ISSN 0103-5657

Revista Brasileira de Ornitologia

Volume 18Número 2

Junho 2010www.ararajuba.org.br/sbo/ararajuba/revbrasorn

Publicada pela

Sociedade Brasileira de OrnitologiaSão Paulo - SP

Diurnal variation in transect counts of birds in a cerrado landscape in the state of São Paulo, Brazil

Vagner Cavarzere1 and Gabriel Parmezani Moraes2

1 Correspondingauthor.DepartamentodeZoologia,UniversidadedeSãoPaulo.RuadoMatão,Travessa14,nº 101,05508‑900,SãoPaulo,SP,Brasil.E‑mail:cavarzere@usp.br.

2 DepartamentodeCiênciasBiológicas,UniversidadeEstadualPaulista.RuaEngenheiroLuizEdmundoCarrijoCoube,14‑01,17033‑360,Bauru,SP,Brasil.E‑mail:g_moraes@hotmail.com.

Recebido:15/01/2010.Aceitoem:04/07/2010.

ReSuMo: Variação diurna de aves amostradas por transecções em áreas de cerrado de São Paulo, Brasil.Por incluíremoperíododemaioratividadedasaves,asmanhãs sãoconsideradasmelhoresqueas tardesparaamostragemdestegrupo.EmboramuitospesquisadorestenhamrealizadoestudossobrevariaçõestemporaisnadetecçãodeavescomPontosdeEscutaemregiõestemperadas,transecçõeslinearesehabitatsneotropicaisforampoucoexploradosnatentativaderesponderasmesmasquestões.AmetodologiadetransecçõesfoiutilizadaparatestarahipótesedequeasavessãoregistradasigualmentedurantedoisperíodosdodiaemumambientedecerradopredominantementeabertonosudestedoBrasil.Emboranemsempredemaneirasignificativa,onúmerodeespéciesedeindivíduosfoiconsistentementemaiorduranteasmanhãs,corroborandoaidéiadequeasavespodemsermaisdetectáveisduranteesseperíodododia.Noentanto,algumasfamílias,assimcomoumapequenaporcentagemdeespécies,demonstrarampreferênciasdedetecçãoduranteumdosdoisperíodosanalisados.Osresultadossugeremque,noambienteestudado,asmanhãssãomaisadequadasparaaamostragemdeumnúmeromaiordeespéciesedeindivíduos,porémtáxonsespecíficosvariamseuspadrõesdedetecçãoedevemseravaliadosanteriormenteàsamostragens.

PalaVRaS-ChaVe:Abundância,detecçãovisual,padrõesdedetecção,neotrópico,períodododia,riqueza.

aBStRaCt: Thegeneralstatementthatbirdsarerecordedmoreoftenonmorningthanonafternooncountsisquitecommonand widespread among ornithologists. Although many investigators have reported temporal variations in bird detections usingPointCountsintemperateregions,fewresearchesregardingthesameobjectiveshavebeenconductedinNeotropicalhabitatsorusedtransectcountsasfieldmethod.WeusedtransectcountstotestthehypothesisthatbirdsareevenlyrecordedbetweentimesofdayinapredominantlyopenCerradolandscapeinsoutheasternBrazil.Althoughnotalwayssignificantly,thenumberofspeciesandindividualswereconsistentlygreaterduringthemorningcounts,whichcorroboratesthefactthatbirdscanbemoredetectableduringthistimeofday.However,afewfamiliesaswellasasmallpercentageofspeciesweremorelikelytoberecordedduringeitheroneof thetwoperiodsweanalyzed.Ourresults suggest thatmorningcountsshoulddetecthighernumberofbothspeciesandindividualsinourstudyarea,butspecifictaxashowdistinctpatternsofdetectionwhichshouldbeacknowledgedpriortosampling.

Key-WoRDS:Abundance,Neotropicalregion,patternsofdetection,richness,timeofday,visualdetection.

Point Counts were developed by Blondel et  al.(1970) to estimate bird community densities and werespecifically designed for temperate habitats. Many au‑thors have investigated hourly variation of birds usingthis method (Skirvin 1981, Blake 1992, Betini 2001)andmostoftheseapproacheshavesetagloballyacceptedtrend,inwhichcensusarebestemployedwhenbirdsaremoreactiveandthereforemoredetectableduringthefirsthoursoftheday.Somedeviationsfromthispatterncanoccur(Dawson1981,VernerandRitter1986)andmuchworkisstillneededtoimprovebirdcountingtechniques(Blake1992).

Few studies on diurnal variation in detections ofNeotropical birds have been carried out to date (Blake1992,Lynch1995,Betini2001)andinvestigationsusing

transect counts are even rarerorwere either conductedinthenorthernhemisphere(Järvinenet al.1977,Shields1977, Blake et  al. 1991) or only recently in the Neo‑tropics(Antunes2008).Whilemanyinvestigationshavefocusedforestedhabitats(Blake1992),openformationshaveneverbeentheaimofstudiesontemporalvariationsoftropicalbirds.

On situations in which surveys must be sampledinopenareas,suchasCerradoformations(Eiten1972),transectsaremoreindicatedthanPointCountssincetheyare based on visual rather than aural detections (Bibbyet al.1993,Develey2003).Asmethodsarerecklesslyusedwithoutprevioustestsconcerningitsefficiencyorhabitattype(Betini2001),itisimportantoneinvestigatesfactorsthatinfluencebirdcountcensuses(Blake1992).

aRtIGo RevistaBrasileiradeOrnitologia,18(2):97-101 Junhode2010

Herewepresentourresultsondiurnalvariationofbirdsusingtransectcountsduringsixmonths inapre‑dominantlyopenCerrado landscape in the stateof SãoPaulo,Brazil.Wespecificallywantedto:(A) testthehy‑pothesis thatbirds,measuredasnumberof species andindividuals,areevenlyrecordedduringtwodistinctperi‑odsoftheday(morningsandafternoons)and(B) exam‑inewhetherparticulartaxaexhibitspecificpreferencetoberecordedmoreoftenduringoneoftheseperiods.

MethoDS

Study area

WeconductedthisresearchattheJardimBotânicoMunicipaldeBauruandatthecampusoftheUniversi‑dadeEstadualPaulista (UNESP), located in thecityofBauru(22°20’S49°00’W),nearthecenterofthestateofSãoPaulo,Brazil.ThestudyareasbelongtotheCerradoBiome and includemature cerradão (closed‑canopydryforest),seasonalsemideciduousforestandalluvialforest(Cavassanet al.1984).

The UNESP campus has anthropogenic environ‑ments suchasorchardsandBrachiaria sp.grassy fields.Therearealsolakesandregeneratingsecondgrowth,allreachable by roads or routes. Dry periods occur fromApriltoSeptemberandrainyseasonlastsfromOctobertoMarch(Cavassanet al.1984).

Bird count

We established two transect lines (1.5  km  ±  0.3;mean ± SD)1.5 kmapartrunningacrossallenvironmentsin the study areasbutpredominantly rangingoveropenhabitats.Observers(VCandGPM)walkedslowlyatcon‑stantspeed(5 km/h)andrecordedeverybirdseenorheardwithinanestimated100 mofthetransectcenterline.Inordertoavoidcountingthesamebirdtwice,onespecieswasneverregisteredafterifitwasheardagaininfrontoftheobserversunlessitwasundoubtedlyanotherindividual.Observationsweremadewithbinoculars(8 × 42)andvo‑calizationsweretaperecordedwithaPanasonicRQ‑L31.Thesamedirectionoftravelwasfollowedonbothroutes,whichweresampledtwiceduringtheday,onceduringthemorningsandonceduringtheafternoons.Transectscom‑prisedthesametypeofforestformation(cerradão)althoughwespent80%oftransectcountinginopenhabitats.

Wecountedbirdsevery15daysfromSeptember‑De‑cember2005andcomplementedoursamplingeffortvis‑itingthelocalitiesonamonthlybasisduringJanuaryandFebruary2006.Westartedalltransectsfrom5to10 minbeforesunriseandcontinueduntilmid‑day,exceptforafewoccasionswhenweinterruptedthemat10:00.

analyses

We conducted 20 visits (subsamples) to the areasduringthemonthswecountedthebirdsandweanalyzedthesamenumberofsubsamples(n = 10) formorningsand afternoons. As subsamples were not independent,whichprecludestheindependencepremisetousepara‑metric tests, we used values obtained from the differ‑encesbetweenmorningsandafternoonstoanalyzedata.These values indicate how many more species or indi‑viduals could be detected during the morning counts.Afternormalitytests,weusedatwo‑tailedpairedt‑testtoexaminethedifferencesinnumbersofspeciesandindi‑vidualsbetweentimesofday.Weexaminedvariationsforeachrouteseparatelyandpoolingdatafrombothspatialreplicates.

Wedecided to excludenon‑passerines andanalyzeonlythepasserinebirdsoncetheirhomerangeissmallerand their foragingbehavior ismoreappropriate for thequestions we elaborated. Non‑passerines’ home rangesaremuchwiderandsomeofthespeciesrecordedinourstudyareasprobablydonotusethemasarestrictedhabi‑tatyear‑round.

Finally,weusedagoodness‑of‑fitG‑testtocomparethedistributionofthenumberofobservationsoffami‑liesandspecieswithmorethan150and50detections,respectively,betweentimesofday.Thenullhypothesiswasthatdetectionswereevenlydistributedbetweenperiods of day and showed no particular temporalassociation.

Forabundanceanalysisweconsideredabundanceasthenumberofindividualscountedper100 hsothatdif‑ferencesinsamplingeffortcouldbefactoredout.Theal‑phalevelfortestsofsignificancewasα = 0.05.Thevaluesinthe‘Results’sectionrefertomeanandSD.

FIGuRe 1:Meandifferences innumberofspeciesandindividualsrecordedon two transect counts inBauru, SãoPaulo,Brazil.Errorbarsrepresent± SE.

98 DiurnalvariationintransectcountsofbirdsinacerradolandscapeinthestateofSãoPaulo,BrazilVagnerCavarzereandGabrielParmezaniMoraes

RevistaBrasileiradeOrnitologia,18(2),2010

ReSultS

Werecordedatotalof144speciesin9045contactsduring31.5 hattheJardimBotânicoandduring22.5 hattheUNESPcampus.TheoverallspeciesrichnessandabundanceweregreateronmorningcountsexceptforthespeciesrichnessattheUNESPcampus;thepercentageofspeciesexclusivelyrecordedduringtheperiodsofthedaywasquitesimilarforbothsites(Tables 1 and 2).

Consideringbothspatialreplicates,morningcountsconsistently recorded more species and individuals (ex‑cept for individuals recorded on subsample 6) than af‑ternoon counts (Figure 1).At the JardimBotânico,werecorded10.45 ± 8.10morespeciesduringthemorningthan during the afternoon counts and 93.82  ±  59.42more individuals in the same situation.At theUNESPcampusthesevalueswere2.27 ± 4.28and32.55 ± 45.53,respectively, whereas pooled data revealed 6.80  ±  4.06

morespeciesand34.75 ± 23.46moreindividualsforthemorningtime.

DifferencesinspeciesrichnessandabundanceweresignificantlygreateronmorningsattheJardimBotânico,althoughthesevariablesdidnotdiffersignificantlyattheUNESP campus; when pooling data, only richness wassignificantlydifferentbetweentimesofday(Table 2).Ad‑ditionally,somespeciesmayhaveindividuallyaccountedformorerecordsduringaspecificperiodoftheday,soweranouranalysesoncemorewithoutthem.Asaresult,thenumber of individual detections increased significantlyonmorningcountstheUNESPcampus(tabundance = 2.79;df = 9;P = 0.021).

Whenanalyzingthefamilieswithmorethan150de‑tections,onlyone(Emberizidae)didnottendtobemore

taBle 2:Resultsofdifferences in species richness andnumberofindividuals between morning and afternoon counts in Bauru, stateofSãoPaulo,Brazil,revealedbytwo‑tailedpairedt‑tests.Degreesoffreedom= 9forall.

Jardim Botânico uNeSP Pooled datat P t P t P

Richness 4.18 0.002 2.14 0.061 5.30 0.000Abundance 2.75 0.023 1.26 0.240 2.08 0.285

taBle 1:SpeciesrichnessandnumberofindividualsrecordedonmorningandafternooncountsfortwoseparatetransectroutesattheJardimBotânicoandattheUNESPcampusinBauru,stateofSãoPaulo,Brazil.

total Mornings % exclusive % afternoons % exclusive %Jardim Botânico

Richness 140 130 93 17 12 123 88 10 7Abundance 4080 2298 56 1782 44

uNeSPRichness 95 85 89 10 11 89 94 6 6Abundance 4965 2573 52 2392 48

taBle 3:Listoffamilieswithmorethan150detectionsinBauru,stateofSãoPaulo,Brazil.Boldedones tended tobe recordedmoreoftenduringthemorningcounts,whileanasteriskindicatesfamiliesrecordedmoreoftenon afternoon counts, as revealedbygoodness‑of‑fitG‑tests.Numbersofindividualsper100 hofobservationsaregiven.Degreesoffreedom= 1forall.

Familiestime of day

G PMorning afternoon

Thamnophilidae 235.2 59.3 112.48 0.000Tyrannidae 757.4 585.2 22.15 0.000Vireonidae 196.3 114.8 21.59 0.000Hirundinidae* 277.8 431.5 33.57 0.000Turdidae 200.0 161.1 4.20 0.041Emberizidae 227.8 224.1 0.03 0.862

taBle  4: List of species with more than 50 detections in Bauru,stateofSãoPaulo,Brazil.Boldedones tended tobe recordedmoreoftenduringthemorningcounts,whileanasteriskindicatesspeciesrecordedmoreoftenon afternoon counts, as revealedbygoodness‑of‑fitG‑tests.Numbersofindividualsper100 hofobservationsaregiven.Degreesoffreedom= 1forall.

Speciestime of day

G PMorning afternoon

Thamnophiluspelzelni 90.7 18.5 65.38 0.000Furnariusrufus 59.3 42.6 0.00 0.098Elaeniaflavogaster 55.6 40.7 0.03 0.130Pitangussulphuratus 83.3 48.1 2.89 0.002Myiodynastesmaculatus 63.0 31.5 5.74 0.001Megarynchuspitangua 51.9 44.4 1.62 0.450Empidonomusvarius 53.7 53.7 6.17 1.000Tyrannusmelancholicus 100.0 88.9 4.57 0.419Tyrannussavana 77.8 66.7 2.43 0.355Antilophiagaleata 92.6 61.1 0.48 0.011Vireoolivaceus 125.9 77.8 2.12 0.001Pygochelidoncyanoleuca* 231.5 359.3 176.18 0.000Turdusleucomelas 148.1 103.7 0.05 0.005Mimussaturninus 74.1 66.7 3.81 0.532Thraupissayaca 55.6 37.0 0.22 0.053Volatiniajacarina 137.0 109.3 1.50 0.076Basileuterusflaveolus 155.6 68.5 18.75 0.000Molothrusbonariensis 68.5 75.9 14.11 0.538Euphoniachlorotica 70.4 51.9 0.06 0.093Passerdomesticus 111.1 105.6 8.93 0.706

99DiurnalvariationintransectcountsofbirdsinacerradolandscapeinthestateofSãoPaulo,BrazilVagnerCavarzereandGabrielParmezaniMoraes

RevistaBrasileiradeOrnitologia,18(2),2010

detectableduringonespecificperiodofthedaywhiletheHirundinidaewereregisteredmoreoftenduringtheaf‑ternooncounts (Table 3).Of the20 specieswithmorethan50detections,sevenweremorelikelytoberecordedduringthemorningcountsandtheBlue‑and‑whiteSwal‑lowPygochelidoncyanoleucawastheonlyspeciesrecordedmoreoftenonafternoons(Table 4).

DISCuSSIoN

Ourresultspartlysupportedthegeneral statementthat birds are more detectable during the mornings(Shields 1977, Skirvin 1981, Pizo et  al. 1997) becausea significant higher number of species was recorded atthe Jardim Botânico and when pooling data. Howeveritmaybenotsuggestedthatmorningsarealwaysmorefruitful to conduct transect counts in open Neotropi‑calhabitats.Thesereasonsaretwo‑fold:(1) theUNESPsiteandthedifferenceinnumberofindividualsdetectedwhenpoolingdatastronglydeviatedfromthesefindingsand (2)  most SE intervals in Figure  1 crossed the zeroline (non‑significant differences). This indicates that,although morning counts do consistently record morespecies and individuals, suchdifferencesmaybenot asevidentaspreviouslyconsidered.

If one looks at the overall number of species re‑cordedonmorningandafternooncountsinTable 1hewillprobablytendtoconsiderthefirstperiodmoreap‑propriatetoconductbirdsurveys.ButafurtherglimpseatFigure 1clarifiesthatonlyfewmorespeciescouldberecordedduringthemorningthanduringtheafternooncounts.Thus,morningsmaybeusedtorapidlyassessthegreatmajorityofbirdspeciesofanareawhileafternoonsmaybeusedforthesamepurposeifsomewhatgreateref‑fortisspent.Unfortunately,wedidnotteaseapartvisualfrom aural detections, which could detect distinct pat‑ternsforourresults.

Järvinenet al.(1977)investigateddiurnalvariationin detections of birds using transect counts in Finlandand reported significant greater number of species anddetectionsformornings.Blake(1992)didfindrichnesspeaks on late afternoons using Point Counts in CostaRica,buthisresultsstillyieldedhighernumbersofspe‑ciesdetectedduringthebeginningofthedays.Althoughwithoutcontrollingfordifferencesinsamplingeffortbe‑tweentimesofday,Antunes(2008)concludedthatinaSãoPaulosemideciduousforestbirdsaremoredetectableduring themorningcountsusing transects.Comparingthese three investigations with our own findings, thisUNESP’srouteseemedtorepresentanoutlieranditmayhavebeenbiasedbycountingbirdsinpartiallyanthropo‑genichabitats.

Also our results probably reflected the variabil‑ityobservedfor fewspecies that individuallyaccounted

100 DiurnalvariationintransectcountsofbirdsinacerradolandscapeinthestateofSãoPaulo,BrazilVagnerCavarzereandGabrielParmezaniMoraes

RevistaBrasileiradeOrnitologia,18(2),2010

for more records during a specific period of day suchas the Planalto Slaty‑Antshrike Thamnophilus pelzelni,the Helmeted Manakin Antilophia galeata and the Fla‑vescentWarblerBasileuterusflaveolus,threeofthesevenspeciesrecordedmoreoftenduringthemorningcounts(Table  4).This is further supported by the fact that ifsuchspecieswereremovedfromouranalysis,thenumberof individualdetectionswould increase significantlyonmorningcountsattheUNESPcampus.

Foragingbehaviorandbiologymaydrivesomedi‑urnalpatternsofdetection inour studyareas.TheHi‑rundinidaeconsumeinsectswhilesoaring.Thisbehaviormatchwiththetendencytobedetectedmoreoftendur‑ingtheafternooncounts.Also,thedependenceonforest‑edhabitatsseemedtoplayanimportantrolehere.Mostof the species recordedmoreoftenduring themorningcountswereforest‑dependentoratleastsemi‑dependentonthesehabitats(Pale‑breastedThrushTurdusleucome-las;Silva1995).TheonlyexceptionwastheGreatKis‑kadeePitangussulphuratus,typicalofopenenvironmentsbutstillrecordedmoreoftenduringthemorningcounts.Thesepatternsmayindicatethatforestsshouldbesam‑pledduringthemorningsasbirdstendtobemoredetect‑able during this period. Aerial‑foraging birds, however,maybesampledduringtheafternoonsforbirdsshouldbemoredetectablelaterintheday.

Therefore, sampling procedures must be designedaccordingtothescopeoftheinvestigation(Blake1992).At community levels,morningsmaybe censused alonebecause afternooncounts should record lowernumbersofspeciesandindividuals.However,familiesorspecies‑specific preferences or behavior must be taken into ac‑count before conducting field surveys. Forest‑indepen‑dent tyrannids, swallows and finches, for example, arerecordedmoreonafternoons.

OurresultsapplytooneCerrado site inSãoPauloand morning counts, although not as evident as previ‑ously considered, did consistently record more speciesand individuals than afternoon counts. Some speciesweremorelikelytoberecordedduringeitheroneofthesetimesofday.Disagreementsfoundwithindiurnalvaria‑tion in species richness and abundance using the samemethoddeservesfurtherinvestigationandmuchmoreef‑fortmustbespenttoimprovethedesigningofcountingtechniquesinNeotropicalhabitats.

aCKNoWleDGMeNtS

LuizCarlosdeAlmeidaNeto,directoroftheJardimBotânicoMunicipal de Bauru, stimulated our research and his employeesoffered much assistance. Andreli C. Dalbeto, Carolina D. FerreiraandFernandaG.Macielhelpedusduringsomefieldtrips.VítorQ.Piacentini kindly reviewed an early version of this manuscript andJames R. Roper helped with the statistical treatments. Anonymousreviewersgreatlycontributedfortheimprovementofthisresearch.

ReFeReNCeS

antunes, a. Z. (2008). Diurnal and seasonal variability in birdcountsinaforestfragmentinsoutheasternBrazil.Rev.Bras.Zool.,25:228‑237.

Betini, G. S. (2001). Amostragem de aves por pontos numa florestaestacionalsemidecidual,SãoPaulo,Brasil.Dissertaçãodemestrado.Piracicaba:EscolaSuperiordeAgricultura“LuizdeQueiroz”.

Bibby, C. J.; Burgess, N. D. and hill, D. a. (1993). Bird censustechniques.SanDiego:AcademicPress.

Blake, J. G.; hanowskig, J. M.; Niemi, J. and Collins, P. t. (1991).Hourly variation in transect counts of birds. Ornis Fennica,68:139‑147.

Blake, J. G. (1992).TemporalvariationinPointCountsofbirdsinalowlandwetforestinCostaRica.TheCondor,94:265‑275.

Blondel, J.; Ferry, C. and Frochot, B. (1970). La mèthod desindicesponctuelsd’abondance(IPA)oudesrevelésd’avifaunepar“stationsd’écoute”.Alauda,38:55‑71.

Cavassan, o.; Cesar, o. and Martins, F. R. (1984).FitossociologiadavegetaçãoarbóreadaReservaEstadualdeBauru,EstadodeSãoPaulo.Rev.Bras.Bot.,7:91‑106.

Dawson, D. G. (1981).Countingbirdsforarelativemeasure(index)ofdensity.Stud.AvianBiol.,6:12‑16.

Develey, P. F. (2003).Métodosparaestudoscomaves,p. 153‑168.In:L.Cullen,R.RudranandC.Valladares‑Pádua(orgs.).Métodos

deestudosembiologiadaconservaçãoemanejodavidasilvestre.Curitiba:EditoraUFPR.

eiten, G. (1972). The Cerrado vegetation of Brazil. Bot. Rev.,38:201‑341.

Järvinen, o.; Väisänen, R. a. and haila, y. (1977). Bird censusresultsindifferentyears,stagesofthebreedingseason,andtimesofday.OrnisFennica,54:108‑118.

lynch, J. F. (1995). Effects of Point Count duration, time‑of‑day,andauralstimuliondetectabilityofmigratoryandresidentbirdspecies inQuintanaRoo,Mexico,p. 1‑6.In:C.J.Ralph,J.R.SauerandS.Droege(eds).Monitoringbirdpopulationsbypointcounts.Albany:GeneralTechnicalReport.

Pizo, M. a.; Simão, I. and Galleti, M. (1997).Daily variation inactivityandflocksizeoftwoparakeetspeciesfromsoutheasternBrazil.WilsonBull.,109:343‑348.

Shields, W. M. (1977).The effect of time of day on avian censusresults.TheAuk,94:380‑383.

Silva, J. M. C. (1995). The birds of the Cerrado region, SouthAmerica.Steenstrupia,21:69‑92.

Skirvin, a. a. (1981).Effectoftimeofdayandtimeofseasononthenumberofobservationsanddensityestimatesofbreedingbirds.Stud.AvianBiol.,6:271‑274.

Verner, J. and Ritter, l. V. (1986). Hourly variation in morningpointcountofbirds.TheAuk,103:117‑124.

101DiurnalvariationintransectcountsofbirdsinacerradolandscapeinthestateofSãoPaulo,BrazilVagnerCavarzereandGabrielParmezaniMoraes

RevistaBrasileiradeOrnitologia,18(2),2010