ContentslistsavailableatSciVerseScienceDirect
Ecological
Engineering
jou rn a l h o m e pa g e:w w w . e l s e v i e r . c o m / l o c a t e / e c o l e n g
Cyprinid
swimming
behaviour
in
response
to
turbulent
flow
Ana
T.
Silva
a,∗,
Christos
Katopodis
b,
José
M.
Santos
a,
Maria
T.
Ferreira
a,
António
N.
Pinheiro
caForestResearchCentre,InstitutoSuperiordeAgronomia,TechnicalUniversityofLisbon,TapadadaAjuda,1349-017,Lisboa,Portugal bKatopodisEcohydraulicsLtd.,122ValenceAvenue,Winnipeg,MB,CanadaR3T3W7
cCEHIDRO,InstitutoSuperiorTécnico,TechnicalUniversityofLisbon,Av.RoviscoPais,1049-001Lisboa,Portugal
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:Received4August2011
Receivedinrevisedform30March2012 Accepted8April2012
Available online 5 May 2012 Keywords: Iberianbarbel Swimmingbehaviour Pool-typefishway Turbulence Eddies
AcousticDopplerVelocimeter(ADV)
a
b
s
t
r
a
c
t
Turbulenceisacomplexphenomenonwhichcommonlyoccursinriverandfishwayflows.Itisa diffi-cultsubjecttostudy,especiallybiologically,yetturbulencemayaffectfishmovementsandfishpassage efficiency.Studiesonquantifyingfishresponsestoturbulence,particularlywithinfishways,arelacking. Thisstudyinvestigatedtheswimmingbehaviourof140adultIberianbarbel(Luciobarbusbocagei)oftwo size-classes(smallfish:15≤TL<25cm,largefish:25<TL≤35cm)underturbulentflowconditions cre-atedbythreesubmergedorificearrangementsinanexperimentalpool-typefishway:(i)offsetorifices,(ii) straightorificesand(iii)straightorificeswithadeflectorbarof0.5bolocatedat0.2Lfromtheinletorifices,
whereboisthewidthofthesquareorificesrangingfrom0.18to0.23mandListhepoollength(1.90m).
Watervelocityandturbulence(turbulentkineticenergy,Reynoldsshearstress,turbulenceintensityand eddysize)werecharacterizedusinga3DAcousticDopplerVelocimeter(ADV)andwererelatedwithfish swimmingbehaviour.Theinfluenceofturbulentflowontheswimmingbehaviourofbarbelwasassessed throughthenumberofsuccessfulfishpassageattemptsandassociatedpassagetimes.Theamountoftime fishspentinacertaincellofthepool(transittime)wasmeasuredandrelatedtohydraulicconditions. Thehighestratesofpassageandthecorrespondinglowesttimeswerefoundinexperimentsconducted withoffsetorifices.Althoughsize-relatedbehaviouralresponsestoturbulencewereobserved,Reynolds shearstressappearedasoneofthemostimportantturbulencedescriptorsexplainingfishtransittime forbothsize-classesinexperimentsconductedwithoffsetandstraightorifices;furthermore,swimming behaviouroflargerfishwasfoundtobestronglyaffectedbytheeddiescreated,inparticularbythoseof similarsizetofishtotallength,whichweremainlyfoundinstraightorificeswithadeflectorbar arrange-ment.Theresultsprovidevaluableinsightsonbarbelswimmingbehaviouralresponsestoturbulence, whichmayhelpengineersandbiologiststodevelopeffectivesystemsforthepassageofthisspeciesand otherswithsimilarbiomechanicalcapacities.
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
Freshwatersystemsareoneofthemostmodifiedecosystems
onearth(Saundersetal.,2002),asaresultofanthropogenic
over-exploitation (Arthington and Welcomme, 1995). For a number
ofdecades,thenaturalflowregimesofriversweresubstantially
alteredbytheburgeoningconstructionofdams,weirsandother
barriers,whichblockedanddelayedmigrationoffish,leadingtoa
dramaticdecline ofmanyspecies(CowxandWelcomme,1998;
Knaepkenset al.,2007).Thus, ecological continuity,in
particu-lar longitudinalconnectivity,has becomea majorfactor in the
∗ Correspondingauthorat:DepartamentodeEngenhariaFlorestal,Instituto Supe-riordeAgronomia,TapadadaAjuda,1349-017Lisboa,Portugal.
Tel.:+351213653380;fax:+351213653338. E-mailaddress:[email protected](A.T.Silva).
restorationofregulatedriversystems,with,considerableefforts
devoted to thedevelopment and improvement of fishpassage
structuresformultiplespeciesandlife-stages.Despiteadvancesin
biologicallyorientedfishwayresearchoverseveraldecades,efforts
weremainly focusedonanadromous fishspecies (i.e.fish that
spendmostoftheirlives intheseaand migratetofreshwater
tospawn) (Baras et al., 1994; Gowans et al., 2003; Katopodis,
2005).Considerable lack of knowledge existsonthe suitability
andefficiencyoffishpassdesignsforpotamodromousspecies(i.e.
migratoryfishwhosemigrationsoccurentirelywithinfreshwater)
(Katopodis,2005),whicharethepredominantgroupofmigratory
fishencounteredinMediterraneanrivers(Barasetal.,1994;Lucas
andFrear,1997;Santosetal.,2005).
Poolandweirfishwaysareoneofthemostcommontypesoffish
pass(FAO/DVWK,2002)usedtofacilitateupstreammovementof
fish.Althoughtheirdesigncriteriaarereasonablywell-understood
forthepassageofsalmonids(Bell,1986;Clay,1995;Eadetal.,2004;
0925-8574/$–seefrontmatter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecoleng.2012.04.015
Nomenclature
ae areaoftheeddy
Ao areaofthesubmergedorifice
Ap totalareaofthepool
B poolwidth
bo widthofthesubmergedorifice
g accelerationofgravity(9.81m2/s)
hm poolmeanwaterdepth
L poollength
Lex maximumlongitudinaldiameteroftheeddies
Ley maximumtransversaldiameteroftheeddies
Llf lengthofthelargerfish
Lsf lengthofthesmallerfish
PV volumetricpowerdissipationinapool
Q fishwaydischarge
T samplingperiod
TI turbulenceintensity
TKE turbulentkineticenergy
TL fishtotallength
U averagelocalvelocity
TLsf smallfishtotallength
TLlf largefishtotallength
u2
rms1,u2rms2,u2rms3 rootmeansquarevaluesofu,
v
,wu fluctuatingcomponentofvelocityatsamplingtime
T
Vo maximumflowvelocityatorifice
x,y,z coordinates
h waterleveldifferencebetweentwoadjacentpools
waterdensity
−u
iuj Reynoldsshearstressatthexyplane
Reynoldsshearstress
xy Reynoldsshearstressatthexycomponent.
Katopodis,2005), thusfarthebiological criteriatoensure
suit-abilityofthesedevicesforcyprinidandotherfreshwaterspecies
arenotwellunderstoodandrequirefurtherresearch(Katopodis,
2005).Fishpassageefficiencyforvariousfishwaytypesbuiltina
varietyoflocationswasfoundtobelowerfornon-salmonidsfor
thanforsalmonids(Noonanetal.,2011).Improvementson
fish-waydesigncriteriaareknowntodependontheunderstandingof
howhydrodynamiccuesareusedbyfishtoguidefine-scaleswim
pathselection.Hencevariationonhydrodynamicvariablessuchas
velocity,waterdepthandturbulencewhichisstronglyinfluenced
bypooldimensions,designconfiguration,sizeoforificesornotches,
aswellasfishwayslopeanddischarge,aredeterminingfactorsin
fishpassageefficiency,fishswimmingabilityandbehaviour(Silva
etal.,2010,2011).
Turbulenceisdefinedasthethree-dimensional,timeand
spa-tially dependent heterogeneous motion of rapid flow velocity
fluctuationswhichresultfromthesuperpositionofchaotic
cor-ticalflowsofmultiplestrengthsandsizes(Kirkbride,1993).This
hydraulicvariableaffectsfishswimmingcosts(Liaoetal.,2003),
captureefficiency (Enders et al.,2003), habitat selection(Cotel
etal.,2006)andfishdensity(Smithetal.,2006).Dependingonits
magnitude,turbulencemayattractorrepelfish,triggeringor
pre-ventingtheirmigratorymovements.Ithasbeenwelldocumented
thatturbulenceincreasesthecostoffishlocomotion(Pavlovetal.,
1982;Endersetal.,2003)andatextremelyhighlevels,causesbody
injuriesorevenfishmortality(Cadaetal.,1999;Odehetal.,2002).
Nonetheless,recentstudieshavealsoshownthatfishmightreduce
locomotorycostsbyexploitingtheenergyofvorticesgeneratedby
watermovingpastphysicalstructuresorpropulsivemovementsof
otherfishes(Liao,2007;Przybillaetal.,2010),ortheirown(Liao
etal.,2003).Theeffectsofturbulenceonfishenergeticcostshave
beendeemedtwofold.Inthepastfewdecades,severalofthemost
importantturbulentdescriptorsformigratingfishhavebeen
iden-tified: turbulentkineticenergy(TKE),Reynolds shearstress(),
turbulenceintensity(TI),strain,eddylengthscale,orientation,and
vorticity(Pavlovetal.,2000;TriticoandCotel,2010;Silvaetal.,
2011).Turbulentkineticenergy,whichcorrespondstothekinetic
energyassociatedwithfluctuatingvelocityatagivenpoint(Rodi,
1980),wasshowntoaffectfishswimmingperformance(Odehetal.,
2002;Silvaetal.,2011)byincreasingswimmingcosts(Endersetal., 2005).Endersetal.(2005),whofocusedondevelopingaswimming
costmodelforjuvenileAtlanticsalmon(Salmosalar)by
estimat-ingthetotalcostsofswimminginarespirometer,foundthattotal
swimmingcostsincreasedwiththeincrementofturbulentkinetic
energy.
Reynoldsshearstressisdefinedbythecontinuoustransferof
momentumbetweenadjacentviscouswatermassesofdifferent
velocities,intersectingormovingneareachother(Tennekesand
Lumley, 1972).Thegenerated force,exertedparalleltothefish
body,hasastrongimpactonfishswimmingperformanceand
sta-bility(Odehetal.,2002;Silvaetal.,2011)andatextremelyhigh
levelsitmaycausesevereinjuryormortality(Cadaetal.,1999;
Odehetal.,2002).Furthermore,theimpactofthistensiondepends
onitsorientationtowardsthebodyofthefish,asdemonstrated
bySilvaetal.(2011),whofoundthatthelongitudinalcomponent
ofReynoldsshearstress(xy)affectsfishthemost.Another
com-monturbulentdescriptoristurbulenceintensity,whichisdefined
asanormalizedmeasureofvariationinvelocitymagnitude
rel-ativetothelocalaveragespeedwherethemeasurementismade
(Pavlovetal.,2000;Odehetal.,2002),Thismaydecreasefish
swim-mingspeedforvalueshigherthat2/3ofthetotallengthofthefish
asfoundby(Lupandin,2005)inhisstudyontheeffectof
turbu-lenceintensityintheswimmingspeedofperch(Percafluviatilis)of
differentsize-classes.
Turbulenteddiesaretypicalvortical structuresthatoccurin
unsteadyflows.Thesecoherentrotatingstructuresinthefluidare
oftendescribedbytheirdiameter,orientationandrateofrotation
or vorticity.These typeof turbulentstructures,result byeither
connecting toaninterface(solid–liquidsuchasthebedsurface
orbetweentwofluidssuchastheair-waterinterface)andmaybe
visualizedascylindersoffluidspanningtwointerfaces,orby
con-nectingtothemselvesinatoroidalor‘vortexring’shape(Triticoand
Cotel,2010).Theseturbulentstructuresareasubjectofintensive
studyinthefieldofhydraulics(seeShahandTachie,2009among
others),yetverylittleinformationexistsontheeffectsofeddieson
theswimmingbehaviouroffish.Theeffectsofvortexstructureson
animalbehaviour,suchasbirdmigratorypatterns,hasalong
his-toryofresearchinaerodynamics,particularlyonsoaringbirdsthat
canmaintainflightwithoutwindflappingbytakingadvantageof
risingaircurrents(Hedenström,1993;Ristrophetal.,2011;Sapir
etal.,2011).Similarhydrodynamicresearchontheeffectsofeddies
infishswimmingperformance,ismorerecent(Pavlovetal.,2000;
Lupandin,2005;Liao,2007).Eddydiameter,vorticityand
orienta-tionrelativetothefish,asbeingprimaryvariables,wereshown
tostronglyaffectfishorientation,stabilityandswimmingspeed
(TriticoandCotel,2010).Intheirstudyontestingtheeffectsof
tur-bulenceonthestabilityandswimmingspeedofcreekchub,Tritico
andCotel(2010)foundthatstabilitychallengesonfishoccurred
for eddy diameters larger than 50–75% of the fish body total
length,andthatcriticalswimmingspeedreducedmorein
turbu-lentflowdominatedbylargehorizontaleddiesthanlargevertical
eddies.
This study was conducted with Iberian barbel (Luciobarbus
Peninsula,occurringinalmostalloftheriverbasinsinnorthernand
centralPortugal(Lobón-Cerviá,1982;Geraldesetal.,1993;Santos
etal.,2005).Themigratorypathwaysforfeedingandspawning
purposesofthisrheophilicpotamodromousspecies(i.e.
freshwa-terfishthatpreferwatercurrents),includeriverswithturbulent
flowswhichoccuroverconsiderabledistances(Smith,1991).Yet,
thereisstillverylittleinformationonhowthisspeciesrespondsto
thekinematicdescriptorsofturbulenceandtothecharacteristics
ofthetypicalvorticalstructures(eddyshape,size,vorticity,
orien-tation)thatoccurinunsteadyflows.Abetterunderstandingofthis
biological–physical interplay would provide important insights
on theinfluence of turbulence onthe swimming behaviour of
thisspeciesandotherspeciesofsimilarbiomechanicalattributes,
whichiscriticalinformationtoimprovetheefficiencyoffishways
forthesespecies,allowingthelong-termsustainabilityofsuch
pop-ulations.
Theprimaryobjectiveofthisstudywastoinvestigatetheeffects
ofturbulence(TKE,Reynoldsshearstress,TIandeddysize)onthe
swimmingbehaviouroftwo-size-classesofIberianbarbel,under
turbulent conditions createdby three different orifice
arrange-mentsinanexperimentalpool-typefishway.Itwashypothesized
thatdifferentorificearrangementsgeneratedissimilarturbulent
conditions,which impactfishswimming behaviourand
perfor-mancedifferently,influencingfishpassageefficiencyandtransit
times. Asecondary hypothesis wasthat anintra-specific
varia-tiononbehaviourcouldbeidentifiedbetweenthetwosize-classes
ofbarbel,andthiswouldlikelybedependentontheswimming
capacitytorespondtothedifferentturbulentdescriptors.Thelast
hypothesiswasthatthemagnitude,orientation,andsizeofeddies
maystronglyinfluencebehaviour,swimmingstabilityand
perfor-mancewithpossiblefishsizedifferences.
2. Materialsandmethods
2.1. Experimentalsetup
Adult Iberian barbel (N=140) were captured by means of
electrofishing(ElectrocatchInternational,SarelmodelWFC7 HV,
Wolverhampton,UK)withlowvoltage(250V)attheRiver
Sor-raia,the largesttributary ofthe RiverTagus (centralPortugal).
Sampling wasperformed duringMay 2009,since April–May is
thenaturalreproductivemigrationseasonforthisspecies(Santos
et al., 2005).Testfish ranged from15 to 35cm and according
totheirtotallength(TL)wereseparatedintotwodifferent
size-classes: small adults (N=70, 15≤(TL)<25cm, mean±SD (cm):
19.60±2.60)andlargeadults(N=70,25≤TL<35cm,mean±SD
(cm):28.95±3.46).Fishwerekeptinfilteredandaerated
hold-ingtanks(1.45m×0.70m×0.80m)ata densityof20pertank,
to recoverfrom handlingand transport stress,for at least one
week prior tothe experiments. Fishwere fed daily withpond
sticks (Tetra Pond)until 24hprior to experimentation.
Experi-mentswereconductedinafull-scalepool-typefishwayprototype
attheHydraulics andEnvironmentDepartment oftheNational
Laboratory for Civil Engineering (LNEC), in Lisbon. It consisted
of a rectangular open-channel (10.0m×1.0m×1.2m),
featur-inganupstream(2.6m×1.0m×1.2m) anda downstreamtank
(4.0m×3.0m×4.0m).Silvaetal.(2011)providemoredetailson
fishhandlingandtestequipment.
TotesthowIberianbarbelrespondtodifferentlevelsof
tur-bulence, fishwere tested during Mayto July2009 in different
hydraulicconditions(Table1)createdin anexperimental
pool-typefishwayby,varyingflowdischargeandorificearrangement:
(a)offsetorifices(i.e.orificesalternatingfromsidetosideinthe
pools);(b)straightorifices(i.e.orificesalignedononesideofthe Table
1 Summary of the hydraulics conditions associated to the three experimental designs tested: submerged orifices area (Ao ), head drop between pools ( h ), volumetric power dissipation (Pv ), pool mean water depth at 25, 50 and 80%. Number and fish size of individuals used in the experiments. Design Exp. Variables Ao (m 2) h (m) Q (l/s) Pv (W/m 3) hm (m) 25% hm (m) 50% hm (m) 80% hm (m) Small adults: 15 ≤ TL < 25 cm Large adults: 25 ≤ TL < 35 cm N Mean ± SD (cm) N Mean ± SD (cm) Offset orifices E1 0.03 0.16 38.5 37.0 0.89 0.22 0.44 0.71 10 19.07 ± 1.76 10 28.87 ± 2.59 E2 0.04 0.16 47.5 47.2 0.86 0.21 0.43 0.68 10 19.85 ± 2.49 10 28.44 ± 3.11 E3 0.05 0.16 62.7 63.1 0.85 0.21 0.42 0.68 10 19.74 ± 2.14 10 28.67 ± 2.89 Straight orifices E4 0.03 0.16 50.0 48.5 0.88 0.22 0.44 0.70 10 20.03 ± 2.12 10 26.82 ± 3.83 E5 0.04 0.16 71.8 69 0.89 0.22 0.44 0.70 10 18.35 ± 2.11 10 28.57 ± 5.31 Straight orifices with a bar E6 0.03 0.16 38.5 37.0 0.88 0.23 0.45 0.72 10 20.15 ± 2.49 10 30.61 ± 2.46 E7 0.05 0.16 62.7 63.1 0.84 0.22 0.44 0.70 10 20.18 ± 4.53 l0 30.98 ± 2.40
Fig.1. Schematicoforificearrangementstested:(a)offsetorifices;(b)straightorifices;(c)straightorificeswithabar.
pools);and (c) straightorifices witha deflectorbarprotruding
0.5bo (0.09,0.10and0.12cm)fromthepoolsideandlocatedat
0.20Lfromtheinletorifice,whereboisthewidthofthesquare
orificesandListhepoollength(Fig.1).Silvaetal.(2010)intheir
studyonthepassageefficiencyofoffsetandstraightorificesfor
Iberianbarbelinapool-typefishwayfoundthatstraightorifices
createajetflowofveryhighvelocitywhichmakesupstreamfish
movementsdifficult.Adeflectorbarbetweentwoconsecutive
ori-ficeswasusedinthepresentworkasanattempttoreducesuch
criticalhydraulicconditionsforfish.Tothisend,boththewidth
and the location of the barwere choices based onthe results
found by Alvarez (2009). This study focused on the hydraulic
characterizationofturbulenceinpool-typefishwayswith
differ-ent orifice arrangements, in which the former combination of
widthandlocationofthebarwaslikelytocreatehydraulic
condi-tionsmorefavourableforupstreamfishmovements.Experiments,
which lasted1.5h, wereconductedusing twoadultfish
simul-taneously,oneof each size-class. Fishwere randomlyremoved
fromtheholdingtanksandheldatthedownstreamtankfor
accli-mationforatleast1hpriortoexperimentation.Apanelwitha
finemeshscreenlocatedwithinthedownstreamtankprevented
fishfromenteringtheflumebeforeexperimentsbegan.Toavoid
possiblebiasin theoutcome oftheexperiments, asa resultof
learningbasedeffects(Mallen-Cooper,1994),eachfishwasused
onlyonceinthestudy.Intotal70trialswereconductedbytesting
sevendifferenthydraulicconditions(Table1)with10replicates
each.
2.2. Experimentaltrials
Experimentswereperformedwithuniformflow(i.e.identical
depthatequivalentpointswithineach ofthesixpools), witha
headdropbetweentwo consecutivepools(h)of0.16m. This
correspondstoapotentialvelocity(Vo)of1.77m/s,basedon
cal-culationfromtheformulaV0=
2gh=1.77 m/s, wheregis
theaccelerationduetogravity(9.81m/s2).Measurementsofwater
surfaceelevationsforeachexperimentwereconductedbyusinga
transparentgridattachedtotheflumesidewallwhichdetermined
meanflowdepth(hm)ineachpool.Thevelocityfieldsintheflume
werecharacterizedbymeansofaNORTEKAS3DAcousticDoppler
Velocimeter(ADV)atafrequencyof25Hz.Atthisrate,2500
instan-taneousmeasurementsofvelocitywererecordedforeachsample
point(n=48)forasamplingperiodof90s,whichisthesampling
period determinedbySilva et al.(2010)to provideconvergent
statisticsofmeanwatervelocityandturbulenceparameters.
Mea-surementsweretakenonhorizontalplanesparalleltotheflume
bottomat25%,50%and80%ofthepoolmeandepthinthe2nd
downstreampool,toensuredevelopedflowoccurred.These
mea-surementswereusedtodetermine:time-averagedvelocities,flow
patterns,fluctuatingvelocities,turbulentkineticenergy,Reynolds
shearstressandturbulenceintensity.
In this study,turbulent kineticenergy (TKE), wascalculated
accordingtoRodi(1980)as:
TKE= 1 2(u 2 rms1+u2rms2+u2rms3) with, urms=
u2=⎡
⎢
⎢
⎣
1 T T 0 u2dt⎤
⎥
⎥
⎦
1/2 (1) where,u2rms1,u2rms2andu2rms3arerespectivelytherootmeansquare
values of thefluctuating components of velocity onthex, y, z
orthogonalcoordinatesystemanduisthefluctuatingcomponent
ofvelocityatthesamplingtimeT.Toevaluateandcompare
val-uesbetweenthedifferenthydraulicconditionstested,aswellasto
developuniversalexpressions,turbulentkineticenergywas
nor-malizedusingthepotentialvelocity(
TKE/Vo).Reynolds shearstress(), whichderivesfromthethree
nor-malstresstermsintheReynolds-averaged-Navier–Stokes(RANS)
equation(TennekesandLumley,1972)wasdefinedfortheplane
XYas:
xy=−uiuj, with i/=j (2)
whereisthewaterdensity(1000kg/m3)andi,jarethetwo
com-ponentsintwo-dimensionalspace.Reynoldsshearstressforthe
XYplanewasnormalizedbyusingthepotentialvelocity(xy/Vo2).
Turbulenceintensity,aturbulentdescriptorcommonly
quanti-fiedinstudiesusingfish(TriticoandCotel,2010)wascalculatedby
usingthefollowingequation:
TI=
(1/2)(u2 rms1+u2rms2+u2rms3) U2(3)
whereUistheaveragelocalvelocity.
Tostudythepossibleinfluenceofeddiesobservedinthethree
experimentalconfigurationsonfishswimmingbehaviour,
maxi-mumlongitudinal(Lex)andtransversal(Ley)diameterofthe
eddiescreatedwithinthepoolweredetermined.Streamlines
out-linededdycontourswhich weredefinedbytracingthevelocity
vectors,andtheirrespectivelongitudinalandtransversal
diame-tersweremeasuredinthehorizontalplaneat0.25hm.Aqualitative
Fig.2. Topologyofmeanflowinthepoolforexperiments,E2,E4andE6:(a)offsetorificeconfigurationatz=0.25hmand(b)atz=0.80hm;(c)straightorificeconfigurationat z=0.25hmand(d)atz=0.80hm;(e)straightorificewithabarconfigurationatz=0.25hmand(f)atz=0.80hm.Flowfromtheorificeentersatthebottomleftofthediagram.
and transversal (Ley)horizontaleddy diameters and thetotal
lengthofthefishtested(TL)wasalsodetermined.
Flowcharacterizationwassupplementedbydirectobservations
andvideorecordingoffishmovementswithinthe2nddownstream
pooltoidentifypatternsoffishmovements.Furtherdetailsonthe
videorecordingsystemcanbefoundinSilvaetal.(2011).
2.3. Videoanalysis
Movementsoffishwererecordedandsubsequentlyanalyzed
byusingtheIVisionLabviewsoftwarefromNationalInstruments
(http://www.ni.com).Toaidvideoanalysis,a1.90mlongby1.00m
widereferencegridcontaining15contiguoussequentially
num-beredcells(each,0.38mlength×0.33mwidth)wasplacedinthe
poolwithinthecamerafieldofview.Fromthevideorecordings,fish
locationandthetimespentbyfishineachcellofthegrid(transit
time)weredeterminedandfurtheranalyzedincombinationwith
thehydraulicdata.Anindividualwasconsideredtooccupyone
cell,whenmorethanhalfofitsbodylengthwaswithinthecell’s
boundaries.Fishbehaviourwithinthepoolwasalsoanalyzedand
relatedwiththesizeofobservededdies.
2.4. Statisticalanalysis
DataweretestedfornormalitywiththeShapiro–Wilktest.
Non-parametrictestswereconductedforviolationsoftheassumption
ofnormalityandhomogeneityofvariance.Mann–WhitneyRank
Sumtestwasusedtodetectsignificantdifferencesonvelocityand
turbulencebetweenhydraulicregionscreatedineach
experimen-talconfiguration,andtosearchforsignificantdifferencesonthe
numberofsuccessfulpassagesandtherespectivetimestakenby
fishineachsize-classandconfiguration.Toassessthehypotheses
thatforeachfishsize-class,therateofsuccessandtherespective
timeofpassageweredifferentamonghydraulicconfigurations,the
Kruskal–WallisANOVAwasapplied.Possiblecorrelationsbetween
fishtransittime ineachcelland therespectivemeanvalues of
velocity,TKE,TIand Reynoldsshear stress,in each
Table2
Mann–WhitneyU-testcoefficientsobtainedwhentestingfordifferencesonthevelocity,TKE,ReynoldsshearstressandTIamonghydraulicregions(A,B,C)inthethree experimentaldesigns.
Design Experiments Regions Hydraulicvariables
v(m/s) TKE(m2/s2) −uv(N/m3) TI
Offsetorifices E1 A–B 5.11*** 5.28*** 3.83*** 0.38(ns)
E2 A–B 5.04*** 5.02*** 3.42*** −2.76**
E3 A–B 5.30*** 4.29*** 3.36*** −2.53*
Straightorifices E4 A–B 4.39*** 4.79*** 4.61*** −0.65(ns)
E5 A–B 5.22*** 5.53*** 5.44*** −1.18(ns)
Straightorificeswithabar E6 A–B 4.48*** 3.24*** 0.33(ns) 2.47*
A–C 3.81*** 2.47** −0.08(ns) 3.24** B–C −0.84(ns) −0.53(ns) −0.21(ns) 0.53(ns) E7 A–B 3.38*** 1.47(ns) −0.49(ns) −3.91*** A–C 3.58*** 3.52*** −2.08* −1.90(ns) B–C −0.70(ns) 1.41(ns) −2.32* 1.50(ns) ns,notsignificant. *SignificantinP<0.05. **SignificantinP<0.01. ***SignificantinP<0.001.
Multiple stepwise forward regressions were employed for fish
transittime asthe dependentvariableand velocity and
turbu-lentparametersasindependentvariables.Tomeetthenormality
requirementsofparametricanalysis,data werelog(x+1)
trans-formed.Durbin–Watsonstatistics(D)(DurbinandWatson,1951)
wereusedtotestforfirst-orderautocorrelationintheresidualsof
eachconfiguration.Analyseswereperformedusingdatacollected
atz=0.25hm,as,inallexperiments,fishpreferentiallymovedclose
tothebottomofthefishway(seeSection3.2.2).Thestatistical
pro-cedurewascarriedoutusingSTATISTICA(STATSOFT,INC.,2000)
forasignificancelevelof˛=0.05.
3. Results
3.1. Turbulentflow
3.1.1. Flowtopologyandvelocities
Inallexperiments,watervelocitywasfoundtobelowernear
thewatersurface(z=0.80hm)thannearthepoolfloor(z=0.25hm)
(Fig.2).Inexperimentsconductedwithoffsetorifices,nearthe
bot-tomoftheflume(z=0.25hm;Fig.2a)twodistinctareascouldbe
defined:(a)amainflow(regionA)withhighvelocities(mean±SD
(m/s):0.56±0.37;Table2)travellingbetweentwo consecutive
orifices occupying 41% of the total area of the pool (Ap) and
a recirculation region (region B) with reverse flow direction
(counter-clockwise)andsignificantlylowervelocities(mean±SD
(m/s):0.20±0.07;Table2)andb)largecounter-clockwiseeddies
whichextendthrough59%ofAp,nexttothemainflow(Fig.3a).
Nearthesurface(z=0.80hm;Fig.2b),auniformrecirculationarea
(region A)withlower velocities(mean±SD (m/s): 0.25±0.09;
Table2)couldbefound.Inexperimentsconductedwiththestraight
orificearrangementaprimaryflow(regionA)travellingdirectly
betweenthetwostraightorificeswithhighvelocities(mean±SD
(m/s):0.47±0.22;Table2;33%ofAp)andacontiguous
recircula-tionarea(regionB;67%ofAp)ofmuchlowervelocities(mean±SD
(m/s):0.16±0.06m/s;Table2)and largeeddies(Fig.3b),were
createdatthedeepestlevel(z=0.25hm;Fig.2c).Nearthesurface
(z=0.80hm)thisorificearrangementcreatedacounter-clockwise
recirculation area(region A; Fig. 2d) withnegligible velocities
(mean±SD: 0.16±0.06; Table 2). From all three experimental
designs tested at the deepestwater level (z=0.25hm), average
watervelocitywaslowerinexperimentsconductedwithstraight
orificeswithadeflectorbar(0.23m/s).Inthiscase,threedifferent
regionscouldbedistinguished:aprimaryflow(regionA;Fig.2eand
f)ofhighvelocities(mean±SD(m/s):0.36±0.22;Table2)flowing
fromtheupstreamorificetowardsthedeflectorbarandthen
turn-ingtotheoppositesidewalltowardstothedownstreamorifice,
occupyingc.40%ofAp;asecondaryregioninthecounter-clockwise
direction(regionB;Fig.2eandf)oflowervelocities(mean±SD
(m/s):0.14±0.19;Table2)locatedbetweenthemainflow and
theupstreamcross-wall,andathirdregionofclockwiserotation
(regionC;Fig.2eandf),ofverylowvelocities(mean±SD(m/s):
0.16±0.05;Table2)createdimmediatelybelowthedeflectorbar,
constrainedbetween0.42L–0.74Land0.12B–0.5Bandoccupying
c.25%ofAp.Significantdifferencesinwatervelocitywerefound
betweentheprimaryflow andregionBand C(Table2).Itwas
observedthatinthisconfigurationgeneratededdieshada
nega-tivelyskeweddistribution(Fig.3c)andtheirsizevariedthemost,
whencomparedtotheeddiesgeneratedbyoffsetandstraight
ori-fices.Depth-relatedvariationsofflowpatternwerenotfoundin
theexperimentsinvolvingstraightorificeswithabar(Fig.2eand
f).
3.1.2. Turbulence
Fig.4,givesthecontoursofnormalizedturbulentkineticenergy
forthethreeexperimentaldesignstested.TKEstronglyvariedwith
waterdepth,decreasingnearthewatersurface.Inallexperiments
thehighestvaluesofTKEwerecreatednearthebottomoftheflume
(z=0.25hm)alongtheprimaryflow(seeSection3.1.1)decreasing
intheremainingareas.SignificantdifferencesonTKEwerefound
betweenthemainflowandregionBand/orCinalltheexperiments
(Table2).The straightorifice arrangementwitha deflectorbar
seemedtohavecreatedthehighestvaluesofsqrt(TKE)(0.47Vo),
whencomparedwiththosefoundinexperimentsconductedwith
offsetandstraightorifices,0.29Voand0.32Vo,respectively(Fig.4a,
cand e).Therange forsqrt(TKE)was0.04Vo to0.47Vo forthe
straightorificearrangementwithadeflectorbar,thelargest
vari-ation observed betweenthethree configurationstested. Atthe
deepestpoollevel(z=0.80hm),sqrt(TKE)waslowestinall
con-figurations (Fig.4b,d and f).Atthis depthsqrt(TKE) variation
was similarto theexperiments conducted withoffset (0.15Vo)
andstraight(0.12Vo)orificesinwhichthehighestsqrt(TKE)
val-ues(0.32Vo)werereachednearthecross-walloppositetheorifice
(Fig.4b and d).In theexperimentswithstraight orificesanda
deflectorbar,sqrt(TKE)wasclearlyhigher(0.10Vo)upstreamthan
downstreamofthebar,decreasingtowardsregionCtovaluesofc.
Fig.3.Eddiescontoursatz=0.25hmforthethreeexperimentaldesignstested:(a)offsetorificeconfiguration,(b)straightorificeconfigurationand(c)straightorificewith abarconfiguration.Flowfromtheorificeentersattheupperleftcorner.
ContoursofnormalizedReynolds shearstressfortheXYare
plottedinFig.5.Itisclearthatthisnormalizedhorizontal
compo-nentoftheReynoldsshearstressdecreasednearthewatersurface
whereitreachednegligiblevaluesc.<0.001V2
o (Fig.5b,dandf).
Thiseffectwashighlypronouncedintheexperimentswithstraight
orifices(Fig.5candd)andstraightorificeswithabar(Fig.5eandf).
Intheexperimentswithstraightorifices,maximumabsolutevalues
ofthishydraulicvariable(0.02V2
o)werefoundatz=0.25hminthe
mainflowarea(seeSection3.1.1)anddecreasedtowardsthe
oppo-siteside-wall,wheretheyremainedalmostnegligible(Fig.5c).A
similarvariationoftheXYcomponentofReynoldsshearstresswas
alsoobservedintheexperimentswithoffsetorifices(Fig.5a).
Sig-nificantdifferencesonthishydraulicparameterwerefoundamong
thedifferenthydraulicregionscreatedinexperimentswith
off-setandstraightorifices(Table2).Theintroductionofadeflector
barbetweentwostraightorificesinducedthehighestvariations
ofthishydraulicparameteratthedeepestwaterlevel,wheretwo
distinctareasofdifferentReynoldsshearstressvaluescouldbe
dis-tinguished:(a)anareabetweentheupstreamcross-wallandthe
deflectorbar(0<L<0.20),wherethehighestvaluesofhorizontal
Reynoldsshearstresswerereached(<0.015V2
o)and(b)anarea
locatedbetweenthedeflectorbarandthedownstreamcross-wall
(L>0.20),inwhich verylowlevelsofhorizontalReynolds shear
stresswerefoundc.<0.002V2
o.
MapsofTIwerecreatedforthethreeexperimentaldesignsat
z=0.25hmandz=0.80hm(Fig.6).Foralltheexperimentaldesigns
testednearthebottomofthepool(z=0.25hm),thestraightorifice
arrangementwithadeflectorbargeneratedthehighestvaluesofTI
observedc.4.24inthemainflowarea(seeSection3.1.1).TIvalues
decreasedtowardstheremaininghydraulicareas(regionBandC)
andreachednegligiblevalues(c.<0.06),leadingtosignificant
dif-ferencesonTIamongthesehydraulicregions(Table2).Though
variationsofmaximumTIfromthebottomtothesurfacewerenot
clearatthemostsuperficiallevel(z=0.80hm),maximumTIvalues
werealsofoundtobehigherinexperimentswithstraightorifices
anda deflectorbar(4.20)whencompared toexperimentswith
offsetorstraightorificearrangements(2.51and3.04,respectively).
3.2. Fish
3.2.1. Fishpassageefficiency
Amongthethreetestedconfigurations,thehydraulicconditions
createdbytheoffsetorificearrangementresultedinasignificantly
higherrateofpassageforfishofbothsizeclasses(smallerfish:
61%;largerfish:78%)whencomparedtostraightorifices(smaller
fish: 25%;largerfish: 30%)or straight orifices witha deflector
bar(smallerfish:55%;largerfish:15%)(Kruskal–WallisANOVA:
P=0.0001). Althoughinexperiments conductedwithoffsetand
straight orifices largerfish exhibited higher passageefficiency,
size-relateddifferencesontherateofsuccesswerenotfoundto
besignificantforeachoftheseconfigurations(Mann–Whitney
U-test:offsetorifices,Z=−0.82,P=0.409;straightorifices,Z=−0.34,
P=0.726).Thisresultisincontrastwithexperimentsforstraight
orificeswithadeflectorbar.Inthiscase,arelationshipwasclearly
evident(Mann–WhitneyU-test:Z=2.61,P=0.009)asahigh
num-berofsmallerfishpassed(55%),comparedtoalownumber(15%)
ofthelargerfishwhichweresuccessful(Fig.7a).
Thetimetakenbyfishtosuccessfullynegotiatethefishway
var-iedwiththeexperimentalconfigurationandfishsize-class(Fig.7b).
Theoffsetorificearrangementseemedtobemoresuitableforboth
fishsize-classesforupstreampassage,asevidentbytheshortest
transittimes(mean±SE(min):7.18±0.83).Fishtransittimesin
experimentswithstraightorifices(mean±SE(min):7.81±2.54)
andstraightorificeswithabar(mean±SE(min):7.57±2.12)were
longerforbothsizeclasses.Smallerfishexperiencedmoredifficulty
Fig.4.DimensionlessTKEforexperiments,E2,E4andE6:(a)offsetorificeconfigurationatz=0.25hmand(b)atz=0.80hm;(c)straightorificeconfigurationatz=0.25hmand (d)atz=0.80hm;(e)straightorificewithabarconfigurationatz=0.25hmand(f)atz=0.80hm.Flowfromtheorificeentersatthebottomleftofthediagram.
orifices, leading to longer transit times (mean±SE (min):
13.2±4.61) compared to offset orifices (mean±SE (min):
7.78±1.36)andstraightorificeswithadeflectorbar(mean±SE
(min):6.18±2.13).Nevertheless,nosignificantdifferencesonthe
time ofpassageofsmallfishwere foundamong configurations
(Kruskal–Wallis ANOVA: P=0.093). In contrast, the time taken
by larger fish to successfully ascend the fishwaywas strongly
dependentonthetested configuration(Kruskal–WallisANOVA:
P=0.005).Indeed,largerfishtookmuchlongertoascendthe
fish-wayinexperiments withstraight orificesanda bar(mean±SE
(min):12.66±6.06)thanwithoffset(mean±SE(min):6.55±0.94)
orstraightorifices(mean±SE(min):3.33±0.84).Size-related
dif-ferenceswerealsofoundinexperimentswithstraightorificeswith
adeflectorbar(Fig.7b),wheresmallerfishclearlytooklesstimeto
ascendthefishway(Mann–WhitneyU-test:Z=2.33,P=0.019).
3.2.2. Fishswimmingbehaviourresponsestoturbulentflow
kinematics
Inalltheexperiments,fishwerefoundtopreferareasoflower
turbulenceasevidentbythehighest%oftime[(transittimes/total
timeinthepool)×100]spend byfishineachflowregion.Both
size classes spent the following % time in region B of each
orifice arrangement: offset (small fish: 78%; large fish: 71%);
straight(smallfish:87%;largefish:92%).Incontrast,respective
%timesspentinregionAwere:offset(smallfish:22%;largefish:
29%);straight(smallfish13%;largefish:8%).Althoughless
evi-dent, suchtrend wasalsoobserved in thestraight orificewith
a deflectorbararrangement,as shownby the%times spentin
regionB(smallfish:9%; largefish:6%)andC (smallfish:63%;
largefish:48%)comparedtothosespentinregionA(smallfish:
28%;largefish:47%).Theeffectofhydraulicsonfishswimming
behaviourwasdeterminedbasedonthecorrelationsestablished
amongtransittimesoffishineachcellofthepoolandthe
respec-tivehydraulickinematicconditions(Table3).Sincefishmovements
preferentiallyoccurrednearthebottomoftheflume,previous
anal-yses(Silvaetal.,2011,2010)includedhydraulicconditionsonly
at z=0.25hm. Configuration-related patternsbetween fish
tran-sit time and hydraulics were foundfor species size-classes. In
experimentsconductedwithoffsetorifices,fishbehaviourforboth
size-classesseemedtobeinfluencedbyhydraulicconditions,
par-ticularlyTI,TKEandReynoldsshearstress.Swimmingbehaviourof
smallerfishappearedtobestronglyinfluencedbythesevariables
(Table3).Underthehighvelocitiesandturbulencelevelscreatedby
Fig.5.DimensionlessReynoldsshearstress(XYplane)forexperiments,E2,E4andE6:(a)offsetorificeconfigurationatz=0.25hmand(b)atz=0.80hm;(c)straightorifice configurationatz=0.25hmand(d)atz=0.80hm;(e)straightorificewithabarconfigurationatz=0.25hmand(f)atz=0.80hm.Flowfromtheorificeentersatthebottomleft ofthediagram.
Table3
Spearmanranktestresultsobtainedtotestforpossiblecorrelationsbetweenthetransittimeoffishineachcellandhydraulics.
Configuration Fishsize-class Dependentvariable N Spearmanrankr P-value
Offsetorifices Smallerfish TKE(m2/s2) 20 −0.42 0.003**
|−uv|(N/m2) 20 −0.43 0.003**
TI 20 −0.45 0.001**
Largerfish |−uv|(N/m2) 20 −0.33 0.025*
Straightorifices Largerfish |−uv|(N/m2) 20 0.37 0.039*
Straight+bar orifices
Smallerfish |−uv|(N/m2) 20 0.52 0.002**
Largerfish v(m/s) 20 −0.38 0.034*
|−uv|(N/m2) 20 0.4 0.026*
Fourdifferentfactorswereanalyzed(velocity,TKE,ReynoldsshearstressandTI)butonlythosewithsignificantvalues(P<0.05,Zar,1996)areshown. *P<0.05.
**P<0.01. ***P<0.001.
theircapacitytoholdposition,driftingandbeingdisplacedtothe
downstreamadjacentpool,sonorelationbetweentransittimeof
smallerfishandtheexistinghydraulicconditionswasfoundforthis
experimentaldesign.Incontrast,behaviouroflargerfishappeared
tobesignificantlyinfluencedbyReynoldsshearstress(Spearman
rankcorrelation:r=0.37,P=0.039)(Table3).Thishydraulic
param-eterwasalsocorrelatedwithfishtransittimeforbothsize-classes
Fig.6. Turbulenceintensityforexperiments,E2,E4andE6:(a)offsetorificeconfigurationatz=0.25hmand(b)atz=0.80hm;(c)straightorificeconfigurationatz=0.25hm and(d)atz=0.80hm;(e)straightorificewithabarconfigurationatz=0.25hmand(f)atz=0.80hm.Flowfromtheorificeentersatthebottomleftofthediagram.
Table4
Variablesenteredinmodelofforwardstepwiseregressionanalysisexplainingfish’stransittimeofbothsize-classesinthetesteddesigns.
Design Variable F-test r2 D
Offsetorifices Smallerfish 1.86 TI 7.77** 0.15 |−uv| 4.96* 0.24 Largerfish 1.43 |−uv| 4.66* 0.09 Straightorifices Largerfish 1.25 |−uv| 7.06* 0.2
Fourdifferentfactorswereanalyzed(velocity,TKE,ReynoldsshearstressandTI)butonlythosewithsignificantvalues(P<0.05)areshown.F,teststatistic;r2,coefficientof determination;D,Durbin–Watsonstatistics.
*P<0.05. **P<0.01. ***P<0.001.
rankcorrelation:smallfish,r=0.52,P=0.002;largefish,r=0.40,
P=0.026).Thissuggeststhatinallexperimentaldesigns,Reynolds
shearstressmayhavebeenthemostinfluentialhydraulicvariable
onfishbehaviour.Indeed,fromthehydraulicvariablesconsidered,
thiswasthemostimportantparameterwiththemostsignificant
resultsinexplainingvariationinfishtransittime(Table4).Whereas
bothReynoldsshearstressandTIexplained39%ofthevariation
intransittimeforsmallerfishinexperimentswithoffsetorifices,
Reynoldsshearstresswastheonlyfactorsignificantlycorrelated
Fig. 7.(a) Fish passage success (%); (b) mean successful passage time (mean±standarderror)(min)forsmalladults(15≤TL<25cm)andlargeadults (25≤TL<35cm);barbelacrossallexperiments.Offsetorifices(E1,E2,E3),straight orifices(E4andE5)andstraightorificeswithabar(E6andE7).
andstraightorifices(20%).Therewasnosignificant
autocorrela-tion(0.52<D<2.01, ˛=0.05)intheresiduals ofeachregression
(Table4).
3.2.3. Fishresponsetoeddysize
Theratiobetweenmaximumlongitudinal(Lex)and
transver-sal(Ley)horizontaleddydiametersandthetotallengthofthefish
tested(TL)isplottedinFig.8.Thedistributionofeddysizeswas
negativelyskewed,withahighernumberoflargereddiesandfew
smalleddies,particularlyinexperimentswithoffsetandstraight
orifices(Fig.3andTable5).Underthesetwoexperimental
condi-tionsthesizeofthecreatededdies,inmostcases,washigherthan
fishlengthforbothfish-sizes(Lex/TL>1andLey/TL>1)(Fig.8a
andb).Inthestraightorificearrangement,smallerfishwerefound
todrifttowardsthemainflowandbeendraggedtothe
immedi-atelydownstreampool.Incontrast,thelargerfish,whichhavea
higherswimmingcapacitythanthesmallerones(Silvaetal.,2011),
wereabletore-establishtheirorientationandsuccessfullyascend
thefishway.Nevertheless,underthestraightorificearrangement,
Fig.8.Relationbetweenmaximumlongitudinal(Lex)andtransversal(Ley) diam-etersoftheeddyandthetotallengthofthefish(TL)forsmall(TLsf)(+)andlarge fish(TLlf)()inoffsetorifices(inred),straightorifices(inblue)andstraightorifices withabar(ingreen).(Forinterpretationofthereferencestocolourinthisfigure legend,thereaderisreferredtothewebversionofthisarticle.)
somelargerspecimenswerenotabletorecovertheirorientation
in thepool and subsequently fell backdownstream. In
experi-mentsconductedwithstraightorificeswithadeflectorbar,where
awider rangeofeddieswithdifferentsizeswascreated(Fig.3
andTable5),fishswimmingbehaviourwaslikelyaffectedmoreby
thesehydraulicstructures.Herein,strongsize-relateddifferences
onswimmingbehaviouralresponsestoeddysizewereobserved
betweensmallerandlargerfish.Whenfacingeddieslargerthan
theirsize (Lex/TLsf>1 and Ley/TLsf>1) (Fig.8c), smallerfish
wereobservedtodisplayoneoftwobehaviours:(a)swimsteadily
throughtheeddies,exhibitingaswimbehaviourcharacterizedby
largerlateralbodyamplitudesandcurvaturesthanswimmingin
areaswithveryhighvelocities(mainflowregion);(b)become
dis-orientatedbutrapidlyadjusttheirbodystabilityandfindingtheir
routetosuccessfullyascendthefishway.Someoftheeddies
gener-atedwereofsimilarsizetothesizeofthelargerfish(Lex/TLlf≈1
andLey/TLlf≈1)(Table5andFig.8d).Undersuchconditionsfish
wereseen tospread theirpectoralfins inan attemptto
stabi-lizetheirbodyposition.Theseresponsesincreasedthehydraulic
resistanceoftheirbody,therebydecreasingfishswimming
perfor-mance.Thisisevidentfromthelongertransittimesestimatedto
ascendthefishwaywhenstraightorificeswithadeflectorbarwere
used,comparedtoshortertransittimeswhenstraightoroffset
ori-ficearrangementswereemployed(Fig.7b).Nevertheless,mostof
thefishweredraggedtothedownstreampoolasaresultofthehigh
magnitudesoftheforcesactingontheirbodyandtherespective
energycostsassociatedtorestorestability.Inareaswitheddysize
smallerthanfishlength(Lex/TLlf<1andLey/TLlf<1;regionBin
Fig.8d),largerfishwereobservedtoswimsteadilythrougheddies
towardstheupstreamorifice.Whenfacingareaswitheddiesbigger
thantheirsize(Lex/TLlf>1andLey/TLlf>1;Fig.8e),somelarger
fishgotdisorientatedandfellbacktothedownstreampool,while
otherswerefoundtorecovertheirrouteandsuccessfullymove
forward.
4. Discussion
This study focused on the analysis of the effects of
hydro-dynamicturbulentflowkinematicsontheswimmingbehaviour
Table5
Characterizationoftheeddiescreatedinthethreeexperimentaldesigntested:numberofidentification(#),location,maximumlongitudinal(Lex)andtransversal(Ley) diametersandarea(ae)oftheeddy.Ratiobetweenlongitudinalandtransversaldiameterandsmaller(TLsf)andlargerfish(TLlf)totallength.
Design # Zone Lex(m) Ley(m) ae(m2) Smallerfish(sf) Largerfish(lf)
Offsetorifices 1 RegionB 0.45 0.58 0.82 Ley/TLsf>1
Lex/TLsf>1 Ley/TLlf>1 Lex/TLlf>1 2 RegionB 0.64 0.78 1.57 3 RegionB 1.26 0.83 3.28 4 RegionB 1.47 0.85 3.92 5 RegionB 1.72 0.86 4.64
Straightorifices 1 RegionB 0.29 0.33 0.30 Ley/TLsf>1
Lex/TLsf>1 Ley/TLlf>1 Lex/TLlf>1 2 RegionB 0.78 0.62 1.52 3 RegionB 1.18 0.76 2.82 4 RegionB 1.61 0.80 4.04 5 RegionB 1.81 0.80 4.55 Straightorifices withabar 1 RegionB 0.23 0.21 0.15 Ley/TLsf>1 Lex/TLsf>1 1<Ley/TLlf≤1 1<Lex/TLlf≤1 2 RegionB 0.25 0.44 0.35 3 RegionA 0.50 0.40 0.63 4 RegionA 0.82 0.70 1.80 5 RegionA 1.00 0.83 2.61 6 RegionA 1.13 0.90 3.19 7 RegionC 0.40 0.30 0.38
turbulentconditionsweregeneratedbythethreedifferentorifice
arrangements(offset,straightandstraightorificeswithadeflector
bar)ina fullscaleexperimentalpool-typefishway.Fish
perfor-mancewithinthefishwaywassimilartothebehaviourcommonly
displayedbythisspeciesundernaturalconditions,asfishmoved
primarilynearthebottomoftheflume(z=0.25hm).Thisis
typi-calbehaviourexhibitedbythisspecieswhenmovingfreelyand
volitionallyinitsnaturalhabitat,asaresultofvertical
segrega-tionassociated withtrophic (bottomfeeders) and reproductive
adaptations(epi-benthicfish,i.e.fishthatspawningravelbeds)
(Collares-Pereiraetal.,1995;Martínez-CapelandGarcíadeJalón,
1999).Furthermore,thehighest passagesuccessand the
corre-spondinglowertransittimesobservedinexperimentswiththe
offset orificearrangement occurred undervelocities commonly
encountered by Iberian barbel when exploring for feedingand
spawningpurposes (Martínez-Capelet al., 2009). Thisfits well
withthephysiologicalstrategyofthisspeciestominimizeenergy
expenditureinmaintainingpositioninthewatercolumnbecause
oftheirlimitedswimmingability(Doadrio,2001).Likewise,the
strongtrendbyfishtoavoidareasofhighturbulencelevelswas
demonstrated bythehighnegativecorrelations foundbetween
thetransittimeoffishineachcellandtherespectivevaluesof
TKE,ReynoldsshearstressandTIinalltheexperimental
configu-rationstested.Avoidinghighturbulenceisabehaviourcommonly
observedwiththisandmanyotherfishspeciesandemergesasan
attempttominimizeenergyexpendituretosustainstability.
Thebehaviouralresponseoffishtothehydrodynamic
hetero-geneityofflowsinaquaticsystems,suchasturbulenceresulting
from the interactions of gravity and wind-driven currents in
water–airandwater–waterinterfaces,aswellaswatermovingpast
physicalstructures(Webb,2004),stronglydependsonthespatial
andtemporalperturbationmagnitudeoftheforcesactingonthe
fish,thetimeoffishexposure,species,lifestageandindividualsize
(Lupandin,2005;Coteletal.,2006).Resultsfromthepresentstudy
attesttothisrelationship,asReynoldsshearstresswasthemain
factoramongallthehydraulicvariablesanalyzed,whichexplained
passagesuccessandfishtransittimesinexperimentswithoffset
andstraightorificearrangements.Shearstresslevelsthatarelow
enoughtoavoidinjurymaystillleadtofishdisorientationand/or
displacement,asaresultoftheoverlapofdragtothrustforces
andthehighenergyexpenditurerequiredtogeneratethrust(Odeh
etal.,2002;Silvaetal.,2011).Odehetal.(2002)determinedthat
rainbowtrout(Onchorhynchusmykiss),Atlanticsalmon(S.salar)
andhybridbass(Moronesaxatilis×Moronechrysops),exposedto
levelsofReynoldsshearstressabove50N/m2foraperiodof10min
may sufferinjuries but notmortality.Cadaet al. (1999)report
shearstressvalues≥700N/m2causeinjuriesormortalitiesinfish.
In thepresent experimentsReynolds shearstressvalues varied
from0.02N/m2 to73.4N/m2.Asexpectedfishdamagewasnot
witnessedbutfishdisorientationanddisplacementwasobserved.
Althoughtheeffectsofsingleturbulencedescriptors(TI,TKE,
or Reynoldsshearstress)havebeenstudied(Odeh etal.,2002;
Silva et al., 2011), these may yield only partial effects on the
swimmingcapacityandbehaviouroffish.Considerationofother
ormultipleturbulencedescriptorsmayprovideamorethorough
understandingoftheeffectsofhydrodynamicsonfishbehaviour
andmovements.Thebehaviouroffishinexperimentsconducted
withstraightorificeswithabarclearlyatteststothisconcept,as
noneof theabovelistedsingledescriptorsappeared toexplain
fishbehaviourinthisexperimentaldesign,pointingtowards
addi-tionalhydraulicmechanisms.Infact,whenswimminginturbulent
flows,fishareexposedtoacomplexsystemofforcesactingupon
theirbodythatcancausetranslationaland/orrotational
displace-ments(TriticoandCotel,2010;Liao,2007).Thismayresultfrom
theactionoftypicalvorticalstructurescreatedinturbulentflows
thatareknowntoplayasignificantroleinfluidflowphenomena
asmomentum,massandheattransfer(Pope,2000).Thus,thehigh
levelsofturbulencefoundintheoffsetwithabarexperimental
setup(seeSection3.1.2),ledtothehypothesisthatthedistribution
ofeddysizescouldbetheprimaryturbulentvariableaffectingfish
swimmingbehaviour.
As theentire length of theeddy approaches thefish width
forhorizontaleddiesorthefishdepthforverticaleddies(Tritico
and Cotel,2010), thereisarobustcorrelationbetweenfishsize
andturbulencescale,asclearlyevidentthroughthesize-related
behaviouraldifferencesfoundinthethreeexperimental
configu-rationswherefishfacededdiesofdifferentsize.Ifaneddyislarger
thanthetotallengthofafish,itsbalanceshouldnotbeaffected,
although fishmayexperiencedisorientationand even
displace-ment(Lupandin,2005;TriticoandCotel,2010;WebbandCotel,
2010),asobservedinfishofbothsize-classesinexperiments
con-ductedwithoffsetandstraightorifices.Nevertheless,ifaneddy
sizeismuchlargerthanfishtotallength,fishorientationmight
notbedisturbed,asobservedinsomesmallfishinexperiments
withstraightorificeswithabar.Inaddition,whenfacingeddies
smallerthantheirsize,fishmayswimsteadilythroughthem(Liao,
2007).Thisbehaviourisacomplexphenomenonthatresultsfrom
processes.Suchabilityallowsfishtoexploreturbulentareasand
greatlyenhancepropulsiveefficiencybyextractingenergyfrom
eddies,thusdecreasingtheenergyexpenditurerequiredto
gen-eratethrust(Liao,2007).Thisinterestingcapacityofanimalsto
reducelocomotoryenergycostsiswellknowninsoaring
migra-torybirds(Hedenström,1993)andinsects(Ristrophetal.,2011)
whichusetheenergyofaircurrentstopropeltheminturbulent
air.Thesamestrategyisusedbyfishwhileswimminginturbulent
flow,whichmightexplainthebehaviourexhibitedbylargerfish
whilesteadilyswimmingthroughsmallereddiesinexperiments
involvingstraightorificeswithabar.Incontrast,bothstabilityand
swimmingperformanceoffishareknowntobestronglyaffected
byeddykinematicsofsizesimilartotheirbodylength(Triticoand
Cotel,2010).
Itisknownthatfishabilitytoholdpositionresultsfromtheir
facultytogenerateenoughthrusttobalancedragforces(Webb,
1988),whichimpliesacomplexinterplaybetweenbiomechanic
andphysiologicalprocesses.Severalauthorshavepointedtofin
activityasoneofthemainmechanismsthatpermitsfishto
main-tainstabilityincomplexflows(BiolyandMagnan,2002;Webband
Cotel,2010).Pectoralfinscanbeparticularlyeffectiveandtheir
movementscanvarywidelyduringswimminginturbulentflows
(Liao, 2007).The observed spread ofthe pectoralfinsof larger
fish,inareaswitheddies ofsimilarsizetotheirbody lengthin
experimentsconductedwithstraightorificeswithadeflectorbar,
clearlyillustratedthatfinsprovideanimportantmechanismfor
fishtoself-correctandre-establishstabilityduringswimmingand
manoeuvring(WebbandCotel, 2010).Thisbehaviouris usually
associatedwithadecreaseinfishswimmingperformancefroman
increasein hydraulicresistance(TriticoandCotel,2010),which
mayexplainthehigherpassagetimesfoundforlargerfishinthis
experimentaldesign.Undersuchhydraulicconditionsfishcanalso
experienceangularrotationsoftheirbody,whichstronglydepend
oneddyorientation(verticalorhorizontal),andiscommonly
asso-ciatedtoanincreaseinoxygenconsumption(TriticoandCotel,
2010)whichalsodecreasesfishswimmingperformance.Itis
pos-tulatedthattheeffectsofeddyorientationonfishvaryaccordingto
speciesmorphologyandlocomotion(Liao,2007;TriticoandCotel,
2010;WebbandCotel,2010).It isalsobelieved thathorizontal
eddiesmightinducestrongerimpactsonfishswimmingcapacity
whencomparedtoverticaleddies,asdemonstratedbyTriticoand
Cotel(2010).Thestabilizingcontrolsystemoffishwouldbe
bet-tersuitedforcounteringyawingratherthanpitchingperturbations
(Weihs,2002;Webb,2006).
Inthepresentresearch,itisclearthatincertaincircumstances,
turbulencemaybedeemedafeatureofhydrodynamic
environ-mentsthatisabenefitratherthana constraint.Furthermore,it
wasobservedthatfishhadatendencytopreferentiallyexplore
areasoflowlevelsofturbulenceandwereclearlyabletodetect
specificrangesofturbulence.Thisreveals,notsurprisingly,that
theycanbequitesensitivetoturbulentcues(Pavlovetal.,2000;
Liao,2007),whichcaninfluencefishbehaviouralroutinesand
habi-tatchoices(Webb,2002).Theeffectsofcomplexunsteadyflows
onfishswimmingperformanceandbehaviourhasrecentlybeen
investigated (Triantafyllouet al.,2002; Liao et al.,2003; Smith
andBrannon,2005).Yet,thereisastronggapbetweenthespatial
and temporalresolution offlows atwhichhydraulicsare
mod-elled and at which fish respond(Liao, 2007; Webb and Cotel,
2010).Although,hydrodynamicshasalonghistoryandisa
well-studiedsubject(TennekesandLumley,1972;Pope,2000;Shahand
Tachie,2009),thenatureandcharacterizationofturbulentflow
structuresisstilllacking.Thus,thereisacriticalneedto
compre-hendandevaluatethehydrodynamicfeaturesofturbulentflows
thatwillleadtoabetterunderstandingonhowturbulencemight
impactthefish’sswimming capacityand behaviour.Theuseof
physicalmodelsand advanced hydrometricsystems, aswellas
computationalfluiddynamics(CFD)areneededtoallowamore
comprehensiveunderstandingofphysicalphenomena,aswellas
predictand analyzethelevelsofturbulence in unsteadyflows.
Equally,technologicaladvancementsthatallowfishobservation
andreal-timemeasurementsofphysiologicalcostsofswimming
arenecessarytounderstandthebiologicalinfluenceofturbulent
flowsonfish behaviour.Electromyogram(EMG)studies, which
havebeenshowntoadequatelymonitorthephysiological
swim-mingeffortoffish(Mateusetal.,2008),associatedwithunderwater
video verification of swimming kinematics, may help to
iden-tifyfishpreference/responsetodifferentlevelsandstructuresof
turbulence.Thus,researchcorrelatingmeasurementsofflow
kine-matics and analyses of the vortical turbulent structures (eddy
size, strength, orientation and vorticity) with metabolic
swim-mingcosts, to better understandthe impact of turbulence on
fishbehaviouris needed.Todate, datafromstudiesconducted
inlaboratorysettingshaveconclusivelycontributedtoimproved
understandingbetweenflowhydrodynamicsandfishswimming
behaviour,stressing theaccuracy ofsuchresearchefforts.Asis
evidentfromthisstudy,similarityexistsbetweenfishbehaviour
observedinexperimentsandundercommonlyfoundnatural
con-ditionsforthesamespecies.Hence,comprehensiveecohydraulic
laboratorystudieswhichincludebothfishobservationsand
hydro-dynamics,whereconfoundingvariablesarepossibletocontroland
manipulate,arelikelytobeagoodapproachtoincreaseknowledge
inthisarea.
Theresultsshowthatstationholding,abilitytomaintain
pos-ture,andswimmingbehaviourarestronglyaffectedbyturbulence,
particularlybyReynoldsshearstressandthepresenceof
turbu-lenteddiesincloseproximitytofish.Althoughbiologicaltestingis
neededandexperimentsareunderway,strategicallyplacedand
properly sized artificialelements, such as blocks, hemispheres
or cylinders (Shamloo et al., 2001; Katopodis, 2002) within a
fishwaymaygeneratefavourablehydraulicconditions,including
appropriatelevelsofshearstress,aswellaseddyorientationand
size.Theplacementofmorenaturalsubstratesmayalsoimprove
migrationopportunitiesfordifferentspeciesinfishwaysorriver
restorationprojects.Suchsubstrateswouldbeexpectedto
gen-erate more nature-like areas of different levels of turbulence,
as wellas eddy sizes and orientations, which fish may useto
volitionallymigrateovera rangeof hydraulicconditions.These
conditionsmayallowfishtominimizeswimmingenergeticcosts
byeithertakingadvantageoftheenergyassociatedwitheddies
forpropulsionorusingareasoflowturbulencetorest.
Further-more,addingartificialormorenaturalsubstratesmaybeavery
lowcostalternative,withlowmaintenancerequirements,aslong
asitissecuredfromwashing-out.Thefindingsofthisstudyare
believedtoprovidevaluableinsightsonswimmingbehavioural
responsesofepi-benthiccyprinidstothehydrodynamic
kinemat-icsand vortical structuresof turbulent flows.Nevertheless,the
effects ofturbulence onfish swimmingbehaviourwere shown
tobeimportant,thusaccurateknowledgeofitseffectson
differ-entfishspeciesisessentialtoimprovefuturefishwayandstream
restorationprojects.
5. Conclusions
Theinfluenceofturbulenceandflowkinematicsonthe
swim-mingbehaviourofa potamodromouscyprinidwasstudiedin a
laboratoryecohydraulicflume.Fishresponseswereobservedand
associatedwithturbulencevariablescreatedinapool-type
fish-waybythree orificearrangements(offset,straight and straight
andflowkinematics(velocity,turbulentkineticenergy,Reynolds
shearstressandturbulenceintensity)werecharacterizedusinga
3DADVandtheconsequentresponsesonswimmingbehaviourand
upstreammovementsofIberialbarbeloftwodifferentsize-classes
wereanalyzed.Althoughsize-related fishbehaviouralresponses
tohydraulics werefound, Reynolds shear stressand eddy size
appearedasthemainfactorsexplainingfishswimmingbehaviour.
FishwerefoundtoavoidareasofhighReynoldsshearstressand
whenfacingeddiesofsimilarsizetotheirtotallength,experienced
disorientationandlossofstability,leadingtohighlevelsofenergy
expenditure,lowratesofpassagesuccessandlongertransittimes.
TheresultsshowthatReynoldsshearstressandeddysizeboth
ofwhicharestronglydependentonfishwaypoolgeometric
fea-tures,couldbeabarrierforupstreamfishmigration.Minimizing
fishdisorientationandlossofstabilityattributedtoeddiesmay
increasefishpassagesuccess,efficiencyandshortentransittimes.
Ofthethree orificearrangements,theoffsetconfiguration
con-trolledReynolds shearstressand eddysizebest,producing the
mostfavourablehydraulicconditions,andwasthemostefficient
forpassageofbothsizesofbarbel.Introducingabarontheside
ofthepoolhadapositiveeffectonpassageofsmallerbarbelbut
anegativeeffectonthelargerones.Abetterunderstandingofthe
relationshipbetweenfishswimmingbehaviourandhydraulic
con-ditionscouldresultfromtheability tomeasure andaccurately
modelturbulence,improvingbothspatialandtemporalresolution
atwhichhydraulicsareanalyzedandatwhichfishareobserved
torespond.Computermodelscouldbeagoodapproachtomore
faithfullyreproducedifferentflowconditionsand predictlevels
ofturbulence.Nonetheless,suchmodelsshouldfirstbeverified
toensuretheircapabilityandaccuracy.Ecohydrauliclaboratory
studiesfocusingonturbulenceandbiomechanismswhich
differ-entspeciesandsizescanusetoexploitunsteadyflowcouldprovide
valuableinsights.Suchinsightsandinformationrelatetofish
pas-sageeffectivenessandareimportantforthesustainabilityoffish
populations.
Acknowledgements
WethankthestaffoftheNationalLaboratoryforCivil
Engi-neering (LNEC), Lisbon, for the valuable assistance during this
study.ThanksareextendedtoTeresaAlvarezandPauloBranco
fortheirsupportintheexperimentalandfieldwork.Fundingfor
thestudywasprovidedbytheTechnologyandScienceFoundation
(FCT)intheformofaPhDscholarshipundertakenbyAnaT.Silva
(BD/19859/2004).
References
Alvarez,T.P.,2009.Caracterizac¸ãodaturbulênciaempassagensparapeixespor bacias sucessivascom escoamentoatravés de orifíciosde fundo.Tese de MestradoemEngenhariaCivil,InstitutoSuperiorTécnico,Lisboa.
Arthington,A.H.,Welcomme,R.L.,1995.Theconditionoflargeriverssystemsof theworld.In:Armantrout,N.B.,WolotiraJr.,R.J.(Eds.),ConditionoftheWorld’s AquaticHabitats.WorldFisheriesCongress,SciencePublishers,Lebanon,New Hampshire,pp.44–75.
Baras,E.,Lambert,H.,Philippart,J.C.,1994.Acomprehensiveassessmentofthe fail-ureofBarbusbarbusspawingmigrationsthroughafishpassinthecanalized RiverMeuse(Belgium).Aquat.LivingResour.7,181–189.
Bell,M.C.,1986.FisheriesHandbookofEngineeringRequirementsandBiological Criteria.FishEng.Res.Prog.U.S.ArmyCorpsofEng.,NorthPacificDiv.,Portland, OR,p.290.
Bioly,P.,Magnan,P.,2002.Relationshipbetweenindividualvariationin morpho-logicalcharactersandswimmingcostsinbrookchar(Salvelinusfontinalis)and yellowperch(Percaflavescens).J.Exp.Biol.205,1031–1036.
Cada,G.F.,Carlson,T.,Ferguson,J.,Richmond,M.,Sale,M.,1999.Exploringtheroleof shearstressandsevereturbulenceindownstreamfishpassage.In:Brookshier, P.A.(Ed.),Proceedingsofwaterpower’99.Hydro’sFuture:Technology,Markets, andPolicy.AmericanSocietyofCivilEngineers,Reston,VA,10pp.
Clay,C.H.,1995.DesignofFishwaysandOtherFishFacilities,2nded.Lewis Publish-ers,BocaRaton,LA,248pp.
Collares-Pereira,M.J.,Magalhães,M.F.,Geraldes,A.M.,Coelho,M.M.,1995. Ripar-ianecotonesandspatialvariationoffishassemblagesinPortugueselowland streams.Hydrobiologia303,93–101.
Cotel,A.J.,Webb,P.W.,Tritico,H.,2006.Dobrowntroutchooselocationswithreduce turbulence?Trans.Am.Fish.Soc.135,610–619.
Cowx,I.G.,Welcomme,R.L.,1998.Rehabilitationofriversforfish.FishingNewBooks, Oxford.
Doadrio,I.,2001.AtlasylibrorojodelospecescontinentalesdeEspa ˜na.Museo NacionaldeCienciasNaturales,Madrid.
Durbin,J.,Watson,G.S.,1951.Testingforserialcorrelationinleastsquaresregression II.Biometrika38,159–178.
Ead,S.A.,Katopodis,C.,Sikora,G.J.,Rajaratnam,N.,2004.Flowregimesandstructure inpoolandfishways.J.Environ.Eng.Sci.3,379–390.
Enders,E.C.,Boisclair,D.,Roy,A.G.,2003.Theeffectofturbulenceonthecostof swimmingforjuvenileAtlanticsalmon(Salmosalar).Can.J.Fish.Aquat.Sci.60, 1149–1160.
Enders,E.C.,Boisclair,D.,Roy,A.G.,2005.Amodelofthetotalswimmingcosts inturbulentflowforAtlanticsalmon(Salmosalar).Can.J.Fish.Aquat.Sci.62, 1079–1089.
FAO/DVWK,2002.FishPasses—Design,DimensionsandMonitoring.FAO,Rome. Geraldes,A.M.,Coelho,M.M.,Collares-Pereira,A.M.,1993.Régimealimentaireet
croissancedubarbeauibérique.Barbusbocagei(Steind1865)danslaRivère Sorraia(Portugal).Cahiersd’Ethologie13(2),179–180.
Gowans,A.D.,Armstrong,J.D.,Priede,I.G.,McKelvey,S.,2003.MovementsofAtlantic salmonmigratingupstreamthroughafish-passcomplexinScotland.Ecol. Freshw.Fish.12,177–189.
Hedenström,A.,1993.Migrationbysoaringorflappingflightinbirds:the rel-ativeimportanceofenergycostandspeed.Phil.Trans.R.Soc.Lond.B342, 353–361.
Katopodis, C.,2002.Nature-mimickingfishways:conceptsandpractical appli-cations. Keynote lecture. In:Proceedings of the2nd NordicInternational SymposiumonFreshwaterFishMigrationandFishPassage,Evaluationand Development,20–22September2001,Reykjavik,Iceland.Environmental Sci-encesandTechnology,UniversityofOulu,Finland,pp.87–93.
Katopodis,C.,2005.Developingatoolkitforfishpassage,ecologicalflow manage-mentandfishhabitatworks.J.Hydraul.Res.43,451–467.
Kirkbride,A.D.,1993.ObservationsoftheInfluenceofBedRoughnessonTurbulence StructureinDepthLimitedFlowsOverGravelBeds.Turbulence:Perspectiveson FlowandSedimentTransport.Wiley,Chichester,UK.
Knaepkens,G.,Maerten,E.,Eens,M.,2007.Performanceofapool-and-weirfishpass forsmallbottom-dwellingfreshwaterfishspeciesinaregulatedlowlandriver. Anim.Biol.57(4),423–432.
Liao,J.C.,Beal,D.N.,Lauder,G.V.,Triantafyllou,M.S.,2003.Fishexploitingvortices decreasemuscleactivity.Science302,1566–1569.
Liao,J.C.,2007.Areviewoffishswimmingmechanicsandbehaviorinalteredflows. Phil.Trans.R.Soc.B362,1973–1993.
Lobón-Cerviá, J.,1982. Populationanalysisofthe Iberian nose(Chondrostoma polylepisStein,1865)intheJaramaRiver.VieMilieu32,139–148.
Lucas,M.C.,Frear,P.A.,1997.Effectsofaflow-gaugingweironthemigratory behaviourofadultbarbel,ariverinecyprinid.J.FishBiol.50,382–396. Lupandin,A.I.,2005.Effectofflowturbulenceonswimmingspeedoffish.Biol.Bull.
32,461–466.
Mallen-Cooper,M., 1994.Swimmingability ofadultgolden perch, Macquaria ambigua (Percichthyidae), andadult silver perch, Bidyanus bidyanus (Ter-aponidae),inanexperimentalvertical-slotfishwayaust.J.Mar.FreshwaterRes. 45,191–198.
Martínez-Capel,F.,GarcíadeJalón,D.,Werenitzkt,D.,Baeza,D.,Rodilla-alamá,M., 2009.MicrohabitatusebythreeendemicIberiancyprinidsinMediterranean rivers(TagusRiverbasin,Spain).Fish.Manag.Ecol.16,52–60.
Martínez-Capel,F.,GarcíadeJalón,D.,1999.Desarrollodecurvasdepreferencia demicrohabitatparaLeuciscuspyrenaicusyBarbusbocageiporbuceoenelrio jarama(CuencadelTajo).Limnetica17,71–83.
Mateus,C.S., Quintella,B.R.,Almeida, P.R.,2008.Thecriticalswimmingspeed ofIberianbarbelBarbusbocageiinrelationtosizeandsex.J.Fish.Biol.73, 1783–1789.
Noonan,M.J.,Grant,J.W.A.,Jackson,C.D.,2011.Aquantitativeassessmentoffish pas-sageefficiency.FishFish.,http://dx.doi.org/10.1111/j.1467-2979.2011.00445.x. Odeh,M.,Noreika,J.F.,Haro,A.,Maynard,A.,Castro-Santos,T.,2002.Evaluationof theEffectsofTurbulenceontheBehaviorofMigratoryFish.BonnevillePower Administration,Portland,OR,FinalReporttotheBonnevillePower Administra-tion,Contract00000022,Project200005700.
Pavlov,D.S.,Lupandin,A.I.,Skorobogatov,M.A.,2000.Theeffectsofflowturbulence onthebehavioranddistributionoffish.J.Ichthyol.40,S232–S261.
Pavlov,D.S.,Skorobagatov,M.A.,Shtaf,L.G.,1982.Thecriticalcurrentvelocityoffish andthedegreeofflowturbulence.Rep.USSRAcad.Sci.267,1019–1021. Pope,S.B.,2000.TurbulentFlows.CambridgeUniversityPress.
Przybilla,A.,Kunze,S.,Rudert,A.,Bleckmann,H.,Brucker,C.,2010.Entrainingin trout:abehaviouralandhydrodynamicanalysis.J.Exp.Biol.213,2976–2986. Ristroph,L.,Bergou,A.T.,Guckenheimer,J.,Wang,Z.J.,Cohen,I.,2011.Paddlingmode
offorwardflightininsects.Phys.Rev.Lett.106,178103.
Rodi,W.,1980.TurbulenceModelsandtheirApplicationinHydraulics.IAHR,Delft, theNetherlands(IAHRMonograph).
Santos,J.M.,Ferreira,M.T.,Godinho,F.N.,Bochechas,J.,2005.Efficacyofa nature-like bypass channel in a Portuguese lowland river. J. Appl.Ichthyol. 21, 381–388.
Sapir,N.,Horvitz,N.,Wilkelski,M.,Avissar,R.,Mahrer,Y.,Nathan,R.,2011. Migra-tionbysoaringorflapping:numericalatmosphericsimulationsrevealthat turbulencekineticenergydictatesbee-eaterflightmode.Proc.Biol.Sci.278, 3380–3386,http://dx.doi.org/10.1098/rspb.2011.0358.
Saunders,D.L.,Meeuwig,J.J.,Vicent,A.C.J.,2002.Freshwaterprotectedareas: strate-giesforconservation.Conserv.Biol.16,30–41.
Shah,M.K.,Tachie,M.F.,2009.Properorthogonaldecompositionofpressuregradient flowsdownstreamofatransversesquarerib.AIAAJ.47(11),2616–2631. Shamloo,H.,Rajaratnam,N.,Katopodis,C.,2001.Hydraulicsofsimplehabitat
struc-tures.IAHRJ.Hydraul.Res.39(4),351–366.
Silva,A.T.,Santos,J.M.,Ferreira,M.T.,Pinheiro,A.N.,Katopodis,C.,2010.Passage effi-ciencyofoffsetandstraightorificesforupstreammovementsofIberianbarbel inapool-typefishway.RiverRes.Appl.,http://dx.doi.org/10.1002/rra.1465. Silva,A.T.,Santos,J.M.,Ferreira,M.T.,Pinheiro,A.N.,Katopodis,C.,2011.Effectsof
watervelocityandturbulenceonthebehaviourofIberianbarbell(Luciobarbus bocagei,Steindachner1864)inanexperimentalpool-typefishway.RiverRes. Appl.27,360–373.
Smith,R.J.F.,1991.Socialbehaviour,homingandmigration.In:Winfields,I.J.,Nelson, J.S.(Eds.),CyprinidFishes—Systematics,BiologyandExploitation.Chapman& Hall,London,pp.509–529.
Smith,D.L.,Brannon,E.L.,2005.Responseofjuvenilerainbowtrouttoturbulence producedbyprismatoidalshapes.Trans.Am.Fish.Soc.134,741–753. Smith,D.L.,Brannon,E.L.,Shafii,B.,Odeh,M.,2006.Useoftheaverageandfluctuating
velocitycomponentsforestimationofvolitionalrainbowtroutdensity.Trans. Am.Fish.Soc.135(2),431–441.
STATSOFT,INC,2000.StatisticaforWindows(ComputerProgramManual).Statsoft, Inc,Tulsa.
Tennekes,H.,Lumley,J.L.,1972.AFirstCourseinTurbulence.TheMITPress. Triantafyllou,G.S.,Techet,A.H.,Zhu,Q.,Beal,D.N.,Hover,F.S.,Yue,D.K.P.,2002.
Vorticitycontrolinfish-likepropulsionandmaneuvering.Integr.Comp.Biol. 42,1026–1031.
Tritico,H.M.,Cotel,A.J.,2010.Theeffectsofturbulenteddiesonthestabilityand criticalswimmingspeedofcreekchub(Semotilusatromaculatus).J.Exp.Biol. 213,2284–2293.
Webb,P.W.,1988.Simplephysicalprinciplesandvertebrateaquaticlocomotion. Am.Zool.28,709–725.
Webb,P.W.,2002.Controlofposture,depth,andswimmingtrajectoriesoffishes. Integr.Comp.Biol.42,94–110.
Webb,P.W.,2004.Responselatenciestoposturaldisturbancesinthreespeciesof teleosteanfishes.J.Exp.Biol.207,955–961.
Webb,P.W.,2006.Stabilityandmaneuverability.In:Shadwick,R.E.,Lauder,G.V. (Eds.),FishPhysiology.ElsevierPress,SanDiego,pp.281–332.
Webb, P.W., Cotel, A.J.,2010. Turbulence: does vorticity affect the structure andshapeof bodyandfin propulsors?Integr.Comp.Biol. 50 (6),1155– 1166.
Weihs,D.,2002.Stabilityversusmaneuverabilityinaquaticlocomotion.Integr. Comp.Biol.42,127–134.
Zar,J.H.,1996.BiostatisticalAnalysis.PrenticeHallInternational,Inc.,NewJersey, USA.
