.3 m to 5.3 m of slope length for each event (Fig. 1). The thick-ness for cornstalk buffer stripwas the diameter of one plant of cornstalk,

nt rainfall events: (1) no buffer, (2) buffer strip applied before third rain, (3) buffer strip

420 X. Xu et al. / Catena 149 (2017) 417–425

approximately 5 cm. Mulching density was controlled to be 1 t ha−1 forwhole soil bed surface or in otherwords 10 t ha−1 for the cornstalk buff-er strip mulching area.

2.3. Experimental procedure

Before each experimental run, a pre-soak rain was applied at30 mm h−1 for around 25 min to the soil surface at a 3° slope gradientuntil surface runoff occurred. The purpose of pre-soaking was to ensurecomparative uniformity of the surface soil moisture and roughness con-dition. A plastic sheet covered the soil bed to prevent soil moistureevaporation and allow the soil water equilibrating with depth for 24 hafter pre-soaking rain and prior to the experimental runs.

During each rainfall event, once runoff occurred, sediment and run-off samples were taken succinctly at the soil pan outlet for the first 1 or2min and then adapted stepwise to 2-min intervalswhen the dischargereached steady state. During the rainfall events, surface runoff velocity(Vs) and flow depth were measured at five slope sections (1, 3, 5, 7and 9 m of slope length) along the soil bed at 6-min intervals. Vs wasmeasured by using KMnO4 dye tracer method. The time of tracer move-ment to a prescribed distance (0.5 or 1.0 m) was determined based onthe color-front propagation using a stopwatch. Flow depth was mea-sured perpendicularly to the surface using a thin ruler and read to0.1 mm precision. As a cornstalk buffer strip was set at 4.3 m to 5.3 mof slope length in the rainfall events with a buffer, flow velocities at5 m of slope length were measured from 5.3 m to 6.3 m of slope length.

The runoff/sediment collection containers used in this experimentwere 15-l buckets and the runoff was weighed with a platform scale.After sufficient time for sediment settling, samples were decanted,dried in an oven at 105 °C for 48 h until the weight did not change,and then weighed to calculate runoff and sediment yields.

2.4. Data analysis

Vs was used to estimate the mean flow velocity (V) by the formula:

V ¼ kVs ð1Þ

where Vs is surface flow velocity (cm s−1); V is mean flow velocity(cm s−1); k is a coefficient that is 0.67 for laminar flow, 0.7 for transitionflow, and 0.8 for turbulence flow (Li et al., 1996).

The Reynolds number (Re) and the Froude number (Fr)were used toreflect the flow regime and calculated as follows:

Re ¼ VRν

Fr ¼ VffiffiffiffiffiffigR

p ð2Þ

where v is kinematic viscosity (cm2 s−1) determined at the test temper-ature (t) by ν= 0.01775 / (1 + 0.0337 t + 0.000221 t2); R is hydraulicradius (cm); g = 980 cm s−2.

The Darcy-Weisbach friction coefficient (f) characterizing the flowretardation was calculated by Eq. (3) (Abrahams et al., 1986):

f ¼ 8gRJ

V2 ð3Þ

where J is surface slope (mm−1) calculated as the tangent of the slopedegree.

Shear stress was calculated according to Eq. (4) (Foster et al., 1984):

τ ¼ γRJ ð4Þ

where τ is shear stress (Pa); γ is the gravity of water (N m−3).Yang (1973) presented the concept of unit stream power according

to the equation of sediment transport. Unit stream power was

calculated by Eq. (5):

P ¼ dydt

¼ dxdt

dydx

¼ VJ ð5Þ

where P is unit stream power (m s−1).Unit energy of cross section (E) was the sum ofwater potential ener-

gy and kinetic energy and calculated by Eq. (6) (Zhang et al., 2015):

E ¼ aV2

2gþ h ð6Þ

where h is flowdepth (cm); a is correction coefficient for kinetic energy.Comparisons of flow hydraulic and hydrodynamic characteristics at

different positions were conducted using least significant difference(LSD) test, and the values were statistically significant at the 95%confidence.

3. Results and discussions

3.1. Runoff and sediment yield

Cornstalk buffer strip delayed the runoff occurrence time and ad-vanced the runoff ending time (Table 1). Compared with the controlrun, buffer strips delayed the runoff occurrence by 0.22 to 0.29 minand advanced the runoff ending time by 0.11 to 2.34min. Air dried corn-stalk buffer strips on the soil bed surface retarded the surface runoff. Theupslope runoff was intercepted, restricting convergence to down sloperunoff channels which delayed the runoff occurrence time and short-ened the runoff ending time. Due to pre-soaking rain, surface runoffreached a steady rate quickly. The cornstalk buffer strip retarded runoffinitiation and runoff rate, thereby promoting infiltration and changingthe hillslope hydrological condition.

Runoff volume for the continuous buffer run (Run 4: all three rainfallevents with buffer) was decreased by 8.79mm compared to the controlrun with no buffer strip. Runoff volume of other two experimental runsin which the buffer strip was applied for one of three rainfall events(Run 2: third rainfall events with buffer and Run 3: second rainfallevents with buffer) were only decreased by 3.59 and 2.74 mm, respec-tively. Infiltration volume of continuous buffer run was increased by77.4% compared with the control run, and the infiltration volumes ofthe other two experimental runs with a buffer were also increased by23.9% to 31.2%. Cornstalk buffer strip obviously enhanced the infiltrationvolume but the effects on runoff volume appeared minor due to themagnitude of the runoff volume.

The continuous buffer run (4) experienced a 29.1% sediment yieldreduction compared with the control run, while the intermittent bufferruns (2 and 3) only had a 2.0% and 9.1% reduction in sediment yield, re-spectively. Moreover, the early buffer run (3) had a larger reduction insediment yield than late buffer run (2). As a result, soil conservationmeasures such as cornstalk buffer strip should be applied on the farm-land as soon as possible to get better results, which corresponds wellwith other studies (Bradford and Huang, 1994; Prosdocimi et al.,2016a,b).

Excess infiltration runoff caused by high intensity rainfall and lowsoil surface infiltration capacity is the most common runoff regime onthe Loess Plateau. This situation was also obtained in severe soil erosionon loessial hillslopes (Huang et al., 2003). Cornstalk buffer strip in-creased the infiltration rate and reduced the runoff amount, whichcould alleviate the excess infiltration runoff regime to a certain extentand reduce soil particle detachment and transport by runoff. In thearea where soil was mulched with cornstalk, raindrop impact on thesoil surface were eliminated which also contributes to the reduction ofsediment yield (Marzen et al., 2015; Xiao et al., 2015). By shorteningthe slope length for runoff convergence and changing the hillslope hy-drological condition, cornstalk buffer strip reduced the sediment yield

Table 1Average total runoff, infiltration and sediment yield for each rainfall event.

Experimental

run

Rainfall

event

Runoff threshold

for soil erosion

initiation b/a

(mm min–1)

Runoff

occurrence

time (min)

Runoff

ending time

(min)

Runoff

volume

(mm)

Infiltrated

volume

(mm)

Sediment

yield (kgm–2)

1. Control

run

1 0.241 0.30(0.02) 33.98(0.34) 43.92(0.77) 6.08(0.77) 13.27(0.88)

2 0.206 0.23(0.01) 35.97(0.40) 46.60(0.95) 3.40(0.95) 17.55(1.56)

3 0.193 0.30(0.02) 36.58(0.12) 48.04(0.98) 1.96(0.98) 14.59(1.28)

Total 138.56(1.75) 11.44(1.75) 45.41(1.31)

2. Late

buffer run

1 0.282 0.30(0.03) 34.15(0.27) 43.30(1.02) 6.70(1.02) 13.60(0.61)

2 0.185 0.28(0.00) 35.64(0.59) 45.36(0.72) 4.64(0.72) 18.40(1.14)

3 0.196 0.58(0.05) 36.47(0.18) 46.31(0.46) 3.69(0.46) 12.51(0.75)

Total 134.97(2.04) 15.03(2.04) 44.51(1.34)

3. Early

buffer run

1 0.254 0.30(0.00) 34.12(0.24) 43.80(0.94) 6.20(0.94) 13.63(0.70)

2 0.266 0.50(0.02) 35.47(0.32) 45.31(0.62) 4.69(0.62) 16.70(1.12)

3 0.128 0.30(0.02) 35.17(0.57) 46.71(0.73) 3.29(0.73) 10.95(0.52)

Total 135.82(1.57) 14.18(1.57) 41.28(1.22)

4. Continuous

buffer run

1 0.310 0.53(0.03) 32.62(0.41) 39.30(1.04) 10.70(1.04) 9.46(1.21)

2 0.241 0.52(0.00) 33.63(0.20) 45.14(0.98) 4.68(0.98) 12.94(0.28)

3 0.260 0.52(0.01) 35.21(0.08) 45.27(0.38) 4.73(0.38) 9.77(0.62)

Total 129.71(1.96) 20.29(1.96) 32.17(1.64)

Note: rows within gray background were the rainfall event with cornstalk buffer strip

421X. Xu et al. / Catena 149 (2017) 417–425

and controlled soil erosion on loessial hillslope. Similar erosion controlresults were also found using other materials and application methods(Gilley et al., 1986; Jordán et al., 2010; Montenegro et al., 2013).

3.2. Relationship between runoff and sediment

A linear relationship between runoff and sedimentwas described bystudies on the loessial hillslope (Zheng et al., 2008, 2012). Fig. 2 showsthe relationships between the sediment rate and the runoff rate in dif-ferent rainfall events in this study which were described by:

SR ¼ aRR−b ð7Þ

where SR is the sediment rate (g min−1 m−2); RR is the runoff rate(mm min−1 m−2); a and b are coefficients. The coefficient a was ableto reflect the relationship between sediment rate and runoff rate,whose physical meaning could be described as the ability of runoff todetach and transport sediment. The values of a illustrate the soil erosionintensity induced by different runoff rates and soil surfacemanagement.At the same time, the ratio of b to a (b/a)was able to illustrate the runoffrate required to initiate the soil loss, its value can be recognized as therunoff rate threshold for soil erosion initiation (Table 1).

For the control run without a buffer strip, the a value in the secondrainfall event was much larger than that in the first and third rainfallevent (Fig. 2). This can be explained by the changing of water erosionpatterns and the rill development stages (Berger et al., 2010; Shen etal., 2015). Splash and sheet erosion were the dominate erosion process-es before rills formed in thefirst rain, so the sediment yieldwas relative-ly small. Headcut retreat of the original constructed rill alongwith side-wall collapse and deep-cut erosion accelerated the rill developmentprocesses during the second rain, which resulted in the greatest

sediment yield among the three rainfall events. During the third rain,rills eroded down to the plow panwhich restricted further rill develop-ment and as a result the sediment yield in the third rain was smallerthan that in second rain.

Compared with the control run without buffer, obvious decreases ofa values and increases in the runoff rate threshold (b/a values) could beseen in continuous buffer run (Table 1). The values of a in the three re-spective successive rainfall eventswith cornstalk buffer strip was 10.2%,18.6% and 24.8% lower than those without cornstalk buffer strip. Thethreshold values in three respective rainfall events with buffer were29.2%, 14.3%, and 36.8% higher than those without buffer. It can be con-cluded that more runoff was required to initiate soil erosion after corn-stalk buffer strip was applied. As a result, the effectiveness of cornstalkbuffer strip on reducing soil erosion can be illustrated by these two co-efficients. A shift to a higher threshold by applying a buffer strip not onlyresults in later initiation of runoff but also less sediment yield. The bufferstrip wasmost affective at increasing the threshold when applied early.

3.3. Mulching impacts on flow hydraulic parameters and hydrodynamicmechanisms

3.3.1. Flow velocity along the slopeAverage flow velocity during the rainfall events increased when the

slope length was greater (Fig. 3). At 1 m of slope length positions, sheetflow velocities varied from 9.64 to 13.94 cm s−1 with a standard devia-tion of 2.30 cm s−1, while rill flow velocities varied from 16.83 to24.63 cm s−1 with a standard deviation of 3.31 cm s−1. Relativelysmall standard deviations showed the good repeatability of experiment,this could be also proved by the flow velocities variations at 3m of slopelength. At 5 m of slope length where cornstalk buffer strips were ap-plied, average sheet flow velocity of interrill areas in rainfall events

Fig. 2. Runoff-sediment linear relationships for each rainfall event. Note that runs with a buffer strip are in gray color.

422 X. Xu et al. / Catena 149 (2017) 417–425

with a buffer were 5.62 cm s−1 greater than those in rainfall eventswithout a buffer; while the average rill flow velocity in rainfall eventswith a buffer were 6.54 cm s−1 smaller than in rainfall events withouta buffer. At the 7 m of slope length, flow velocity showed the sametrend as that at the 5 m of slope length, i.e. cornstalk buffer strip dis-persed the upslope runoff which transformed concentrated flow inrills to interrill sheet flow. As the slope length increased, sheet flow ve-locity and rill flow velocity at 9m of slope length in rainfall events with-out buffer were both larger than those in rainfall events with buffer.

Fig. 3.Average sheetflow and rillflow velocity along the slope. Note that flowvelocities incovered treatment at 5 m of slope length were measured from 5.3 m to 6.3 m of slopelength.

Rill flow dominates sediment delivery on steep slopes (Peng, et al.,2015), and rill areas receive the runoff and sediment delivered frominterrill areas (Knapen et al., 2007). As a result, rill flowmerits more at-tention in soil erosion control. Similar result was also obtained by Wenet al. (2014) in the Mollisol region of northeastern China, showing thatrill flow velocities with a buffer were smaller than those without a buff-er, which reduced soil erosion.

3.3.2. Hydraulic and hydrodynamic characteristicsFlow hydraulic and hydrodynamic characteristics below cornstalk

buffer stripmulching position (5.5 to 6mof slope length)were calculat-ed and displayed in Table 2. Compared with the rainfall events withoutbuffer, sheet flow velocity significantly increased but rill flow velocitydecreased in rainfall events with a buffer. According to the values ofRe and Fr, it can be concluded that both sheet flow and rill flow in thisslope section represented transition flow conditions; sheet flow wasrapid while rill flow was streaming flow. Furthermore, f of sheet flowdecreased 9.7% while f of rill flow increased 12.4% after cornstalk bufferstrips were applied. Other hydrodynamic characteristics, e.g., τ, P and E,corresponded well to flow conditions, their values for sheet flow in-creased 27.3%, 15.9% and 28.0%, respectively after cornstalk buffer stripswere applied; while on the other hand, their values for rill flow de-creased 25.1%, 13.5% and 25.5%, respectively, due to the influence ofthe buffer strips.

Within rills, cornstalk buffer strips numerically increased the flowresistance, and reduced the unit stream power, and significantly re-duced the shear stress and unit energy of cross section. These factorscombined to reduce the soil erosion. Compared with the sheet flow,concentrated flow definitely contributes more to slope erosion(Knapen et al., 2007; Al-Hamdan et al., 2012), so the total erosionamount reduced even though the sheet flow velocity was significantlyincreased by the buffer strip.

Table 2Average total flow hydraulic and hydrodynamic characteristics below the cornstalk buffer mulching position (5.3 to 6.3 m of slope length).

Flowregime

Surfacetreatment

Average flow velocity V(cm s−1)

Reynoldsnumber Re

Froudenumber Fr

Darcy–Weisbachfriction f

Average shearstress τ (Pa)

Unit stream power w(m s−1)

Unit energy of crosssection E (cm)

Sheetflow

Withoutbuffer

24.31(3.26)b 1121.0(364.6)c 1.21(0.19)a 2.17(0.84)bc 15.31(4.27)c 0.088(0.012)ab 0.75(0.16)c

With buffer 27.90(1.34)a 1648.0(633.7)bc 1.24(0.16)a 1.96(0.51)c 19.49(6.67)c 0.102(0.005)a 0.96(0.23)cRill flow Without

buffer24.51(3.50)ab 3820.8 (756. 6)a 0.60(0.09)b 7.32(2.18)ab 51.98(5.77)a 0.089(0.013)ab 2.04(0.24)a

With buffer 21.20(5.00)b 2486.8(789.6)b 0.61(0.13)b 8.23(4.61)a 38.91(6.60)b 0.077(0.018)b 1.52(0.37)b

Note: different letters (a, b, c) indicate that hydraulic and hydrodynamic characteristics within a column are significantly different at 0.05 level.

423X. Xu et al. / Catena 149 (2017) 417–425

3.3.3. Flow regime zoningAccording to the open channel flow theory, Re= 500 and 5000 are

the thresholds that separate laminar flow, transition flow and turbulentflow, while Fr = 1 separates subcritical from supercritical flow. In thelog-log plot of runoff velocity and hydraulic radius, three boundarylines (Re = 500, Re = 5000, and Fr = 1) were drawn to separatesheet flow (Fig. 4a) and rill flow (Fig. 4b) into six flow regimes (Zhang

Fig. 4. Flow regime zoning of sheet flo

et al., 2014; Xu et al., 2015b). In treatments without cornstalk buffer,sheet flow never showed turbulent flow regime while concentratedflow in rills reached turbulent conditions. After cornstalk buffer stripwas applied, sheet flow shifted from subcritical to supercritical flowand from laminar to transition and/or turbulent flow, and some datawith buffer reached supercritical-turbulent flow zone. In contrasts, rillflow switched fromsupercritical-turbulentflow to subcritical-transition

w (a) and concentrated flow (b).

Fig. 5. The relationship between sediment rate and shear stress, unit stream power, as well as unit energy of cross section.

424 X. Xu et al. / Catena 149 (2017) 417–425

flow when the cornstalk buffer applied. According to Zhang et al.(2014), flow regime is decided by a large range of factors in which sur-facemulching is oneof themost important and often leads towards sub-critical and laminar flow conditions which contributed to reducing thesediment yield. In our study, buffers led rill flow from turbulent to tran-sition conditions without reaching fully laminar conditions.

3.3.4. Hydrodynamic understanding of sediment yieldBased on previous studies focusing on calculating the critical hydro-

dynamic parameters to initiate raindrop erosion and rill erosion (An etal., 2012; Shen et al., 2016), the critical values, such as shear stress,unit stream power, and unit energy of cross section, for initiating soilerosion could be obtained by establishing relationships between sedi-ment yield and excess shear, stream power, and unit energy properties.As illustrated in Fig. 5, sediment yield increased linearly with the in-crease of shear stress, unit stream power, and unit energy of cross sec-tion. In rainfall events without buffer, the linear relationship could bedescribed as follows:

Sr ¼ 0:417 τ–1:207ð Þ R2 ¼ 0:91;n ¼ 14� �

Sr ¼ 248:24 P–0:0291ð Þ R2 ¼ 0:80;n ¼ 14� �

Sr ¼ 16:46 E–0:161ð Þ R2 ¼ 0:82;n ¼ 14� � ð8Þ

In rainfall events with buffer, the linear relationship could beexpressed as:

Sr ¼ 0:396 τ–5:916ð Þ R2 ¼ 0:88;n ¼ 10� �

Sr ¼ 176:17 P–0:0346ð Þ R2 ¼ 0:84;n ¼ 10� �

Sr ¼ 11:04 E–0:222ð Þ R2 ¼ 0:86;n ¼ 10� � ð9Þ

where Sr is sediment rate (kg min−1); τ is shear stress (Pa); P is unitstream power (m s−1); E is unit energy of cross section (cm).

The intercept, i.e., Sr =0, represents the critical hydrodynamic forcerequired to initiate soil loss. For rainfall events without buffer, criticalshear stress, critical unit stream power, and critical unit energy ofcross section were 1.207 Pa, 0.0291 m s−1, and 0.161 cm, respectively.In rainfall events with buffer, the corresponding values calculatedwere 5.916 Pa, 0.0346m s−1, and 0.222 cm, respectively. The results in-dicated that cornstalk buffer strip caused the critical shear stress, thecritical unit stream power, and the critical stream power for soil erosioninitiation to increase by 390.1%, 18.9%, and 37.9%, respectively. Bufferstrip could be used on hillslopes to reduce the sediment yield by en-hancing the critical hydrodynamic forces required to initiate soilerosion.

4. Conclusions

Rainfall simulation experiments under high intensity erosive rainfallevents (100 mm h−1) and high inclination (20°) were conducted to in-vestigate the impact of cornstalk buffer strips on the processes of

initiation of soil erosion in order to enhance the understanding oftheir hydrodynamic impacts. The results showed that 1) cornstalk buff-er strips delayed runoff occurrence time and advanced the runoff end-ing time, it also reduced the total runoff amount and increased theinfiltration volume. Continuous buffer tests had a reduction of 29.1% intotal sediment yield compared with the control run without a bufferstrip. Early application of a buffer had a larger reduction in sedimentyield than late buffer application; 2) a linear relationship was fitted be-tween runoff and sediment. According to the coefficients of runoff-sed-iment relationship, it could be concluded that cornstalk buffer stripdecreased the sediment concentration and increased the runoff ratethreshold for soil erosion initiation; 3) while the buffer strip increasedthe sheet flow velocity in interrill areas, it decreased the rill flow veloc-ity and promoted a shift in rills from supercritical turbulent towardssubcritical laminar flow conditions, which is meaningful for reducingsoil erosion. Cornstalk buffer strip can increase the flow resistance, re-duce shear stress, unit streampower, unit energy of cross section, there-by reducing soil erosion; 4) compared with non-buffer events, thecritical shear stress increased from 1.207 to 5.916 Pa, the critical unitstream power increased from 0.0291 to 0.0346 m s−1, and the criticalunit energy of cross section increased from 0.161 to 0.222 cm, respec-tively in events with buffer strip. As a result, cornstalk buffer strip en-hanced the critical hydrodynamic forces required for the initiation ofsoil erosion and reduced the hydrodynamic forces operating on thesoil surface which contributed to the reduction of hillslope sedimentyield.

Acknowledgments

This study was supported by the National Natural Science Founda-tion of China (No. 41271299), Special-Funds of Scientific Research Pro-grams of State Key Laboratory of Soil Erosion and Dryland Farming onthe Loess Plateau (A314021403-C2), and China Scholarship Councilfunds. We thank the editor and reviewers for their suggestions, whichgreatly improved our paper.

References

Abrahams, A.D., Parsons, A.J., Luk, S.H., 1986. Resistance to overland flow on deserthillslopes. J. Hydrol. 88, 343–363.

Al-Hamdan, O.Z., Pierson, F.B., Nearing, M.A., Stone, J.J., Williams, C.J., Moffet, C.A., Kormos,P.R., Boll, J., Weltz, M.A., 2012. Characteristics of concentrated flow hydraulics forrangeland ecosystems: implications for hydrologic modeling. Earth Surf. Process.Landf. 37 (2), 157–168.

An, J., Zheng, F.L., Lu, J., Li, G.F., 2012. Investigating the role of raindrop impact on hydro-dynamic mechanism of soil erosion under simulated rainfall conditions. Soil Sci. 177(8), 517–526.

Arnaez, J., Lasanta, T., Ruiz-Flaño, P., Ortigosa, L., 2007. Factors affecting runoff and erosionunder simulated rainfall in Mediterranean vineyards. Soil Tillage Res. 93, 324–334.

Barling, R.D., Moore, I.D., 1994. Role of buffer strips in management of waterway pollu-tion: a review. Environ. Manag. 18 (4), 543–558.

Berger, C., Schulze, M., Rieke-Zapp, D., Schlunegger, F., 2010. Rill development and soilerosion: a laboratory study of slope and rainfall intensity. Earth Surf. Process. Landf.35 (12), 1456–1467.

Blavet, D., De Noni, G., Le Bissonnais, Y., Leonard, M., Maillo, L., Laurent, J.Y., Asseline, J.,Leprun, J.C., Arshad, M.A., Roose, E., 2009. Effect of land use and management onthe early stages of soil water erosion in French Mediterranean vineyards. Soil TillageRes. 106, 124–136.

425X. Xu et al. / Catena 149 (2017) 417–425

Bradford, J.M., Huang, C.-h., 1994. Interrill soil erosion as affected by tillage and residuecover. Soil Tillage Res. 31, 353–361.

Brown, L.C., Foster, G.R., Beasley, D.B., 1989. Rill erosion as affected by incorporated cropresidue and seasonal consolidation. Trans. ASAE 32 (6), 1967–1978.

Cai, Q.G., 2001. Soil erosion and management on the Loess Plateau. J. Geogr. Sci. 11 (1),53–70.

Cassol, E.A., Cantalice, J.R.B., Reichert, J.M., Mondardo, A., 2004. Interrill surface runoff andsoil detachment on a sandy clay loam soil with residue cover. Pesq. Agrop. Brasileira39 (7), 685–690.

Cerdà, A., 1997. Soil erosion after land abandonment in a semiarid environment of south-eastern Spain. Arid Soil Res. Rehabil. 11, 163–176.

Cerdà, A., 1998. Effect of climate on surface flow along a climatological gradient in Israel: afield rainfall simulation approach. J. Arid Environ. 38, 145–159.

Cerdà, A., Ibáñez, S., Calvo, A., 1997. Design and operation of a small and portable rainfallsimulator for rugged terrain. Soil Technol. 11, 163–170.

Chen, Y., Wang, K., Lin, Y., Shi, W., Song, Y., He, X., 2015. Balancing green and grain trade.Nat. Geosci. 8 (10), 739–741.

de Lima, J.L.M.P., Singh, V.P., 2002. The influence of the pattern of moving rainstorms onoverland flow. Adv. Water Resour. 25, 817–828.

de Lima, J.L.M.P., Singh, V.P., de Lima, M.I.P., 2003. The influence of storm movement onwater erosion: storm direction and velocity effects. Catena 52, 39–56.

Döring, T.F., Brandt, M., Heβ, J., Finckh, M.R., Saucke, H., 2005. Effects of straw mulch onsoil nitrate dynamics, weeds, yield and soil erosion in organically grown potatoes.Field Crop Res. 94, 238–249.

Foster, G.R., Huggins, L.F., Meyer, L.D., 1984. A laboratory study of rill hydraulics: II. Shearstress relationships. Trans. ASAE 27 (3), 797–0804.

Gilley, J.E., Kottwitz, E.R., 1995. Darcy-Weisbach roughness coefficients for surfaces withresidue and gravel cover. Trans. ASAE 38 (2), 539–544.

Gilley, J.E., Finkner, S.C., Spomer, R.G., Mielke, L.N., 1986. Runoff and erosion as affected bycorn residue: part I. Total Losses. Trans. ASAE 29 (1), 157–160.

Guo, T., Wang, Q., Li, D., Zhuang, J., Wu, L., 2013. Flow hydraulic characteristic effect onsediment and solute transport on slope erosion. Catena 107, 145–153.

He, J., Li, X., Jia, L., Gong, H., Cai, Q., 2014. Experimental study of rill evolution processesand relationships between runoff and erosion on clay loam and loess. Soil Sci. Soc.Am. J. 78 (5), 1716–1725.

Huang, M., Zhang, L., Gallichand, J., 2003. Runoff responses to afforestation in a watershedof the Loess Plateau, China. Hydrol. Process. 17 (13), 2599–2609.

Hussein, J., Yu, B., Ghadiri, H., Rose, C., 2007. Prediction of surface flow hydrology and sed-iment retention upslope of a vetiver buffer strip. J. Hydrol. 338 (3), 261–272.

Iserloh, T., Fister, W., Seeger, M., Willger, H., Ries, J.B., 2012. A small portable rainfall sim-ulator for reproducible experiments on soil erosion. Soil Tillage Res. 124, 131–137.

Iserloh, T., Ries, J.B., Arnáez, J., Boix-Fayos, C., Butzen, V., Cerdà, A., Echeverría, M.T.,Fernández-Gálvez, J., Fister, W., Geißler, C., Gómez, J.A., Gómez-Macpherson, H.,Kuhn, N.J., Lázaro, R., León, F.J., Martínez-Mena, M., Martínez-Murillo, J.F., Marzen,M., Mingorance, M.D., Ortigosa, L., Peters, P., Regüés, D., Ruiz-Sinoga, J.D., Scholten,T., Seeger, M., Solé-Benet, A., Wengel, R., Wirtz, S., 2013a. European small portablerainfall simulators: a comparison of rainfall characteristics. Catena 110, 100–112.

Iserloh, T., Ries, J.B., Cerdà, A., Echeverría, M.T., Fister, W., Geiβler, C., Kuhn, N.J., León, F.J.,Peters, P., Schindewolf, M., Schmidt, J., Scholten, T., Seeger, M., 2013b. Comparativemeasurements with seven rainfall simulators on uniform bare fallow land. Z. FürGeomorphol. Suppl. 57, 11–26.

Jordán, A., Zavala, L.M., Gil, J., 2010. Effects of mulching on soil physical properties andrunoff under semi-arid conditions in southern Spain. Catena 81 (1), 77–85.

Kamara, C.S., 1986. Mulch-tillage effects and soil loss properties on an ultisol in the humidtropics. Soil Tillage Res. 8, 131–144.

Knapen, A., Poesen, J., Govers, G., Gyssels, G., Nachtergaele, J., 2007. Resistance of soils toconcentrated flow erosion: a review. Earth-Sci. Rev. 80 (1), 75–109.

Lal, R., 1976. Soil Erosion Problems on an Alfisol in Western Nigeria and Their Control:Mulching Effect on Runoff and Soil Loss. I.I.T.A. Monograph No. 1. International Insti-tute for Tropical Agriculture, Ibadan, Nigeria.

Lassu, T., Seeger, M., Peters, P., Keesstra, S.D., 2015. The Wageningen rainfall simulator:set-up and calibration of an indoor nozzle-type rainfall simulator for soil erosionstudies. Land Degrad. Dev. 26, 604–612.

León, J., Echeverría, M.T., Badía, D., Martí, C., Álvarez, C.J., 2013. Effectiveness of woodchips cover at reducing erosion in two contrasted burnt soils. Z. Für Geomorphol.Suppl. 57, 27–37.

Li, Y., Shao, M., 2008. Infiltration characteristics of soil water on loess slope land under in-termittent and repetitive rainfall conditions. Chin. J. Appl. Ecol. 19 (7), 1511–1516 (inChinese with English abstract).

Li, G., Abrahams, A.D., Atkinson, J.F., 1996. Correction factors in the determination ofmeanvelocity of overland flow. Earth Surf. Process. Landf. 21 (6), 509–515.

Li, J., Wang, L.X., Shao, M.A., Fan, T.L., 2002. Simulation of maize potential productivity inthe Loess Plateau region of China. Acta Agron. Sin. 28 (4), 555–560 (in Chinese withEnglish abstract).

Li, G., Zheng, F., Lu, J., Xu, X., Hu, W., Han, Y., 2016. Inflow rate impact on hillslope erosionprocesses and flow hydrodynamics. Soil Sci. Soc. Am. J. 80 (3), 711–719.

Marzen, M., Iserloh, T., Casper, M.C., Ries, J.B., 2015. Quantification of particle detachmentby rain splash and wind-driven rain splash. Catena 127, 135–141.

Montenegro, A.A.A., Abrantes, J.R.C.B., De Lima, J.L.M.P., Singh, V.P., Santos, T.E.M., 2013.Impact of mulching on soil andwater dynamics under intermittent simulated rainfall.Catena 109, 139–149.

Mulumba, L.N., Lal, R., 2008. Mulching effects on selected soil physical properties. Soil Till-age Res. 98 (1), 106–111.

Nearing, M.A., Norton, L.D., Bulgakov, B.A., Larionov, G.A., West, L.T., Dontsova, K.M., 1997.Hydraulics and erosion in eroding rills. Water Resour. Res. 33 (4), 865–876.

Peng, W., Zhang, Z., Zhang, K., 2015. Hydrodynamic characteristics of rill flow on steepslopes. Hydrol. Process. 29 (17), 3677–3686.

Prosdocimi, M., Cerdà, A., Tarolli, P., 2016a. Soil water erosion on Mediterraneanvineyards: a review. Catena 141, 1–21.

Prosdocimi, M., Jordán, A., Tarolli, P., Keesstra, S., Novara, A., Cerdà, A., 2016b. The imme-diate effectiveness of barley straw mulch in reducing soil erodibility and surface run-off generation in Mediterranean vineyards. Sci. Total Environ. 547, 323–330.

Reichert, J.M., Norton, L.D., 2013. Rill and interrill erodibility and sediment characteristicsof clayey Australian Vertosols and a Ferrosol. Soil Res. 51, 1–9.

Rodrigo Comino, J., Brings, C., Lassu, T., Iserloh, T., Senciales, J., Martínez Murillo, J., RuizSinoga, J., Seeger, M., Ries, J., 2015. Rainfall and human activity impacts on soil lossesand rill erosion in vineyards (Ruwer Valley, Germany). Solid Earth. 6, 823–837.

Rodrigo Comino, J., Iserloh, T., Lassu, T., Cerdà, A., Keesstra, S.D., Prosdocimi, M., Brings, C.,Marzen, M., Ramos, M.C., Senciales, J.M., Ruiz Sinoga, J.D., Seeger, M., Ries, J.B., 2016a.Quantitative comparison of initial soil erosion processes and runoff generation inSpanish and German vineyards. Sci. Total Environ. 565, 1165–1174.

Rodrigo Comino, J., Iserloh, T., Morvan, X., Malam Issa, O., Naisse, C., Keesstra, S.D., Cerdà,A., Prosdocimi, M., Arnáez, J., Lasanta, T., Ramos, M.C., Marqués, M.J., Ruiz Colmenero,M., Bienes, R., Ruiz Sinoga, J.D., Seeger, M., Ries, J.B., 2016b. Soil erosion processes inEuropean vineyards: a qualitative comparison of rainfall simulation measurementsin Germany, Spain and France. Hydrology 3, 6.

Rodrigo Comino, J., Ruiz Sinoga, J.D., Senciales González, J.M., Guerra-Merchán, A., Seeger,M., Ries, J.B., 2016c. High variability of soil erosion and hydrological processes inMed-iterranean hillslope vineyards (Montes de Málaga, Spain). Catena 145, 274–284.

Shen, H.O., Zheng, F.L., Wen, L.L., Lu, J., Jiang, Y.L., 2015. An experimental study of rill ero-sion and morphology. Geomorphology 231, 193–201.

Shen, H.O., Zheng, F.L., Wen, L.L., Han, Y., Hu, W., 2016. Impacts of rainfall intensity andslope gradient on rill erosion processes at loessial hillslope. Soil Tillage Res. 155,429–436.

Shi, Z.H., Yue, B.J., Wang, L., Fang, N.F., Wang, D., Wu, F.Z., 2013. Effects of mulch cover rateon interrill erosion processes and the size selectivity of eroded sediment on steepslopes[J]. Soil Sci. Soc. Am. J. 77 (1), 257–267.

Smets, T., Poesen, J., Knapen, A., 2008. Spatial scale effects on the effectiveness of organicmulches in reducing soil erosion by water. Earth-Sci. Rev. 89 (1), 1–12.

Turmel, M.S., Speratti, A., Baudron, F., Verhulst, N., Govaerts, B., 2015. Crop residue man-agement and soil health: a systems analysis. Agric. Syst. 134, 6–16.

USDA NRCS, 1999. Soil Taxonomy: A Basic System of Soil Classification for Making andInterpreting Soil Surveys. Agric. Handbook 436. 2nd edition. U.S. Government Print-ing Office, Washington DC.

Wang, H., Gao, J.E., Zhang, M.J., Li, X.H., Zhang, S.L., Jia, L.Z., 2015. Effects of rainfall inten-sity on groundwater recharge based on simulated rainfall experiments and a ground-water flow model. Catena 127, 80–91.

Wen, L.L., Zheng, F.L., Shen, H.O., Gao, Y., 2014. Effects of corn straw mulch buffer in thegully head on gully erosion of sloping cropland in the black soil region of NortheastChina. J. Sediment. Res. 6, 73–80 (in Chinese with English abstract).

Wilson, G.V., McGregor, K.C., Boykin, D., 2008. Residue impacts on runoff and soil erosionfor different corn plant populations. Soil Tillage Res. 99 (2), 300–307.

Xiao, L., Hu, Y., Greenwood, P., Kuhn, N.J., 2015. A combined raindrop aggregate destruc-tion test-settling tube (RADT-ST) approach to identify the settling velocity of sedi-ment. Hydrology 2, 176.

Xu, X.M., Zheng, F.L., Qin, C., Wu, H.Y., 2015a. Erosion control effects of cornstalk mulchingon loess hillslope with gully. Trans. CSAM 46 (8), 130–137 (in Chinese with Englishabstract).

Xu, X.M., Zheng, F.L., Wu, H.Y., Qin, C., 2015b. Impacts of cornstalk mulching buffer stripon rill erosion and its hydrodynamic character. Trans. CSAE 31 (24), 111–119 (in Chi-nese with English abstract).

Yang, Z.T., 1973. Incipient motion and sediment transport. Trans. ASAE 99 (10),9198–9314.

Zhang, G.H., Shen, R.C., Luo, R.T., Cao, Y., Zhang, X.C., 2010. Effects of sediment load on hy-draulics of overland flow on steep slopes. Earth Surf. Process. Landf. 35 (15),1811–1819.

Zhang, K.D., Wang, G.Q., Sun, X.M., Wang, J.J., 2014. Hydraulic characteristic of overlandflowunder different vegetation coverage. Adv.Water Sci. 25 (6), 825–834 (in Chinesewith English abstract).

Zhang, L.T., Gao, Z.L., Yang, S.W., Li, Y.H., Tian, H.W., 2015. Dynamic processes of soil ero-sion by runoff on engineered landforms derived from expressway construction: acase study of typical steep spoil heap. Catena 128, 108–121.

Zhao, G., Mu, X., Wen, Z., Wang, F., Gao, P., 2013. Soil erosion, conservation, and eco-envi-ronment changes in the loess plateau of China. Land Degrad. Dev. 24 (5), 499–510.

Zheng, M.G., Cai, Q.G., Cheng, Q.J., 2008. Modelling the runoff-sediment yield relationshipusing a proportional function in hilly areas of the Loess Plateau, North China. Geo-morphology 93 (3), 288–301.

Zheng, M.G., Yang, J.S., Qi, D.L., Sun, L.Y., Cai, Q.G., 2012. Flow-sediment relationship asfunctions of spatial and temporal scales in hilly areas of the Chinese Loess Plateau. Ca-tena 98, 29–40.

Zhou, P.H., Wang, Z.L., 1987. Soil erosion rainfall standard in the Loess Plateau. Bull. SoilWater Conserv. 7 (1), 38–44 (in Chinese with English Abstract).