Surveillance of the chikungunya vector Aedes albopictus (Skuse) in Emilia-Romagna (northern Italy): organizational and technical aspects of a large scale monitoring system

108 Journal of Vector Ecology June2011

Surveillance of the chikungunya vector Aedes albopictus (Skuse) in Emilia-Romagna (northern Italy): organizational and technical aspects of a large scale monitoring

system

MarcoCarrieri1,AlessandroAlbieri1,PaolaAngelini2,FlaviaBaldacchini2,ClaudioVenturelli3,SilviaMascaliZeo3,andRomeoBellini1*

1Medical and Veterinary Entomology, Centro Agricoltura Ambiente “G. Nicoli”, Crevalcore (BO), Italy2Emilia-Romagna Region Public Health Service, Bologna, Italy

3Public Health Department, Cesena (FC), Italy

Received 3 September 2010; Accepted 6 January 2011

ABSTRACT:Thechikungunyavirusoutbreakthatoccurredin2007innorthernItaly(Emilia-Romagnaregion)promptedthedevelopmentofalargescalemonitoringsystemofthepopulationdensityofAedes albopictus (Skuse,1894),comparableattheprovincialandmunicipallevels.In2007,eggdensitydatapresentedanaggregateddistribution(VMR>1)andTaylor’spowerlawwasappliedtocalculatetheminimumnumberofovitrapsneededtoobtaintheprefixedprecisionlevels:D=0.2intheareaswherethechikungunyaepidemicoccurredandD=0.3inalltheotherurbanareas>600ha.Theestimatedminimumovitrapnumberwasthenusedtosetupamonitoringnetworkattheregionalscaleinseason2008(May-October).In242municipalities2,741ovitrapswereactivatedandthe2008sampleddatashowedasimilaraggregateddistributionasin2007.TheadequacyofthemonitoringdesignwasevaluatedbyrecalculatingtheTaylor’scoefficientsandtheminimumovitrapnumberforeachurbanarea>600hausingthe2008eggdensitydata.Thecomparisonbetweenthetwoestimatesshowedthattheminimumovitrapnumbercalculatedin2007wasunderestimatedby2.7%inweeks22-41butwasoverestimatedby29.4%ifreferringtotheperiodofhighestpopulationdensity(weeks27-37).Thelowcostoftheproposedmonitoringsystem,basedontheuseoffortnightlycheckedovitraps,couldmakeiteconomicallysustainableeveninanon-epidemicseason.Journal of Vector Ecology 36 (1): 108-116. 2011.

Keyword Index: Aedes albopictus,ovitraps,monitoringsystem,chikungunya,Taylor’spowerlaw.

INTRODUCTION

Aedes albopictus (Skuse,1894)wasprobablyintroducedintoEurope through the tire trade,andwasfirstdetectedin Albania in 1979 (Adhami and Murati 1987), then inItaly (Sabatini et al. 1990, Dalla Pozza and Majori 1992),France(SchaffnerandKarch2000),SerbiaandMontenegro(Petrić et al. 2001), Switzerland (Flacio et al. 2004),Belgium(Schaffneretal.2004),Greece (Samanidouetal.2005),Spain(Arandaetal.2006),Croatia(Klobučaretal.2006),Slovenia,andBosniaandHerzegovina(ScholteandSchaffner2007).

In the Emilia-Romagna region of northern Italy, thefirst detection of Ae. albopictus occurred in Bologna anddates back to 1994 (Carrieri, unpublished data), then thespeciesspreadallover theregion.Anoutbreakcausedbychikungunyavirus(CHIKV),neverpreviouslyreportedinEurope, occurred in the Emilia-Romagna region in 2007with247identifiedcasesbetweenJulyandSeptember2007(Angelinietal.2007,Rezzaetal.2007,Bonilaurietal.2008).ChikungunyavirusbelongstothefamilyTogaviridae genusAlphavirus and is mainly transmitted by Ae. aegypti (L.)in Africa, India, and in southeast Asia. CHIKV epidemicoutbreakssuchasthosethatoccurredonRéunionIslandin2005,2006,and2007(Renaultetal.2007),Gabon(Vazeilleetal.2008a),India(Saxenaetal.2006),Mauritius(Ramchurn

etal.2008)in2006,andItalyin2007wereinsteadcausedbyAe. albopictus.Viralsequencesanalysisrevealedavirusmutation that increases the transmissibility of CHIKV bythis species (Tsetsarkin et al. 2007, de Lamballerie et al.2008,Vazeilleetal.2008b).

Since 2006, the Public Health Service of the Emilia-RomagnaregionhasdevelopedamonitoringsystemforAe. albopictus withtheaimofobtaininginformationaboutitstemporalevolution(Angelini2006).The2007chikungunyaoutbreak prompted the Emilia-Romagna Public HealthDepartment to develop a monitoring system aimed atobtaining spatially comparable data on species density attheprovincialandmunicipallevelandamongmajorurbanareas.

Historically,themonitoringofspeciesbelongingtothesubgenus Stegomyia (Ae. aegypti and Ae. albopictus) wasachieved through the use of ovitraps (Focks 2003). Thismethod provides several advantages over other methods,such as high sensitivity (it can detect the presence of theinsect even at low densities), ease of field management,achievable even by unskilled staff, and low material costs(Bellinietal.1996).

The number of eggs that can be found within eachovitrap is affected by the skip-oviposition behavior offemales and by environmental conditions (Hawley 1988).Therefore, the number of ovitraps to be placed and the

Vol.36,no.1 Journal of Vector Ecology 109

choice of their location are two important issues to beaddressedforareliableestimateofthepopulationdensityin an urban area and it is crucial to obtain comparableinformationforthevectorsurveillance(Craigetal.2006).

The optimal number of ovitraps varies according tothe stage of colonization of the region and to the abilityofthespeciestodisperse.Thedensityofthespeciesatthebeginningofthecolonizationispatchyandaggregatedanditspatternofdispersaldependsonhabitatfeaturesandonweatherconditions.Inordertoachievethesamereliabilitylevel, the monitoring protocol requires more ovitrapsin areas at the initial step of colonization (low mosquitodensityandhighdataaggregation)withrespecttotheareasat a mature step of colonization (high mosquito densityand more uniform spatial dispersion). The ratio betweenthe variance and the mean of the number of eggs laid intheovitraps, indicatedasVMR(VarianceMeanRatio), isoftenusedasadensitydistribution-relatedparameterthatprovidesinformationonthedispersalpatternofthespecies:VMRisequalto1whenthespeciesisrandomlydispersed;lower than 1 when species dispersal is uniform; higherthan1whenspeciesdistributionisaggregatedorclustered(Service1993).

Thevariance(s2)isrelatedtothemeandensity(m)byTaylor’spowerlaw(Taylor1961,1984,Kuno1991),whichhasbeenlargelyusedtoquantifytheaggregationdegreeandthestatisticallysignificantsamplesizeforinsectmonitoring:

(1) s2=a *mb

whereb isaconstantforthespeciesandmeasuresdataaggregationsimilartoVMR,i.e.,whenb isgreaterthan1itindicatesthatdataareaggregated;a isaconstantdependingonenvironmental conditions;m is themeaneggsdensityvalue.

The equation currently used to define the minimumsampleunitsizeforanurbanareais:

(2) N=[Zα/2/D]2*a *mb-2

where Z is the Standard Normal Distribution Valueforagivenprobability(Buntin1994);D isthemonitoringprecisionlevelrequired.Accordingtotheliterature,D=0.1isconsideredasufficientvalue(SouthwoodandHenderson2000), while 0.2<D<0.3 has been considered optimal forthebinomialsamplingofAe. aegypti byMogietal.(1990);m isthemeaneggsdensityvalue;a andb aretheTaylor’scoefficients.

Theadequacyandreliabilityofthemonitoringsystemcanbeevaluated,accordingtoService(1993),bymeasuringtheRelativeVariation(RV),i.e.,theratio:

Standarderrorofthemeannumberofeggs/ovitrap/weekMeannumberofeggs/ovitrap/week

SouthwoodandHenderson(2000)definedanRV=0.25as usually adequate for most extensive sampling surveys,although in certain intensive programs an RV=0.1 may

berequired.HighlyaggregatedmosquitopopulationswilllikelyproduceahighRV.

TheaimofthisstudywastodevelopalargescaleAe. albopictus monitoringnetwork,achievableatlowcostandwell-designedtoestimatetheinfestationlevelinlargeurbanareas(exceeding600ha),wherethehealthriskofmosquito-bornediseaseswashigher.Toachieve this aim, themeanegg density data collected by means of ovitraps in 2007wereusedtoassesstheaggregationdegreeofthemosquitopopulation through the application of Taylor’s power law.Thisallowedustodefinetheminimumsamplesizeadequatetosetupamonitoringdesignfor2008sufficienttoensureahighdegreeofaccuracyattheprovincialandmunicipalscale.Thereliabilityof themethodanditsefficiencywereassessed in 2008 by measuring the RV and recalculatingTaylor’scoefficients.

MATERIALSANDMETHODS

Monitoring system organizationIn order to develop a homogeneous approach to the

health problems caused by Ae. albopictus over the wholeregion, the General Directorate for Health and SocialPoliciesofEmilia-RomagnahaspromotedaRegionalGroupforthesurveillanceandcontrolofAe. albopictus(Figure1).Itsaimwastosharetheknowledgeofallparticipantsinthemonitoringnetwork(localhealthunitsandmunicipalities),in particular among the members of the Scientific Groupthat provided expertise in entomology, epidemiology,meteorology,andinformatics.

The habitat characteristics of the survey station arefundamentaltotheeffectivenessoftheovitraps(Craigetal.2006).Therefore,inordertomaximizethestandardizationoftheenvironmentalparametersandtoavoiddifferencesinthe attractiveness among ovitraps (depending on relevantenvironmental characteristics of the station, i.e., shadingdegree),thechoiceoftheplaceandthepositioningoftheovitrapswaslefttoskilledtechnicians.

Emilia-RomagnaRegion

Public Health Department

Regionalgroup for technical

coordination

ScientificCentres

AnalysislaboratoriesEggs ovitraps

counting

MunicipalitiesMonitoring

management

Internet data publication

www.zanzaratigreonline.it

Figure 1. Aedes albopictus monitoring system organization in the Emilia-Romagna Region.

Figure1.Aedes albopictusmonitoringsystemorganizationintheEmilia-RomagnaRegion.


Theroutinemanagementoffieldovitrapswasconductedby municipal technicians. The Provincial LaboratoriesNetwork of the Regional Agency for EnvironmentalProtection(ARPA),theUniversitiesofParmaandFerrara,andtheMuseumofNaturalHistoryofParmawereinvolvedintheclassificationandcountingofthecollectedeggs.

In2008,theovitrapmonitoringwasbegunonweek20andcontinueduntilweek44.Datawerepublishedweeklyon a dedicated website (www.zanzaratigreonline.it) thatallowedthegenerationofreportsaccordingtothreelevelsof access: all registered users could get basic statistics ontheweeklymeanofAe. albopictus populationdensityandonthetrendoftheinfestationattheprovincialscale(firstlevelofaccess);municipalityoperatorscouldalsogetdatastatisticsrelated to their territory(second levelofaccess),and regional group members had access to data at theprovincialandmunicipalscaleandalsotothedataprovidedbyeachsingleovitrap(thirdlevelofaccess).ThecostsofthemonitoringsystemareshowninTable1.

Study areaIn 2008, the Emilia-Romagna region included 341

municipalities subdivided into nine provinces for a totalinhabited area of 22,122 km2(Table 2), with a populationof4,275,843andadensityof193/km2(ISTAT2009).Onlythe municipalities at altitudes below 500 m a.s.l. (70%of the total number of municipalities in the region) wereconsidered in the monitoring program, as the climaticconditionsofnorthernItalydonotallowthedevelopment

of high Ae. albopictus populations at higher altitudes(unpublisheddata).

Thenumberofmunicipalitiesinvolvedperprovinceinthemonitoringprogram,theirterritorialextension,andthenumberofovitrapsactivatedin2008arereportedinTable2. Within the 242 municipalities, 543 urban areas wereidentifiedandmonitoredfora totalsurfaceof104,973haandapopulationof1,202,223.Twentyurbanareasextendedovermorethan600ha,whilealltheothersweresmaller.

Ovitrap placement and managementHomogenoussurveyofeachinhabitedareawasobtained

by using GIS (Geographic Information Systems – ESRIArcView) todivideeacharea intoanumberofquadrantsequal to the number of ovitraps needed to be activated.EachovitrapwasgeoreferencedbyGPS-equippedpalmtopsand labeled by an identification code. The stations werekeptfixedforthewholeseason.Withineachquadrant,theovitrapwasplacedinagreen,shaded,andeasilyaccessiblearea.Itwaspositionedontheground,withafreespaceofatleast1maboveit.Eachovitrapconsistedofablackplasticpot(capacity:400ml;upperdiameter:8cm),filledtoabout2/3ofitsheightwithabout285mlofdechlorinatedwater.One12.5x2.5cmstripofmasonitewasfixedtotheovitrapwith a metal clip as egg deposition substrate. The weeklycheckof theovitrapsprovided for the replacementof thedeposition substrate and of the dechlorinated water, aftera careful cleaning in order to remove any eggs. Masonitestrips were then delivered to the Provincial Laboratory

Activities CostsEggcounting €80,000.00Ovitrappositioningandconsultants €33,960.00Routineovitrapmanagement(inchargeofthemunicipalities)* €409,864.00Total €523,824.00

Table1.Costsofthe2008monitoringprogramoftheEmilia-RomagnaRegion.

*Theglobalcosthasbeenestimatedconsidering7€perovitrapand26collectionturns.

Table2.Numberofmunicipalities,surfaceextension,andnumberofovitrapsperprovinceinthe2008monitoringnetwork.

Province No.municipalitiesinvolved Inhabitedmonitoredsurface(ha) No.ovitraps

Bologna 28 17,631 445

Forlì-Cesena 29 13,305 550

Ferrara 26 10,510 199

Modena 29 15,794 322

Piacenza 37 7,499 164

Parma 34 11,204 201

Ravenna 18 9,716 379

ReggioEmilia 21 11,686 231

Rimini 20 7,628 250

Total 242 104,973 2,741


NetworkandtotheUniversitiesofParmaandFerraraforclassificationandcounting.

Minimum ovitrap numberTheeggdensitydataobtainedfrom479ovitrapsduring

weeks27-37ofthe2007seasonwereusedtocalculatethecoefficients a and b of Taylor’s power law by regressingthe variance and the mean number of eggs/ovitrap/weekonalogarithmicscale.Eggdensitydatawereavailablefor12mainurbanareas>600haandfor44urbanareas<600ha.Fortheremainingtownswithurbanarea<600ha,forwhich local data were not available, Taylor’s coefficientswerecalculatedbyaveragingeggdensitydatacollectedbyalltheovitrapsactivatedintheregion.

In2008,thea andb coefficientsofTaylor’spowerlawcalculatedintheseason2007allowedtheactivationoftheminimum ovitrap number according to the desired levelsofaccuracy:D=0.2fortheurbanareas largerthan600ha(N=4), where the CHIK cases occurred in summer 2007;D=0.3fortheurbanareaslargerthan600hawhereCHIKepidemic did not occur (N=16) (Figure 2). In the smallurban areas (<600 ha, N=523) where it was not possibleto activate the proper number of ovitraps (often due toinsufficientfundavailability),anumberofovitrapsrangingfrom1to30werepositioned.

Testing the adequacy of the monitoring designIn the2008season the followingdatawerecollected:

i)meannumberofeggs/ovitrap/weekfromweek22to41;ii) mean number of eggs/ovitrap/week in the period ofmaximumdevelopment(fromweek27to37);iii)meanrateofoperatingovitraps(numberofworkingovitrap/numberofweeksofoperation)calculatedforeachurbanarea>600ha.

The adequacy of the number and distribution of theovitrapsofthemonitoringdesignappliedin2008(settledonthebaseofthe2007monitoringdata),wasassessedbymeansofthefollowingsteps:

1). Recalculation of the Taylor’s power law coefficients:the a and b coefficients were recalculated for each urbanarea>600haonthebaseoftheeggdensitydatacollectedfromweek22toweek41.

2). Recalculation of the adequate number of ovitraps:thenewTaylor’scoefficientswereusedtocalculatehowmanyovitraps should have been activated in the same area inordertoobtainthedesireddegreeofprecision.

3). Calculation of the percentage of error of the 2007 estimate:forthesamelocality,thecomparisonbetweenthenumberofovitrapscalculatedin2007andrecalculatedin2008allowedustoevaluatethepercentageoferrorofthe2007estimate.

4). Trend of RV:calculationofthetrendofthecoefficientofRelativeVariationof themeaneggdensitywithineachurbanarea.Allmeandataarepresentedwiththeirstandarddeviations.

RESULTS

Calculation of the minimum ovitrap number in 2007The VMR of the mean number of eggs/ovitrap/week

in 2007 was always higher than 1, indicating a clustereddistributionofthemosquitopopulation.ItwasthuspossibletocalculatetheTaylor’scoefficientsa andb, bytheempiricalfunctionalrelationshiponalogarithmicscale,betweenthevariance(s2)andthemean(m)ofsamples(Figure3)andtocalculate theoptimalovitrapnumber foreachurbanareaaccording to the desired level of accuracy (Figure 4 andTable3).

Monitoring management and results in 2008 In 2008 the monitoring network was activated in all

of the nine Provinces of the Emilia-Romagna region and242municipalitiesthatwereinvolved.Twentyurbanareasextendedovermorethan600ha(Table3),while523urbanareashadanextensionlowerthan600ha.Atotalof2,741ovitrapswaspositioned.Inthe20urbanareasof>600ha,1,396ovitrapswereplaced,while1,349ovitrapsweresetoffinthe523smallerurbanareas(datanotshown).

Figure 2. Details of the study area. PC, Piacenza; PR, Parma; RE, Reggio Emilia; MO, Modena;BO, Bologna; FE, Ferrara; RA, Ravenna; FC, Forlì-Cesena; RN, Rimini.

km0 20 40 km

BOMO

REPR

PCRA

FCRN

FE

2007 CHIKV municipalities

Administrative borders

Emilia subregion D = 0.3

Romagna subregion D = 0.2

Figure2.Detailsofthestudyarea.PC,Piacenza;PR,Parma;RE,ReggioEmilia;MO,Modena;BO,Bologna;FE,Ferrara;RA,Ravenna;FC,Forlì-Cesena;RN,Rimini.

Figure3.Taylor’spowerlawregressionlineforthenumberof eggs/ovitrap/week of Ae. albopictus. Data collected bymeansofthe523ovitrapsplacedinthe20urbanareas>600hain2007.

0.1

1.0

10.0

100.0

1,000.0

10,000.0

100,000.0

0.1 1.0 10.0 100.0 1,000.0

Mean

Var

ianc

e

Figure 3. Taylor’s power law regression line for the number of eggs/ovitrap/week of Ae. albopictus. Data collected by means of the 523 ovitraps placed in the 20 urban areas > 600 ha in 2007.


Figure4.Relationshipbetweenthemeaneggsdensityintheovitraps and the minimum ovitrap’s number adequate forthe monitoring protocol according to the desired level ofaccuracy.Calculationsweremadeonfielddatacollectedin2007in20urbanareas.

No. eggs/ovitrap

0 50 100 150 200 250 300 350 4000

100

200

300

400

500

600

700

800

900

1,000

Min

imum

num

ber o

f ovi

trap

D = 0.3D = 0.2D = 0.1

Figure 4. Relationship between the mean eggs density in the ovitraps and the minimum ovitrap’s number adequate for themonitoring protocol according to the desired level of accuracy. Calculation were made on field data collected in 2007 in 20urban areas.

Table3.Detailsofthemonitoringnetworkfortheurbanareas>600ha,meaneggdensity,andvalidationofthemonitoringdesign.

Monitoringnetworkdetails Monitoringresults Monitoringdesignvalidation

Urbanareas>600ha

No.ovitraps2007

No.ovitraps2008

%operatingovitraps

Eggdensitydata(eggs/ovitrap/week)

Taylor’scoeff.basedon2008eggdensity

Minimumnecessarynumberofovitraps

Weeks Mean±SD a b calculatedon2008eggdensitydata

%differencecomparedto2007

estimate

Piacenza 20 52 100.0 22-41 61.4±97.73 16.0 1.4 57.3 9.227-37 88.7±112.26 45.9 -13.3

Parma 22 90 100.0 22-41 18.0±28.90 17.4 1.2 83.8 -7.427-37 25.2±33.67 65.0 -38.5

ReggioEmilia 25 56 99.7 22-41 35.5±51.72 20.2 1.2 53.4 -4.927-37 56.2±59.41 37.3 -50.1

Correggio - 50* 98.2 22-41 49.5±60.32 13.9 1.3 35.7 -40.127-37 73.9±65.35 26.7 -87.3

Modena - 60* 97.8 22-41 43.4±60.87 14.1 1.4 54.7 -9.727-37 67.6±69.38 41.2 -45.6

Carpi 18 50 96.9 22-41 60.5±71.42 21.9 1.2 36.4 -37.427-37 86.9±74.74 27.3 -83.2

Formigine 17 25 96.4 22-41 74.2±82.43 31.8 1.1 27.5 9.127-37 112.0±84.36 18.9 -32.3

Mirandola - 50* 92.6 22-41 69.0±90.62 12.4 1.4 46.6 -7.327-37 102.7±100.23 37.1 -34.8

Bologna 150 58+58** 75.9 27-37 43.2±56.68 31.7 1.1 55.1 -5.327-37 63.2±63.40 39.8 -45.7

Casalecchio - 50* 79.0 22-41 49.3±74.31 15.4 1.3 50.1 0.227-37 87.5±89.66 34.3 -45.8

Imola - 46* 51.0 31-41 33.9±55.84 20.2 1.3 82.1 44.031-37 44.9±61.96 68.1 32.5

Ferrara 73 65 100.022-41 54.2±94.08

27.5 1.496.2 32.4

27-37 81.4±109.55 74.6 12.9

Cervia - 58* 100.0 22-41 63.2±90.55 17.2 1.4 56.8 -2.127-37 96.8±102.56 43.7 -32.7

Faenza - 50* 98.7 22-41 51.1±65.99 16.6 1.3 53.0 5.727-37 72.6±73.10 42.1 -18.8

Lugo 17 35 100.0 22-41 58.8±71.68 30.0 1.2 46.0 23.927-37 82.9±77.82 34.7 -0.9

Riccione - 50* 95.822-41 84.1±109.10

10.8 1.555.5 9.9

27-37 133.9±155.28 44.5 -12.4

Forli 46 141*** 96.7 22-41 60.0±84.75 26.5 1.3 132.8 -6.227-37 93.6±97.23 96.4 -46.3

Cesena 55 145*** 93.2 22-41 51.8±80.50 28.7 1.3 166.2 12.827-37 78.1±93.10 124.1 -16.8

Ravenna 59 106*** 98.8 22-41 48.2±75.05 26.6 1.2 111.2 4.727-37 72.0±85.12 80.4 -31.8

Rimini 60 101*** 96.6 22-41 72.7±109.10 13.7 1.5 134.0 24.627-37 115.5±125.54 104.7 3.5

*Nodataavailablein2007;theminimumovitrapnumbertobeplacedin2008wasestimated(seetext).**DuetotheextensionofthetownofBologna,twosamplingareaswereconsidered.Theminimumnumberofovitrapscalculatedin2008isreferredtoeachurbanarea(i.e.,110.2ovitrapsistheminimumnecessarynumber).***MainurbanareasforwhichthedesiredlevelofprecisionDwasfixedasD=0.2.ForalltheothersthedesiredlevelofprecisionwasfixedasD=0.3.


Thepercentageoftheoperatingovitraps(i.e.,regularlypositioned and functioning at the weekly check) in theurban areas that extended over more than 600 ha rangedfrom 75.9 to 100.0% (Table 3), with a mean of 96.4%. Ina few cases, organizational problems caused a significantdelayofthemonitoringstart,thusreducingthepercentageofworkingweeks(i.e.,inthecaseofImola).

The whole season (weeks 20-44 included) mean eggdensity calculated over all the ovitraps placed in the 20urbanareas>600ha(numberofvalidsamples:54,365)was49.2±80.8eggs/ovitrap/week.

Thehighestmeaneggdensityintheweeks22-41wasrecorded in Riccione (84.1 ± 109.1 eggs/ovitrap/week)whilethelowestwasregisteredinParma(18.0±28.9eggs/ovitrap/week). During the period of maximum mosquitopopulationincrease(weeks27-37),theeggdensitiesrangedfrom 25.2 ± 33.7 eggs/ovitrap/week in Parma to 133.9 ±155.3 in Riccione, and the average over the whole regionwas77.7±95.7eggs/ovitrap/week.

Theseasonaldynamicanalysis(Figure5)showedfourpeaksofeggslaidintheovitraps,thefirstinearlyJune(23rdweek), thesecondoneat thebeginningofJuly(week27),thethirdoneatmid-August(week33),andthefourthonein mid-September (week 37), as expected. Occasionally,eggsofAe. geniculatusOlivierwerefoundintheovitraps,but,becauseoftheaimofthisstudy,onlyAe. albopictuseggswerecounted.

Testing the monitoring design adequacyRecalculation of Taylor’s power law coefficientsin 2008.

Taylor’scoefficientsb, recalculatedin 2008andreportedinTable 3, indicated a different aggregation level accordingtothearea,beinglowerforthedataoftheinhabitedareasof Formigine and Bologna (b=1.1), Carpi, Parma, ReggioEmilia, Lugo, and Ravenna (b=1.2) in comparison to theaggregationlevelmeasuredforPiacenza,Ferrara,Modena,Mirandola, and Cervia (b=1.4), Rimini and Riccione(b=1.5).

Recalculation of the adequate number of ovitraps. InTable3,theminimumnumberofovitrapsbasedontheeggdensityregisteredin2008andthepercentdifferenceswithrespecttotheminimumovitrapnumberscalculatedinthe

2007arereported.Considering the time period between the weeks 22-

41 for six out of 20 urban areas, the percent differencesbetweentheminimumnecessaryovitrapnumbersprovidedby the estimates 2007 and the 2008 recalculations werebelow 6%. In particular, to obtain the prefixed levels ofaccuracyofthemonitoringsystem,inCasalecchio,Faenza,and Ravenna up to 6% more ovitraps would have beenneeded,whileinReggioEmilia,Bologna,andCerviafewerthan6%exceedingovitrapswereactivated.Foreighturbanareas, higher percent differences were found; in some ofthemupto32.4%moreovitrapswouldhavebeenneeded(asinFerrara),whilefortheremainingsixurbanareasupto40.1%moreovitrapswerepositioned(Table3).

Duringtheperiodofmaximumactivity,intheweeks27-37,theminimumovitrapnumbercalculatedin2007andpositionedin2008wasadequatetoensurethedesiredlevelsofaccuracyof themonitoringsystem.Intwourbanareas(Lugo and Rimini), the variations between the 2007 and2008calculationswerebelow6%.For16urbanareas, theminimumovitrapnumbercalculatedin2008rangedfrom-12.4% (Riccione) to -87.3% (Correggio) with respect tothe2007calculations,showingthatthenumberofovitrapspositionedin2008hadbeenlargelyoverestimatedforthisperiod(Table3).

Trend of RV. At the regional scale, from week 27 to37, theeggdensitydatacollectedthroughthemonitoringnetwork showed a good mean Relative Variation (RV <0.25), thus demonstrating that the number of ovitrapspositioned in theurbanareas>600hahadbeencorrectlycalculatedtoobtainthedesiredmonitoringprecisionlevel,fixedatD=0.2forthechikungunyaareasandatD=0.3fortheotherareas(Figure6).

Attheprovincialscale,ovitrapnumbersinurbanareas>600hawerehighenoughtoobtainhighlyreliablemeanRelative Variation (RV <0.2) in all Provinces. Moreover,

Figure5.SeasonalpopulationdynamicofAedes albopictusbasedoneggdensityin2008.

21 23 25 27 29 31 33 35 37 39 41

Week

0

20

40

60

80

100

120

No.

eggs

/ ovi

trap/

wee

k

Mean± SE ± 0.95 Conf.Int.

Figure 5. Seasonal population dynamic of Aedes albopictus based on egg density in 2008.

Week

RV

Other areas D = 0.3

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

0.22

0.24

0.26

27 28 29 30 31 32 33 34 35 36 37

Chikungunya areas D = 0.2

27 28 29 30 31 32 33 34 35 36 37

Median

25%-75%

Min-Max

Figure 6. Weekly RV average calculated for the Chikungunya areas (required precision level D = 0.2) and in the other areas (requiredprecision level D = 0.3) from week 27 to 37.

Figure6.WeeklyRVaveragecalculatedforthechikungunyaareas(requiredprecisionlevelD=0.2)andintheotherareas(requiredprecisionlevelD=0.3)fromweek27to37.


the number of all the ovitraps positioned within eachprovincial territory (urban areas > 600 ha + urban areas< 600 ha) allowed an even higher level of accuracy (D =0.1),withtheweeklytrendoftheRV<0.1formostoftheseason,asdepictedinFigure7fortheProvincesinwhichthechikungunyaoutbreakoccurredin2007,andinFigure8fortheotherProvinces.

Considering only the ovitraps positioned in theprovincial capital urban areas of Ravenna, Rimini, andForlì-Cesena,thenumberofactivatedovitrapsobtainedthedesiredprecisionlevelD=0.2atleastinthemiddleofthesummerseason(week26-38;datanotshown).

DISCUSSION

The monitoring system applied in the 2008 seasonprovidedtheactivationof2,741ovitrapsininhabitedareasofEmilia-Romagnaregionwithanaltitudelessthan500m.Inweeks22-41,thesizeoftheovitrapmonitoringnetwork,activatedandtestedinthe2008seasononthebaseofthe2007 mean egg density data, was underestimated by only2.7% (intended as the minimum number of ovitraps).This means that the ovitrap numbers needed to maintaina precision of D=0.2 throughout the Ae. albopictus development season in the towns where the epidemic ofchikungunyaoccurredin2007(Ravenna,Rimini,Forlì,andCesena),andD=0.3inalltheothertowns>600hashouldhavebeenequalto1,434insteadofthe1,396calculatedonthe2007monitoringsampledata.

Inparticular,intheurbanareasinfestedbyCHIKV,itshould have been necessary to increase the whole seasonmonitoringnetworkonanaverageof9%,asthenumberofovitrapsappearedtobetoolowinCesena(underestimatedby 13%, 145 were activated instead of 166) and Rimini(underestimated by 25%, 101 ovitraps instead of 134). IntheotherurbanareasforwhichaprecisionlevelD=0.3wasrequested,themonitoringnetworkappearedgenerallytobeadequate,evenif,insomecases,itwouldhavebeenusefultoincreasethenumberofovitraps(i.e.,inLugo,from35to46ovitraps).

From week 27 to week 37, during the period ofmaximum abundance of egg laying females (i.e., ofmaximum egg density in the ovitraps), the minimumnumberofovitrapsneededfor thesame levelofaccuracywaslowerforallthestudiedareaswithrespecttothewholedevelopmentalperiod.Infact,in17urbanareasoutof20thenumberofovitrapspositionedin2008exceededbyanaverage 29.4% the number that would have ensured thedesiredlevelofaccuracyofthemonitoringprogram,bothin theareaswhere thechikungunyaoutbreakoccurredorwhereitdidnotoccur,andthismeansthat1,048ovitrapswould have been required to gain the prefixed levels ofaccuracy.Thenumberofovitrapsneededforanyprefixedlevel of accuracy of the monitoring program is lower ifmosquitodispersionintheurbanareastendstobeuniform(i.e.,theTaylor’scoefficientbislow)thanitisifthemosquitopopulationhasspreadaccordingtoaclustereddistribution(highbvalue).

Manyenvironmentalvariables,suchasthedimensionof the town, the incidence of small premises or quarterswithlargebuildings,thepresenceofgreenareas,thedegreeofmaintenanceofthepremisesandthecourtyards,andthenumberofprivateandpubliccatchbasinsaffectthedensityanddispersionofAe. albopictusintheurbanenvironment.Inadditionthecontrolprogramsadoptedbythedifferentmunicipalitiesaffectthereliabilityoftheovitrapmonitoringdifferently.

In the central part of the reproductive season, theovitrapmonitoringnetworkprovideddatawithlowRelativeVariation (RV<0.2), thus indicating that the minimumovitrap numbers in the urban areas >600 ha could be

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

20 24 28 32 36 44Week

RV

RA RN FC

40

Figure 7. RV trend in 2008 for the provinces where Chikungunya epidemic occurred in 2007. All the ovitraps positionedwithin the provincial territory were considered. RA: Ravenna; RN: Rimini; FC: Forlì-Cesena.

Figure 7. RV trend in 2008 for the provinces whereChikungunyaepidemicoccurredin2007.All theovitrapspositionedwithintheprovincialterritorywereconsidered.RA:Ravenna;RN:Rimini;FC:Forlì-Cesena.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

20 24 28 32 36 44Week

RV

PC PR REMO BO FE

40

Figure 8. RV trend at the provincial scale in 2008 for the provinces where Chikungunya did not occur in 2007. All theovitraps positioned within the provincial territory were considered. PC: Piacenza; PR: Parma; RE: Reggio Emilia;MO: Modena; BO: Bologna; FE: Ferrara.

Figure 8. RV trend at the provincial scale in 2008 for theprovinces where Chikungunya did not occur in 2007. Alltheovitrapspositionedwithintheprovincialterritorywereconsidered. PC: Piacenza; PR: Parma; RE: Reggio Emilia;MO:Modena;BO:Bologna;FE:Ferrara.


further reduced, lowering the maintenance costs (Albierietal.2010).Carrierietal.(unpublisheddata)showedthatin urban areas in Italy a correlation exists between themeannumberofeggslaidbyAe. albopictus femalesintheovitrapsandthemeanadultpopulationdensitycalculatedby entomological survey, such as the Pupal DemographicSurvey (PDS). PDS is a survey method based on theestimation of female populations by pupal density and isusedinmostepidemiologicalstudiesonAe. aegypti (Focks2003). In Italy, Ae. albopictus colonizes mainly the catchbasins in both public and private areas (Carrieri et al.,unpublished data). Catch basin sampling is highly time-consuming,andthedevelopmentofamonitoringnetworkbased on PDS would require unsustainable costs for aroutinesurveillanceprogram.

The monitoring design described in this paperprovided reliable monitoring datasets for most of thesummerseason,andthesmoothedmapobtainedthroughtheinterpolationofdatawithanRV<0.3(Albieri2010)willallow a general estimate of the egg density in those areaswhere themonitoring isnotproperlyorganized.Thecostof the monitoring program was around 0.13 Euro/person(=523,824/4,086,874 total costs/number of inhabitants inthe region). Management costs accounted for 78% of thetotal (409,864-523,824 €), while egg identification andcounting accounted for 15% of the total costs (80,000-523,824 €). Only 6.5% of the total costs were invested inscientific supervision and technical assistance for carefulplanning of the territorial organization of the monitoringnetwork and for ensuring a highly standardized ovitrappositioning.

The low cost of this monitoring program based onthe use of ovitraps makes the system sustainable even ina non-epidemic season. In addition, the availability ofcontinuoussetsofdataonmosquitopopulationsprovidesinformation on real-time development of the infestationandcanthereforebeknowledgereadilyavailableincaseofanepidemicemergency.Thenetworkhasprovidedusefulinformation to implement systems topreventandcontrolthe outbreak to arboviruses such as chikungunya anddengue,andcreatedaWEB-GISapplicationthatfacilitatesthe analyses of the monitoring data by the regionalcoordinatinggroup(Albierietal.2010,Bellinietal.2010).

TheorganizationalschemecoordinatedbytheEmiliaRomagna Public Health Department has helped managethe system without major problems. In particular, datawererapidlyanddirectlytransferred,fromthefieldtotheInstitutionschargedforthecontrolprogramactivation,andalsotothecitizens.Thesystemprovedtobehighlyefficientandpreparedtofaceanyaccidentalarrivalofviralinfectionreservoirs,whichcouldcauseoutbreaksifthediseasevectorinsect is present at high density levels. The occasionalcollectionofAe. geniculatuseggsshowedthattheovitrapsarealsosuitablebreedingsitesforothermosquitospecies,suchasAe. aegyptiandAe. japonicus Theobald,whichhaveneverbeenfounduptonow.

Medicalsurveillanceaccompaniedbyanentomologicalsurveillanceisessentialtopreventthespreadofarboviruses

and toevaluate theriskof theviraldiseaseoutbreak.Thedevelopmentofanefficientmonitoringnetworkisalsoanessential tool forverifying theeffectivenessof thecontrolmeasures.

Theresultsofthepresentstudysupportthedevelopmentof a new monitoring system based on fortnightly ovitrapchecking(Carrierietal.2009)inordertoreducethecostsofthemonitoringnetwork.

Acknowledgments

We thank the “Emilia-Romagna Team for Aedes albopictus surveyandcontrol”forthepreciouscollaboration.Inparticular,wearegratefultoRossiL.(Emilia-RomagnaRegional Agency for Environmental Protection), LeisM. (University of Ferrara), and Mezzadri M. (Museumof Natural History of Parma) for supervision during eggcounting.

REFERENCESCITED

Adhami,J.andN.Murati.1987.PresencedumoustiqueAe. albopictus enAlbanie.Revist.Mjekesore1:13-16.

Albieri, A., M. Carrieri, P. Angelini, F. Baldacchini, C.Venturelli, S. Mascali Zeo, and R. Bellini. 2010.QuantitativemonitoringofAedes albopictus(Skuse)inEmilia-Romagna,NorthernItaly:clusterinvestigationandgeostatisticalanalysis.Bull.Insectol.63:209-216.

Angelini,P.2006.Comeleconoscenzeacquisitenell’ambitodelprogettoregionalepossonofornireuoveindicazionioperative.In:Emilia-RomagnaRegion(ed.)Verso una strategia di lotta integrata alla Zanzara tigre.pp.125-127.CentroStampaGiuntaRegioneEmilia-Romagna,Bologna,Italy.

Angelini, R., A.C. Finarelli, P. Angelini, C. Po, K.Petropulacos,P.Macini,C.Fiorentini,C.Fortuna,G.Venturi,R.Romi,G.Majori,L.Nicoletti,G.Rezza,andA.Cassone.2007.AnoutbreakofChikungunyafeverintheprovinceofRavenna,Italy.EuroSurveill.12:pii=3260.

Aranda, C., R. Eritja, and D. Roiz. 2006. First record andestablishment of the mosquito, Aedes albopictus, inSpain.Med.Vet.Entomol.2:150-152.

Bellini, R., M. Carrieri, G. Burgio, and M. Bacchi. 1996.Efficacy of different ovitraps and binomial samplinginAedes albopictus surveillanceactivity.J.Am.Mosq.Contr.Assoc.12:632-636.

BelliniR.,S.MascaliZeo,C.Venturelli,A.Albieri,M.Carrieri,P.Angelini,F.Baldacchini, andS.Fontanari. 2010.Aweb-basedGISforAedes albopictusandChikungunyavirussurveillance inEmilia-RomagnaRegion.p.102.Proc. 2nd Int. EDEN Conf. 2010, Montpellier, France.http://international-conference2010.eden-fp6project.net/

Bonilauri,P.,R.Bellini,M.Calzolari,R.Angelini,L.Venturi,F. Fallacara, P. Cordioli, P. Angelini, C. Venturelli, G.Merialdi,andM.Dottori.2008.ChikungunyavirusinAedesalbopictus,Italy.Emerg.Infect.Dis.14:852-854.


Buntin,G.D.1994.Developingaprimarysamplingprogram.In: L.P. Pedigo and G.D. Buntin (eds.) Handbook of Sampling Methods for Arthropods in Agriculture. pp.99-115.CRCPress,Inc.,BocaRaton,Florida.

Carrieri, M., A. Masetti, A. Albieri, B. Maccagnani, andR. Bellini. 2009. Larvicidal activity and influenceof Bacillus thuringiensis var. israelensis on Aedes albopictus oviposition in ovitraps during a two-weekcheckintervalprotocol.J.Am.Mosq.Contr.Assoc.25:149-155.

Craig,R.W.,S.A.Long,R.C.Russell,andS.A.Ritchie.2006.Optimizing ovitrap use for Aedes aegypti in Cairns,Queensland,Australia.J.Am.Mosq.Contr.Assoc.22:635-640.

DallaPozza,G.andG.Majori.1992.FirstrecordofAedes albopictus establishment in Italy. J.Am.Mosq.Contr.Assoc.8:318-320.

Flacio, E., P. Lüthy, N. Patocchi, F. Guidotti, M. Tonolla,and R. Peduzzi. 2004. Primo ritrovamento di Aedes albopictus inSvizzera.Boll.Soc.TicineseSc.Natl.92:141-142.

Focks, D.A. 2003. A review of entomological samplingmethods and indicators for dengue vectors. WHO,Geneva,Switzerland.40pp.

Hawley,W.A.1988.ThebiologyofAedes albopictus.J.Am.Mosq.Contr.Assoc.4:1-40.

ISTAT.2009.Rapporto annuale: la situazione nel paese nel 2008.RTIPoligraficaRuggieros.r.l.andA.C.M.s.p.a.,Avellino.428pp.

Klobučar,A.,E.Merdić,N.Benić,Ž.Baklaić,andS.Krćmar.2006.FirstrecordofAedesalbopictus inCroatia.J.Am.Mosq.Contr.Assoc.22:147-148.

Kuno,E.1991.Samplingandanalysisofinsectpopulations.Annu.Rev.Entomol.36:285-304.

de Lamballerie, X., E. Leroy, R.N. Charrel, K. Tsetsarkin,S. Higgs, and E.A. Gould. 2008. Chikungunya virusadaptstotigermosquitoviaevolutionaryconvergence:asignofthingstocome?Virol.J.5:33-36.

Mogi, M., C. Choochote, C. Khamboonruang, and P.Suwanpanit. 1990. Applicability of presence, absenceand sequential sampling for ovitrap surveillance ofAedes (Diptera: Culicidae) in Chiang Mai, northernThailand.J.Med.Entomol.27:509-514.

Petrić,D.,I.Pajović,A.IgnjatovićĆupina,andM.Zgomba.2001. Possible establishment of Aedes albopictus (Skuse, 1894) in Serbia and Montenegro. p. 85. Proc.14th Europ. Conf., Soc.Vector Ecol., Bellinzona,Switzerland.

Ramchurn,S.K.,S.S.D.Goorah,D.Mungla,B.Ramsurrun,V.Pydiah,andA.Summun.2008.Astudyofthe2006chikungunyaepidemicoutbreakinMauritius.InternetJ.Med.Update3:11-21.

Renault,P., J.L.Solet,D.Sissoko,E.Balleydier,S.Larrieu,L.Filleul,C.Lassalle,J.Thiria,E.Rachou,H.deValk,D. Ilef, M. Ledrans, I. Quatresous, P. Quenel, and V.Pierre.2007.Amajorepidemicofchikungunyavirus

infectiononReunionIsland,France,2005–2006.Am.J.Trop.Med.Hyg.77:727-731.

Rezza,G.,L.Nicoletti,R.Angelici,R.Romi,A.C.Finarelli,M. Panning, P. Cordioli, C. Fortuna, S. Boros, F.Magurano, G. Silvi, P. Angelini, M. Dottori, M.G.Ciufolini,G.Majori, andA.Cassone.2007. Infectionwith chikungunya virus in Italy: an outbreak in atemperateregion.Lancet.9602:1840-1846.

Sabatini,A.,V.Raineri,G.Trovato,andM.Coluzzi.1990.Aedes albopictus in Italia e possibile diffusione dellaspecie nell’area del mediterraneo. Parassitologia 32:301-304.

Samanidou-Voyadjoglou,A.,E.Patsoula,G.Spanakos,andN.C. Vakalis. 2005. Confirmation of Aedes albopictus (Skuse)(Diptera:Culicidae)inGreece.Eur.Mosq.Bull.19:10-12.

Saxena,S.K.,M.Singh,N.Mishra,andV.Lakshmi.2006.ResurgenceofchikungunyavirusinIndia:anemergingthreat.EuroSurveill.11:pii=3019.

Schaffner, F. and S. Karch. 2000. First report of Aedes albopictus (Skuse,1894)inmetropolitanFrance.C.R.Acad.Sci.III.323:373-375.

Schaffner,F.,W.VanBortel,andM.Coosemans.2004.Firstrecord of Aedes (Stegomyia) albopictus in Belgium. J.Am.Mosq.Contr.Assoc.20:201-203.

Scholte,E.-J.andF.Schaffner.2007.Waitingfor the tiger:establishment and spread of the Aedes albopictus mosquito inEurope. In:W.Takken.andB.G.J.Knols(eds.) Emerging Pests and Vector-Borne Diseases in Europe. Book series: Ecology and control of vector-borne diseases, Vol. 1, Wageningen AcademicPublishers,Wageningen,TheNetherlands.

Service, M.W. 1993. Mosquito Ecology: Field Sampling Methods.2nded.Elsevier,London.1,007pp.

Southwood,T.R.E.andP.A.Henderson.2000.Thesamplingprogramme and the measurement of description anddispersion.In:Ecological Methods.pp.7-62.BlackwellScienceLtd.,Oxford.

Taylor, L.R. 1961. Aggregation variance and the mean.Nature189:732-735.

Taylor, L.R. 1984. Assessing and interpreting the spatialdistributions of insect populations. Annu. Rev.Entomol.29:321-357.

Tsetsarkin, K.A., D.L. Vanlandingham, C.E. McGee, andS. Higgs. 2007. A single mutation in Chikungunyavirusaffectsvectorspecificityandepidemicpotential.PLoS Pathog. 3(12): 418 e201. doi:10.1371/journal.ppat.0030201.

Vazeille, M., S. Moutailler, F. Pages, F. Jarjaval, and A.Failloux. 2008a. Introduction of Aedes albopictus in Gabon: what consequences for dengue andChikungunya transmission?Trop.Med. Int.Hlth.13:1176-1179.

Vazeille, M., C. Jeannin, E. Martin, F. Schaffner, and A.Failloux.2008b.Chikungunya-ariskforMediterraneancountries.ActaTrop.105:200-202.

Documents

Surveillance of the chikungunya vector Aedes albopictus (Skuse) in Emilia-Romagna (northern Italy): organizational and technical aspects of a large scale monitoring system