Managementproceduresrequiredtoincreasechironomid avaiLabilitytowadersfeedingonartificiallagoons
remainunclear
A.J.GREEN*and G.M.HILTONt
‘Doflana Biol9gicalStation, AvenidatieMariatulsas/n,PabellbndelPerd,41013Sevilla,Spain:andtApplied OrnithologyUnit,DivisionofE,wironsnentalEvolutionaryBiology,GrahatnKerrBuilding,Glasgow University,G128QQ.UK
Summary
1.Rehfisch(1994)madevarious recommendationsforhowtoincreasechironomid bioniassanditsavailability towadersinartificiallagoons.Wearguethathisstudies formaninsufficientbasisforsomeoftheserecommendations.
2.He suggested protocols for depth manipulationbased on amodel predicting
biomass in terms of depth and depth squared which wa*not ameaningful rep resentationof hisdata. This model should therefore not heusedas a guidefor managingwaterlevels.
3.Hepredicted that binmasswould peakatadepthofahout117cm,yetheonly
studieddepthsofupto54cm.Weuseddatafromalarge,permanent,hrackishlake toillustrate howbiomassatshallowdepthscannot beusedtopredictthatfoundat greaterdepths.
4.Hisaimofasimplemodelallowingmanagers topredictchironomid biomasson thebasisofdepthinartificiallagoonswasunrealistic,owingtowidevariation inthe biomass—depthrelationship withinandbetweensites.
5.Thevalueofhis‘invertedsombrero’designforawaderlagoonneedstobedemon
stratedempirically.He advocated regulardrying-outoflagoonsfollowedbyimmedi
aterefilling,afloodingcycle that maynotmaximizechironomidbiomass.
Key-words:Chironomidae,man-made lagoons,habitat management, waterdepth.
Journalof Applied&ology(1998j35,9—12
Rehflsch(1994)provided ausefulstudyofthebiotic andabioticfactorstntluencingchironomidbiomass inartificial,brackishpoolsmanagedspecificallyfor wadersatBlacktoftSandsintheHumberEstuary, England.However,weareconcernedthatbewent toofarwhensuggesting‘easilyapplicableroutinesfor increasingthebiomassandproductivityofa pre dominantlychironomidfauna’(p 383).We believe thatsome ofhisrecommendationsforimproving reservemanagementinordertoincreasechironomid availabilitytowadersaremisleading,andmayinduce managersto adoptineffectivemeasures.Wefeelthat thedatasetutilizedandtheanalysespresentedby Rehfisch(1994)areinsufficienttojustifyallhiscon clusions
Rehflsch(1994) presentedaninappropriateanalysis
oftherelationshipbetweenchironomidbiomassand waterdepth.HeusedmuLtipleregression todemon strateconclusivelythat,inhisstudysite,totalchi ronomidbiomassincreaseswith increasingdepthand
decreasingorgantcmattercontentofthesubstrate.He thenwentonto presentasimplemodelofmore immediateusetoareservemanagerforcontrolling thewaterdepthineachlagoon’(p383), whichrelied onwaterdepthastheonlyindependentvariable.In Table10,hepresentedaquadraticmodelrelatingtotal biomassb(gm)towaterdepthw(cm):
b=032467+01j239w—00%43w’.
Theaboveequationdidnotprovideausefulfitto theavailabledata,sincethepartial effectofI4’ was nowherenearstatisticallysignificant(t=0’43,103 d.i,F>05)andthepartialeffectofwwasonly weakly significant (P<O’05).Usingthestandard errorsofTable10,wecalculated95%confidence intervalsfortheregression coefficients,whichindi catedthatthedepthatwhichbiomasspeakedcould lieanywherefrom16cmtoinfinity.Theabovemodel therefore forms no sound basis for predicting real
9
valuesof bat BacktoftSands or anywhere elseA rnorcsatisfactory model for Rchfisch’data wasa simpleregressionbetweenbandw,whichhepresented inTable4(r 047,P <0000!),
Accordingto the above quadratic equation,b
peaked atw=117cm1decreasing againatgreater
depthsAswehyeseen,neitherthecxthencenor posftIOnOc’flti11tThngpotfltlSMlppOr[edSLaL
precisely whatRehfisch(1994)attempted,albeit fora
verydifferentwetland
Evenwhensamplingoverawiderdepth range,we believethatamodelassimpleas that presentedby Rehfisch(1994)relatir.gbtowandi?cannotprovide reservemanagers with an effective guidetoman- ipulatingwaterdepth,evenatthesitewhereIFusmodel wasdeveloped.AmoicophaicatedapproachiS
yReh sh(1994)gvIIbyronrquirdccurft,r hconi&rthknn.rnnd
mending that‘reservemanagementshouldaimtokeep the lagoonsfloodedtotheirmaximaldepthsuntilused toattractfeedingwaders,asuptoabout ll7m,the greaterthemeanwaterdepthover theyearthegreater theberithicinvertebratebioniass’(pp.394—395).Even if the abovequadratic equationwas statistically robust,iiwouldstill beimprudentto assumethat therewasaturningpoint at around117cm, since Rehflsch(1994)wasonly abletocollectsamplesover thedepthrange0—54cm.Ashehimselfpointedout4
‘predictionsshouldonly bemadewithintherangeof
water depths foundinthestudy area,asatgreater depthsovercrowding,reduced oxygenleveLsdueto therniodlines,andthepresenceof newpredatorcom munities,suchasfishwhich requirewater depths of overl7intoestablishthemselves(Street1989),would actaslimitingfactors’(p.394).
ToiiltstratehowanaJysesofdatafromshallow
depthscannotbeusedtopredictchironomidbiomass atgreaterdepths,weanalyseddataonchironomid biomass fromBurdurLake,alarge,deep,permanent, brackishlakeinTurkey(Greenetat.1996).From10
Februaryto4March1993,benthicinvertebrateswere
sampLedtwiceateach20-rnintervalalong200-mtran sectsperpendicular to theshoreat nine‘ocations aroundthelake(Greenetat. 1996).Sampleswere
seasonalfluctuationsatagivensiteintherelationship between w and b (includingthedepthat which/, peaks),whichhavebeendemonstratedinawiderange ofwetlandtypes(Forsyth1986;SephtonPaterson
1986;Kajak1988;CarterMurphy1993).Forexam ple,Rehfisch’smodeltakesnoaccountoftheimpor tantinfluenceofdepthfluctuations(i.e.functionsof pastdepth)on chironomidbiomass.Rehfisch(1994) rightlyarguedthatthetimingoflagoonfloodinghas arnalorinfluenceontLespeedwithwhichch.ironomid biomassincreases,owingtotheseasonalityofchi ronomidoviposition.However,hismodelsrelatingb towandw2onlyincludedannualmeansofbandw (oneforeachoftwostudyyears)ateachsampling point,ignoringthetimingandextentofdepthfluc tuationsandtheimportanceofseasonalandannual variations.
Atagivenstudysite,therelationshipbetweenband
wcanalsochangewithftuctuationsinthestructureof thechironomidcommunity.McLachlan(1970)found thatfluctuations in waterlevelsatLakeKaribawere accompanied bymajorchangesinthespeciescom positionofbenthicchironomid community,andin therelationshipbetweenwandb.Likewise,Rehflsch (1994)recorded changesin thechironornidcom munityathisstudysite,sinceChironatnusan.’utlarius
fromaboatrisinganErkmann Grabaridwas intheproressofralanioing thelagoonsarid
washed insievesof1mmandO25mmsquaremesh. Drymassofchironomidlarvaeineachsamplewas
measureëafterstoragein70%methanolfor1—8days. Sampleswerecollectedfromdepthsofupto21m (Fig.1),andfittingan equivalentquadraticmodelto thewholedatasetproducedasatisfactoryfittothe data,inwhichbothwand w2werehighlysignificant (Table1).Accordingtothisequation,biomasspeaked attheturningpointw=112in,andthendeclined, Usingthestandarderrorstocalculateconfidence intervalsfortheregressioncoefficientsofTabLeI, the truevalueatwhichbiomasspeakedwasfoundwithin
therange73—177inwithaprobabilityof095
However,fittingasimilarmodeltoareduceddata setofoursamplestakenatshallowdepthsof0—2in producedacompletelydifferent,U-shapedrelation shipin whichneitherw norw1hadstatisticallysig
,nificantpartialeffects,andbiomassreachedamini
mum attheturningpointw=44in,with ahigher biomassatlowerandhigherdepths(Table1). This illustrateshowfittingaquadraticmodelatshallow depthsdoesnothelptopredicttherealrelationship between w andbatgreaterdepths However,thisis
uncommonduringthefirststudyyear’(p,387),but muchmoreabundantduringthesecondandfinal studyyear.However,Rehfisch(1994)didnotsay whetherornottheseobservedchangesaffectedthe relationshipbetweenwandb.
Rehfisch(1994)argued that his simpleniodel
developedatBLacktoftSandscouldbeusedasabasis forlagoonnianagenientatothersites,atleastinthe UK(p.396).However,evenasophisticatedmodel developedatonesitetopredictchironomidbiomass onthebasisofdepthwouldnotbereadilyapplicable tootherwetlands.Foragivencommunityofchi ronomidspecies,bisinfactdirectlydeterminedby oxygensupply,temperatureandorganicfoodsupply (J3rinkhurst1974;Kajak1988), aswellasbysalinity (Velasquez1992),sedimentstabilityandwaveaction (McLachLanMcLacLiia.n.9ii9, Lis.degaard&Jonas son1979;Forsyth1986;SephtonPaterson1986).It isthecorrelationsbetweenthesevariablesandwwhich determinetherelationshipbetweenw andb.Asa result,thisrelationshipislikelytovarygreatly,even betweendifferentshallowlagoonsintheUK.The variationacrossa rangeofwetlandtypesisillustrated
Depth (m)
Fig.1.ChIrononud biornassftgm2)1 againstdepth(m)atBardurLake,Turkey,inFebruary—March1993.
Table1.Multipleregressionoftotal biornassofchironomidlarvae6 (grn-2)againstwaterdepthw (in)atBurdurLake, presentedforalldata(w=0—21m)and forshallowdepths(0—2in).Thedependentvariableistransformed(6525)toremove heteroscedasticity
Intercept
w11
A[ldata / 02339±004537 / 0*1434±O-0l436’ / 01)06383—±0*0007878*4180 / O4132*44
2m / 0*6542±01800 / O3O28±O3638 / 0-034l6±Ol582 / 62 / Ol327
Theinterceptandpartialregressioncoefficientsaregiven±SE.iiandvrefertothe overallmodel.Significancelevelsaregiven
forthepartialcoefficientsofwandi+?,andforrF<00001;
P005.
-
bythewaythatbhas beenfoundtopeakatdepths from0’66in(VodopichCowell1984)to16m(Gra hamBurns1983).Furthermore,6doesnotpeakat interniecliatedepths inallwetlands.Iwakuma,Ueno
Nohara (1993)found no relationship betweenw and6oeradepth rangeof2—12in,whereasForsyth (1986)found6reachedatroughat5m.
Wealsoquestion Rehfisch’s(1994)recoinmenda tionforan‘invertedsombrero’designfora wader lagoon, witha deep,permanent, central reservoir designedtoboostchironomid biomassinshallowsby promotingrecolonizationofnewlyfloodedgroundvia larvalmigrationfromdeepwater.Thevalueofsuch areservoirneedstobedemonstratedempirically,since adult oviposition mayplayamuchmore important roleinthelarvalcolonizationofnewlyfloodedareas thanlarvalmigration fromdeeperareas(McLachlan
1970, Sephton Paterson 1986) Rehflsch (1994)
found differencesinthedepthdistribution ofsix different tp&itteddeuitivorouschtronomids.‘flit valueofareservoiristhereforepartlydependenton theextenttowhichdeeperwaterspeciescolonizeshal lowwaters, animportantissuenotaddressed inhis paper.Sucha reservoircouldconceivablylowerlarval biomass in the shallows because of its unknown impactonpredators orcompetitorsofchironornids.
Rehfisch(1994)also recommended thatdried lagoonsshould‘berefilledasquicklyaspossible’(p.
396)afterdryingout,thusrefloodingdrymud,Hedid
not fullyconsider the‘pros and cons’ofreflooding temporary lagoonsw:thorwithout allowingtimefor the dry basin to becolonised by terrestrial plants. Allowingsufficienttimefor thedevelopmentofter restrialvegetationmayincrease 6uponreflooding, owingtotheincreased nutrient supplyfromtherot tingvegetation (Kajak 1988).Such an effectledto apeak in6in reflooded shallows at Lake Kariba, Zimbabwe(McLachlan 1970). Thepotential for increasingchironomidbiomassbyallowingtimefor drywaterlagoons tobecolonized byterrestrial veg etation beforerefloodingshould be investigated, althoughmanywaderspeciesprefer tofeedinbare mudthanamongst rottingvegetation.
Furthermore,Rehfisch’s (1994)stated viewthat drying out periods areessential as they[maintain] achitonomid-dominattônüycoomzetcommunity’ (pp 395—396)isnotapplicable toallwetlands.Pater son Fernando(1969)found that, inadrawdown reservoir,6wasalwayslowerinreffoodedareasthan inpermanentlyfloodedareas.Chironomidsaredomi nantin thebenthosofBurdurLakedespiteits per manentnature(Greenetat.1996).Highsalinitiescan
besufficienttostimulate chironomi4-dominated Benthic benthic communitiesinshallowlagoonsintheabsence chironamidsin ofdryingout periods(Velasquez1992).
Acknowledgements
Threeanonymousrefereesprovidedconstructivecriti cisrnsof an earlier version of this manuscript. The workat Burdur Lake wascarried out together with A D FoxandBarryHughes,withthehelpofmany othem(.seeGreenerat.19%).Webenefttedfromdis cussionswithMikeC.Bell.
References
Brtnkhurst,RO.(1974)TheBent/sosofLakes. Macmillan
Press,London.
Carter,CF.Murphy,P.M.(1993)Themacrotnvertebrate faunaof Laugh NeaghtoughNeagh(edsR.B.Wood& RN. Smith),pp.339—367 KtuwerAcademic,Nether lands.
Forsyth,Di.(1986)Distnbutionandproductionof Clii
rononsusineutrophicLakeNgapouri.NewZealandJour
nalofMarineandFreshwaterResearch,20,47.—M. GrahsatL,k.kBurns,CW.çV3tt’?toductionan.ttcAowj
ofbenthic‘Ohironomidlarvae(J3iptera)inLakeHayes, NewZealand,awarni-monomicticeutrophiclake.inter
nanonode Revuedergesamlen HydrobtologieuriSHydro
graphie,68,351—377.
Green,A.J,Fox,AD.,Yarar,M.,Hilton,0,Hughes,B.
Salathe,T.(1996)Threats to Burdur Lakeseosystem, Turkeyand itswaterbirds, particularlythewhite-headed
duck Ox.vuraleucocephala.BiologicalConservation,76,
241—251
twakuma, T.,Ueno,WAc Nohara,S.(1993)Distribution and populationdynamics ofChironomidac(Diptera)in LakeYunoko,Japan.JapaneseJournalofLunnokgy,54,
199—212.
Kajak, Z.(1988)Considerationson benthos abundancein freshwaters, its factorsand mechanisms Jnternarionale Revuedesgeso.niten1-fydrobtologie u.ndHydrographee,78,
5—19.
Lindegaard,D.Jonasson,PM.(1979) Abundance,popu
lationdynamics and productionof zaobentbosinLake
Myvatn,Iceland. Qikos,32,202—227.
McLaclilan, A.J.(1970)Someeffectsofannual fiuctiations inwLterlevelon.thelarvalchironrimid colllmunitiesof LakeKariba JournalafAniMal1Ecolngy,39,79—90.
McLachlan,AJ.McLachlan, S.M. (1969)The bottom
fauna andsedimentsinadryingphaseofasalineAfrican lake(LakeChilway,Malawi).Hydrobiologia,34,401—403. Paterson, C.G. Fernando, C.H. (1969) The macro invertebratecolonizationoasmallreservoir ineastern Canada. Verhandlungen derinrernadonalen vereinigung
Lisnnoiogie,1’),126—136
Rehllsch, M.M. (1994)Man-niadelagoons and how their attractivenesstowadersmightbe increasedbyman ipulatingthebiomassofaninsectbenthos Journal of ApplIedEtalogy,31,383—401.
Sephton.,T.W.Paterson, C.G.(1986)Production of the
chirononsidProcladmsbelliesinanannualdrawdown re
servoiFreshwaterBiology,16,721—73.
Street,M.(1989)PondsandLakesforWddfowl.TheGame
Conservancy,Fordingbrtdge.
Velasquez,C R (1992)Managingartificialsalipansasa
waterbird habitat species’responses to water level manipulation.ColonialWaterbirds,15,43—55.
Vodopich,0.8.CoweIl,B.C.(1984)Jnteractionoffactors
governingthedistributionofapredatoryaquatic insect.
Ecolcgy6539—52.
Received8June1996;revisionreceived28Deceinbes1996
©1998British EcologicalSociety, JournalofApplied
&ology,
9—12