<%BANNER%>

UFIR



Risk perception and evacuation decisions of Florida tourists under hurricane threats: a stated preference analysis
CITATION PDF VIEWER
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/IR00001201/00001
 Material Information
Title: Risk perception and evacuation decisions of Florida tourists under hurricane threats: a stated preference analysis
Series Title: Matyas, C., S. Srinivasan, I. Cahyanto, B. Thapa, L. Pennington-Gray, and J. Villegas, 2011: Risk perception and evacuation decisions of Florida tourists under hurricane threats: a stated preference analysis. Nat. Hazards, 59, 871-890.
Physical Description: Journal Article
Creator: Matyas, Corene
Publisher: Natural Hazards
Publication Date: 2011
 Notes
Abstract: Though most hurricane evacuation studies have focused on residents, tourists are also a vulnerable population. To assess their perceptions of risk and evacuation likelihood under different hurricane conditions, we surveyed 448 tourists visiting central Florida. Respondents viewed four maps emulating track forecast cones produced by the National Hurricane Center and text information featuring variations of storm intensity, coast of landfall, centerline position relative to the survey site, time until landfall, and event duration. We performed chi-square tests to determine which hurricane conditions, and aspects of tourists such as their demographics and previous hurricane experience, most likely influenced their ratings of risk and evacuation likelihood for respondents located on Pinellas County beaches or inland near Orlando, FL. Highly rated scenarios featured a Category 4 hurricane making landfall along the Gulf Coast with the centerline passing over the sampling site. Overall, tourists that indicated the highest risk and evacuation ratings were not previously affected by a hurricane, had a trip duration of less than 6 days, and had checked for the possibility of a hurricane strike before departure. However, results for other tourist attributes differed between tourists in coastal and inland locations. We found that although somewhat knowledgeable about hurricanes, tourists misinterpreted the track forecast cone and hurricane conditions, which led to a lower perception of risk and subsequent likelihood to evacuate. Tourists, particularly those from outside of Florida, need to be better educated about the risks they face from hurricanes that make landfall.
Acquisition: Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Corene Matyas.
Publication Status: Published
 Record Information
Source Institution: University of Florida Institutional Repository
Holding Location: University of Florida
Rights Management: All rights reserved by the submitter.
System ID: IR00001201:00001

Downloads

This item is only available as the following downloads:

1matyas2011NatHaz ( PDF )


Full Text

PAGE 1

ORIGINALPAPER RiskperceptionandevacuationdecisionsofFlorida touristsunderhurricanethreats:astatedpreference analysis CoreneMatyas € SivaramakrishnanSrinivasan € IgnatiusCahyanto € BrijeshThapa € LoriPennington-Gray € JorgeVillegas Received:3December2010/Accepted:23March2011 SpringerScience+BusinessMediaB.V.2011 Abstract Thoughmosthurricaneevacuationstudieshavefocusedonresidents,tourists arealsoavulnerablepopulation.Toassesstheirperceptionsofriskandevacuationlikelihoodunderdifferenthurricaneconditions,wesurveyed448touristsvisitingcentral Florida.Respondentsviewedfourmapsemulatingtrackforecastconesproducedbythe NationalHurricaneCenterandtextinformationfeaturingvariationsofstormintensity, coastoflandfall,centerlinepositionrelativetothesurveysite,timeuntillandfall,and eventduration.Weperformedchi-squareteststodeterminewhichhurricaneconditions, andaspectsoftouristssuchastheirdemographicsandprevioushurricaneexperience,most likelyinuencedtheirratingsofriskandevacuationlikelihoodforrespondentslocatedon PinellasCountybeachesorinlandnearOrlando,FL.Highlyratedscenariosfeatureda Category4hurricanemakinglandfallalongtheGulfCoastwiththecenterlinepassingover thesamplingsite.Overall,touriststhatindicatedthehighestriskandevacuationratings werenotpreviouslyaffectedbyahurricane,hadatripdurationoflessthan6days,andhad checkedforthepossibilityofahurricanestrikebeforedeparture.However,resultsfor othertouristattributesdifferedbetweentouristsincoastalandinlandlocations.Wefound thatalthoughsomewhatknowledgeableabouthurricanes,touristsmisinterpretedthetrack C.Matyas( & ) DepartmentofGeography,UniversityofFlorida,3141TurlingtonHallBox117315, Gainesville,FL32611,USA e-mail:matyas@u.edu S.Srinivasan DepartmentofCivilandCoastalEngineering,UniversityofFlorida,Gainesville,FL,USA I.Cahyanto B.Thapa L.Pennington-Gray TourismCrisisManagementInstitute,EricFriedheimTourismInstitute, UniversityofFlorida,Gainesville,FL,USA I.Cahyanto B.Thapa L.Pennington-Gray DepartmentofTourism,Recreation,andSportManagement,UniversityofFlorida, Gainesville,FL,USA J.Villegas DepartmentofBusinessAdministration,UniversityofIllinoisatSpringeld,Springeld,IL,USA 123 NatHazards DOI10.1007/s11069-011-9801-0

PAGE 2

forecastconeandhurricaneconditions,whichledtoalowerperceptionofriskandsubsequentlikelihoodtoevacuate.Tourists,particularlythosefromoutsideofFlorida,needto bebettereducatedabouttheriskstheyfacefromhurricanesthatmakelandfall. Keywords Hurricanes Tourists Evacuation Riskperception Florida 1Introduction HurricanesaffectFloridamorethananyotherUSstate(Elsneretal. 2004 ).Whilestorm surgeanddamagingwindsaremajorconcernsalongthecoastline,peoplelocatedinland arealsovulnerabletohurricaneforcewinds,tornadoes,andoodingfromheavyrainfall (Rappaport 2000 ).Giventhathurricanescancausewidespreaddestruction,efcient evacuationstrategiesarecriticalforsavinglives.Manyrecentstudieshaveexamined evacuationbehaviorsofresidentsinresponsetohurricanes(e.g.,Arlikattietal. 2006 ; CutterandFinch 2008 ;Kangetal. 2007 ;Kusenbachetal. 2010 ;Senkbeiletal. 2010 ; SmithandMcCarty 2009 ;Zhangetal. 2007 ).Thesestudieshaveshownthathurricanerelatedconditionssuchasintensityandprojectedtrackandpersonalattributes(i.e., experiencingfalsealarms,owningpets,income,andage)affecttheperceptionsofriskand evacuationdecisionsofresidents. However,especiallyinthecontextofFlorida,itisimportanttospecicallyaddress theevacuationbehavioroftourists.AccordingtoVisitFlorida( 2010 ),theState receivesapproximately80.9million(88.1%aredomestic)visitorsannually,manyof whomarriveduringthehurricaneseasonthatspansJune1November30.Touristsare avulnerablepopulationforseveralreasons(PhillipsandMorrow 2007 )includingthe factthattheymaylackknowledgeabouttheriskspresentedbyhurricanes.Also,they aretypicallyinunfamiliarsurroundingsandarewithoutthenormalsupportsystemsof theirhomecommunity(BurbyandWagner 1996 ;WorldTourismOrganization 1998 ). Yet,thefewsignicantcontributionsontheevacuationoftouristsduringhurricanes havebeenpublishedlargelybyoneauthor(Drabek 1991 1993 1994 1996 )whose workprincipallyfocusedonevacuationstrategiesandpoliciesfromasupplyperspective.Fewstudieshaveexaminedthedemandperspective(i.e.,thetourists'preferencesandintentions),andtoourknowledge,noempiricalstudiesoftourists perceptionsandevacuationpreferencesunderhurricanethreatshavebeenpublished and/oravailable. Thisstudyisarsttowardacomprehensiveunderstandingoftheriskperceptionand evacuationdecisionsoftouristsinresponsetohurricanesthatmakelandfall.Weexamined touristswhovisitedcentralFloridatodeterminehowknowledgeabletheywereabout hurricanes,howmanyhadpreviousexperiencewithhurricanes,andwhatfactorsinuencedtheirperceptionsofriskandevacuationplanswhenahurricanewasexpectedto affectthearea.Ourthreemainhypotheseswereasfollows: (1)Location:Touristsonthecoastwillperceiveahigherriskandbemorelikelyto evacuatethantouristslocatedinland. (2)Hurricanecharacteristics:Theinformationthatatouristreceivesaboutthehurricane characteristics,suchasstormintensity,thetrackofthestorm,thetimetolandfall,and landfalllocationstronglyinuencestheirperceptionsofriskandevacuation decisions. NatHazards123

PAGE 3

(3)Touristattributes:Theattributesofthetourists,suchastheirknowledgeand experienceabouthurricanes,modeoftransportation,andagestronglyinuencetheir perceptionsofriskandevacuationdecisions. Toempiricallytestthehypotheses,statedpreference(SP)surveyswereadministeredto 448touriststhathadvisitedtheOrlandoarea(inland)andtheClearwater-StPetersburg region(coastal)inPinellasCounty,Florida(Fig. 1 ).Eachtouristviewedfourdifferent scenariosthatdepictedahurricaneforecasttomakelandfallinFloridaandaffecttheir currentlocation.Foreachscenario,theyindicatedthelevelofriskinvolvedwithremaining intheirpresentlocationandthelikelihoodtoevacuate.Weemployednon-parametricchisquaretests(Wilks 1995 )todeterminethathurricaneconditionsandtouristattributes producedstatisticallysignicanthigherorlowerratingsofriskandthelikelihoodto evacuatethanexpected. 2Relevantliterature Thecommunicationofriskforaspeciceventstartsfromanoriginalsourcethattransmits informationeitherdirectlytotheultimatereceiverorthroughintermediariessuchasmedia (Burnsideetal. 2007 ;LindellandPerry 2004 ;LindellandPrater 2007 ).IntheUSA,the NationalHurricaneCenter(NHC)issuesstatementsaboutthepotentialforahurricaneto Fig.1 SurveysitelocationsinPinellasandOrangeCounties,Florida NatHazards123

PAGE 4

affectacertainarea(Rappaportetal. 2009 ;Sheets 1990 ).TheNHCusesamapreferredto asthetrackforecastcone(NHC 2011 )toconveythelikelihoodthatatropicalcyclonewill passnearagivenlocation.Thesizeoftheconeisbasedupontrackforecasterrorsduring thepast5years,notthesizeofthestorm,andtheNHCexpectsthecenterofthestormto belocatedwithinthecone6070%ofthetime.Astheeffectsofhurricanesmayextend hundredsofkilometersfromtheircenter(NHC 2011 ),peoplelocatedoutsideofthecone maystillbeatriskfromthestorm.Unfortunately,Broadetal.( 2007 )foundthattheCone ofUncertainty''asitisreferredtobythepubliciswidelymisunderstoodbyFlorida residents astheyplacetoomuchfocusonthelocationofthecenterline.Yet,theirstudy concludesthatthemajorityofresidentsindicatethatthisgraphicplaysamajorroleintheir evacuationdecisions.Touriststhatresideoutsideofhurricane-proneregionsmaybeeven morelikelytoincorrectlyinterprettheirriskofexperiencinghurricaneconditionsafter viewingthetrackforecastcone.Inadditiontoviewingthismapontelevisionorthe Internet,touristsmaygaininformationrelatedtotheestimatedplaceoflandfallandstorm track,projectedintensity(thespeedofthemaximumsustainedone-minutewind),andtime untillandfallfromsourcessuchashotelstafforothertourists.Thus,itisimportantto gaugethereactionsoftouriststosuchmapsandtounderstandthemethodbywhich hurricane-relatedinformationissought. Someresearchershavestudiedtheimpactofriskperceptionsontourists'decisions accordingtothetypeofrisk.Examplesincludenaturalriskssuchasanearthquake(Huan etal. 2004 ),terrorism(Urielyetal. 2007 )orexposuretohealthriskssuchasSARS(Lau etal. 2004 ).Otherresearchershaveapproachedthestudyofriskbytestingtheeffectson decisionmakingofdiverserisks(Floydetal. 2004 ;Kozaketal. 2007 ;LeppandGibson 2008 ;ParkandReisinger 2010 ;Thapaetal. 2008 ).Kozaketal.( 2007 ),forinstance,found thatpotentialtourists'perceptionsofhighriskofinfectiousdiseaseoraterroristact inuencethelikelihoodtochangeplanstovisitaplacefortouristicreasons,butthiseffect wasnotfoundforotherriskssuchasnaturaldisasters.Overall,thekeyndingofthe reviewedresearchsuggeststhatingeneralhigherperceptionsofriskleadtoreduced intentiontovisit.However,resultsofsuchstudiesalsoindicatethatindividualtraitssuch asgender,education,andcountryoforiginhavealargeimpactonthelevelsofrisk perceptionsaswellastherelationshipbetweenriskandbehavioralintention(foradetailed reviewseeParkandReisinger 2010 ).Basedontheliterature,themajorityofrisk-related studiesintourismfocusonpre-traveldecisionswhenpotentialtouristsareevaluating likelihoodsofriskandtheirimpactontheselectionofdestination.However,analysesof tourists'decision-makingprocessincaseofeventsthatthreatentheirhealthorsafetywhile onvacationarerelativelyscarce.Thepurposeofthisstudyistoofferarststeptoclose thisgapbyfocusingontourists'decision-makingprocessesaftertheyhavebeenexposed toamessagethatexplainsthenatureofacrisisornaturaldisaster. Moststudieshaveexaminedtheactionsofresidentsduringnaturaldisasters,yet evacuationdecisionsoftouristsareformulatedunderdifferingcircumstancescomparedto thoseofresidents.Previousresearchhasshownthataccesstoavehicleisimportantfor evacuation(e.g.,CutterandSmith 2009 ;Eisenmanetal. 1997 ),butmanytouristsmayy intotheregionand/ormayrentavehiclebutbeunfamiliarwithevacuationroutes.For example,inFlorida,almostanevensplitofvisitorsarrivebyairversusnon-airtransport (car,RV,train,bus,orother)(VisitFlorida 2010 ).Inastudyofinternationaltourists, So nmezandGraefe( 1998 )foundthatperceptionsofriskorsafetycanalterrational decisionmakingasitpertainstotravelmodesandchoiceswhileinadestination.Also, touristsmayperceivelocationsthathavebeenaffectedbypreviousnaturaldisasters(i.e., MiamiorKeyWest)asmoreriskythandestinations,whichhavenotbeenhitbypast NatHazards123

PAGE 5

naturaldisasters.Thus,touriststhatvisitaninlandcitysuchasOrlando,Florida,may perceivethelocationtobelessatriskfromahurricanestrikewhencomparedtothecoast, andhence,maybelessinclinedtoevacuategiventhesamescenarioasacoastaltourist. Additionally,WestandOrr( 2007 )foundthathavingchildrenandlivingnearthecoast wereimportantfactorsintheevacuationdecisionsofresidents,butgenderandracewere not.However,thecompositionoftouristgroupsmaydifferfromresidentpopulationsin termsoftheirgroupsize,thenumberofchildrenpresent,ethnicbackgrounds,income, education,andage. Communicatorsofhurricane-relatedinformationcannotassumethatallmembersofthe generalpublichavetherequisitepriorknowledgetocomprehendtheinformationand formulateaplanofactioninresponsetohurricaneforecastsinthesamemanner.Thismay beespeciallytrueincomparisonwithresidentsandtourists.Residentsofhurricane-prone regionsaremorelikelytounderstandhurricane-specicinformationastheyareexposedto hurricane-relatedterminologybeforeandduringthehurricaneseasonandmorefrequently whenahurricanehasthepotentialtostriketheirarea(DanielsandLoggins 2007 ;Dowand Cutter 2002 ).Previousexperiencewithhurricaneshasbeenlinkedtolowerevacuation ratesofresidentpopulations(Rinconetal. 2001 ;Kusenbachetal. 2010 ),yetother researchershavefoundthatthosepreviouslyinvolvedinhurricaneevacuationsaremore willingtoevacuateforfuturestorms(Baker 1991 ;Riadetal. 1999 ).Sincetherehasbeena paucityofresearchthathasconsideredtouristresponsestohurricanes,itisimportantto gaugetheirlevelofknowledgeandexperienceandtodeterminehowitinuencestheir evacuationdecisions. 3Surveydesign Dataonevacuationdecisions(i.e.,behavioralaction)canbecollectedthroughrevealed preference(RP)orstatedpreference(SP)surveys.RPsurveysinvolveelicitingdataon actual actions,whereasSPsurveysinvolveelicitingbehavioral intentions ofrespondents underhypotheticalscenarios(Hessetal. 2006 ;KroesandSheldon 1988 ;Wardman 1988 ). Inthecontextofhurricaneevacuations,allstudiestodatehavefocusedonresident populationsandmostemployedtheRPmethod(e.g.,Baker 1991 ;SmithandMcCarty 2009 andstudiesreferencedtherein).ThefewstudiesemployingSPtechniques(inthe contextofresidentpopulations)includeLazoetal.( 2010 ),Whitehead( 2003 2005 ), Whiteheadetal.( 2000 ),andFu( 2004 ).Astouristsrepresentatransientpopulation, collectingRPdatafromtouristsaboutactualhurricaneexperiencesmaybeevenmore difcultthansimilardatafromresidents.Despitethefactthattheintentionsoftourists revealedinanSPsurveymaydifferfromtheiractionswhenahurricanelandfallactually occurs,theSPapproachisappropriatetoexaminetheintendedresponsesoftouriststo hurricane-relatedscenariosandhencewasadoptedinthisstudy. Fivesurveysiteswereselectedbasedontheprobabilityofhighvolumesoftourists (Fig. 1 ).TheOrlandoarearepresentedtheinlanddestination,andwecollected304surveys fromtwosites:(1)theFloridaMalland(2)WyndhamBonnetCreekResortinLakeBuena Vista.Wecollected144surveysfromPinellasCounty,whichwereadministeredatthree GulfofMexicobeaches:(1)SheratonSandKeyResort,(2)ClearwaterBeach/Pier60,and (3)St.PeteBeach.ThesurveyswereconductedduringJuly22August9,2009.No hurricaneswerepresentwithintheAtlanticOceanbasinduringorpriortothisperiodasthe rstnamedstormof2009formedonAugust11.Ascreeningquestionwasemployedto identifytourists,whichwedenedasthosethathadtravelledgreaterthan50milesasthis NatHazards123

PAGE 6

isthedenitionutilizedbythestateofFlorida(FloridaStatute125.104).Weveriedthis distancebyplottingthelocationofeachzipcodesuppliedbytherespondentwithin ArcGISandcalculatingthestraight-linedistancebetweenthecoordinatesofthesurvey locationandthoseofthecentroidofthepolygonenclosingthezipcode.Onlyoneadult lledoutthe27-questionsurveypergroup. Whencompletingthe3-pagesurvey,participantsrstindicatedinformationpertaining totheirtriplogistics(e.g.,frequencyofvisits,modeoftransportation,groupcomposition). Thesecondsectionassessedknowledgeabouthurricanesthroughfourtrue/falsequestions (Fig. 2 ).Welaterdeterminedwhichrespondentsansweredallfourquestionscorrectly,as wellasthosewhoindicatedthattheydidnotknow.Thethirdsectionassessedhurricane experienceandpreparednessbyaskingwhethertouristsortheirfamilyorfriendshadbeen affectedbyahurricaneortyphoonandwhethertheyhadcheckedforthepossibilityofa hurricaneoccurringduringtheirvisit.Theythenindicatedthesourcesofinformationthat theywerelikelytousetolearnabouthurricanesduringtheirtrip(Fig. 2 ).Finally, demographicinformationwascollectedsuchasgender,age,income,andzipcodeor countryofresidence,whichwerelatercategorizedasresidingwithinFloridabutoutsideof a50-mileradiusofthesurveylocation,outsideofFloridabutwithintheUSAandoutside oftheUSA.Allvariablesutilizedintheanalysisarelistedwiththeirpossibleresponsesin Table 1 Areviewoftheliteraturedemonstratesthatthelocationofastormanditsintensityare importantconsiderationsforresidentswhenmakingthedecisiontoevacuate;evacuation likelihoodincreasesforstrongerstormsthathitclosertotherespondent'slocation(Baker 1991 ;DowandCutter 2002 ;SmithandMcCarty 2009 ).Ourstudyaimedtoexamine Fig.2 Portionofthesurveyaskingtouristsabouttheirhurricaneknowledge,experience,andinformation sourcestheyarelikelytouse NatHazards123

PAGE 7

whethertouristswouldrespondsimilarly.WithrespecttoSP,touristswerepresentedwith fourdifferentscenariosdepictingahurricaneforecasttomakelandfalloverFlorida.Each scenarioconsistedofamapdesignedtorepresentanactualtrackforecastconeissuedby Table1 VariablesacquiredthroughsurveyresponsesandtheircategoricalgroupingsNameDescriptionPossiblevalues CoastFloridalandfallcoast1 = Atlantic,0 = Gulf LandtimeTimetolandfall1 = 24h,0 = 48h CategoryS.S.Categoryatlandfall1 = Category1,0 = Category4 DurationWindduration1 = Short,0 = Long TrackTrackpassagerel.tosite1 = Offsetfromsurveysite,0 = Oversurveysite RiskPerceivedlevelofrisk0 = Donotknow,1 = Notatall ƒ 5 = Verymuch EvacuateLikelihoodofevacuation0 = Donotknow,1 = Veryunlikely ƒ ..5 = Verylikely FVisitFirstvisittodestination0 = No,1 = Yes NVisitNo.FLvisitspast5years1 = 01,2 = 23,3 = 45,4 = GT5 NDaysNumberofdaysintrip1 = 05,2 = 610,3 = 1116,4 = GT16 AccomAccommodations1 = Hotel,2 = Campground,3 = Resort,4 = Friend, 5 = Bed/breakfast,6 = Other GroupSizeNumberofpersonsingroup1 = 12,2 = 3,3 = 4,4 = GT4 ChildChildreningroup0 = No,1 = Yes ModeAPPlanewasused0 = No,1 = Yes ModePTPublictransportationused0= No,1 = Yes ModePVPersonalvehiclewasused0 = No,1 = Yes ModeRVRentalvehiclewasused0 = No,1 = Yes QrightAllknowlquestionsright0 = No,1 = Yes Qunknown1 ? quest.didnotknow0 = No,1 = Yes AffectPAffectedbyahurricane personally 0 = No,1 = Yes AffectFFriends/FamilyAffected0 = No,1 = Yes InfoTVInformationfromTV0 = No,1 = Yes InfoHotInformationfromhotelstaff0 = No,1 = Yes InfoWCInformationfromWeather Channel 0 = No,1 = Yes InfoWebInformationfromInternet0 = No,1 = Yes InfoTourInformationfromothertourists0 = No,1 = Yes InfoNewsInformationfromnewspaper0 = No,1 = Yes Hurricane Possibility Checkedforthepossibility ofahurricane 0 = No,1 = Yes GenderGender1 = Male,2 = Female AgeAgein20091 = 1829,2 = 3039,3 = 4049,4 = 5059,5 = 60 ? EducationHighestlevelofeducation1 = Highschool,2 = Bachelor.,3= Masters,4 = Other EthnicEthnicity1 = Caucasian,2 = AfricanAmerican,3 = Hispanic, 4 = Asian,5 = PacicIslander,6 = NativeAmerican, 7 = Mixedrace,8 = Other IncomeHouseholdincome20081 = LT$24K,2 = 2435,3 = 35-50,4 = 5075, 5 = 75100,6 = 100125,7 = GT125 ResideAggregatelocation1 = Florida,2 = USA,3 = International NatHazards123

PAGE 8

theNHCwiththesurveylocationmarkedasastar,andcorrespondingtextcontaining informationabouttheintensityandduration.Forexample,aCategory4hurricaneprojectedtomakelandfallin24halongtheGulfCoastofFloridawiththecenterlinepassing directlyoverPinellasCountysurveysitesandanexpecteddurationof18h(Fig. 3 ).All scenariosfeaturedthevesamplingsiteswithinthewhitecone.Eachscenariofeatureda combinationofvedifferentconditions,witheachconditionhavingtwopossiblevariations(Table 2 ). Tolimitsurveylength,eachtouristviewedfourtotalscenarios.Thescenariosalternated betweentheGulfandAtlanticlandfallssothattherespondentwasabletovisuallyperceivethe differencesbetweenthesuccessivescenarios.Theconditionthatdifferedamongthelandfalls alongthesamecoastwasthesurveysite'spositionrelativetothecenterlineofthecone. Variationoneshowedthecenterlinepassingdirectlyoverthesurveysite,whilevariationtwo depictedthesurveysitewithinthecone,butoffsetfromthecenterlineby5080km.Again,to limitsurveylength,allfourscenariosincludedonlyoneofthetwopossiblevariationsforthe remainingthreeconditions.Theseconditionswereintensityatlandfall(SafrSimpson Category1or4(SimpsonandSafr 1974 )),timeuntillandfall(approximately24or48h),and durationofhurricane-forcewinds(shortorlong).Ashort(long)durationwas3(12)hfora Category1and6,(18)hforaCategory4hurricane,whichcorrespondtoaveragesizesand forwardvelocitiesofhurricanesfoundinclimatologicalstudies(seeKimballandMulekar 2004 ;Matyas 2010 ).Respondentswereaskedtorateeachscenarioonascaleof15based Fig.3 Scenarionumber25of32possiblescenariosincludinggraphicsdesignedtocloselysimulatethose issuedbytheNHCandtexttoindicatestormintensityandduration.Thisscenariowashypothesizedtoyield thehighestratingsforriskperceptionandlikelihoodtoevacuatefortouristssurveyedinPinellasCounty NatHazards123

PAGE 9

Table2 Scenariospresentedtotouristswithtwovariationseachoflandfallcoast,track,timeuntillandfall, intensity,andduration Scenario number Landfall coast Track center line Timeto landfall (h) SafrSimpson categoryat landfall Hurricane forcewind duration(h) Number ofratings Average risk Average evacuation likelihood 1GulfOver2413593.303.29 2AtlanticOver2413593.053.00 3GulfOffset2413613.103.05 4AtlanticOffset2413612.983.17 5GulfOver4813503.183.12 6AtlanticOver4813502.922.90 7GulfOffset4813482.912.86 8AtlanticOffset4813493.082.88 9GulfOver2446513.573.56* 10AtlanticOver2446483.093.13 11GulfOffset2446503.183.24 12AtlanticOffset2446503.003.04 13GulfOver4846473.73*3.57* 14AtlanticOver4846483.353.28 15GulfOffset4846483.163.35 16AtlanticOffset4846493.223.08 17GulfOver24112423.292.85 18AtlanticOver24112433.023.02 19GulfOffset24112433.143.07 20AtlanticOffset24112413.002.95 21GulfOver48112443.092.63* 22AtlanticOver48112473.062.60* 23GulfOffset48112472.48*2.85 24AtlanticOffset48112462.802.53* 25GulfOver24418533.90*3.19 26AtlanticOver24418513.523.22 27GulfOffset24418523.563.04 28AtlanticOffset24418523.153.26 29GulfOver48418483.423.26 30AtlanticOver48418453.333.29 31GulfOffset48418433.143.00 32AtlanticOffset48418443.213.09 Eachtouristviewedoneofthegroupsoffourscenariosseparatedbytheblankline.Thenumberofratings foreachscenarioandtheaverageratingsforperceptionofriskandlikelihoodtoevacuatearealsopresented foreachscenario;responsesofzeroareexcluded *Denotesvalues1.5standarddeviationsaboveorbelowthemeanforeachcolumn NatHazards123

PAGE 10

uponthelevelofrisktheyperceivediftheyweretoremainattheircurrentlocationandthe likelihoodthattheywouldevacuate(Fig. 4 ).Alternatively,theycouldindicatethattheydid notknow.Weexpectedthatthescenariosmostlikelytohavethehighestratingswouldhavea Category4intensity,landfallalongtheclosestcoast,thecenterlinepassingoverthelocation, 24huntillandfall,andalongduration. 4Proleofrespondents Mosttouristshadsomeknowledgeabouthurricanes,buttheirknowledgewasincomplete. While63%ofresponsestothefourknowledgequestionswerecorrect,only25%correctly answeredallfourquestions(Table 3 ).Question10b(Fig. 2 )wasmostfrequentlyanswered incorrectly,and46%indicatedthattheydidnotknowatleastonequestion.Nearlythreequartersofthetouristsindicatedthattheyhadnothadanypriorexperiencewithhurricanes (Table 3 ).Ofthosewithpastexperience,49%indicatedthattheydidnotevacuateand Fig.4 Portionofthesurveyaskingtouriststoratetheirresponsestoeachofthefourhurricane-related scenarios Table3 Percentageoftourists thatindicatedaresponsetoyes Variable%Yes FVisit34.2 Child66.8 ModeAP66.7 ModePT2.7 ModePV28.8 ModeRV19.1 Qright24.5 Qunknown45.6 AffectP25.8 AffectF46.7 InfoTV84.7 InfoHot33.5 InfoWC44.5 InfoWeb42.7 InfoTour12.4 InfNews21.8 HurricanePossibility30.3 NatHazards123

PAGE 11

remainedatthedestination.Alargerportionoftherespondents(47%)indicatedthattheir familyorfriendshadbeenaffectedbyahurricaneortyphoon(Table 3 ).Only30%ofthose surveyedhadcheckedforthepossibilityofahurricanestrikepriortocurrenttripdeparture (Table 3 ).Notsurprisingly,alargerpercentageofthosewhocheckedforahurricanestrike answeredallfourknowledgequestionscorrectly(38%)comparedtothosethatdidnot (22%). Thedemographicvariablesshowedthatrespondentswerealmostevenlydistributed betweenmalesandfemales(Table 4 ).Agegenerallyfollowedanormaldistributionwith mostrespondentsaged4049(Table 4 ).Respondentswerefairlywelleducated(33% holdingaBachelor'sdegreeand25%holdingaMaster'sorhigherdegree)andwere relativelyafuentwith22%reportingincomeofabove$125,000.Only9%notedtheir incometobelessthan$24,000.Withrespecttorace/ethnicity,Caucasiansrepresentedthe largestgroup(Table 4 ).Themajorityoftouristswereinternational(48%),while38%were fromUSAotherthanFlorida(Table 4 ).Accordingtotheresultsofadditionalchi-square tests,knowledgeandexperienceexhibitastatisticallysignicantrelationshipwithresidence( p values \ 0.05).Morethan90%ofFloridatouristsansweredquestion10a(Fig. 2 ) correctly,whileapproximately65%ofUSnon-Floridaandinternationaltouristsanswered correctly.Fortheotherthreequestions,fewerinternationalrespondentsansweredcorrectly thanexpectedwithmoreindicatingthattheydidnotknowtheanswerascomparedtoUS tourists.Floridarespondentsrecordedthehighestnumberofcorrectanswers.While77% oftouristsfromFloridaindicatedthatafriendorfamilymemberhadexperienceda hurricaneortyphoon,only56%(34%)oftouristsfromtheUSA(othercountries) respondedafrmativelytothisquestion. MorethanhalfoftherespondentshadvisitedFloridabeforeandmosthadvisitedthe statemultipletimes(Table 4 ).Only5%ofrespondentstravelledalone,while32%travelledinagroupoffour(Table 4 ).One-thirdofthetouriststravelledwithoutchildren (Table 3 ),whilethosetravelingwithchildrenmostoftenhadtwochildren.Themajority (67%)ofrespondentsutilizedairtransportationduringtheircurrenttrip,with29%usinga personalvehicleand19%arental(Table 3 ).Nearlyone-thirdoftouristsindicatedthat theirlengthofstaywasbetween6and10or11and16days,and74%stayedinahotelor resort.Localtelevisionwasselectedby85%ofrespondents,followedbytheWeather Channel(45%)andtheInternet(43%)astheinformationsourceslikelytobeusedtolearn abouthurricanesduringatrip(Table 3 ).Theseresultsaresomewhatsimilartothoseof Zhangetal.( 2007 )whointerviewedresidentsafterHurricaneRita(2005)anddetermined Table4 PercentofresponsestoeachtouristattributecategoryaslistedinTable 1 Variable1(%)2(%)3(%)4(%)5(%)6(%)7(%)8(%) NVisit31.432.516.919.3 NDays20.932.133.014.0 Accom38.50.235.50.211.114.5 GroupSize22.915.631.929.6 Gender50.949.1 Age18.114.732.821.812.5 Educate31.733.224.810.7 Ethnic74.36.78.34.00.20.43.32.6 Income8.75.410.016.919.517.422.1 Reside13.937.748.3 NatHazards123

PAGE 12

thathurricaneforecastsweremostoftenobtainedfrombroadcastTVandtheWeather Channel.However,Zhangetal.'s( 2007 )studyshowedradioandpersonalnetworkstobe muchmoreimportantthantheInternet,whiletheseinformationsourceswerenotas frequencyindicatedbytouristsinthisstudy. 5Riskperceptionandlikelihoodofevacuation Toevaluatewhichofthe32scenariosyieldedthehighestandlowestperceptionsofrisk andlikelihoodtoevacuate,theaverageratingforeachscenariowascalculated.Scenarios withvaluesfallingoutsideof1.5standarddeviationsfromthemeanweredeemedtohave exhibitedthehighestandlowestratings.Next,weemployedchi-squareteststodetermine whichvariablesproduceddifferencesintheresponseratings.Eachtestfeaturedverows correspondingtotheresponseratingsof15,andcolumnscorrespondingtothevariations ofeachhurricaneconditionortouristattribute.Thenullhypothesisreectedalackof signicantdifferencebetweenthedistributionofresponseratingsbetweenoramongthe groups.Iftheresulting p -valuewasequaltoorlessthan0.05,weacceptedthealternate hypothesis,indicatingthatthevariablehadaneffectontheriskperceptionofthetourists and/ortheirlikelihoodtoevacuate.Itshouldbenotedthattheintentwastounderstandthe majorassociationsorrelationships,andourmethodsdonotallowtheestablishmentof causallinksbetweenanyofthefactorsandtheriskperceptionorevacuationdecisionsand thevariousexplanatoryfactorsofinterest. Thetwochi-squaretestscomparedperceptionofriskandlikelihoodtoevacuatefor inlandtouristsandforcoastaltourist,andthentwoadditionaltestscomparedtheperceptionofriskforbothregionsandthelikelihoodtoevacuateforbothregions.Foreachof thehurricanecharacteristicsandtouristattributes,theratingsoftheOrangeandPinellas Countytouristswereevaluatedinseparatetests,andriskperceptionwasevaluatedseparatelyfromevacuationlikelihood.Fourofthetouristattributes(useofpublictransportation,accommodation,obtaininginformationfromlocaltelevision,andethnicity)assome ofthecategoriesproducedfewerthanveresponsestooneormoreratings.Thus,fourchisquaretestswereperformedforeachofthevehurricanecharacteristicsandeachofthe remaining24touristattributes(Table 1 ).Groupswithmoreratingsof4and5than expectedweredeemedmorelikelytoevacuateortoperceiveahigherlevelofriskthan groupswithmoreratingsof1or2. 5.1Effectofdestinationlocation(coastalversusinland) Theresultsofthersttwochi-squaretestsindicatethattheratingsforriskperceptionand evacuationlikelihooddifferedbetweenthetwosurveylocations.Forevacuation ( p -value \ 0.0001),morePinellasCountytourists(averagerating3.08)indicatedratingsof 4or5thanexpected;therefore,theyweremorelikelytoevacuatethantheOrlandotourists (averagerating2.88).Thisallowsustoacceptthesecondpartofourrsthypothesis.Yet, thetestforriskperception( p -value \ 0.000)producedanunexpectedresult.Although manytouristsinPinellasCountyratedtheirlevelofrisk35,manymorethanexpected indicatedaratingof1ornotatallrisky''.Thus,thestatisticalsignicanceoftherisk perceptiontestcomesfrommoretouristsindicatingratingsof1and5thanexpectedin PinellasCounty;whereaswehypothesizedthatfewerPinellasCountytouristswould indicatearatingof1thanexpected.Theseresultsindicatethatadifferenceexistsbetween atourist'sperceptionofriskandtheirlikelihoodtoevacuate.Wealsondthatevenifthe NatHazards123

PAGE 13

perceivedlevelsofriskarethesameintwolocations,thelikelihoodofevacuationmightbe different.Theresultsofthechi-squaretestssupportthisnding.Statisticallysignicant differencesexistedintheratingsforriskcomparedtothelikelihoodtoevacuate( p -value PinellasCounty0.023,OrangeCounty \ 0.000). Hurricanewindsexperiencedinthecoastallocationsarefasterthanthoselocated fartherinland,andsuchcoastallocationsarealsoatriskfromstormsurge.Hence,anyone locatedonacoastthatiswithinthewhiteconeisatriskfromaland-fallinghurricane. Thus,itisimportanttoexplorewhyone-thirdoftheratingsof1andone-halfoftheratings of0wereassociatedwithscenariosthatdepictedahurricanemakinglandfallalongthe GulfCoastwithin80kmofthesurveylocation.Itispossiblethatthesetouristsviewed scenarioscontainingalowintensity,anoffsetcenterline,and/orhad48huntillandfalland feltthattheywerenotatriskatthistimeorwereunsureoftheirrisk.Yet,recenthistory hasdemonstratedthathurricanetracksandintensitycanchangeevenwithin24hof landfall.Theerrorsassociatedwithforecastsofhurricaneintensityhavenotshownsignicantimprovementsduringthepast15years(Elsberryetal. 2007 ),andahurricane forecasttobeatCategory1intensitycouldintensifypriortolandfallandcausemore damagethanexpected,whileaCategory4hurricanecouldweakenintoaminimalhurricanepriortolandfall.In2004,HurricaneCharleyrapidlyintensiedfromCategory2to Category4withinseveralhoursofmakinglandfallalongFlorida'sGulfCoast(Leeand Bell 2007 ).Eventhoughtheprojectedstormtrackshiftedduringthistime,thepointof landfallhadbeenwithintheconebutoffsetfromthecenterlinefor48hpriortolandfall. CasessuchasCharleydemonstratethatallofourGulfCoastscenariosposedsomeriskto PinellasCountytourists.Thepresenceofriskratingsof0and1suggeststhattouristsneed tobebettereducatedabouttherisksassociatedwithremainingatacoastallocationwhena hurricaneisexpectedtomakelandfallnearby. 5.2Effectofhurricanecharacteristics AccordingtoTable 2 ,scenariosthathavethehighestratingsdepictCategory4hurricanes makinglandfallalongtheGulfCoastwiththecenterlinepassingoverthesamplingsite. ScenarioswiththelowestratingsallfeatureCategory1hurricanesmakinglandfallwithin 48h.Althoughscenarios21and22receivedsomeoftheloweraverageratingsforlikelihoodtoevacuate,theyreceivedamuchhigheraverageratingforriskperception.Scenario17alsoreceivedamuchhigherratingforriskthanforlikelihoodtoevacuate,which demonstratedthattheperceptionofriskislikelynottheonlyfactortoinuencethe decisiontoevacuate.Oneunexpectedresultisthattheshort-durationeventsreceived higherratingsthandidlong-durationevents(Table 2 ).Thisndingsuggeststhateithera tourist'sperceptionofeventdurationisnotanimportantfactorintheirdecision-making process,orthatdurationmightnothavebeenprocessedcompletelyasthisinformationwas printedatthebottomofthepageandmighthavebeenviewedaftertheperceptionofrisk anddecisiontoevacuatehadalreadybeenmadebasedontheotherhurricaneconditions. Anexaminationofthe p -valuesofthechi-squaretests(Table 5 )allowedustorankthe hurricaneconditionsintermsoftheirrelativeinuenceonriskperceptionandthelikelihoodofevacuationasfollows:(1)intensity(2)coastwherelandfalloccurs(3)centerline position,and(4)timeuntillandfall.Thedurationofhurricaneforcewindswasnot includedasthenullhypothesiswasacceptedforeachtest.Wefoundtheintensityofthe hurricaneatthetimeoflandfallwasthemostimportantscenarioconditionasallfourofthe p -valuesthatresultedfromtheintensitytestswerelessthan0.05.Ourresultsareinconcert withsurveysofcoastalresidents(DowandCutter 2002 ;SmithandMcCarty 2009 )that NatHazards123

PAGE 14

alsoidentiedstormintensityasanimportantfactorthatinuenceddecisionstoevacuate. Overall,touriststhatperceivedahighlevelofrisk(rating4or5)andthosemostlikelyto evacuate(rating4or5)viewedmorescenarioscontainingCategory4hurricanesthan Category1hurricanes.Althoughonly5%PinellasCountytouristsindicatedthattheydid notknowoftheirriskorlikelihoodtoevacuate,77%ofthesetouristsviewedCategory4 scenarios.Again,thisndingsuggeststhattouristsneedtobebettereducatedonthe potentialdangersassociatedwithremainingnearthecoastwhenahurricanelandfallis expected,particularlywhenthestormisforecasttohaveahighintensity. Theresultsthatexaminedcoastoflandfall,tracklocation,andtimeuntillandfallwere statisticallysignicantonlyforPinellasCounty.Sincetheyweresurveyedwhilestanding ontheGulfCoast,itisnotsurprisingthatmoretouristsindicatedhigherratingswhenthe hurricanewasprojectedtomakelandfallalongtheGulfratherthantheAtlanticCoast. TwiceasmanytouristsratedtheirlikelihoodtoevacuateforaGulflandfallasa5rather thana4,whileratingsforAtlanticlandfallsweredistributedfairlyevenlyamongallve ratings.Asimilarresultwasachievedforthelikelihoodtoevacuatewhenthecenterlineof theconewasprojectedtopassdirectlyoverthesurveyregion;nearlytwiceasmany touristsindicatedaratingof5whencomparedtoanyotherrating.Asallsurveylocations werelocatedwithinthecone,wherethemaximumsustainedwindsoftheapproaching stormareexpectedtooccur6070%ofthetime,alteringthepositionofthecenterline shouldnothaveproducedadifferenceinratingsthatwasstatisticallysignicant.Thefact thatastatisticallysignicantdifferencewasfoundimpliesthattouristslikeresidents (Broadetal. 2007 )mayplacetoomuchemphasisonthecenterlineandmisinterprettheir riskuponviewingthetrackforecastconemap.Whenthehurricanewasprojectedtomake landfallwithin24h,more(fewer)touristsselected3(1)thanexpected,whichaccounted forthestatisticallysignicantdifferencebetween24and48hlandfalltimes.Giventhat evacuationleadtimesformajorurbanareascanbe4872h(Regnier 2008 ),itisimportant tonotethattouristsmaynotbeaslikelytoreactattheselongerleadtimes.Overall,these resultsindicatethatoursecondhypothesiswassupportedandaccepted. 5.3Effectoftouristattributes Wearealsoabletoacceptourthirdhypothesis.Forsixofthe24touristattributes,allfour chi-squaretestsproducedstatisticallysignicantresults(Table 6 ),andtheseincluded demographics,triplength,andpreviousexperiencewithhurricanes.Touristsperceiving thehighestlevelsofriskandmostlikelytoevacuateinbothPinellasandOrangeCounties hadatripdurationlessthan6days;werenotpreviouslyaffectedbyahurricane;didnot Table5 Resultsofchi-squareteststodeterminewhichhurricaneconditionshadthegreatestinuenceon theperceptionofriskandlikelihoodtoevacuatefortouristssurveyedinPinellasandOrangeCounties Pinellas countyrisk perception Pinellascounty likelihoodto evacuate Variation withhigher ratings Orange countyrisk perception Orangecounty likelihoodto evacuate Variation withhigher ratings Intensity 0.0100.000 Cat.4 0.0330.001 Cat.4 Coast 0.0000.001 Gulf0.5040.807 Track 0.0470.007 Over0.6140.856 Landtime 0.008 0.07824h0.1170.527 Duration0.8820.8830.2470.307 Boldvaluesindicateresultsthatarestatisticallysignicantat a = 0.05 NatHazards123

PAGE 15

haveafamilymemberthatwaspreviouslyaffectedbyahurricane;andhadcheckedforthe possibilityofahurricanestrikepriortocurrenttripdeparture.Twooffourchi-squaretests producedresultssignicantat a = 0.05insevenoftheremainingvariables,whichinclude beingonarstvisittothedestination,gender,travelingwithchildren,notknowingthe knowledgequestions,travelingbyairplane,andobtaininginformationfromhotelstaffor theWeatherChannel(Table 6 ).Below,weprovideadetaileddiscussionofthemost importantfactorsasrevealedbythechi-squareresults.Itisimportanttonotethatour explanationsareplausiblereasonsforthestatisticalrelationships,andfuturestudiesshould aimtotestthesemoredirectlywithappropriatelywordedquestionsinthesurveys. Ashorttripand/orthistripbeingtherstvisittothedestinationmayleadtoan increasedperceptionofriskandlikelihoodtoevacuateastouristsdonothaveenoughtime tofamiliarizethemselveswiththenewenvironmentandthesupportsysteminthedestination.Previousresearchhasshownthatthissituationleadstoanxiety,whichisone consequenceofbeingexposedtopotentialrisks(DowlingandStaelin 1994 ;Reisingerand Mavondo 2005 ).Ontheotherhand,touristswhofrequentthedestinationaremorefamiliar withtheirsurroundings,whichmayreducetheiranxietylevel.Thus,theywheretond informationandcanappropriatelydecidewhatactionshouldbetakenintheeventofa hurricane.VogtandFesenmaier( 1998 )assertedthattouristsaversetoriskanduncertainty aremorelikelytoengageinstrategiesthatreduceriskanduncertaintysuchassearching informationextensivelyduringthedecision-makingprocess.Theirndingshelptoexplain whytouristswhocheckedforthepossibilityofahurricanestrikebeforedepartingontheir tripweremorelikelytoevacuate. Ourndingthatprevioushurricaneexperiencewaslinkedtoalowerevacuationlikelihoodinbothregions(Table 6 )isinterestingparticularlyinrelationtotheresident Table6 Resultsofchi-squareteststodeterminewhichattributesofthetouristshadthegreatestinuence ontheperceptionofriskandlikelihoodtoevacuatefortouristssurveyedinPinellasandOrangeCounties Pinellas countyrisk perception Pinellascounty likelihoodto evacuate Higher/ lower rating group Orange countyrisk perception Orangecounty likelihoodto evacuate Higher/ lower rating group Reside 0.0000.003 3/2;3/1 0.0000.000 2/3;1/3 NDays 0.0020.002 1/2;1,4/3 0.0000.000 1/4;2/4 Age 0.0000.000 5/1;5/2 0.0000.000 5/3;1/4 AffectP 0.0370.007 0/1 0.0000.000 0/1 HurricanePossibility 0.0020.034 1/0 0.0440.000 1/0 AffectF 0.0000.024 0/1 0.0480.042 0/1 Child 0.000 0.0571/0 0.006 0.8231/0 FVisit 0.035 0.1741/00.053 0.000 1/0 Gender0.2310.0911/0 0.0000.002 1/0 InfoHotel0.064 0.000 0/10.866 0.000 1/0 Qunknown0.5140.651 0.0090.000 1/0 InfoWC 0.050 0.5711/00.672 0.003 1/0 ModeAP0.0670.2560/1 0.0330.008 0/1 ModeRV0.192 0.000 0/10.4300.499 ModePV0.1360.2921/00.119 0.038 1/0 Boldvaluesindicateresultsthatarestatisticallysignicantat a = 0.05 NatHazards123

PAGE 16

population.Lazoetal.( 2010 )andLindelletal.( 2005 )foundthatpriorexperience exhibitednoeffectontheevacuationdecisionsofresidents,andRiadetal.( 1999 )and Burnsideetal.( 2007 )foundthatresidentsevacuatingpreviouslywereextremelylikelyto evacuateagain.Wefoundthatthe25%oftouristswhohadbeenpreviouslyaffectedbya hurricaneortyphoonperceivedthescenariosthattheyviewedaslessriskyandwereless likelytoevacuatethantouristswithoutthispriorexperience.Inaddition,nearlyhalfofthe touristssurveyedindicatedthattheirfamilyand/orfriendshadbeenaffectedbyahurricane ortyphoon.Thesetouristswerealsolesslikelytoevacuateandperceivedalowerlevelof riskwhencomparedtothosewhodidnotknowsomeonethathadbeenaffected.Our resultsconcurwiththosefromstudiesofresidentsconductedbyKusenbachetal.( 2010 ) andRinconetal.( 2001 ).Onepossibleexplanationisthatprevioushurricanesencountered bytouristswereweakorshort-lived,orthestormsthatnotpassnearenoughtotheir locationtocausesignicantdamage.Anotherpossibilityisthattouristsfacedproblems duringpastattemptstoevacuatesuchascrowdedorclosedairports,makingthemless willingtoevacuateforfuturehurricanes.TheaforesaidpossibilityisparallelwithJohnson andTversky's( 1983 )ndingthatpeopleoftenusepastsituationsasananchorinthe decisionmakinginthenewsituationtheyfaceandthatindividual'sexperiencewith particularriskscouldtransfertotheindividualresponsetootherrisks. Anexaminationofboththelocationofthetourist'shomeresidenceandagealso producedstatisticallysignicantresultsforriskperceptionandevacuationlikelihood. However,thegroupsdifferedbetweenthetwosurveycounties.Internationaltourists surveyedinPinellasCountyindicatedthehighestriskandweremostlikelytoevacuate, whiletouristsinOrangeCountyfromtheUSAbutoutsideofFloridaperceivedthehighest riskandthosefromFloridaweremostlikelytoevacuate.Oneexplanationforthisresultis thatinternationaltouristsperceivehurricanestobeathreatprimarilytocoastallocations andmaynotunderstandthelevelsofdamagethatcanalsooccurinland.Previoushurricane evacuationstudieshaveyieldedmixedresultswhenconsideringage(Baker 1991 ).Our resultsindicatethattouristsaged50andhigherprovidedthehighestratingsoverallin PinellasCountyandforriskperceptioninOrangeCounty,yettouristsaged1829were mostlikelytoevacuatefromOrangeCounty.Someauthorsreportthatolderpopulations arelesslikelytoevacuaterelativetoyoungerpopulationsduetolowerincomeorinability totravelduetohealthreasons(Baker 1979 ;Drabek 1986 ;Eisenmanetal. 1997 ).However, thetouristswesurveyedwerecapableoftravelfrombothahealthandincomeperspective; thiscouldbeonereasonwhyoldertouristsindicatedhigherratingsthanmightbeexpected. Themixedresultsobtainedforbothhomeresidenceandagesuggestthatfurtherresearchis requiredtomorefullyunderstandthekeyfactorsinuencingriskperceptionandevacuationlikelihoodsoftourists. Somepreviousstudieshaveshownthatwomen(Lindelletal. 2005 ;Riadetal. 1999 ; SmithandMcCarty 2009 )andhouseholdswithchildrenunder18yearsofage(Dashand Gladwin 2007 ;Lindelletal. 2005 ;SmithandMcCarty 2009 )aremorelikelytoevacuate, whileotherstudieshavenotfoundthistobethecase(Baker 1991 ;Burnsideetal. 2007 ; Lazoetal. 2010 ).Inourstudy,morementhanwomenweresurveyedinOrangeCounty, whiletheoppositewastrueinPinellasCounty.Womentendedtoofferhigherratingsfor bothriskperceptionandevacuationlikelihoodthanmeninbothregions,however,thechitestresultsweresignicantat a = 0.05onlyinOrangeCounty(Table 6 ).Onepossible explanationforthisoutcomemaybethepresenceofchildren.Morechildrenaccompanied OrangeCountytouriststhatPinellasCountytourists.Achi-squaretestexaminingthe numberofchildrenineachgroupaccordingtowhetherthetouristsurveyedwasmaleor femaleproduceda p valueof0.0002.Thisindicatedthatmorewomenthanexpected NatHazards123

PAGE 17

travelledwithchildren,whilemorementhanexpectedtravelledwithoutchildren.So,it maybeduetothepresenceofchildrenintheirtravelgroupsthatgenderwashighly statisticallysignicantonlyforOrangeCounty.Interestingly,ourstudyshowsthatthe presenceofchildrendoesnotcorrespondwithhigherratingsforthelikelihoodofevacuation,buttouristswithchildrenindicatedhigherratingsofriskperceptionthanexpected inbothregions(Table 6 ).Oneexplanationmaybethatthosewithchildrenmayhavea higherperceptionofriskbutdonotknowwheretogoorwhattodoandtherefore,they electtostayinplace.However,ouruni-variateanalysismethodsdonotallowustofurther explorethecomplicatedinteractionsbetweenfactorsthatmayinuencethedecisionsof thetourists,thusfutureresearchintothisissueiswarranted. Modesoftransportationexhibitedstatisticallysignicantcorrelationsinfouroftwelve chi-squaretests(Table 6 ).Touriststhateworrentedavehicleindicatedlowerratings thanexpectedascomparedtotouriststhatdidnotusethesemodesoftransportation.Onthe otherhand,touriststhatusedtheirpersonalvehicleindicatedahigherlikelihoodto evacuatethanwasexpectedcomparedtotouristswhodidnottravelintheirownvehicle. Eisenmanetal.( 1997 )foundthatresidentpopulationswithoutaccesstomodesoftransportationundertheircontrolwerelesslikelytoevacuate.Touristswhorentvehiclesand/or relyonairlinesfortransportationmightincurunexpectedcostsandfacelonglineswhen attemptingtochangetheirtravelplans,whichmightleadtothedecisionnottoevacuate. Thesetouristsmayneedtobebettereducatedabouttheirrelocationoptionsshoulda hurricaneoccurduringtheirtrip.Severalpreviousstudieshavefoundthattrafccongestionisnotamajorconcernforarearesidentsduringevacuation(DowandCutter 2002 ; Lindelletal. 2005 ).Althoughtouristsrentingvehiclesprovidedlowerratings,tourists travelingintheirownvehicleindicatedhigherratingsofevacuationlikelihoodsandperceptionsofriskthanthosethatdidnottravelintheirownvehicle.Overall,theseresults suggestthatrelianceontransportationnotinone'spersonalcontrol,ratherthanfearsof trafccongestion,couldlowertheevacuationrates,whiletheexibilityofrelianceona personalvehiclemayenabletouriststoevacuate. 6Conclusionsandfutureresearch Astouristsareavulnerablepopulationwhennaturaldisastersoccur,wesurveyedtourists visitingthebeachesofPinellasCountyandinlandlocationsnearOrlando,Florida.We collectedinformationonthehurricane-basedknowledgeandexperienceofthetourists alongwiththeirtriplogisticsandbasicdemographics.Forthestatedpreferenceportionof thesurvey,eachtouristindicatedtheirperceptionoftheriskassociatedwithremainingat theirlocationandlikelihoodthattheywouldevacuateuponviewingfourdifferenthurricanescenarios.Eachscenarioincludedatrackforecastconedepictingahurricaneexpected tomakelandfalloverFlorida,andallsurveysiteswerelocatedwithinthewhitecone.In additiontoseeinginformationrelatedtothelocationoflandfall,thetimeuntillandfall,the centerlineoftheconepassingeitheroverortothesideofthesurveysite,informationabout stormintensityanddurationwerealsoprovided.Ourkeyresultsinclude: € Theratingsforperceptionofriskdidnotmatchpreciselythoseforlikelihoodto evacuate. € Coastaltouristsindicatedahigherlikelihoodtoevacuatethandidinlandtourists. € Touristsperceivedahigherlevelofriskandweremorelikelytoevacuatefora Category4thanforaCategory1hurricane,regardlessoftheirlocation. NatHazards123

PAGE 18

€ TouristslocatedinPinellasCountyperceivedthehighestlevelofriskandweremost likelytoevacuateforahurricanemakinglandfallalongtheGulfCoastwiththe centerlinepassingovertheirlocationwithin24h. € Touristswhoperceivedthehighestlevelsofriskandmostlikelytoevacuateinboth PinellasandOrangeCountieswerenotpreviouslyaffectedbyahurricane;didnothave afamilymemberthatwaspreviouslyaffectedbyahurricane;hadatripdurationless than6days;andhadcheckedforthepossibilityofahurricanestrikebeforedeparting ontheirtrip. € Touristswhoeworrentedavehicleindicatedthattheywerelesslikelytoevacuate, whilethosewhoutilizedtheirownvehiclesindicatedahigherriskperception(Pinellas County)orlikelihoodtoevacuate(OrangeCounty). € Touristswithchildrenindicatedahigherperceptionofrisk,butthelikelihoodof evacuationdidnotdiffersignicantlyforthosewhotravelledwithorwithoutchildren. Whilethisempiricalstudyrepresentsarststeptowardunderstandingtheriskperceptionsandevacuationdecisionsoftourists,itisimportanttoemphasizethattheactual behaviorsoftouristswhenhurricanesaffecttheareamaydifferfromtheirresponsesonthe SPsurveys.Althoughlimitedinsamplesize,numberofpotentialscenarios,andgeographiclocations,severalkeyinsightswereobtained.Ourresultssuggestthatmosttourists havesomeknowledgeabouthurricanesbutmaybemisinterpretingthetrackforecastcone. Asthismisinterpretationmeansthattheymaynotbeawareofthetrueriskthattheyfaceif theyremaininacoastallocation;touristsneedtobebettereducatedaboutthesegraphics. Ourndingsthattouristsmaybemisinterpretingthetrackforecastconeandthattourists withprevioushurricaneexperiencearelesslikelytoevacuatearesimilartothosereported bystudiessurveyingresidents(Broadetal. 2007 ;Kusenbachetal. 2010 ;Rinconetal. 2001 ).Thissuggeststhatsimilarsurveysshouldbeadministeredtoresidentsaswellas additionaltouriststomoreaccuratelycomparetheresponsesfromthesetwopopulations. Futuresurveysshouldalsoincludequestionsthatmorespecicallyaddresshowtourists andresidentsinterpretthetrackforecastconeandthevehurricaneconditions.Ourresults indicatethattouristswouldmostoftenconsultlocaltelevisiontoobtainhurricane-related information,butmanywouldcheckmultiplesourcesincludingtheWeatherChannel.Future surveyscouldalsobeconductedatatimewhenahurricaneisnearFloridatodetermine whethertheincreasedmediacoveragemightallowforamoreaccurateinterpretationofthe informationbytourists,whichcouldthenleadtoamorerealisticperceptionofriskand higherlikelihoodtoevacuatecoastallocations.Landfalltimesthataremoreconsistentwith theleadtimesneededtoevacuatetheregioncouldalsobeutilizedinfuturestudies.Acknowledgments WethanktheEricFriedheimFoundationfortheirnancialcontributionthatallowed ustoconductthisstudy.Wealsothanktheteamofstudentassistantswhohelpedtoadministerthesurveys andtwoanonymousreviewersfortheirhelpfulcomments.ReferencesArlikattiS,LindellMK,PraterCS,ZhangY(2006)Riskareaaccuracyandhurricaneevacuationexpectationsofcoastalresidents.EnvironBehav38:226247 BakerEJ(1979)Predictingresponsetohurricanewarnings:areanalysisofdatafromfourstudies.Mass Emerg4:924 BakerEJ(1991)Hurricaneevacuationbehavior.IntJMassEmergDis9:287310 BroadK,LeiserowitzA,WeinkleJ,SteketeeM(2007)MisinterpretationsoftheConeofUncertainty''in Floridaduringthe2004hurricaneseason.BullAmMeteorolSoc88:651667.doi: 10.1175/ bams-88-5-651 BurbyRJ,WagnerF(1996)Protectingtouristsfromdeathandinjuryincoastalstorms.Disasters20:4960 NatHazards123

PAGE 19

BurnsideR,MillerDS,RiveraJD(2007)Theimpactofinformationandriskperceptiononthehurricane evacuationdecision-makingofgreaterNewOrleansresidents.SociolSpect27:727740.doi: 10.1080/02732170701534226 CutterSL,FinchC(2008)Temporalandspatialchangesinsocialvulnerabilitytonaturalhazards.ProcNat AcadSciUSA105:23012306.doi: 10.1073/pnas.0710375105 CutterSL,SmithMM(2009)Fleeingfromthehurricane'swrath:evacuationandthetwoAmericas. Environment51:2636 DanielsRJ,LogginsGM(2007)Conceptualizingcontinuouscoverage:astrategicmodelforwall-to-wall localtelevisionweatherbroadcasts.JApplCommRes35:4866 DashN,GladwinH(2007)Evacuationdecisionmakingandbehavioralresponses:individualandhousehold. NatHazardsRev8:6977 DowK,CutterSL(2002)Emerginghurricaneevacuationissues:HurricaneFloydandSouthCarolina.Nat HazardsRev3:1218 DowlingG,StaelinR(1994)Amodelofperceivedriskandintendedrisk-handlingactivity.JConsumerRes 21:119135 DrabekTE(1986)Humansystemresponsestodisaster.Springer,NewYork DrabekTE(1991)Anticipatingorganizationalevacuations:disasterplanningbymanagersoftourist-orientedprivaterms.IntJMassEmergDis9:219245 DrabekTE(1993)Variationsindisasterevacuationbehavior:publicresponsesversusprivatesector executivedecision-makingprocesses.Disasters16:104118 DrabekTE(1994)Disasterevacuationandthetouristindustry.In:Environmentandbehaviormonograph no.57.NaturalHazardsResearchandApplicationInformationCenter,Boulder,CO.p282 DrabekTE(1996)Disasterevacuationbehavior:touristsandothertransients.In:Environmentandbehavior monographno.58.NaturalHazardsResearchandApplicationInformationCenter,Boulder,CO.p354 EisenmanDP,CordascoKM,AschS,GoldenJF,GlikD(1997)Disasterplanningandriskcommunication withvulnerablecommunities:lessonsfromHurricaneKatrina.AmerJPublicHealth97:109115 ElsberryRL,LambertTDB,BootheMA(2007)AccuracyofAtlanticandeasternNorthPacictropical cycloneintensityforecastguidance.WeatherForecast22:747762.doi: 10.1175/waf1015.1 ElsnerJB,NiuXF,JaggerTH(2004)DetectingshiftsinhurricaneratesusingaMarkovchainMonteCarlo approach.JClim17:26522666 FloydMF,GibsonH,Pennington-GrayL,ThapaB(2004)Theeffectofriskperceptionsonintentionsto travelintheaftermathofSeptember11,2001.JTravelTourMarket15:1938 FuH(2004)Developmentofdynamictraveldemandmodelforhurricaneevacuation.Dissertation,LouisianaStateUniversity HessS,AdlerT,PolakJW(2006)Modelingairportandairlinechoicebehaviorwiththeuseofstated preferencesurveydata.TransportResPartE43:221225 HuanTC,BeamanJ,ShelbyL(2004)No-escapenaturaldisaster:mitigatingimpactsontourism.AnnTour Res31:255273 JohnsonEJ,TverskyA(1983)Affect,generalization,andtheperceptionofrisk.JPersonalSoc45:2031 KangJE,LindellMK,PraterCS(2007)HurricaneevacuationexpectationsandactualbehaviorinHurricane Lili.JApplSocPsych37:887903 KimballSK,MulekarMS(2004)A15yearclimatologyofNorthAtlantictropicalcyclones.PartISize Parameters17:35553575 KozakM,CrottsJ,LawR(2007)Theimpactoftheperceptionofriskoninternationaltravelers.IntJTour Res9:233242 KroesEP,SheldonRJ(1988)Statedpreferencemethods.JTransEconPolicy22:1125 KusenbachM,SimmsJL,TobinGA(2010)Disastervulnerabilityandevacuationreadiness:coastalmobile homeresidentsinFlorida.NatHazards52:7995.doi: 10.1007/s11069-009-9358-3 LauJ,XilinY,TsuiHY,PangE(2004)SARSrelatedpreventiveandriskbehaviourspractisedbyHong Kong-mainlandChinacrossbordertravellersduringtheoutbreakoftheSARSepidemicinHong Kong.JEpidemiolCommunHealth58:988996 LazoJK,WaldmanDM,MorrowBH,ThacherJA(2010)Householdevacuationdecisionmakingandthe benetsofimprovedhurricaneforecasting:developingaframeworkforassessment.WeatherForecast 25:207219 LeeWC,BellMM(2007)Rapidintensication,eyewallcontraction,andbreakdownofHurricaneCharley (2004)nearlandfall.GeophysResLett34:5.doi: 10.1029/2006GL027889 LeppA,GibsonH(2008)Sensationseekingandtourism:touristrole,perceptionofriskanddestination choice.TourManag29:740750 LindellMK,PerryRW(2004)Communicatingenvironmentalriskinmultiethniccommunities.Sage, ThousandOaks,CA NatHazards123

PAGE 20

LindellMK,PraterCS(2007)Ahurricaneevacuationmanagementdecisionsupportsystem(EMDSS).Nat Hazards40:627634.doi: 10.1007/s11069-006-9013-1 LindellMK,LuJC,PraterCS(2005)HouseholddecisionmakingandevacuationinresponsetoHurricane Lili.NatHazardsRev6:171179 MatyasCJ(2010)Associationsbetweenthesizeofhurricaneraineldsatlandfallandtheirsurrounding environments.MeteorolAtmosPhys106:135148.doi: 10.1007/s00703-009-0056-1 NHC(2011)DenitionoftheNHCtrackforecastcone. http://www.nhc.noaa.gov/aboutcone.shtml Accessed21Mar2011 ParkK,ReisingerY(2010)Differencesintheperceivedinuenceofnaturaldisastersandtravelriskon internationaltravel.TourGeogr12:124 PhillipsBD,MorrowBH(2007)Socialscienceresearchneeds:focusonvulnerablepopulations,forecasting, andwarnings.NatHazardsRev8:6168 RappaportEN(2000)LossoflifeintheUnitedStatesassociatedwithrecentAtlantictropicalcyclones.Bull AmerMeteorolSoc81:20652073 RappaportEN,FranklinJL,AvilaLA,BaigSR,BevenJL,BlakeES,BurrCA,JiingJG,JuckinsCA,Knabb RD,LandseaCW,MainelliM,MayeldM,McAdieCJ,PaschRJ,SiskoC,StewartSR,TribbleAN (2009)AdvancesandchallengesatthenationalHurricanecenter.WeatherForecast24:395419.doi: 10.1175/2008waf2222128.1 RegnierE(2008)Publicevacuationdecisionsandhurricanetrackuncertainty.ManageSci54:1628.doi: 10.1287/mnsc.1070.0764 ReisingerY,MavondoF(2005)Travelanxietyandintentionstotravelinternationally:implicationsoftravel riskperception.JTravelRes43:212225 RiadJ,NorrisF,RubackR(1999)Predictingevacuationintwomajordisasters:riskperception,social inuence,andaccesstoresources.JApplSocPsych29:918934 RinconE,LinaresMYR,GreenbergB(2001)Effectofpreviousexperienceofahurricaneonpreparedness forfuturehurricanes.AmerJEmergMed19:276279 SenkbeilJC,BrommerDM,DixonPG,BrownME,Sherman-MorrisK(2010)Theperceivedlandfall locationofevacueesfromHurricaneGustav.NatHazard54:141158.doi: 10.1007/s11069-009-9457-1 SheetsRC(1990)TheNationalHurricaneCenterpast,present,andfuture.WeatherForecast5:185232 SimpsonRH,SafrH(1974)Thehurricanedisaster-potentialscale.Weatherwise27:169186 SmithSK,McCartyC(2009)Fleeingthestorm(s):anexaminationofevacuationbehaviorduringFlorida's 2004hurricaneseason.Demography46:127145 SonmezS,GraefeAR(1998)Determiningfuturetravelbehaviorfrompasttravelexperienceandperceptionsofriskandsafety.JTravelRes37:171177 ThapaB,HollandS,AbsherJ(2008)Perceivedrisk,attitude,knowledgeandreactionarybehaviorstowards wildresamongFloridatourists.In:ChavezD,AbsherJ,WinterP(eds)Firesocialscienceresearch fromthepacicsouthwestresearchstation.Gen.Tech.Rep.PSWGTR-209.USFS:PacicSouthwest ResearchStation,Albany,CA,pp87101 UrielyN,MaozD,ReichelA(2007)Rationalisingterror-relatedrisks:thecaseofIsraelitouristsinSinai.Int JTourRes9:18 VisitFlorida(2010)2009EstimatesofFloridavisitors.Researchdepartment,VisitFlorida,Tallahassee,FL VogtCA,FesenmaierDR(1998)Expandingthefunctionalinformationsearchmodel.AnnTourRes 25:551578 WardmanM(1988)Acomparisonofrevealedpreferenceandstatedpreferencemodelsoftravelbehavior. JTransEconPolicy22:7191 WestDM,OrrM(2007)Race,gender,andcommunicationsinnaturaldisasters.PolicyStuJ35:569586 WhiteheadJC(2003)Onemilliondollarspermile?TheopportunitycostsofHurricaneevacuation.Ocean CoastManag46:10691083.doi: 10.1016/j.ocecoaman.2003.11.001 WhiteheadJC(2005)Environmentalriskandavertingbehavior:Predictivevalidityofjointlyestimated revealedandstatedbehaviordata.EnvResEcon32:301316 WhiteheadJC,EdwardsB,VanWilligenM,MaioloJR,WilsonK,SmithKT(2000)Headingforhigher ground:factorsaffectingrealandhypotheticalhurricaneevacuationbehavior.EnvironHazard 2:133142 WilksDS(1995)Statisticalmethodsintheatmosphericsciences.AcademicPress,SanDiego,CA WorldTourismOrganization(1998)Handbookonnaturaldisasterreductionintouristareas.World MeteorologicalOrganizationandWorldTourismOrganization,Madrid ZhangFQ,MorssRE,SippelJA,BeckmanTK,ClementsNC,HampshireNL,HarveyJN,HernandezJM, MorganZC,MosierRM,WangS,WinkleySD(2007)Anin-personsurveyinvestigatingpublic perceptionsofandresponsestohurricaneRitaforecastsalongtheTexasCoast.WeatherForecast 22:11771190.doi: 10.1175/2007waf2006118.1 NatHazards123