PV Technical Series No. 34 State-Wide Synthesis of Threats

Parks Victoria Technical Paper Series No. 34

State-wide Synthesis of Threats to Natural Values in Victoria’s Marine National Parks and Marine Sanctuaries

Janet M. Carey, University of Melbourne AnthonyBoxshall,ParksVictoria

MarkA.Burgman,Universityof Melbourne Ruth Beilin, University of Melbourne Louisa Flander, University of Melbourne

October2007

EXECUTIVE SUMMARY

Thedeclarationin2002ofacompletesystemofmarinenationalparksandmarine sanctuariesinVictoriacreatedimperativesforrobust,scientificallydefendableapproachesto identifyingthreatstovaluedecologicalattributesoftheparks,settingmanagementpriorities anddevelopingmonitoringsystems.Wedevelopedaninclusiveandtransparentprotocolto providetheopportunityforstakeholderinvolvementintheidentificationofvaluedattributes withintheparks,aswellasintheassessmentofassociatedrisks.Sixteenworkshopswere held at various locations across the state, with a total of 206 stakeholders participating.

Over500hazardswereidentified,eachdefinedintermsofathreattoanominatednatural valueataspecificparkorsanctuary.Somehazardssuchasthepoachingofcommercially valuablespecies,werecommontomanyparksacrossthestate.Otherswerelocation- specific,forexample,thetramplingofmangroveflatsandsaltmarshbyhard-hoofed animals such as cattle and sheep. While many hazards involved predictable, tangible threats such as poaching,pollution,invasivespeciesanddisturbancebyparkvisitors,theapproachalso identifiedanumberoflessobviousthreatsincludinggovernanceissuesandtheecological knowledge-basefortheparks.Collectively,waterqualityissueswereprominentamong stakeholderconcerns.Slightlymorethanhalfofthecontaminantthreatstowaterquality originateinadjacentcatchments.Ofthosenotarisingfromterrestrialsources,58%referred to oil or fuel contamination from maritime activities.

TheworkshopsconsistentlyidentifiedthreatsnotpreviouslyconsideredbyParksVictoriain itsinternalassessments.Halfwerethreatstoecologicalvaluesonawiderscalethanthe parksalone,forexample,catchment-basedissuesofwaterqualityandsedimenttransport.A further27%ofthenewthreatsreferredtogovernanceissuesand16%totheecological knowledge-base of the parks.

ParksVictoriausedinformationgeneratedintheriskassessmentworkshopsasadirect inputtothemanagementplanningprocessthatcoincidedwiththeworkshops.Insome casestheoutputfromworkshopswasusedtorefocussectionsofthemanagementplanson certainthreatsorincludemorediscussionofpossiblemanagementresponses.The informationonthreatsandvaluesalsoinformedthemarineresearchstrategybeing developedduring2006-2007.Ithasallowedadetailedanalysisandprioritizationofthe research gaps in each park, as well as more broadly across the state.

CONTENTS

EXECUTIVESUMMARY ...... 2

CONTENTS ...... 3

1.INTRODUCTION ...... 4

2.STAKEHOLDER WORKSHOPS ...... 5

2.1Participants...... 7

2.2TheWorkshopProcess...... 8

2.3Statewide Synthesis...... 13

3.MAJOR OUTCOMES ...... 15

3.1Threats ...... 17

3.2EcologicalValues...... 20

3.3Sensitivity to classification...... 21

3.4Scoring of Risk Register hazards...... 22

3.5Comparison with internal Parks Victoria assessments...... 23

4.DISCUSSION ...... 27

4.1Management planning...... 27

4.2Research needs of Parks Victoria...... 28

4.3Hazard scoring...... 29

4.4Sensitivity of hazards to classification...... 29

4.5Threats across the state...... 29

4.6Ecological values...... 30

4.7Participation in workshops...... 30

4.8The risk assessment project...... 31

5.ACKNOWLEDGEMENTS ...... 31

6.REFERENCES ...... 32

7.SOFTWARE...... 34

8.PERSONAL COMMUNICATION ...... 34

APPENDIX 1. ...... 35

APPENDIX 2. ...... 42

APPENDIX 3. ...... 48

APPENDIX 4. ...... 50

APPENDIX 5. ...... 51

1.INTRODUCTION

The system of 13 marine national parks and 11 marine sanctuaries in Victoria (Figure 1) was establishedin2002andismanagedbyParksVictoria.Theprimarypurposeofthesystemis toprotectandconserverepresentativesamplesofbiodiversity,ecologicalprocessesand natural features in Victorian coastal waters (Parks Victoria 2003a). The system contributes to Australia’sNationalRepresentativeSystemofMarineProtectedAreas,andrepresentsa majorcontributiontoAustralia'sinternationalarrangementsasasignatoryto the 1993 United Nations’ Convention on Biological Diversity.

Themarinenationalparks(MNPs)arerelativelylargeareasprovidingprotectionfor biodiversityandmaintainingrepresentativeexamplesofnaturalecosystemsinperpetuity.In contrast,themarinesanctuaries(MSs)aresmallerareasprotectingsitesofspecialvaluefor conservation,recreationoreducation(ParksVictoria2003a).Managementresponsibilities forthenewsystemcreatedimperativesforrobust,scientificallydefendableapproachesto identifyingthreats(stressors)totheecologicalvaluesoftheparks,settingmanagement prioritiesanddevelopingmonitoringsystems.Wheremanagementresourcesarelimited, settingprioritiesforthevaluesandthreatsallowsresourcestobeusedmosteffectivelyto achieve policy goals.

Wearedevelopingaprotocolforecologicalriskassessmentintheparksthathasdueregard for the perception of risks by individuals, and ensures that stakeholder values are an intrinsic part of decision-making. Ecological risk assessment has previously tended to assume a level of'expertknowledge'inattributingvalues(O’Brien2000).Thechallengeinmakingour protocolinclusiveandtransparent,istoprovideallstakeholderswithanopportunityto identifyvaluesofconcern,aswellasassessingtheassociatedrisks.Theprotocolfollowsa tieredapproachwherehazardsdeemedtobeofvaryingimportancereceivedifferent treatmentsintheriskassessmentprocess(Figure2).Lowriskhazards(Tier1)are recorded, but are not considered further until the next iteration of the risk management cycle. Higherrisksareaddressedmorethoroughly,perhapsbymonitoringtheecologicaleffectsof thethreat(Tier2).Potentiallyunacceptablerisksmaybetreatedmorerigorouslyagain, perhapsinvokingquantitativemodelling,orbydirectmanagementactionfollowedby monitoringandevaluationoftheoutcomes(Tier3).Theprotocolappliestechnicalrisk analysismethodstoensurethatconceptualmodelsareconsistentwithdataandtheory,and thatmonitoringprovidespertinentinformation.Thisapproachisconsistentwithrisk assessmentprotocolinotherareasofnaturalresourcemanagementsuchaswaterquality (ANZECC2000),irrigation(Hartetal.2005)andfisheriesmanagement(Fletcher2005).It differsfromsomeprotocolsbyprovidingaplaceforstakeholderswithandwithoutformal technicalexpertisetorankhazardsandsetpriorities.Outputsfromtheprojectcontribute directlytothemanagementplansfortheparks(e.g.ParksVictoria2006,2007)andprovide the basis ofa detailed strategy for future marine research across the park system.

Thefirststageoftheriskassessmentprojectwasaseriesofstakeholderworkshopsacross thestate,drawingontheexpertiseofagencystaff,communitygroups,fishers,industry representatives,academicsandparkneighbourstoidentifyhazardsofconcernineachof theparks.Individualreportswereprepareddetailingtheoutcomesofeachworkshop.This reportpresentsasynthesisofinformationfromthe16workshopsonriskstothenatural valuesofthemarinenationalparksandsanctuaries.Anassessmentofthe workshop processusedtoelicitinformationfromparkstakeholdersispresentedinacompanionreport (Careyetal.,inprep.)asisthedetailedinformationfromeachworkshop(Careyetal., 2007).

Figure1 Marine National Parks and Marine Sanctuaries in Victoria, Australia

Figure 2 A tiered approach to ecological risk assessment

2.STAKEHOLDER WORKSHOPS

Eachworkshopwasrestrictedtoasingleday,duetothedifficultyinbringingtogethera diversegroupofstakeholders.Itwashowever,generallypossibleforatleastsome stakeholderstomeettogetherinformallyontheeveningprecedingeachworkshop.This providedthefacilitatorwithanopportunitytodevelopasenseoftheecologicalissueslikely toariseandtheinteractionsamongstakeholdersthatmightaffectthecourseofthe workshop.Insomecases,itwasdeemedappropriatetoassesstwoormoreparksor sanctuariesinasingleworkshop.Thedatesoftheworkshopsandtheparksassessedin each are listed inTable 1.

Table1.Stakeholder workshops.

*AsecondWesternPortworkshopwasheldinFebruary2006tocompletetheassessmentbeguninJuly2004.

The objectives of each workshop were

•toidentifyvaluedattributesofthepark(s)orsanctuary(sanctuaries),andpotential threats to those attributes,

•toranktheimportanceofhazardsusingsubjectiveassessmentsoflikelihoodand consequence,

•to develop a register of major hazards, and

•to outline conceptual models for selected hazards.

Participants

Overall,206individualsparticipatedintheriskassessmentprocess,with13%attending morethanoneworkshop.Attendanceatindividualworkshopsrangedfrom6to23 stakeholders.

Eachworkshopaimedtoattractasdiverseagroupofparkstakeholdersaspossible.To encourageattendancebylocalresidents,workshopswereheldinthevicinityoftheparkor parkstobeassessed.TheexistingParksVictorianetworkofstakeholderswasthestarting pointforidentifyingpotentialworkshopparticipants.Thiswassupplementedwithlocalstaff knowledgeofimportantindividualsandgroupslivingneartheparks,andoccasionallybythe facilitatoridentifyinggapsintherangeofstakeholders.Somestakeholderswerealready involvedinaformalprocessofcommunity-engagementformanagementplanning,while others were groups or individuals who had an interest in the parks, either supportive or not.

Aftertheworkshopserieswascompleted,stakeholderswerebroadlyclassifiedintermsof theirprimaryreasonforparticipating,onthebasisoftheintroductionsmadeatthebeginning ofeachworkshop.Forexample,theattendanceofaparticipantwhointroducedhim/herself as“along-termresidentoftheareawithaninterestinshorebirds”wasclassifiedas “personal/community”,while“IamwithFisheriesVictoria”(aregulatorybody)resultedin classificationas“work-related”.Seventy-ninepercentofattendanceswereprimarilywork- related.However,itwasapparentthatmanystakeholders,particularlythoseatregional workshops,hadmorethanonereasonforparticipating.Alocalparkrangercouldalsobea long-termresidentandamemberofalocalcommunitygroup.Aparkneighbourwithan interest in natural history could also be the operator of a local dive-tourbusiness.

AfurtherbreakdownofattendancesisshowninFigure3,wheresomeindividual stakeholdersmayberepresentedinmorethanonecategory.Thelargestgroupof attendanceswasfromParksVictoria,theagencyresponsibleformanagingtheMarine

NationalParksandMarineSanctuaries.ParksVictoriastaffsplitbroadlyintothosewith regionalplanningorstatewidecorporateresponsibilities(11%and13%ofattendances respectively)andstaffwithafieldpresencewhoalsolivedlocally(15%ofattendances). Othermanagementagencies,parkneighboursandthosefromcommunitygroupsorwitha personal interest in environmental matters were also strongly represented.

Figure 3 Summary of stated reasons for stakeholder participation in risk assessment process. Note that individual stakeholders may be represented in more than one category.

The Workshop Process

TheworkshopsthemselvesfollowedaprocedurepreviouslytrialledwithParksVictoriain terrestrialparksanddescribedbyCareyetal.(2004,2005).Whiletheprocedurewas modifiedslightlyastheseriesofworkshopsprogressed,allworkshopsfollowedthesame general pattern:

•Introductiontoproject and workshop

•Hazard identification

•Hazard ranking

•Summary of workshop and future developments in the project

Introductiontoprojectandworkshop

Participantsweregivenashortintroductiontotheproject,totheparkorsanctuary,andto riskassessmentingeneral.Thelatterincludedtheinfluenceofcognitivebiases,andthe concept of uncertainty in decision-making.

Thecontextoftheassessmentwasthendefined.Thespatialsettingwasusuallyconfinedto theparkorsanctuaryunderconsideration.Thisrestrictionappliedonlytotheecological valuesofconcern,andnottothepotentialthreats,manyofwhicharisebeyondpark boundaries.ExceptionstothisspatialrestrictiononvalueswereWilsonsPromontoryMNP, CornerInletMNPandBunurongMNP,whereothertypesofprotectedareacoverwaters adjacenttotheMNPs,andMarengoMSwhereareasoffullyexposedreefabovehighwater markare,atypically,partofthesanctuary.Thetimeframetobeconsideredwas10yearsin allcases,chosenbecausethisistheintendedlifespanoftheParksVictoriamanagement plans.

Hazard identification

Defininghazardsintermsoftheirpotentialconsequencesforspecificvaluesofconcernisan approachtoecologicalriskassessmentthatavoidsconfusionwhereathreathasthe potentialtoaffectanumberofdifferentvalues.Itfocussestheassessmentonvaluesrather thanonthreatsalone,andhasbeenusedsuccessfullyintheirrigationandminingindustries (Hartetal.2005,Burgman2005).Italsofacilitatesthedevelopmentofmathematicalmodels byexplicitlyidentifyingecologicalvalues,andguidestheidentificationofpotentiallyuseful assessmentendpoints.Intheone-dayworkshops,allstakeholdersinturnnominated hazardsaboutwhichtheywereparticularlyconcerned,definingeachhazardintermsofthe specificecologicalvalue(s)theybelievedtobepotentiallyunderthreat.Thefacilitatorcould ofcourse,havepresentedasetofvaluesforconsiderationattheoutsetofeachworkshop, butthiscouldhavehadtheundesirableeffectofrestrictingdiscussiontothevaluesthought most important by the facilitator.

Hazard identification took place during two brainstorming sessions, the first unstructured and thesecondaidedbychecklistsofthreatsandvaluesidentifiedinotherstudiesorprevious workshops.Thechecklistsweredesignedtopromptafullerconsiderationofhazards,andto identifyvaluesorthreatsthatmayotherwisehavebeenoverlooked.Thelistofnaturalvalues (Appendix1)wasspecifictotheparkorsanctuaryinquestion,andwasdrawnlargelyfrom theEnvironmentConservationCouncilrecommendationsformarineprotectedareas(ECC 2000)andareportonthenaturalvaluesofthemarineprotectedareas(Plummeretal. 2003).Thelistofthreatswasmoregeneric,andfromthesecondworkshoponwards,was basedonthedisturbancesandactivitieslistedintheSoutheastRegionalMarinePlan (SERMP;NOO2002).RationalisingtheSERMP-basedlistandaddingotherthreats identifiedbythefacilitatororinthepreviousworkshopproducedalistof84threatsforthe thirdworkshop.Thislistwasusedatallsubsequentworkshopsandrevisedasappropriate withadditionsfromeach,tonumber145threatsbytheendoftheseries(Appendix2). Combined in a hazard matrix (Figure4) either literally or conceptually, the lists of threats and values generated a large number of potential hazards for consideration by participants.

Inaddition,aconceptualmodelforagenericVictorianmarinenationalparkcontainingboth naturalvaluesandpotentialthreats(Figure5)wasprogressivelydevelopedthroughoutthe workshopseries.Itwaspresentedatworkshops12to15tofurtherpromptthinkingabout

possiblehazards.Furtherconceptualmodelsweredevelopedasnecessarytohelpresolve language-baseduncertaintiesandotherarbitrarysourcesofmisunderstanding,leadingto better defined hazards.

Figure 4Exampleofahazardmatrix,usedasanaidtoidentifyingthreatsof concern in relation to specific ecological values.

Hazard ranking

Thenextstepintheassessmentprocesswastoselectasubsetofhazardsdeemedtobeof mostconcernforthenextstageoftheriskassessment.Thiswasachievedbyinviting participantstovotefortheir“top”hazards,thosescoringthehighestnumbersofvotesbeing promotedtoaformalriskregisterusuallynumberingbetween10and15hazards.Those hazardsnotmakingtheinitialcut-off(Tier1hazards)wererecordedforreassessmentina subsequent iteration of the risk management cycle.

Thesubsetof"top"hazardswasthenranked,broadlyfollowingtheprocessoutlinedinthe AustralianStandardforRiskManagement,AS/NZS4360(SA/SNZ1999).Participantswere dividedintogroups,usuallyofthreeorfourpeople.Eachgroupwasasmixedaspossiblein termsoftheinterestsandexperiencesofitsmembers.Eachgroupsubjectivelyassigned valuestothelikelihoodofthehazardeventuating(i.e.,notsimplythelikelihoodofthethreat occurring)and the severity of the consequences should it do so.

Bothlikelihoodandconsequencewerescoredonascaleof1to5,whereahighervalue indicatedagreaterlikelihoodoramoresevereconsequence.Groupswereprovidedwith definitionsofthelevelsused(Appendix3)inordertominimisedifferencesininterpretation. Ascoreofoneindicatedarareeventoraninsignificantconsequence,whileascoreoffive indicatedanalmostcertaineventoracatastrophicconsequence.Assigningvaluestothe essentiallyqualitativelevelsoflikelihoodandconsequenceallowedcalculationofanumeric riskrating(SA/SNZ1999)astheproductofthelikelihoodandconsequence,withpossible

ParksVictoriaTechnicalSeriesNo.34StatewideSynthesisofthreatstonaturalvalues

Figure5Conceptual model of generic marine national park

scoresrangingfrom1to25(Figure6).Fromtheseriskratings,thehazardswererankedfor eachgroup,fromhighestprioritytolowest.Anyuncertaintysurroundingthesubjective scoreswasexplicitlyincorporatedbyallowinggroupstoplaceboundsontheirestimatesif theywished(e.g.consequence,moderatetomajor).Thiswascarriedthroughtotherisk ratingcalculationbytheuseofintervalarithmetic(e.g.likelihood3to5andconsequence2 to3 results in risk rating6 to 15).

Figure6Risk scores as the product of likelihood and consequence scores.

ThelevelofagreementbetweenpairsofgroupswasthenmeasuredusingSpearman'srank correlation,where+1indicatesperfectagreementinrankorder,-1perfectdisagreement, andvaluesaround0indicatenoparticularpatternbetweenthetwosetsofranks.Theuseof aspecialisedcomputerprogram"SubjectiveRiskAssessment"(Chisholmetal.2005), allowedrapidcalculationanddisplayofriskscoresandrankcorrelationsforallpairsof groups.Withaviewtoidentifyingandpossiblyresolvingofsometypesofdisagreement,the correlationsandthelikelihoodandconsequencescoreswereexaminedtoidentifymajor disagreements,andselectedhazardswerediscussedindetail.Therewasnodemandfor consensusintherankingprocess,theaimbeingtoacknowledgeandretaingenuine differencesofopinion.Followingthediscussion,groupshadtheopportunitytorevisetheir likelihoodsandconsequencesiftheywished.Anychangeswerethenenteredintheranking software and the risks and ranks recalculated.

Theriskposedbyeachhazardwasrecordedintheriskregister,wheretherangeofvalues forlikelihood,consequenceandriskforeachhazardreflectedthefinalrangeofopinion amongthegroupsofparticipants.Riskswerealsocategorisedonafour-pointscalein accordancewithAS/NZS4360(SA/SNZ1999)forconsistencywithpreviousParksVictoria practice(Figure7).Notethatthereisnosimpleconversionfromthe25-pointriskratingscale used here (Figure 6) to the four risk categories.

Figure7RiskcategoriesbasedonlikelihoodandconsequenceinaccordancewithAS4360, Appendix E (SA/SNZ 1999).

Summary of workshop and future developments in project

Attheendoftheday,thefacilitatorpresentedasummaryofthehazardrankingexercise;in thelaterworkshops,thisincludedagraphdisplayingapriorityorderofthehazardstogether withtheuncertaintyassociatedwiththeriskscoreofeach(e.g.Figure 8a).TheseniorParks Victoriamanagerorplannerpresentthenexplainedhowtheworkshopinformationwouldbe used, and outlined later stages of the risk assessment project.

Statewide Synthesis

Over500hazardswerenominatedbystakeholdersduringtheworkshops,thussomepost- workshopsynthesiswasnecessarytoidentifystate-widepatternsinthethreatsandvalues. Thiswasachievedbytheprimaryworkshopfacilitatorbreakingdowneachhazardintoits componentthreat(s)andvalue(s)forinclusioninamatrixidentifyingtheindividualparksin whicheachthreatorvaluewasofconcern.Thefinalmatrixconsistedof114categoriesof threatand45ecologicalvalues.Thebodyofthematrixcontained1482entriesrepresenting parks at which the various combinations of threat and value occurred.

Themannerinwhichthreatsorvalueswerecategorisedandthedegreetowhichcategories were split or lumped, could have a bearing on the results of the synthesis (Florig etal.2001). Toevaluatethepossibleeffectsofthisprocess,acondensedversion(52threatsx31 values)ofthehazardmatrixwasusedtoidentifycommonhazardsandre-estimatethe priorityordersofthreatsandvalues.Scoresforthe52threatsofthecondensedmatrixwere calculatedasthenumberofparksaffectedbyagiventhreat,summedacrossallvaluesina matrix.Thiswasdonefromboththeoriginalmatrixwith45values,andacondensedmatrix with31values.Therankordersofscoresfromthetwomatriceswerethencomparedusing

Figure 8 Examples of risk scores and categories displaying associated uncertainty

Mushroom Reef workshop.

Spearman’srankcorrelation.Asimilarprocesswasappliedtothe31valuesofthe condensedmatrix,comparingthepriorityorderofthosevaluesbasedonthefulllistof114 threats to that based on the condensed list of 52.

Multidimensionalscaling(MDS;Clarke1993)wasappliedatthelevelofindividualparks,to identifyanybroadsimilaritiesamongtheparksintermsofthe114threatcategoriesofthe

full threat x value matrix. A Similarity Percentagesanalysis (SIMPER; Clarke 1993) was then used to determine which threats contributed most to any patterns identified by MDS.

3.MAJOR OUTCOMES

Overall,525hazardsweredefinedduringtheworkshops,41%ofwhichwereofsufficient concerntoparticipantstobepromotedtotheriskregisterforformalscoring(Table2).Some hazardssuchasthepoachingofcommerciallyvaluablespecies,werecommontomany parksacrossthestate.Otherswerelocation-specific,forexample,thetramplingof mangroveflatsandsaltmarshbyhard-hoofedanimalssuchascattleandsheep.Thefulllist ofhazardsisprovidedinAppendix4,andasummarisedversionofthehazardmatrixin Table 3.

Table 2.Count of hazards by workshop and priority.

#AttheShipwreckCoastworkshop,hazardsarisinginthetwobrainstormingsessionswerenotdistinguished

*The8riskregisterhazardsattheEastGippslandworkshopwerescoredseparatelyforeachofthe3parksconsidered,bringingthetotalnumberofhazardsscoredto229.

Dogs orstormwater11

2633

41

11 134 2

Trampling or lackof11121

263 3

216 13 2

external activities21223373

fossicking2 2 1 3 2 1 7 332311

groups31 111111

poaching / theft3124 1 1011114 4 7 1 1 14 261

presence1 1148 1 3 7 1 32

research1 1113

trampling1 6 1 5 51022 11111

vandalism /harassment1221412

Aquaculture74

Boating & shipping2

Fish bait or berley13

unspecified32

1

1

3

1

Aquaculture24 4

Commercial shipping6 10

Recreational boating36

other / unspecified8210

e.g. larval supply332

Nutrients from dredging1

2

212

1 112

Marine sources2 141 14

Terrestrial sources2

Commercial shipping1 111 11111

Oil spill cleanup6

Terrestrial sources1

1 141 4142

Pipeline rupture331

Dredging111

Leaching / groundwater1

Licenced discharges2 21 231

Boating & shipping33

Recreational shooting1

Terr. resource extraction1

Industrial spills/seepage211

Urban stormwater & runoff11 2413

Catchment/agric runoff1142 2 11 15

Sewagedischargefromland22 21 342 2122

Sewerage failure2111

Sewagefromboats/ships343

eAircraft1

eCommercial vessels1 1811 11

eNatural seep11

eRecreational boating4 19 111 112

eTerrestrialorigin1131

Aircraft1

Commercial fishing1

Commercial shipping53 13 3 1241131 6 2 16

Recreational boating2

Seabedresourceextraction3 2613

Terrestrial origin11

Coastal / catchment management32

Farming / forestry13111

Recreational areas1

Commercial shipping16

Released during dredging11 11115

Infrastructure failures13

Lic. discharge/dumping3414

Managment action on land1

Land sources in general3

Urban development1141

Ground/stormwaterorrunoff1 1 1 1 11311141

21111333 3 331 1

2 3318

5 1 1 2 2251127

15 111925 5

5372137 4 11 3

663 1

811124618 3

341

33

1

Man-made effects1 3 3 3 31

Natural1121

411 1

51 1113

1

512

81 52 18 871 311 1

1221 2

1 1 2 2 1112

Global warming21

Not globalwarming12

2

Coastal management1 13

3

Land sources3 229242 5 2 27 1

General maritime3 111 1 10651 6 3 214

Lost fishing gear111 1412 44

Aircraft344

Comm. fishing/shipping33

Dredging operations11

Small boats/ PWCs121112 110 43

Terrestrial vehicles432

Global warming44 1414

Global warming211142

Not due to global warming2

3.1Threats

Whilemanyhazardsinvolvedpredictable,tangiblethreatssuchaspoaching,pollutionand disturbancebyparkvisitors,theapproachalsoidentifiedanumberoflessobviousthreats (Table4).Managementissuesrangedfromthecomplexityofinteractionsamongmultiple managementagenciesoperatingaroundapark,totheabilityofparkmanagerstomarkin- waterparkboundariessuccessfully.Alackofdetailedecologicalknowledgewasa widespreadconcernbecausesuchknowledgewasconsideredimportantforparkmanagers to make informed decisions across many values.

Table 4. The most common threats at stakeholder workshops

(i.e. those arising for the highest number of parks and values)

Thelesstangiblethreatsofgovernanceandecologicalknowledgealsoappearinthe summaryofcommonhazards(i.e.threatsandvaluestogether)presentedinTable5.Hereit isnotablethatworkshopparticipantswereunabletobespecificaboutthevaluesmostatrisk fromtheselesstangiblethreats.Recreationalboating,dogsandtheintroductionofexotic

speciesappearamongmajorconcernswhenthreatsareconsideredinconjunctionwiththe natural values seen to be at risk

Table 5. The most common hazards from stakeholder workshops.

(i.e. those arising for the most parks)

Twenty-fourhazardswereofsufficientconcerntoworkshopparticipantstoachieverisk scoreswithanupperboundof25,themaximumpossible.Thoseofgreatestconcern(i.e. thosewiththehighestlowerbounds)arelistedinTable6,togetherwiththeworkshopat which they were nominated.

Table 6. Most extreme hazards (i.e. with highest upper bound of riskscore)

Collectively,waterqualityissueswereprominentamongstakeholderconcerns.Twenty-four percentofthe1482entriesinthefullthreatxvaluematrixinvolvedsuspendedsedimentsor contaminantssuchasoil,heavymetalsandnutrients.Slightlymorethanhalfofthe contaminantthreatstowaterqualityoriginateinadjacentcatchments(Figure9).Ofthosenot arisingfromterrestrialsources,58%referredtooilorfuelcontaminationfrommaritime activities.

ThreatsvariedwiththetypeofparkanditslocationwithinVictoria.Themultidimensional scalinginFigure10showsacleardistinctionbetweenparksincentralVictoriaandthoseto theeastorwest,andalsobetweenmarinenationalparksandmarinesanctuarieswithinthe centralarea.SIMPERanalysisindicatedthatthepatternapparentinFigure10isnotdriven byjustafewkeythreats.Rather,manythreatseachcontributedasmallamount(<3%)to dissimilaritiesamongthefourgroupsofparks.Noclearpatternofthreatswasfoundwith estimatedvisitationrates(ParksVictoria2005)orwithasubjectivemeasureofpark accessibility generated by thefacilitator.

Figure9Type and origin of contaminants of concern to workshop participants.

Figure10MDSplotofmarinenationalparks(MNP)andmarinesanctuaries(MS)basedon presence/absenceofthreatsofconcerntoworkshopparticipantsandusingBray-Curtis dissimilarities. (stress= 0.18)

3.2Ecological Values

Ecologicalvaluesspecifiedinthehazardsweregroupedinto45categories.Tominimise informationlossinthisprocess,somecategoriesweresubsetsofothers,andvalueswere allocatedtothemostspecificcategorypossible.Of1482entriesinthebodyofthehazard matrix,38%referredtothebroadvaluecategories“habitatsingeneral”,“biotaingeneral”or the“ecosystemingeneral”(Table7).Thisusuallyindicatedaninabilityorunwillingnessof participants to be more specific about values they believed to be under threat.

Table 7. Most common values (i.e. arising for the highest number of parks and threats)

3.3Sensitivitytoclassification

Priorityordersofboththreatsandvaluesfromthefull(114x45)andcondensed(52x31) threatxvaluematricesweresimilar.Arankcorrelationcoefficientof0.992indicatedahigh levelofagreementinrankorderforthreatscoresfromthetwomatrices.Asimilarprocedure withtherankordersof31valuesproducedacorrelationcoefficientof0.988,againindicating ahighlevelofagreement.Commonhazardsfromthecondensedmatrixvaluesincorporated all those listed in Table 5, which was constructed from the full matrix.

3.4Scoring of Risk Register hazards

Forthe229riskregisterhazards,therewasconsiderableuncertaintyassociatedwiththe subjectivescoresforbothlikelihoodandconsequence.Thisstemmedfrombothuncertainty withingroupsinassigningscores,andfromdifferencesofopinionbetweengroups. Likelihood scores spanned on average 2.9 units (st. dev. = 1.0) and consequence scores 2.7 units(st.dev.=0.8)(Figure11).Scoresforlikelihoodtendedtohavehigherupperbounds than those for consequence (Figure 12).

Theuncertaintyassociatedwithlikelihoodandconsequencescorescarriedthroughtothe riskscores,with99%ofscoresbeingintervalsratherthanpointscores.Riskscoreson averagespanned11.5units(st.dev.=4.7),withtheupperboundfallingbetween16and20 in over 50% of cases (Figure 13) The only two hazards with point scores for risk were:

•dogs off-lead disturbing of migratory wading birds (Jawbone MS), risk score = 15.

•exoticspeciesintroducedviaaquacultureaffectingtheecologicalintegrityofpark populations (Discovery Bay MNP), risk score = 4.

Seventy-ninepercentofthe229hazardsformallyscoredhadupperboundscategorisedas “extreme”(Figure 14). The upper bounds of a further 20% were rated as “high”.

Figure11Degreeofuncertaintyassociatedwithlikelihoodandconsequenceforall229risk registerhazards.

Figure12Likelihood and consequence intervals for all 229 risk register hazards.

Figure13Risk scores for all 229 risk register hazards.

Figure14Risk categories for all 213 risk register hazards. Level of risk based on likelihood and consequence in accordance with AS4360, Appendix E (SA/SNZ 1999).

3.5Comparison with internal Parks Victoria assessments

Priortotheseriesofriskassessmentworkshops,ParksVictoriahadundertakenecological threatassessmentsinternallyaspartofitscomplianceandmanagementplanning processes.Acomparisonofoutcomesofthetwosetsofassessmentindicatesthatthe workshopsconsistentlyidentifiedthreatsnotpreviouslyconsidered(Table8).These“new” threatsarelistedinAppendix5.Halfwerethreatstoecologicalvaluesonawiderscalethan theparksalone,forexample,catchment-basedissuesofwaterqualityandsediment transport.Afurther27%ofthenewthreatsreferredtogovernanceissuesand16%tothe ecologicalknowledge-base of the parks.

Table 8. Comparison of threats from previous internal Parks Victoria assessments (Parks Victoria 2003b, 2003c, 2004) with those from Risk Register hazards. (Note that partial matches ofthreats are recognised in this comparison.)

aComparisonmadeonacombinedCompliancePlanthreatlistforallparksorparkcomponents

bFortheseparks,highriskintherelevantCompliancePlanwaseffectively=10ona25-pointscaleofrisk.

cAnidenticalthreatoccurredintwoseparatehazards(i.e.hazardsdifferedonlyinvaluesaffected).dParkswerecomparedindividuallybecausethemajorityofhazardsatthisworkshopwereparkspecific.

eParkswerecomparedindividuallybecauseparksatthisworkshopwerescoredindividually.

4.DISCUSSION

4.1Managementplanning

ParksVictoriausedinformationgeneratedintheriskassessmentworkshopsasadirect inputtothemanagementplanningprocessthatcoincidedwiththeworkshops.Insome casestheoutputfromworkshopswasusedtorefocussectionsofthemanagementplanson certainthreatsorincludemorediscussionofpossiblemanagementresponses.Risk assessmentisanowstandardtoolinthemanagementplanningandcommunity engagementtoolkit(Greenwood,pers.comm.).Throughtheriskassessmentproject,Parks Victoriareinforceditsrecognitionofthestrengthoflocalcommunityinput.Inputtothe managementplanningprocesshasalwaysbeenasbroadaspossibleandisdesignedto capture local knowledge and views.

TheworkshopresultsledtoParksVictoriaandtheresearchteamchoosingasubsetof threats(bothatastatewideandlocallevel)forfurtherassessment.Theaimwasforthe researchteamandagencyjointlytodeveloptractabletheoreticalmanagementresponsesto thesubsetofthreats.Overaperiodofayearormore,thetheoreticalmanagement responsesexploredviamodellingwillbetrialledinadaptivemanagementexperimentsinthe parks.Theresultsofsuchexperimentsenablemanagersandresearcherstorefineboththe theoretical models and the actual management responses.

Thetramplingofintertidalrockplatformswasoneofthecommonhazardsidentifiedandisa candidateforTier2orTier3assessment.Itmayinvolvedifferentgroupsofparkvisitors (Figure15).Thishazardisrelativelywellunderstood,itseffectshavingbeenthesubjectofa numberofecologicalstudiesbothinVictoriaandelsewhere(e.g.BeauchampDowling 1982,KeoughQuinn1998,Milazzoetal.2005).Tramplingmayreducealgalcoveronan intertidalrockplatform,andthismayinturnaffecttheassociatedinvertebratecommunity. Tramplingofintertidalplatformsisalsoamenabletodirectmanagementresponse.Possible actionsincludeeducationalprogramstomodifythebehaviourofparkvisitors,orlimiting accesstoparticularlyvulnerablesectionsofplatformoratspecifichighimpacttimes(e.g. summer,midday,lowtides)(Figure16).Overthesummerof2006-2007,theresearchteam andParksVictoriarangerswilldoanadaptivemanagementexperimenttoassessthe efficacy of possible management responses in minimizing the threat at parks with high visitor useandpotentiallyvulnerableplatforms(BarwonBluff,MerriandRickettsPointMarine Sanctuaries,andthePointLonsdalesectionofthePortPhillipHeadsMNP).Itisimportantto considernotonlytheenvironmentaleffectivenessofanymanagementaction,butalsoits potential impact on visitors' enjoyment of the intertidal platforms.

Managementofsomeothercommonhazardswillbelessstraightforwardthanthatofthe moretangiblehazardssuchastrampling.Forexample,thelackofecologicalknowledge wasaconcernin21ofthe24parksandsanctuaries(Table5).Alongwithissuesof governance,itdoesnotimmediatelyofferatractablemanagementresponsethatcouldbe implemented,testedandrefinedwithinthetimescaleofthisresearch.However,monitoring theecologicaleffectsofvariousthreatsandmodellingthelinksbetweeneffectsanda numberofmanagementresponseswillincreasinglyallowparkmanagerstobasefuturework on available evidence rather than instinctor intuition.

impacts?

largegroups

generalpublic

controlledby

permits

howmany?commercial

operators(education/tour/adventure)

voluntaryguidelines

specific

events

non-controlled

aimtoeducate

localsfriendsoflocals

tide/time/weather

Tramplingof

tourists(seasonal)

aimtoeducate

intertidalrockplatforms

permits

positiveinteraction-rangerparkpresence

researchers

careinmanagement

ownguidelinesself-educationaccesspaths

educationcontrolsenforcementalternativeareas

ParksVictoria

impacts?

surfers

illegalfishers

Figure15Groups ofparkvisitorswhomaycausedamagetointertidalreefassemblagesbytramping(ModelfromSurfCoastworkshop).

educatetheeducators

closedseason

visitorgroups

restrictionson

visitornumbers

codeofpracticeregulation

changestoinfrastructure

(posornegeffect)

e.g.reducecarparking

zoningofsanctuary

pre-visiteducation

specifiedpathways(posornegeffect)

on-sitesignageandinterpretation

intertidalzoneinMushroomReefMarineSanctuary

Figure16Possible management responses to effectoftramplingonintertidalreef assemblages (Modelfrom Mushroom Reef workshop) .

4.2Research needs of Parks Victoria

ThesecondgoaldrivingParksVictoriainvolvementintheriskassessmentprojectwasto gatherinformationtoinformresearchneeds.Thebasicresearchneedsforthesystemwere outlinedintheMarineManagementStrategy2003-2010(ParksVictoria2003a),andamore specificmarineresearchstrategyforthenextfiveyearsisbeingdevelopedduring2006- 2007.Theinformationonthreatsandvaluesgeneratedinthecurrentprojectisallowinga detailedanalysisandprioritizationoftheresearchgapsineachpark,aswellasmore broadlyacrossthestate.Thefinalfutureresearchstrategywillbeastatewidedocumentand willbenecessarilyfocussedonwhatcanbeachievedwithinthelimitsofavailablefunding. Inputfromthe16localworkshopshelpedidentifythreats,gapsandeventopicsforfurther investigationthatmaynothavebeenproposedwithouttheworkshops.Anexampleisthe "lackofecologicalknowledge"threat.Whereastherewasalreadyaprocessforgathering baselinebiologicalinformationanddetailedmappingofmarinehabitatsacrosstheMPA system,thefindingfromtheworkshopshasledtoare-assessmentofhowtoaddressthe gapinatargetedwaytobestmeetmanagementneedsattheparklevel.Theinformation derivedlocallyacrossthestatehasallowedforamoreknowledgeableprioritizationof research needs by Parks Victoria.

4.3Hazard scoring

Theformalscoringofhazardsproducedrangesofvaluesratherthanpointscoresinnearly allcases.Therangesreflectthedifferentopinionsheldbyworkshopparticipants,whichare inturninfluencedbyeachparticipant’spersonalvaluesandperceptions.Itisasimportant to preservegenuinedifferencesofopinionandtakethemintoaccountwhenmaking managementdecisions,asitistoeliminateanydifferencesthatarisethrough misunderstandingandmisinterpretation.Ourprotocolattemptstorecogniseandretainthe former, while minimising the latter.

Awiderangeofscoresmayalsoreflectalackofdatathatwouldpermitmoreinformed estimatesofthelikelihoodofeventsoccurring.Forexample,haditbeenavailable,dataon thenumbersofpeoplewalkingacrossagivenrockplatformatdifferenttimesofyear,and whereontheplatformtheytendedtowalk,wouldinformestimatesofthelikelihoodof tramplingdamagetointertidalbiotaonthatplatform.Effortsshouldbemadetocompileany such data for the nextiteration of the riskmanagement cycle.

4.4Sensitivityofhazardstoclassification

Theanalysesofthedegreeofsplittingofhazardssuggestedthatthisprocesshadonlya minorinfluenceontheresultingprioritiesformanagementofhazards.Theranksofthreat werelittleaffectedbyreclassification,withSpearmanrankcorrelationcoefficientsinexcess of 0.99 resulting from comparisons between classifications.

4.5Threats across the state

Thestate-widesynthesisidentifiedadistinctioninthreatsbetweenparksonthecentral Victoriancoastandthosealongregionalcoastlines(Figure10).Webelievethisreflectsthe factthatthecentralcoastisthefocusofhumanactivityinVictoria.Thestatecapital,

Melbourne,isthelargestpopulationcentreinthestate,withover3.5millionpeople.Itisthe focusofcommercialshipping(Larcombeetal.2002)andprobablyalsoofwater-based recreational activity.1

Eachofthestakeholderworkshopsidentifiedoneormorethreatsnotpreviouslyconsidered duringinternalassessmentsbythemanagementagency.Inkeepingwithaworld-wide tendencyinmarineprotectedareamanagement,theinternalassessmentstendedtohave whathasbeencalledan"insideperspective",ratherthanseeingtheparksinawider context (Cicin-SainBelfiore2005).Weattributemuchofthemoreoutward-lookingassessmentat our workshops to the broader base of participating stakeholders. On average, more than half thestakeholdersatanyoneworkshopcamefromoutsideParksVictoria.Fortheseexternal participants,managementandplanningtasksrelatedtotheparkswerenoteverydayissues. Theseparticipantsbroughtawiderrangeofpersonalexperiences,opinionsandvaluesto the assessment process than would have been expected from a purely internal assessment.

4.6Ecological values

We suggest the reluctance or inability of participants to be specific about ecological values in manyhazarddefinitionsmaybeduepartlytotherelativelypoorstateofgeneralknowledge ofAustralianmarinebiota(Poore1995),andinparticularthemarineinvertebrateswhich comprisethegreatbulkofmarinebiodiversity(Ponderetal.2002).Althoughshallowcoastal watersinsouth-eastofAustraliaaresomeofthebetterstudiedofAustralianwaters,there arestillmanygapsinourknowledge.Thepersistentconcernwith“lackofecological knowledge”isalsoasymptomofthissituation.Manyenvironmentalvaluesidentifiedduring theprocessofselectingareasforinclusioninVictoria’ssystemofmarinenationalparksand sanctuariestendedtobegeneralinnature.Forexample,“thesetwosmallreefsprovidea widevarietyofmicrohabitatsinasmallarea”(ECC2000,p.74)or“extensiveintertidalrock platforms,andsubtidalrockyreefs,uncommonalongtheVictoriancoast,whichextend severalkilometresfromshorebutwhichareinrelativelyshallowwater”(ECC2000,p.69). Detailedknowledgeofthehabitatsorcommunitieswithintheparksispatchyandoften restrictedtoconspicuousgeologicalfeaturesorlargerandcharismaticspecies.However, theinvolvementoflocalparticipantsintheriskassessmentprocessprovidesaheightened educationalandreportingopportunityformanagersandresearchersalike.Community managementreferencegroupsareinplacetosupportthegatheringofinformationonthese sites and through experience or training, to recognize new threats as they occur.

4.7Participationinworkshops

Manypeopleinvitedtotheworkshopswereunabletoattendforvariousreasons,despite workshops being held close to or within the park in question. Many with personal or

1 A 2002 report on marine safety regretted the almost complete lack of data on the usage of recreational boats in Victoria (MUARC 2002)

communityreasonsforwishingtoattendwereunabletodosobecauseofworkorother commitments.Thiswasofparticularconcernbecauseitmayhaveledtotheunder- representationoflocalviewpoints,atleastfromstakeholdersnotworkinginmanagement agencies.Figure3suggeststhatwedidinfactachieveamoderatelevelofbalancein representation,butthisimpressioncanbepartlyattributedtomanyofthe personal/communityparticipantsexplicitlygivingmultiplereasonsforattending.Ahigher proportionofnon-agencyparticipantswouldhavebeendesirableandfutureiterationsofthe processshouldstrivetoincreasetheattendanceofstakeholderswithpersonalor community-basedinterestsintheparks.Thismayrequiredifferentengagementstrategiesto ensure a more extensive representation of interests.

4.8The risk assessment project

Inthefinalstageoftheriskassessmentproject,mathematicalmodels(e.g.Bayesian networks)forseveralofthekeyhazardsidentifiedduringtheworkshopswillbedeveloped. Introducedmarinespeciesandcatchment-relatedcontaminantsinwaterarethreatsof particularinterestforthisfinalstage.Themodelswillincorporateanyexistingdatathatis relevant,andwillbelinkedtomanagementresponseswhereappropriate(seeSection4.1). Theassistanceofknowledgeableparkstakeholderswillbesoughttoensurethatsuch models are consistent with their understanding of park systems.

5.ACKNOWLEDGEMENTS

Wethankthestakeholderswhoattendedtheworkshopsfortheirtimeandenthusiasm.This project is supported by a Linkage grant (LP0561075) from the Australian Research Council.

6.REFERENCES

ANZECC (2000)Australian and New Zealand Guidelines for Fresh and Marine Water Quality. Australia and New Zealand Environment and Conservation Council, and the Agriculture and Resource Management Council of Australia and New Zealand, Canberra.

Beauchamp, K.A. Gowing, M.M. (1982)A quantitative assessment of human trampling effects on a rocky intertidal community.Marine Environmental Research. 7: 279-293.

Burgman, M.A. (2005)Risks and Decisions for Conservation and Environmental Management. Cambridge University Press, Cambridge, UK.

Carey, J.M., Burgman, M.A. & Chee, Y.E. (2004) Risk Assessment and the Concept of Ecosystem Condition in Park Management. Parks Victoria Technical Series No. 13. Parks Victoria, Melbourne.

Carey, J.M., Burgman, M.A., Miller, C. & Chee, Y.E. (2005) An application of qualitative risk assessment in park management. Australasian Journal of Environmental Management. 12: 6-15.

Carey, J.M., Burgman, M.A., Boxshall, A., Beilin, R., Flander, L., Pegler, P. & White, A. (2007) Identificationof Threats to Natural Values in Victoria's Marine National Parks and Marine Sanctuaries. Parks Victoria Technical Series No. 33. Parks Victoria, Melbourne.

Carey, J.M., Beilin, R., Boxshall, A., Flander, L. & Burgman, M.A. (in prep.) Stakeholder Involvement in Hazard Identification in Victoria's Marine National Parks and Marine Sanctuaries. Parks Victoria Technical Series No. 35. Parks Victoria, Melbourne

Cicin-Sain, B. and Belfiore, S. (2005) Linking marine protected areas to integrated coastal and ocean management: a review of theory and practice.Ocean and Coastal Management. 48: 847-868.