This article was downloaded by: [McMaster University]On: 16 October 2014, At: 06:37Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: MortimerHouse, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Hydraulic ResearchPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/tjhr20

Turbulent flow and bed pressure within headcutscour holes due to plane reattached jetsSean J. Bennett a & Carlos V. Alonso ba Department of Geography , University at Buffalo , Buffalo, NY, 14261-0055, USAb Supervisory Research Hydraulic Engineer , USDA-ARS National SedimentationLaboratory , PO Box 1157, Oxford, MS, 38655, USAPublished online: 26 Apr 2010.

To cite this article: Sean J. Bennett & Carlos V. Alonso (2006) Turbulent flow and bed pressure within headcut scour holesdue to plane reattached jets, Journal of Hydraulic Research, 44:4, 510-521, DOI: 10.1080/00221686.2006.9521702

To link to this article: http://dx.doi.org/10.1080/00221686.2006.9521702

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose ofthe Content. Any opinions and views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be reliedupon and should be independently verified with primary sources of information. Taylor and Francis shallnot be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and otherliabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Journal of Hydraulic Research Vol. 44, No. 4 (2006), pp. 510–521

© 2006 International Association of Hydraulic Engineering and Research

Turbulent flow and bed pressure within headcut scour holes due to planereattached jets

Ecoulement turbulent et pression de lit dus au recollement de jets plans dans lesaffouillements avec tête de rigole (headcut)SEAN J. BENNETT, Associate Professor,Department of Geography, University at Buffalo, Buffalo, NY 14261-0055, USA.Tel.: +1 716-645-2722 (ext. 65); e-mail: seanb@buffalo.edu (author for correspondence)

CARLOS V. ALONSO, Supervisory Research Hydraulic Engineer,USDA-ARS National Sedimentation Laboratory,PO Box 1157, Oxford, MS 38655, USA

ABSTRACTSoil erosion remains the principle cause of soil degradation worldwide, and the development and migration of headcuts in rills, crop furrows, andgullies can significantly increase soil losses on hillslopes, upland areas, and agricultural fields. Experiments were conducted to define the time-meanturbulent flow characteristics within fixed headcut scour holes typical of upland concentrated flows and to assess the distribution of these flow andpressure parameters for discrete areas of the scour hole domain. These data show that: (1) flow within headcut scour holes is analogous to planeturbulent reattached wall jets; (2) turbulence maxima are associated with the jet entry, recirculation eddies, and flow reattachment; (3) turbulentvelocities are distributed asymmetrically about the free jet axis within the scour hole; (4) turbulent velocities associated with the reattached wall jetdisplay good similarity collapse when scaled with the jet entry velocity; and (5) distributions of wall pressure near reattachment agree well with asimilarity argument derived for impinging jets. This study supports the use and application of a jet impingement approach for modeling flow andsoil erosion in upland concentrated flows due to headcut development and migration. Moreover, it is suggested that velocity-gradient shear, turbulentshear, and near-bed pressure gradients all are involved in soil erosion within headcut scour holes.

RÉSUMÉL’érosion demeure la cause principale de la dégradation du sol dans le monde entier; le développement et la migration des têtes de rigoles dans lesruisselets, les sillons de récolte, et les caniveaux peuvent augmenter de manière significative les pertes de sol sur les pentes de collines, les secteurs demontagne, et les terres agricoles. Des expériences ont été entreprises pour définir les caractéristiques de l’écoulement turbulent moyenné en tempsdanscertains affouillements avec tête de rigole, typiques des écoulements concentrés en altitude, et pour évaluer la distribution des paramètres d’écoulementet de pression sur des secteurs discrets de ces domaines. Ces données montrent que: (1) l’écoulement dans les affouillements avec tête de rigole estanalogue au recollement des jets turbulents plans; (2) les maxima de turbulence sont associés à l’entrée du jet, aux tourbillons de re-circulation, etau recollement de l’écoulement; (3) les vitesses turbulentes sont distribuées asymétriquement par rapport à l’axe du jet dans l’affouillement; (4)lesvitesses turbulentes associées au recollement du jet présentent une bonne similitude de décroissance une fois rapportées à la vitesse d’entrée du jet;et (5) les distributions de la pression de paroi près du recollement sont conformes à ce qui a été établi pour les jets impactants. Cette étude confortel’idée d’utiliser une approche de jet impactant pour modéliser l’écoulement et l’érosion du sol dans le cas des écoulements concentrés en altitude dusau développement et à la migration de têtes de rigoles. D’ailleurs, on suggère que le cisaillement du gradient de vitesse, le cisaillement turbulent,et les gradients de pression au fond sont impliqués dans l’érosion du sol à l’intérieur des affouillements avec tête de rigole.

Keywords: Headcuts, jets, scour, soil erosion, turbulence, wall pressure.

1 Introduction

Soil erosion remains the principle cause of soil degradationworldwide, and off-site impacts of sedimentation can severelyaffect water quality, ecology, and ecological habitat (Pimentelet al., 1995; Uri and Lewis, 1999). On hillslopes and agriculturalfields, soil erosion occurs in areas of concentrated flow such asrills, crop furrows, and gullies. Within these relatively small chan-nels, localized erosion often occurs due to the development and

Revision received February 2, 2006/Open for discussion until August 31, 2007.

510

upstream migration of headcuts, which are abrupt step-changesin bed elevation (Brush andWolman, 1960; Mosley, 1974; Bryan,1990). The development and migration of headcuts can signif-icantly increase soil losses and sediment yields (Seginer, 1966;Mosley, 1974; Piestet al., 1975; Bryan, 1990; Römkens, 1996,1997).

Recent analytical models for headcut erosion in soils have uti-lized elements of jet impingement theory (Stein and Julien, 1993;Stein et al., 1993; Stein and LaTray, 2002). In these related

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

Turbulent flow and bed pressure within headcut scour holes 511

studies, flow at the headcut brinkpoint and in the scour holedomain, as well as the erosion of soil, all were treated explicitlyas an impinging jet and associated wall jets that would act on theerodible boundary. This hypothesis has been further supportedby the experiments of Bennettet al. (2000) and the analysis ofAlonsoet al. (2002), who showed that actively migrating head-cuts developed in soils have turbulent flow patterns resemblingan impinging jet. Studies such as Robinson (1992), Fogleet al.(1993), and Robinsonet al. (2000) confirmed that a jet impinge-ment model is appropriate below free overfalls typical of gullyheadcuts (0.3 m and larger). However, no study has yet confirmedthat a jet impingement model is the correct analog for headcutscour holes, and no study has sought to collect the velocity, turbu-lence, and pressure fields within headcut scour holes and comparethese directly to similarity arguments for length scales and veloc-ity distributions determined for plane turbulent impinging jets asdiscussed by Rajaratnam (1976).

The overall goals of this research program were to fullycharacterize the velocity, turbulence, and wall pressure withinheadcut scour holes and to assess the applicability and limitsof jet impingement theory to soil erosion in upland concen-trated flows. In a companion paper, Bennett and Alonso (2005)examined the applicability of similarity arguments used to definethe limits, length and velocity scales, and velocity distributionswithin discrete regions of the scour hole domain where both clas-sical free jets and wall jets were observed. These limits shall beused herein to further examine the turbulent characteristics offlow within the headcut scour hole. The objectives of the presentstudy were: (1) to experimentally determine the time-mean tur-bulent flow characteristics within and the distribution of wallpressure along fixed-bed models of headcut scour holes typi-cal of upland concentrated flows; and (2) to compare the datafrom specific regions of the scour hole domain to previous stud-ies of plane turbulent free jets, impinging jets, and wall jets.The data required to achieve these goals would be impossibleto collect in live-bed experiments because: (1) sediment-ladenflows preclude the use of laser-based technology; (2) the rate ofchange of headcut morphology from initial growth to steady-stateis quite rapid (as short as 100 s); and (3) any probe inserted intothe flow would disrupt the flow field and thus the observed soilerosion and headcut behavior. Thus the present study utilized two-dimensional, rigid-boundary conditions replicated from live-bedexperiments.

2 Experimental methods

All experiments were conducted at the USDA-ARS National Sed-imentation Laboratory using a recirculating 5.5-m long tiltingflume (see Bennettet al., 2000 for details). Flow rate was con-trolled by two adjustable intake valves and monitored with apressure transducer connected to an in-line Venturi meter. Themain flow channel was 2 m long and 0.165 m wide, with smoothwalls and sand-roughened beds.

Two wooden models of headcut scour holes were placed intothe flume at a distance of 0.4 m downstream of the false floor.

These forms were replicated from previous live-bed experimentsusing a sandy loam to sandy clay loam soil. The first headcut form(Model 1) is an exact replica of the time-averaged, steady-statebed profile of Run 9 from Bennettet al. (2000), representing thenonventilated, nearly submerged overfall commonly observed inlive-bed experiments for bed slopes of 2% and less (see Bennett,1999; Bennettet al., 2000; Bennett and Casalí, 2001). For thefixed-bed experiment, bed slopeS0 was 1% and unit dischargeqw was 0.0071± 0.00006 m2 s−1, which was slightly lower thanthe live-bed experiment (6% lower upstream mean flow depthdu

and 10% lower upstream mean flow velocityVu as compared tothe live-bed experiment). The second headcut form (Model 2),an exact replica of the instantaneous bed profile from Run 5 ofBennett (1999), represents the partially ventilated, free overfallcommonly observed in live-bed experiments with bed slopes of3% and larger. For this fixed-bed experiment,S0 = 5% andqw = 0.0052± 0.00005 m2 s−1 (9% lowerdu and 12% higherVu as compared to the live-bed experiment). All experimentalparameters are summarized in Table 1.

Velocity measurements were obtained with a 300 mW Argon-ion laser Doppler anemometer (LDA). The LDA was operatedin back-scatter mode using a 400 mm focal-length lens and avelocity resolution of 0.5 mm s−1. The probe head was mountedto a manually-controlled positioning system with a spatial reso-lution of less than 1 mm. Flow velocities were recorded on theflume axial plane of symmetry in the directions parallel and per-pendicular to the flume slope for periods of 60–120 s, spacedapproximately 1–2 mm vertically and 5 mm horizontally, with anaverage data rate of more than 100 Hz. Instantaneous velocity

Table 1 Summary of experimental parameters

Parameter Model 1 Model 2

Bulk hydraulic and physical parametersUnit dischargeqw (m2 s−1) 0.0071 0.0052Standard deviation ofqw (m2 s−1) 0.00006 0.00005Bed slopeS0 (%) 1 5Flow widthW (m) 0.165 0.165SD (m) 0.087 0.140

Upstream flow conditionsDepthdu (m) 0.016 0.010Velocity Vu (m s−1) 0.444 0.520Froude numberFru 1.12 1.66

Brinkpoint flow conditionsDepthdb (m) 0.011 0.009Velocity Vb (m s−1) 0.645 0.578Froude numberFrb 1.96 1.94

Nappe entry flow conditionsThicknessbe (m)a 0.008 0.005Velocity Ve (m s−1)b 0.913 1.007Tailwater heighth (m) 0.017 0.037Angle (θe; ◦) 45 55

Ramp position (given as height above SD)Lower (mm) 10 30Upper (mm) 21 60

aDetermined from Eq. (1).bDetermined from Eq. (2).

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

512 Bennett and AlonsoE

leva

tion

(m)

-0.05

0.00

0.05

0.10

0.15

water surfacebed surface

Ele

vatio

n (m

)

-0.05

0.00

0.05

0.10

0.15

water surfacebed surface

β

β

(a)

(b)

loci of maximumvelocity (free jet axis)

lower boundaryof jet

flow reattachment, sr

+x+y

s−sr

Distance (m)

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40

Ele

vatio

n (m

)

-0.05

0.00

0.05

0.10

0.15

(c)lower ramp

upper ramp

water surfacebed surface

vi

ui

Figure 1 Water surface and bed profiles for the fixed headcut Model 1summarizing the experimental configurations and frames of referenceused to collect and process the data. These plots depict the traces used toresample the irregularly spaced data (a) normal to the loci of maximumvelocity (free jet axis) and (b) normal to the scour-hole boundary. Alsoshown are +x values along the orthogonal to the maximum velocityaxis, the lower boundary of the jet, the point of flow reattachmentsr , thecurvilinear distance downstream along scour hole boundarys − sr , +y

values along the wall-normal trace, and the laser Doppler anemometerorientation used to collect the vertical (vi ) and horizontal (ui) velocitycomponents in (a), two examples of the angleβ relative to the horizontalframe of reference in (b), and the horizontal ramps added to the headcutmodel in (c).

components parallel and perpendicular to the flume axial planeare designated asui andvi (Fig. 1), their time-averaged valuesareu andv, and their turbulent fluctuations about the mean areu′ = ui − u andv′ = vi − v, respectively.

To examine the turbulence characteristics of the jet-flow struc-tures within the scour-hole domain, it was necessary to re-samplethe irregular spaced dataset to obtain measurements orthogonal toselected traces and normal to the free-jet axis and the scour-holeboundary (Fig. 1). This was accomplished by: (1) using a gridalgorithm to re-sample the data over a regular rectangular gridwith a cell size of 1× 1 mm; (2) defining the axis for exami-nation, its angle relative to the dataset’s frame of reference, andthe desired orthogonal coordinates; and (3) re-sampling all datarelative to these orthogonal coordinates (to within 0.3 mm). Amore detailed description of this procedure is given in Bennett

andAlonso (2005). Distances along traces orthogonal to the free-jet axis are measured from the jet axis (x = 0), withx > 0 in thedownstream direction, andx < 0 toward the headcut. Similarly,distances along traces orthogonal to the pool boundary are mea-sured from the wall (y = 0). The position of ay-trace is definedby its curvilinear distances measured along the boundary fromthe point of flow reattachmentsr.

Bed pressure measurements along the headcut scour hole weremade with a 1.4 -mm (O.D.), L-shaped static tube with a flat-tened tip connected via flexible tubing to a free-air manometermounted to the flume sidewall. The static tube was inserted nearthe flume sidewall and placed onto the bed perpendicular to theflow, extending to the centerline position, and multiple datasetswere collected for each experimental configuration. Water levelin the manometer was visually recorded to within 1 mm, andfree-surface water elevations were measured with a point gaugeto within 0.5 mm. Hydrostatic pressure heads were subtractedfrom all wall pressure measurements reported here.

Bed pressure measurements also were obtained along theheadcut scour hole in the presence of two horizontal (flat bed)ramps. These ramps were placed at the following positions abovethe maximum scour depth and parallel to the flume slope: 10(lower) and 21 (upper) mm for headcut Model 1 (Fig. 1c), and 30(lower) and 60 (upper) mm for headcut Model 2 (Table 1). Theramps were used to evaluate the change in wall pressure distribu-tion as the bed profile evolved from the flat surface configurationconsidered previously by Beltaos (1976) to the curved scour holeprofiles studied herein.

3 Results and discussion

3.1 General observations

The time-averaged vectors for flow within the headcut scour holemodels are shown in Figs 2 and 3. The overfall nappe enters thescour pool domain with an entry angle of 45◦ for Model 1 and55◦ for Model 2 (Table 1). The thickness of the nappe upon entryto the scour pool was difficult to measure due to its unsteadiness.Alonsoet al. (2002) presented expressions for the nappe thicknessbe and nappe velocityVe at the entry position based on brinkpointflow depthdb, unit discharge, and entry angleθe:

be = db cosθe (1)

Ve = qw

db cosθe(2)

Based on these relations and the data in Table 1, the following aredetermined:be = 0.008 m andVe = 0.913 m s−1 for Model 1,andbe = 0.005 m andVe = 1.007 m s−1 for Model 2.

Upon entering the scour hole domain, the overfall nappe cre-ates a core of high-velocity fluid that extends toward the bed(Figs 2 and 3). This high-velocity core is deflected by the scourhole curvature and remains in close proximity to the bed, achiev-ing uniform flow conditions upon exiting the measuring section.Two relatively large recirculation zones occur on both sides ofthe high-velocity core. Flow reattachment in both models occursupstream of the maximum scour depth. This flow structure is

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

Turbulent flow and bed pressure within headcut scour holes 513

Distance (m)

Ele

vatio

n (m

)

0 0.1 0.2 0.3 0.40

0.04

0.08

0.12

1m/s

Figure 2 Flow field within headcut Model 1 showing time-averaged vectors. The dotted line shows the approximate location of the air bubble belowthe nappe, and the square symbol shows the location of jet reattachment.

Distance (m)

Ele

vatio

n (m

)

0 0.1 0.2 0.3 0.40

0.04

0.08

0.12

0.16

1m/s

Figure 3 Flow field within headcut Model 2 showing time-averaged vectors. The dotted line shows the approximate location of the air bubble belowthe nappe, and the square symbol shows the location of jet reattachment.

identical to the conceptualization presented by Bennettet al.(2000) and similar to the flow field measured by Robinsonet al.(2000) for relatively large, fixed gully headcuts when tailwaterheight approximated the overfall height.

Figures 2 and 3 demonstrate that flow within headcut scourholes typical of upland concentrated flows is a turbulent reat-tached wall jet. For turbulent impinging jets, whether orthogonalor oblique to the boundary, the free jet axis (or high-velocity core)extends completely to the boundary, creating a point of stagna-tion and flow deflection (e.g., Beltaos and Rajaratnam, 1973). Incontrast, the point of reattachment in the present cases, as deter-mined from the vector data, corresponds to the upper boundary ofthe recirculation zone delimited by the scour hole and the free jet.Turbulent reattached wall jets, as discussed by Rajaratnam andSubramanya (1968) and shown here, have an elongated impinge-ment zone downstream of flow reattachment, where flow evolvesinto a plane turbulent wall jet.

3.2 Whole flow-field observations

Contour plots of select flow and turbulence parameters for head-cut Models 1 and 2 are shown in Figs 4 and 5, respectively. Thedistributions of these parameters are wholly consistent with thetime-averaged vector data presented above.

The high-velocity core near the jet entry is clearly demar-cated by the large values of the velocity magnitude

(√u2 + v2

)

shown in Figs 4(a) and 5(a). The location and trajectory of thishigh-velocity core changes in space within the scour hole. Uponentering the scour hole domain, the high-velocity core is locatedin the center of the pool and it has a bed-directed trajectory.

Further downstream, this high-velocity core shifts toward thebed downstream of the maximum scour depth, assuming abed-parallel trajectory.

Relatively higher turbulence intensities are associated with thefree jet upon entry to the scour hole domain. Maximum valuesof the root-mean-square of the downstream velocity componenturms

(urms =

√u′; where the overbar represents a time-average

)

occur along the upper part of the free jet, in association withthe shear layer separating the submerged jet and the downstreamrecirculation eddy (Figs 3b and 4b). Conversely, maximum values

of thevrms(vrms =

√v

′) occur along the lower part of the freejet, in association with the shear layer separating the submergedjet and the recirculation eddy near the headcut face (Figs 4c and5c). Secondary maxima for bothurms andvrms occur within ornear the large recirculation zone downstream of the jet.

Both Reynolds stressτ (τ = −ρu′v′ whereρ is fluid density;Figs 4d and 5d) and two-dimensional turbulent kinetic energyk [k = 0.5(u′2 + v′2); Figs 4e and 5e] show similar distribu-tions in comparison to the turbulent velocities described above.Maxima in τ (positive values) andk are in close spatial asso-ciation with the upper shear layer of the submerged jet, withsecondary maxima for headcut Model 1 occurring near flow reat-tachment. In both models, large, positive values ofτ dominate thefree jet region of the scour hole domain and downstream of flowreattachment.

3.3 Turbulence within the free jet region

The classical definition of free jets involves neutrally buoyantturbulent jets diffusing in an unbounded medium (Albertsonet al.,

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

514 Bennett and Alonso

Figure 4 Contour plots of the (a) velocity magnitude, root-mean-square of the (b) downstream and (c) vertical velocity components, (d) Reynoldsstress, and (e) turbulent kinetic energy for headcut Model 1.

1950). A free jet within a submerged medium has two distinctregions. The zone of flow establishment extends from the jetorigin (J = 0) to a point along the jet axis where the mixingregion on all sides of the jet has penetrated to the centerline ofthe jet (end of potential core). For impinging jets, the zone ofestablished flow for the free jet extends from the end of the jet’spotential core to a point where the boundary begins to influencethe velocity magnitude and direction (Beltaos and Rajaratnam,1973).

Bennett and Alonso (2005) examined the kinematics of flowassociated with the free jets within these fixed-bed headcut mod-els. They found that: (1) the distribution of free jet velocityVx/Vmx about the jet axis(x − xm)/bu agreed well with the sim-ilarity formulation of Beltaos and Rajaratnam (1973) for planeturbulent impinging jets; and (2) the zone of established flowwithin the free jet extended from 4≤ J/be ≤ 9 for both headcutmodels, whereVmx is the maximum jet velocity along the freejet axis,Vx is the jet velocity at an orthogonal distancex fromthe jet axis (xm), bu is equal to the distancebu = x − xm whereVx = 0.5Vmx, andJ is the curvilinear axial distance from nappeentry.

For the established flow zone of the free jet, the distributionsof urms, vrms, andk, normalized byVe, are shown in Fig. 6 for

both headcut models, plotted as a function of distance across thejet. In both cases, maxima ofurms/Ve occur on the upper part(downstream side) of the free jet shear layer [(x − xm)/bu > 0],maxima ofvrms/Ve occur on the lower part (upstream or headcutside) of the free jet shear layer [(x − xm)/bu < 0], and maximaof k/V 2

e show an asymmetric, bimodal distribution about the jetaxis [(x − xm)/bu = 0]. Further,urms/Ve > vrms/Ve, turbulenceintensities for Model 1> Model 2, and reasonable collapse ofthe turbulence profiles within the zone of established jet flow isobserved.

These distributions of turbulence intensities show a charac-teristic asymmetry as compared to the established flow regionsof unbounded free jets. Ramaprian and Chandrasekhara (1985)summarized experimental measurements of turbulent flow veloc-ities associated with plane turbulent jets, and their LDA mea-surements are shown in Fig. 6. These turbulence intensity curvesdisplay a bimodal symmetry with respect to the jet axis, with aslight dip at(x−xm)/bu = 0 and withurms/Ve > vrmsVe. The tur-bulence parameters reported herein show some agreement withthese previous measurements. Yet the present data clearly areinfluenced by the disproportionate size and vorticity of the recir-culation zones, causing the asymmetric distribution ofurms/Ve

andvrms/Ve.

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

Turbulent flow and bed pressure within headcut scour holes 515

Figure 5 Contour plots of the (a) velocity magnitude, root-mean-square of the (b) downstream and (c) vertical velocity components, (d) Reynoldsstress, and (e) turbulent kinetic energy for headcut Model 2.

3.4 Turbulence within the wall jet region

By definition, a wall jet is a shear flow directed along a wallwhere, due to initially supplied fluid momentum, its streamwisevelocity exceeds that of the external stream and whose velocitydistribution can be described using similarity arguments (e.g.,Verhoff, 1963; Rajaratnam, 1976; Launder and Rodi, 1983).

Bennett and Alonso (2005) examined the kinematics of flowdownstream of reattachment within these fixed-bed headcut mod-els. They found that: (1) the distribution of velocityVy/Vmy withorthogonal distance from the wally/δ agreed well with the sim-ilarity formulation of Verhoff (1963) and Rajaratnam (1976),i.e., flow downstream of reattachment is a classical wall jet;and (2) the established flow zone of the wall jet region extendedover 12≤ (s−sr)/be ≤ 18 for Model 1 and 23≤ (s−sr)/be ≤ 37for Model 2, whereVy is the wall jet velocity measured at a heighty orthogonal to the boundary,Vmy is the maximum wall jet veloc-ity for that profile, andδ is the distance from the bed to the planeabove the height ofVmy at whichVy = Vmy/2.

When usingVe as the scaling velocity, reasonable collapseof the urms, vrms, andk profiles for the established flow region

of the wall jets is observed for various distances from reattach-ment (Fig. 7). Moreover, the distribution of turbulence is bothquasi-linear and symmetric (i.e., nearly identical slopes but withchanged sign) about the levely/δ = 1.5. In both headcut mod-els, urms/Ve, vrms/Ve, andk/V 2

e decrease from a maximum aty/δ = 1.5 toward the wall (y/δ = 0) and toward the recirculationeddy (y/δ ≈ 2.5).

This distribution of turbulence intensity is in contrast to pre-vious studies of wall jets, which suggest that the correct velocityscaling parameter should beVmy rather thanVe (Schneider andGoldstein, 1994; Erikssonet al., 1998; Venåset al., 1999).Figure 8 shows this scaling approximation applied to the presentdataset, along with the experimental data from Erikssonet al.(1998), whose results can be considered representative. The tur-bulence parameters for headcut Model 1 do collapse, but thesedata are not in agreement with the observations of Erikssonet al. (1998), and turbulence intensities for headcut Model 2increase with distance from reattachment. Moreover, the dataof Erikssonet al. (1998) and others show a parabolic distribu-tion for (urms/Vmx)

2 and (urms/Vmx)2 with y/δ, reaching a

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

516 Bennett and Alonso

x−xm/bu

-6 -4 -2 0 2 4 6

k/V

e2

0.0

0.1

0.2

0.3

0.4

5.126.387.719.1310.62R & C (1985)

Headcut face

x−xm/bu

-6 -4 -2 0 2 4 6

v rms/V

e

0.0

0.1

0.2

0.3

0.4

u rms/V

e

0.0

0.1

0.2

0.3

0.4

3.85 J/be4.445.045.656.276.907.548.20R & C (1985)

Headcut faceModel 1 Model 2

Figure 6 Variation of turbulent velocities (urms andvrms) and turbulent kinetic energy (k) along the free jet axis as a function of orthogonal distancealong the axis (x − xm) normalized by the jet entry velocityVe and jet length scalebu for headcut Models 1 (on left) and 2 (on right). Solid line isfor a free jet given by Ramaprian and Chandrasekhara (1985). Data are presented as a function of distance along the free jet (J) normalized by the jetentry thickness (be).

maximum aty/δ = 0.6 to 0.8, which is markedly less thanobserved here.

There are two possible reasons for the apparent differencesin the magnitude and distribution of turbulence scaling whilemaintaining velocity similarity within the established wall jetregions. First, Bennett and Alonso (2005) noted that the momen-tum flux of the wall jet, from the point of reattachment, is notconserved as is typically the case in flat-wall, deeply submergedwall jet flows (e.g., Beltaos, 1976; Rajaratnam, 1976). Thereduction in momentum flux with distance is due to the redi-rection of the momentum vector in response to the scour holecurvature, which is more severe in headcut Model 2. Moreover,Ead and Rajaratnam (2002) observed that for plane turbulentwall jets with a shallow tailwater the rate of decay of themaximum velocity increased as the tailwater height decreasedand the size of the recirculation eddy increased. This was theresult of pressure gradients associated with super-elevation ofthe recirculation eddy downstream of reattachment. Second,

Chan et al. (2003) observed that for wall jets developed onconvex surfaces (albeit opposite to the headcut models herein)the streamwise and spanwise turbulence intensities increasedand the relative height of the maxima for the streamwise tur-bulence intensity increased as wall curvature increased. It isclear that turbulence parameters within headcut scour holeswill be affected by both the relatively low tailwater height andthe spatially-varied curvature of the bed profile, which wouldexplain the lack of similarity collapse shown in Fig. 8 and rendercomparisons to flat-bed, unbounded plane wall jets somewhatproblematic.

3.5 Wall pressure

For plane turbulent impinging jets, the static wall pressurep risesabove hydrostatic at impingement and significant pressure gra-dients are established that cause flow to redirect and becomeparallel to the wall (e.g., Beltaos, 1976; Rajaratnam, 1976).

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

Turbulent flow and bed pressure within headcut scour holes 517y/

δ

0

1

2

3

4

12.3213.6915.0616.4117.76

(s−sr)/be

urms/Ve

0.00 0.05 0.10 0.15 0.20

y/δ

0

1

2

3

423.4224.7125.9627.1728.3329.4430.4932.6834.8737.04

vrms/Ve

0.00 0.05 0.10 0.15 0.20

k/Ve2

0.00 0.05 0.10 0.15 0.20

(s−sr)/be

Model 1

Model 2

Figure 7 Variation of turbulent velocities (urms andvrms) and turbulent kinetic energy (k) orthogonal to the wall jet axisy normalized by the jet entryvelocity Ve and wall jet length scaleδ for headcut Models 1 (upper) and 2 (lower). Dashed line isy/δ = 1.5. Data are presented as a function ofdistance from flow reattachment (s − sr) normalized by the jet entry thickness (be).

y/δ

0

1

2

3

4

12.3213.6915.0616.4117.76EKP (1998)

(vrms/Vmy)2

0.0 0.1 0.2 0.3 0.4

k/Vmy2

0.0 0.2 0.4 0.6 0.8 1.0 1.2

(s−sr)/be

(urms/Vmy)2

0.0 0.1 0.2 0.3 0.4

y/δ

0

1

2

3

423.4224.7125.9627.1728.3329.4430.4932.6834.8737.04

(s−sr)/be Model 2

Model 1

Figure 8 Variation of turbulent velocities (urms andvrms) and turbulent kinetic energy (k) orthogonal to the wall jet axisy normalized by the maximumjet velocity within the profileVmy and wall jet length scaleδ for headcut Models 1 (upper) and 2 (lower). Solid line is for a wall jet given by Erikssonet al. (1998). Data are presented as a function of distance from flow reattachment (s − sr) normalized by the jet entry thickness (be).

Wall pressures in excess of hydrostatic are observed within thescour holes of both headcut models near reattachment, with neg-ative wall pressures occurring along the headcut face (Fig. 9c, f;Table 2). The maximum wall pressure in headcut Model 1 occursats = 0.112 m, wheres is the curvilinear distance from the head-cut brinkpoint. This maximum coincides nearly exactly with thepoint of flow reattachmentsr = 0.113 m (see Fig. 2). For head-cut Model 2, the maximum wall pressure occurs ats = 0.178 m,

which is slightly downstream of flow reattachmentsr = 0.153 m(see Fig. 3).

Within headcut scour holes, three hydrodynamic mechanismsare responsible for soil erosion: (1) high shear stresses due to near-bed velocity gradients (e.g., Rajaratnam and Subramanya, 1968);(2) high near-bed Reynolds stresses due to turbulent fluctuationsin velocity (e.g., Figs 4d and 5d; Rajaratnam and Berry, 1977);and (3) large wall pressure gradients near flow reattachment

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

518 Bennett and Alonso

0.0 0.1 0.2 0.3 0.4

p (m

)

-0.04

-0.02

0.00

0.02

0.04

p (m

)

-0.04

-0.02

0.00

0.02

0.04

p (m

)

-0.04

-0.02

0.00

0.02

0.04(b)

(c)

(a)

Curvilinear distance from brink (m)

0.0 0.1 0.2 0.3 0.4 0.5

No ramp

Upper ramp

Lower ramp

Model 1 Model 2

(e)

(f)

(d)

Figure 9 Variation of bed pressure (p) measured along the headcut scour hole for Model 1 and Model 2 with an upper ramp (a, d), a lower ramp(b, e), and no ramp (c, f), respectively. Different symbols are used to signify multiple datasets collected for each experimental configuration.

Table 2 Summary of bed pressure measurements

Bed condition ps(m) s (m) �p

Model 1No ramp 0.015 0.112 0.293Upper ramp 0.026 0.096 1.900Lower ramp 0.020 0.103 1.046

Model 2No ramp 0.016 0.178 0.302Upper ramp 0.033 0.114 2.644Lower ramp 0.019 0.146 1.028

(e.g., Fig. 9 and Table 2). The introduction of the ramps causesmarked changes in the magnitude and location of these pressuremaxima. As the ramp height decreases, the magnitude of theexcess wall pressure decreases and the location of the maximummoves away from the headcut brink. The pressure gradients nearreattachment, defined here as�p = |dps/ds|, increase signifi-cantly, by as much as a factor of 8, as the impingement regionmoves toward the jet entry location (Fig. 9; Table 2), whichhelps explain the rapid initial growth of the scour hole duringheadcut formation as described by Bennettet al. (2000) andothers.

Beltaos (1976) found that the distribution of wall pressurewithin the impingement region of an oblique jet can be describedby the following similarity argument:

p − pm

ps − pm= exp[−0.693η2] (3)

wherepm andps are the minimum and stagnation (maximum)wall pressures,η = |xl/bp|, xl is the curvilinear distance alongthe wall fromps, andbp is the value ofxl wherep − pm =0.5(ps − pm). Because the location and magnitude ofpm andps

can vary due to jet entry conditions and flow unsteadiness of theoverfall nappe, even within individual datasets, average valuesof pm andps were determined using the scaling approximation

p(s,m)ρg

ρV 2e /2

· SD − h

be, (4)

as defined by Beltaos (1976) whereh is the vertical distance fromthe headcut brink to the tailwater surface.

Figure 10 summarizes all wall pressure data for both headcutmodels using the similarity approach in Eq. (3). In general, Eq. (3)provides a good description of wall pressure distribution nearreattachment (η ≤ 1), although some similarity collapse occursforη ≤ 2 to 3 when moving upstream fromps (toward the headcutface) for the headcut models with no ramp (see Fig. 10c, f). Thereis also a general tendency toward wall pressure similarity as the

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

Turbulent flow and bed pressure within headcut scour holes 519

0 1 2 3 4

0.0

0.2

0.4

0.6

0.8

1.0

(p−p

m)/

(ps−

p m)

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

0 1 2 3 4

Model 1 Model 2

No ramp

Upper ramp

Lower ramp

(a)

(b)

(c)

(d)

(e)

(f)

η ηFigure 10 Variation of bed pressure (p) measured along the headcut scour hole for Model 1 and Model 2 with an upper ramp (a, d), a lower ramp(b, e), and no ramp (c, f), respectively, using the scaling approximations of Beltaos (1976). Different symbols are used to signify multiple datasetscollected for each experimental configuration, and solid curve is Eq. (3).

height of the ramp increases and the wall becomes less curved(see Fig. 10a, d).

The pressure data with and without the ramps in place clearlydepart from the similarity formulation proposed by Beltaos(1976) whenη > 1, especially when moving downstream fromps (Fig. 10). Whilst values of(p − pm)/(ps − pm) do collapseonto a single curve, these values are greater than predicted byEq. (3). This departure may be due to sporadic oscillation of theoverfall nappe or the low tailwater height. During the experiment,the nappe was observed to oscillate sporadically in response to thesuction within the air bubble just under the overfall. This sporadicmovement would have produced instantaneous variations in thebed pressure along the headcut scour hole and affect the measure-ments. As noted by Ead and Rajaratnam (2002), plane turbulentwall jets entering a fluid domain with a low tailwater height expe-rience a water depression near the gate containing the slot anda super-elevation near the upwelling region of the downstreamrecirculation eddy. This super-elevation causes the required pres-sure gradient to drive the return flow above the wall jet, and mayresult in higher than expected values in(p − pm)/(ps − pm)

as shown in Fig. 10. The super-elevation associated with thedownstream recirculation eddy without the ramps is ca. 6 mm forheadcut Model 1 (see Fig. 2) and ca. 3 mm for headcut Model 2(see Fig. 3). In fact, all pressure data show a slight excess pressure

(slightly higher than hydrostatic) well downstream of the scourhole (see Fig. 9), which is nearly identical in magnitude (ca. 3–6 mm) to the super-elevation observed. Thus, the presence andlocation of the ramps, the unsteadiness of the overfall nappe, andthe super-elevation of the downstream recirculation eddy all canimpact the magnitude of the wall pressure and cause deviationsin the application of Eq. (3).

4 Conclusions

Recent models for soil erosion in upland concentrated flows dueto migrating headcuts assumed that the overfall nappe at the head-cut brinkpoint and resulting flow structure within the headcutscour hole were analogous to a plane turbulent impinging jet withtwo wall jets. This assumption was somewhat speculative sincelittle quantitative data were available to support this supposition.

The time-mean data presented herein for fixed models of head-cut scour holes typical of upland areas show conclusively that theflow structure is analogous to a reattached plane turbulent walljet as previously described (Rajaratnam and Subramanya, 1968).Turbulence parameters within the scour hole domain are spatiallydistributed and dependent upon the flow structure. Maxima of the

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

520 Bennett and Alonso

turbulent velocities occur in close association with the shear lay-ers surrounding the free jet near entry, with a peak in theurms

occurring on the upper (downstream) side of the jet and a peak inthevrmsoccurring on the lower (upstream) side of the jet. Maximafor Reynolds stress and turbulent kinetic energy occur along theshear layers of the free jet, near flow reattachment, and within therecirculation eddies. Reasonable collapse of turbulence parame-ters is observed in the established flow regions of both the freejet and wall jet when scaled with the jet entry velocity rather thanthe local maximum velocity.

Wall pressures in excess of hydrostatic occur within the reat-tachment region of the headcut scour holes, and their magnitudesand gradients increase as the boundary, in this case a ramp, movestoward the jet entry location. For headcut Model 1, the maximumwall pressure coincides exactly with the point of flow reattach-ment, whereas for headcut Model 2, this pressure maximumoccurs slightly downstream of reattachment. The distribution ofwall pressure along headcut scour holes near reattachment agreeswith a similarity argument previously derived for impinging jetsnear stagnation, but this distribution is affected by wall curvatureand low tailwater height. It is suggested here that soil erosionwithin headcut scour holes occurs due to velocity-gradient andturbulent shear stresses accompanied by favorable wall pressuregradients.

This study provides experimental confirmation of the turbu-lent flow structure within headcut scour holes typical of uplandareas. Such erosional phenomena can be treated hydrodynami-cally as plane reattached wall jets. This conclusion enables thefurther development and application of jet impingement theoryfor predicting soil erosion processes in rills, crop furrows, andephemeral gullies.

Acknowledgments

G. Gray and D. Wren provided technical support for theseexperiments. Funding for this work was provided by the U.S.Department of Agriculture.

Notation

J = Curvilinear distance from the entry point alongthe free jet axis (m)

Frb, Fru = Froude number at and upstream of thebrinkpoint [–]

SD = Scour depth (m)S0 = Upstream bed slope [–]

Vb, Vu = Average velocity at and upstream of thebrinkpoint (m s−1)

Ve = Average nappe velocity at point of entry to thepool (m s−1)

Vx = Velocity of diffusing jet at the distancex fromthe jet axis (m s−1)

Vmx = Maximum jet velocity along the free jet axis(m s−1)

Vmy = Maximum jet velocity along the walljet axis (m s−1)

Vy = Velocity of wall jet at the distanceyfrom the boundary (m s−1)

W = Flow width (m)be = Nappe thickness at the brinkpoint (m)bp = Curvilinear distance fromps where

p − pm = 0.5(ps − pm) (m)bu = Orthogonal distance from free jet axis

whereVx = 0.5Vmx (m)db, du = Flow depth at and upstream of the

brinkpoint (m)g = Acceleration due to gravity (m s−2)h = Vertical distance from headcut brink to

the tailwater surface (m)k = Turbulent kinetic energy (m2 s−2)

p, ps, pm = Wall pressure, and its stagnation(maximum) and minimum values (m)

�p = Spatial gradient in wall pressure [–]qw = Water discharge per unit width

(m2 s−1)s = Curvilinear distance along the pool

boundary (m)sr = Point of jet reattachment to pool

boundary (m)ui , u, u′, urms = At-a-point, time-mean, fluctuation, and

root-mean-square of the lengthwisevelocity component in the measuringframe of reference (m s−1)

vi , v, v′, vrms = At-a-point, time-mean, fluctuation, androot-mean-square of the orthogonalvelocity component in the measuringframe of reference (m s−1)

x = Orthogonal distance from free jet axis(m)

xl = Curvilinear distance along scour holewall (m)

xm = Centerline jet axis position (m)y = Distance along orthogonal to scour

hole (m)δ = Distance from the bed to the point

whereVy = Vmy/2 (m)θe = Jet entry angle (◦)ρ = Fluid density (kg m−3)η = Non-dimensional curvilinear distance

along scour hole wallτ = Reynolds stress (Pa)

References

1. Albertson, M.L., Dai, Y.B., Jensen, R.A. and Rouse, H.(1950). “Diffusion of Submerged Jets”.ASCE Trans. 115,639–664.

2. Alonso, C.V., Bennett, S.J. and Stein, O.R. (2002). “Pre-dicting Head Cut Erosion and Migration in Concentrated

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014

Turbulent flow and bed pressure within headcut scour holes 521

Flows Typical of Upland Areas”.Water Resour. Res. 38,doi:10.1029/2001WR001173.

3. Beltaos, S. (1976). “Oblique Impingement of Plane Tur-bulent Jets”.J. Hydraul. Div. Am. Soc. Civil Engrs. 102,1177–1192.

4. Beltaos, S. and Rajaratnam, N. (1973). “Plane TurbulentImpinging Jets”.J. Hydraul. Res. 11, 29–59.

5. Bennett, S.J. (1999). “Effect of Slope on the Growthand Migration of Headcuts in Rills”.Geomorphology 30,273–290.

6. Bennett, S.J. and Alonso, C.V. (2005). “Kinematics ofFlow Within Headcut Scour Holes on Hillslopes”.WaterResour. Res. 41, W09418, doi:10.1029/2004WR003752.

7. Bennett, S.J. and Casalí, J. (2001). “Effect of Initial StepHeight on Headcut Development in Upland ConcentratedFlows”.Water Resour. Res. 37, 1475–1484.

8. Bennett, S.J., Alonso, C.V., Prasad, S.N. and Römkens,M.J.M. (2000). “Experiments on Headcut Growth andMigration in Concentrated Flows Typical of Upland Areas”.Water Resour. Res. 36, 1911–1922.

9. Brush, L.M., Jr. and Wolman, M.G. (1960). “KnickpointBehavior in Noncohesive Material: a Laboratory Study”.Geol. Soc. Am. Bull. 71, 59–74.

10. Bryan, R.B. (1990). “Knickpoint Evolution in Rillwash”.In: Bryan, R.B. (ed.), Soil Erosion—Experiments andModels. Catena Suppl. 17, 111–132.

11. Chan, T.L., Zhou, Y., Liu, M.H. and Leung, C.W. (2003).“Mean Flow and Turbulence Measurements of the Impinge-ment Wall Jet on a Semi-circular Convex Surface”.Exp.Fluids 34, 140–149.

12. Ead, S.A. and Rajaratnam, N. (2002). “Plane Turbu-lent Wall Jets in Shallow Tailwater”.J. Engng. Mech. 128,143–155.

13. Eriksson, J.G., Karlsson, R.I. and Persson, J. (1998).“An Experimental Study of a Two-dimensional PlaneTurbulent Wall Jet”.Exp. Fluids 25, 50–60.

14. Fogle, A.W., McBurnie, J.C., Barfield, B.J. andRobinson, K.M. (1993). “Modeling Free Jet Trajectory atan Overfall and Resulting Shear Stress Distribution in thePlunge Pool”.Trans. Am. Soc. Agric. Engrs. 36, 1309–1318.

15. Launder, B.E. and Rodi, W. (1983). “The Turbulent WallJet—Measurements and Modeling”.Ann. Rev. Fluid Mech.15, 429–459.

16. Mosley, M.P. (1974). “Experimental Study of Rill Ero-sion”. Trans. Am. Soc. Agric. Engrs. 17, 909–913.

17. Piest, R.F., Bradford, J.M. and Wyatt, G.M. (1975).“Soil Erosion and Sediment Transport from Gullies”.J. Hydraul. Div. Am. Soc. Civil. Engrs. 101, 65–80.

18. Pimentel, D., Harvey, C., Resosudarmo, P., Sinclair,K., Kurz, D., McNair, M., Crist, S., Shpritz, L., Fitton,L., Saffouri, R. and Blair R. (1995). “Environmental and

Economic Costs of Soil Erosion and Conservation Benefits”.Science 267, 1117–1123.

19. Rajaratnam, N. (1976).Turbulent Jets. Developments inWater Science, No. 5, Elsevier Scientific Publishing Co.,Amsterdam.

20. Rajaratnam, N. and Berry, B. (1977). “Erosion byCircular Turbulent Jets”.J. Hydraul. Res. 15, 277–289.

21. Rajaratnam, N. and Subramanya, K. (1968). “Plane Tur-bulent ReattachedWall Jets”.J. Hydraul. Div. Am. Soc. CivilEngrs. 94, 95–112.

22. Ramaprian, B.R. and Chandrasekhara, M.S. (1985).“LDA Measurements in Plane Turbulent Jets”.J. FluidsEngng. 107, 264–271.

23. Robinson, K.M. (1992). “Predicting Stress and Pressure atan Overfall”.Trans. Am. Soc. Agric. Engrs. 35, 561–569.

24. Robinson, K.M., Cook, K.R. and Hanson, G.J. (2000).“Velocity Field Measurements at an Overfall”.Trans. Am.Soc. Agric. Engrs. 43, 665–670.

25. Römkens, M.J.M., Prasad, S.N. and Helming, K. (1996).“Sediment Concentration in Relation to Surface and Subsur-face Hydrologic Soil Conditions”.Proceedings of the SixthFederal Interagency Sedimentation Conference, Las Vegas,NV, Vol. 2, pp. IX-9–16.

26. Römkens, M.J.M., Prasad, S.N. and Gerits, J.J.P. (1997).“Soil Erosion Modes of Sealing Soils: a PhenomenologicalStudy”.Soil Tech. 11, 31–41.

27. Schneider, M.E. and Goldstein, R.J. (1994). “LaserDoppler Measurement of Turbulence Parameters in a Two-dimensional Plane Wall Jet”.Phys. Fluids 6, 3116–3129.

28. Seginer, I. (1966). “Gully Development and SedimentYield”. J. Hydrol. 4, 236–253.

29. Stein, O.R. and Julien, P.Y. (1993). “Criterion Delineatingthe Mode of Headcut Migration”.J. Hydraul. Engng. 119,37–50.

30. Stein, O.R. and LaTray, D.A. (2002). “Experiments andModeling of Head Cut Migration in Stratified Soils”.WaterResour. Res. 38, doi:10.1029/2001WR001166.

31. Stein, O.R., Julien, P.Y. and Alonso, C.V. (1993).“Mechanics of Jet Scour Downstream of a Headcut”.J. Hydraul. Res. 31, 723–738.

32. Uri, N.D. and Lewis, J.A. (1999). “Agriculture and theDynamics of Soil Erosion in the United States”.J. Sustain.Agric. 14, 63–82.

33. Venås, B., Abrahamsson, H., Krogstad, P.-Å. andLöfdahl, L. (1999). “Pulsed Hot-wire Measurements inTwo- and Three-dimensional Wall Jets”.Exp. Fluids 27,210–218.

34. Verhoff, A. (1963). “The Two Dimensional Turbulent WallJet With and Without an External Stream”.Report 626,Princeton University, Princeton, NJ.

Dow

nloa

ded

by [

McM

aste

r U

nive

rsity

] at

06:

37 1

6 O

ctob

er 2

014