ef17roco.com

ThisisarepositorycopyofEcologicalcriteriaforevaluationcandidatesitesformarinereserves.
WhiteRoseResearchOnlineURLforthispaper:http://eprints.
whiterose.
ac.
uk/345/Article:Roberts,C.
M.
orcid.
org/0000-0003-2276-4258,Andelman,S.
,Branch,G.
etal.
(11moreauthors)(2003)Ecologicalcriteriaforevaluationcandidatesitesformarinereserves.
ECOLOGICALAPPLICATIONS.
S199-S215.
ISSN1051-0761eprints@whiterose.
ac.
ukhttps://eprints.
whiterose.
ac.
uk/ReuseItemsdepositedinWhiteRoseResearchOnlineareprotectedbycopyright,withallrightsreservedunlessindicatedotherwise.
Theymaybedownloadedand/orprintedforprivatestudy,orotheractsaspermittedbynationalcopyrightlaws.
Thepublisherorotherrightsholdersmayallowfurtherreproductionandre-useofthefulltextversion.
ThisisindicatedbythelicenceinformationontheWhiteRoseResearchOnlinerecordfortheitem.
TakedownIfyouconsidercontentinWhiteRoseResearchOnlinetobeinbreachofUKlaw,pleasenotifyusbyemailingeprints@whiterose.
ac.
ukincludingtheURLoftherecordandthereasonforthewithdrawalrequest.
S199EcologicalApplications,13(1)Supplement,2003,pp.
S199–S2142003bytheEcologicalSocietyofAmericaECOLOGICALCRITERIAFOREVALUATINGCANDIDATESITESFORMARINERESERVESCALLUMM.
ROBERTS,1,13SANDYANDELMAN,2GEORGEBRANCH,3RODRIGOH.
BUSTAMANTE,4JUANCARLOSCASTILLA,5JENIFERDUGAN,6BENJAMINS.
HALPERN,7KEVIND.
LAFFERTY,8HEATHERLESLIE,9JANELUBCHENCO,9DEBORAHMCARDLE,10HUGHP.
POSSINGHAM,11MARYRUCKELSHAUS,12ANDROBERTR.
WARNER71EnvironmentDepartment,UniversityofYork,York,YO105DDUK2NationalCenterforEcologicalAnalysisandSynthesis,UniversityofCalifornia,735StateStreet,Suite300,SantaBarbara,California93101-3351USA3ZoologyDepartment,UniversityofCapeTown,Rondebosch7701SouthAfrica4CharlesDarwinResearchStation,SantaCruz,GalapagosIslands,Casilla17-01-3891Quito,Ecuador5DepartamentodeEcologia,PonticaUniversidadCatolicadeChile,Casilla114-DSantiagodeChile,Chile6MarineScienceInstitute,UniversityofCalifornia,SantaBarbara,California93106USA7DepartmentofEcology,Evolution,andMarineBiology,UniversityofCalifornia,SantaBarbara,California93106USA8U.
S.
GeologicalSurvey,MarineScienceInstitute,UniversityofCalifornia,SantaBarbara,California93106USA9DepartmentofZoology,OregonStateUniversity,Corvallis,Oregon93331USA10UniversityofCaliforniaSeaGrant,105EastAnapamu,#5,SantaBarbara,California93101USA11DepartmentofAppliedandMolecularEcology,UniversityofAdelaide,PB1GlenOsmond,SouthAustralia,5064,Australia12NationalMarineFisheriesService,NorthwestFisheriesScienceCenter,2725MontlakeBoulevardEast,Seattle,Washington98112USAAbstract.
Severalschemeshavebeendevelopedtohelpselectthelocationsofmarinereserves.
Allofthemcombinesocial,economic,andbiologicalcriteria,andfewofferanyguidanceastohowtoprioritizeamongthecriteriaidentied.
Thiscanimplythattherelativeweightsgiventodifferentcriteriaareunimportant.
Wheretwositesareofequalvalueecologically,thensocioeconomiccriteriashoulddominatethechoiceofwhichshouldbeprotected.
However,inmanycases,socioeconomiccriteriaaregivenequalorgreaterweightthanecologicalconsiderationsinthechoiceofsites.
Thiscanleadtoselectionofreserveswithlittlebiologicalvaluethatfailtomeetmanyofthedesiredobjectives.
Toavoidsuchapossibility,wedevelopaseriesofcriteriathatallowpreliminaryevaluationofcandidatesitesaccordingtotheirrelativebiologicalvaluesinadvanceoftheapplicationofsocioeconomiccriteria.
Weincludecriteriathat,whilenotstrictlybiological,haveastronginuenceonthespeciespresentorecologicalprocesses.
Ourschemeenablessitestobeassessedaccordingtotheirbiodiversity,theprocesseswhichunderpinthatdiversity,andtheprocessesthatsupportsheriesandprovideaspectrumofotherservicesimportanttopeople.
Criteriathatcapturebiodiversityvaluesincludebiogeographicrepresentation,habitatrepresentationandheterogeneity,andpresenceofspeciesorpopulationsofspecialinterest(e.
g.
,threatenedspecies).
Criteriathatcapturesustainabilityofbiodiversityandsheryvaluesincludethesizeofreservesnecessarytoprotectviablehabitats,presenceofexploitablespecies,vulnerablelifestages,connectivityamongreserves,linksamongeco-systems,andprovisionofecosystemservicestopeople.
Criteriameasuringhumanandnaturalthreatsenablecandidatesitestobeeliminatedfromconsiderationifrisksaretoogreat,butalsohelpprioritizeamongsiteswherethreatscanbemitigatedbyprotection.
Whileourcriteriacanbeappliedtothedesignofreservenetworks,theyalsoenablechoiceofsinglereservestobemadeinthecontextoftheattributesofexistingprotectedareas.
Theoverallgoalofourschemeistopromotethedevelopmentofreservenetworksthatwillmaintainbiodiversityandecosystemfunctioningatlargescales.
Thevaluesofeco-systemgoodsandservicesforpeopleultimatelydependonmeetingthisobjective.
Keywords:biodiversityconservation;ecosystemfunctioning;ecosystemservices;sheriesman-agement;marinereserveselection;reserveevaluationcriteria;reservenetworks.
Manuscriptreceived7January2001;accepted25January2001;featureaccepted21February2002.
ForreprintsofthisSpecialIssue,seefootnote1,p.
S3.
13E-mail:cr10@york.
ac.
ukINTRODUCTIONFormostofthe20thcentury,terrestrialecosystemshavereceivedamuchgreaterconservationprioritythanmarinesystems.
However,increasedaccesstothema-rineenvironmentforscientistsandpublichasledtoS200CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssueTABLE1.
Socialandeconomiccriteriausedtoselectthelocationsofmarineprotectedareas.
ValueCriteriaEconomicNumberofshersdependentontheareaValuefortourismPotentialcontributionofprotectiontoenhancingormaintainingeconomicvalueSocialEaseofaccessMaintenanceoftraditionalshingmethodsPresenceofculturalartifacts/wrecksHeritagevalueRecreationalvalueEducationalvalueAestheticappealScienticAmountofpreviousscienticworkundertakenRegularityofsurveyormonitoringworkdonePresenceofcurrentresearchprojectsEducationalvalueFeasibility/practicalitySocial/politicalacceptabilityAccessibilityforeducation/tourismCompatibilitywithexistingusesEaseofmanagementEnforceabilityNote:InformationsummarizedfromSwedishEnvironmentalProtectionAdministration(Naturvardsverket1995),KelleherandKenchington(1992),NordicCouncilofMinisters(NordiskaMinisterradet1995),SalmandPrice(1995),HockeyandBranch(1997),Agardy(1997),andNilsson(1998).
anappreciationthatconservationisneededasurgentlyintheseaasonland.
Wenowusebetweenaquarterandathirdofthetotalprimaryproductionofthemostproductivemarineecosystemsintheworld(PaulyandChristensen1995).
Thisissimilartotheproportionofterrestrialproductionweappropriate(Vitouseketal.
1997).
Incapturingthisproductionwehavebeensys-tematicallydepletingthehighesttrophiclevelsfrommarinefoodwebs(Paulyetal.
1998),threateningsomeofthemostvaluableandspectacularofthemarinemegafauna,anddisruptingthecompositionandfunc-tioningofmarinecommunitiesandecosystems,fromkelpforests(Daytonetal.
1998)tocoralreefs(Roberts1995).
Theseabeditselfhasbeentransformedbyheavyshinggeardraggedacrossthebottom(WatlingandNorse1998).
Thescaleandseverityoftheseimpactshasledtoasenseofurgencythatmarineecosystemsmustbeprotectedquicklytoavertanextinctioncrisisintheseacomparabletothatoccurringonland(RobertsandHawkins1999).
Atthesametime,therehasbeenagrowingrecognitionthatfullyprotectedmarinere-serves,areasclosedtoallshingandotherdisruptiveorharmfulhumanactivities,willnotonlyprotectspe-ciesandhabitatsbutcouldsignicantlyimprovesh-erymanagement(DuganandDavis1993,Bohnsack1996,Roberts1997,Allisonetal.
1998).
Thiscom-binationofconservationandmanagementbenetshasbeenapotentforceinpushingaheadtheestablishmentofmarinereservesworldwide(NationalResearchCouncil2000).
Manycountrieshavenowinitiatedam-bitiousprogramstosetupprotectedareasinthesea.
However,ifnetworksofmarinereservesaretobein-troduced,managersneedguidanceontheirplacementwhichisrmlygroundedinknowledgeofhowmarineecosystemswork.
Inthispaper,wedevelopaseriesofcriteriathatallowobjectiveassessmentofthebiolog-icalvalueofcandidatereserves.
PreviousevaluationschemesTheevaluationschemeweoutlinebelowisnotthersttohavebeendevisedforreserveselection.
Nilsson(1998)hassummarizedfourdifferentschemesdevel-opedbyKelleherandKenchington(1992),SalmandPrice(1995),theNordicCouncilofMinisters(Nor-diskaMinisterradet1995)andtheSwedishEnviron-mentalProtectionAgency(Naturvardsverket1980).
OtherschemeshavebeendevelopedbyAgardy(1997),HockeyandBranch(1997),andDayandRoff(2000).
Allofthemcombinebiologicalwithsocialandaes-theticcriteria,includingecological,economic,socialandscienticvalues,naturalness,nationalorinterna-tionalimportance,andfeasibility.
Anothersetofcri-teriathathasbecomeincreasinglyrelevanttothesitingofmarinereservesrelatestotheirvaluetosheries.
MostauthorssubsumethisundertheeconomicvaluecategorybutHockeyandBranch(1997)specicallyseparatedsherymanagementfromecologicalandoth-ersocialcriteriawhentheydevelopedasystemforselectionandevaluationofreservesinSouthAfrica.
Welimitconsiderationtobiologicalcriteria,orthoseunderpinnedbyaknowledgeofbiology.
Purelysocialcriteria,suchasaccessibilityforrecreationandaes-thetics,falloutsidethedomainofthispaper.
Table1liststhemanyfactorsinadditiontobiologicalcriteriathatmustbetakenintoaccountwhenselectingreservelocations.
KelleherandKenchington(1992)havesaidthatwhereachoiceexistsbetweenareasthatareallFebruary2003S201CRITERIAFORSELECTIONOFMARINERESERVESecologicallysuitable,thenalselectionshouldbede-terminedbysocioeconomiccriteria,andourphiloso-phyhasbeentodevelopaprocesswhichgeneratessuchchoices.
Reserveswhoseobjectivesincludeconser-vationorsheryfunctionsmusthaveasolidfoundationinbiologytosucceed.
Areservewithlittlebiologicalvaluewillprovidefewbenets,justasabankaccountwithlittlemoneywillyieldalmostnointerest.
Thusbiologicalevaluationshouldgenerallyprecedeandin-formsocialandeconomicevaluationofpotentialre-servesites.
Weanticipatethatthisschemewillfeedintoalargerdecision-makingframeworkformarine-reservenetworks,whereeconomicsandothersocialfactorsalsoplayapart.
Airameetal.
(2003)provideexamplesofhowstakeholderswithsocialandeconom-icconcernscanprovideinputinwaysthatdonotcom-promisethebiologicalintegrityofreserves.
Itisim-portanttostressherethatalthoughweproposebiolog-icalevaluationcomesbeforesocioeconomicassess-ment,stakeholdersmustbeinvolvedfromtheverybeginning.
Theirinputisessentialtothesuccessofreservedesignandestablishment(KelleherandRec-chia1998).
Toachieveconservationobjectives,reservesthatarevaluablefortheirbiodiversityneedtobechosen.
Butthey,orthemanagementschemesinwhichtheyareembedded,mustalsoincorporatetheecologicalpro-cesseswhichsupportthatdiversity.
Toachievesheryobjectives,itisnecessarythatreserveswillbepro-ductivewithintheirboundaries(i.
e.
,eitherselfsus-tainingorreceivingpropagulesfromelsewhere)butwillalsoallowthetransferofthatproductiontoshingareas.
Forthesereasons,theselectioncriteriawede-velopconcentrateontheevaluationofsitesaccordingtotheirbiodiversity,theprocessesthatwillleadtothelong-termsustainabilityofthatdiversity,andprocessesthatwillaidsheriesmanagementandprovideapleth-oraofotherhumanbenets.
Criteriathatcapturebio-diversityvaluesofsitesincludebiogeographicrepre-sentation,habitatheterogeneity,andendemism.
Cri-teriathatcapturesustainabilityofbiodiversityandsh-eryvaluesincludethepresenceofexploitablespecies,vulnerablelifestages,connectivitybetweenreserves,linksamonghabitats,theprovisionofecosystem''ser-vices''forhumans,andothercriteriathatassessthemagnitudeofhumanandnaturalthreatstotheviabilityofcandidateareas.
SELECTIONCRITERIABiogeographicrepresentationCoverageofallbiogeographicregionsisaprereq-uisiteforprotectionofbiodiversity.
HockeyandBranch(1994),Ballantine(1997),DayandRoff(2000),andothershavearguedthatrepresentingallthedifferentbiogeographicregionsinaprotected-areanet-workshouldbeacoreconservationobjective,becauseassemblagesofspecieswillbedistinctineach.
Twoofthecriteriadevelopedinotherschemes,thepresenceofrepresentativespeciesorhabitatsandthepresenceofspeciesattheirrangeboundaries(HockeyandBranch1997,Nilsson1998),relatedirectlyto,andareincorporatedin,ourbiogeographic-representationcriterion.
Therstofthesecriteriahighlightsthefactthattheremayalsobebiogeographicpatternsinthedistributionofhabitats.
Representationofhabitatsisdealtwithmoredirectlyinthecriterionbelow,butregardlessofwhetherconservationeffortisfocusedonspeciesorhabitats,wemustprovideprotectioninallbiogeographicregions.
Thesecondofthesecriteriadrawsattentiontoanotherimportantfacetofbiogeo-graphicrepresentation,theinclusionofreservesintransitionzonesbetweenbiogeographicregions.
Inclu-sionofreservesinbothtransitionandcoreareaswillhelpprotectgeneticdiversity,providingaproxymea-sureforanaspectofbiodiversitythatisrarelyconsid-eredseparately.
Reservesintransitionzonesarealsoideallyplacedtomonitorshiftsindistributionofspe-ciesarisingfromclimatechange(Barryetal.
1995).
Reservesintransitionzonescouldalsohelpbuildresilienceofspeciestoshiftsintheboundariesofbio-geographicregions,suchascanbecausedbyregimeshifts.
Forexample,coldtemperateshspeciesintheSouthernCaliforniaBighthavesufferedlong-termre-ductionsinrecruitmentduetoanomalouslywarmcon-ditions(RoemmichandMcGowan1995,Loveetal.
1998).
Recoveryfollowingrestorationofmorenormalwatertemperaturesmaybeslowbecauseintensiveex-ploitationhasdepletedspawningstocks(Loveetal.
1998).
Byprotectinglong-livedadults,reservesinthisregioncouldhavehelpedsustainspawningstocksthroughlengthyrecruitmentlowsandtherebyfosteredfasterrecovery.
HabitatrepresentationandheterogeneityOncebiogeographicregionshavebeendenedandagreementreachedontheneedtoconserverepresen-tativesectionsofeachregion,thenextdecisioniswherethereservesshouldbeplacedwithinregions.
Oneofthemostimportantcriteriaistoensureprotec-tionofallthemajorhabitatspresent(DayandRoff2000).
Thisrequiresmuchlessinformationthanifspe-ciesthemselveswereusedasabasisforreservese-lection(Wardetal.
1999).
Habitatscanbedenedinseveralways.
Forexample,somearedenedprimarilyontheirphysicalcharacteristics(e.
g.
,rockyshoresareintertidalandhaveahard-rocksubstrate)whileothersaremainlyrecognizedbytheirbiologicalattributes(e.
g.
,mangrovesareonlymangrovesbecauseofthedominantroleoftheirtrees;seeDayandRoff2000formoreexamples).
Representingallhabitatsisanessentialobjectiveforanetworkofreserves.
However,habitatheterogeneitycanbeusedtoguidetheselectionofindividualreserveunits.
IftwoormorerivalareasarebeingconsideredS202CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssueFIG.
1.
Therelationshipbetweenhabitatdiversityandthenumberofinvertebratespeciesrecordedintwoareas(Ma-putoland,asubtropicalareaontheeastcoastofSouthAfrica,andGroenrivier,onthewestcoastofSouthAfrica).
Ineacharea,datawereobtainedfrom10quadratsamplesof0.
5m2foreachhabitat,andthenumberofspeciesplottedcumulativelyforeachofthe10quadrats.
Habitatswiththehighestdiversitywereplottedrst,followedbyhabitatswithsuccessivelylowervalues.
Withtheadditionofeachhabitat,specieswereexcludedfromthecountiftheyhadalreadyappearedinthecountforprevioushabitats.
asreserves,oneoftheprimefactorsfortheirevaluationshouldbehabitatheterogeneity(thenumberofhabitatstheycontain;HockeyandBranch1997).
Asthenumberofhabitatswithinitincreases,sodoesthevalueofasiteforareserve.
Thisactsasaproxyformaximizingthenumberofspeciesprotected.
Inmanypartsoftheworldwedonothavetheluxuryoftakingthetimetodocument,onasufcientlynespatialscale,whatspe-ciesarepresent(letalonetheirrelativeabundance).
Theidealofcomprehensivecomparisonsamongsitesbasedonspeciesrichnessordiversityisanunrealizabledream.
Underthesecircumstances,habitatheteroge-neitybecomestheonlyalternativemeasurethatcanbeapplied.
Asanexample,Fig.
1showstherelationshipbe-tweenhabitatdiversityandspeciesrichnessfortwodifferentlocalitiesonthecoastofsouthernAfrica.
Twothingsemerge.
First,asmorehabitatsareaddedtoaconservedarea,speciesrichnessprogressivelyincreas-es.
Therelationshipisnotlinearbecauseallhabitatsarenotequallyspeciesrich,andmostsharesomespe-cies.
Second,differentlocalitiesdifferintheiroverallspeciesrichness,eveniftheypossessasimilarrangeofhabitats.
Tropicalareasarealmostinvariablymorespeciosethantemperateareas.
Thismeansthatdifferentbiogeographicregionscannotbecompareddirectly:habitatheterogeneitycanonlyvalidlybecomparedbe-tweensiteswithinregions.
Astretchofcoastthatcombinesalargenumberofhabitatswillhaveseveraladvantagesasasiteforareserve.
First,itwillincludeagreaterdiversityofspe-ciesthananareacomprisingonlyasinglehabitat.
Sec-ond,ithasaghtingchanceofconservinganarrayoffunctioningecosystems.
Third,wheninterconnectionsoccurbetweenecosystemsorhabitats,conservingsev-eralinasingleareahelpsprotecttheowsbetweensystems,ratherthanrelyingoninputsfromseparateprotectedareasplacedatadistance.
Forexample,manyspeciesundergodifferentphasesoftheirlifecyclesindifferenthabitats,movingbetweenthemastheyde-velop(Appeldoornetal.
1997).
Linksamongmangroveandsea-grassnurseryareasandcoralreefsarewellknown(OgdenandGladfelter1983,Roberts1996,Na-gelkerkenetal.
2000).
TheHolChanMarineReserveinBelizelinksallthreehabitats,althoughatpresentonlythecoralreefinitisfullyprotected(CarterandSedberry1997).
Byusinghabitatheterogeneityratherthanspeciesrichnessasameasureofthedesirabilityofanarea,oneovercomestheproblemthatdifferenthabitatsarein-trinsicallydifferentinspeciesrichness.
Toexemplify,coralreefsalwayshaveahigherdiversityofspeciesthansandybeaches,butthataloneshouldnotmakethemanymorevaluableintermsofconservationpri-ority.
Also,thereisthenonscienticbutimportantfactthatphysicallyheterogeneousareasarethosethatmostpeopleregardasaestheticallyappealing.
Onecaution-arynotemustbeaddedwhenrankingsitesaccordingtohabitatheterogeneity.
Ifasmallareaisdividedintomanyhabitats,thereistheriskthateachwillbetoosmalltobeviable(aviablehabitatisonewhichsup-portspopulationscapableoflong-termpersistence).
February2003S203CRITERIAFORSELECTIONOFMARINERESERVESWhenevaluatinganarea,commonsensedictatesthattiny,unsustainablehabitatsshouldbeexcludedfromconsideration.
Exactlywhatreservesizeistoosmallisadifcultquestiontoanswerandrequiresmoreem-piricalandtheoreticalstudy(seelater).
HumanthreatsTheestablishmentoffullyprotectedmarinereservesintrinsicallyacknowledgesthethreatposedbycon-sumptiveorextractivehumanuseofmarineresourcesandthebycatchand/orhabitatdamageoftenassociatedwiththeseuses.
Anumberofotherthreatsmayseri-ouslyaffecttheviabilityofamarinereserveandpre-cludeitfrommeetingitsintendedobjectives.
Forthesereasons,considerationofpast,present,andforeseeablefutureinuenceofhumanactivitiesonacandidatesiteisimportant.
Humanthreatsoccuronavarietyofspa-tialandtemporalscalesanditmaybepossibletomit-igatesomeofthemthroughcomplementarycoastal-managementefforts.
Otherthreatsmaynotbemiti-gatable.
Tobeeffectivelyemployedasacriterionforselectingmarinereserves,mitigatableandnonmitigat-ablehumanthreatsshouldbeidentiedandquantied,wherepossible.
Inmanycases,asitemaybeexposedtomorethanonethreat.
Forexample,proximitytourbancentersmayconferincreasedshing,visitationpressure,andgreatercontamination.
Fishingpressuremayhavebothpositiveandnega-tiveimplicationsfortheconsiderationofreservesites.
Sometypesofshingpressuremightincreasethepri-orityofasiteasareservebecausetheirunwantedef-fectscouldbelessenedorovercomebyprotection.
In-deed,sustainingsheriesisagoalofreservesandthereisevidencethatreservescanimproveyieldsinadjacentareas(Alcala1988,Castilla1999).
However,veryin-tenseshingmightexcludeasitefromconsideration.
Fishingpressuremaycausehabitatshiftsthroughtro-phiccascadesoractualphysicaldamage(Salaetal.
1998,WatlingandNorse1998).
Areasthatcanpoten-tiallyrecoverafterprotectionfromintensiveshingshouldbeincluded,butnotthosewhereseriousstruc-turaldamagemeansthisisunlikely.
NaturalcatastrophesEveryenvironmentexposesorganismstostressesthatreducetheircapacitytoreproduce.
Inextremecas-esthesestressesleadtodeath.
Theimpactofthesestresseswilldependontheirmagnitude,duration,andfrequency.
Therearefourcircumstancesinwhichstressesarelikelytoexceedthetolerancelimitstowhichorganismsareadapted:(1)ifthestressispro-longed;(2)ifitreachesextremes;(3)ifitoccurswithgreatrapidity;or(4)ifitsoccurrenceistooinfrequentforadaptationtoit.
Alltheseeventsareunpredictable,andincludeinfrequentdiseasesthatdecimatepopula-tions(MillerandColodey1983,LaffertyandKuris1993),earthquake-inducedupliftofrockyshores(Cas-tilla1988),irregularoods(Branchetal.
1985,1990),abrupttemperatureshiftsassociatedwithintermittentupwelling,unusuallyprolongedlowtides,andlowox-ygenlevelsinducedbyharmfulalgalblooms(Halle-graef1993,MatthewsandPitcher1996,HopnerandOelschla¨ger1997).
Afeatureofallthesedisturbancesisthatreserveswillgrantnoprotectionandsoareaswithfrequentepisodiccatastrophesareleastattractiveasreserves.
Itisoftenpossibletoidentifystretchesofcoastwherecatastrophiceventsareconcentrated.
Forexample,eruptionsofhydrogensuldeareregulareventsonthenorthernBenguelaCurrent,particularlyoffWalvisBay,andleadtoperiodicmassmortalitiesofsh(Co-penhagen1953).
Furthersouth,therearetwoareasontheSouthAfricancoastrenownedforperiodicredtides.
Thesebecomesointensetheydepleteoxygenandevengeneratehydrogensulde,leadingtomassmortalities.
Atoneofthesesites,ElandsBay,redtidesledtothreesuccessive''walkouts''ofrocklobstersinasingleyear,withover2000Mgofrocklobstersbe-comingstrandedontheshoreinanattempttoescapelow-oxygenwaters(Pitcher1999).
Attheothersite,St.
HelenaBay,therehavebeenthreemassmortalitiesoversixyears,includingthedeathofover400Mgofrocklobsters.
Ironically,St.
HelenaBayispreciselythesiteselectedforarock-lobstersanctuary.
Quiteclearly,areasthatarefocalpointsforepisodiccatastrophesmakepoorcandidatesforreserves,sincespecieswillhavetorecolonizefromelsewherefollow-ingdisturbance.
Themorefrequentthedisturbance,thelessdesirableasite(Allisonetal.
2003).
Onewayofovercomingtheproblemofcatastrophesistospreadriskthroughreplicationofreserves.
Dataonthefre-quencyandlocationofepisodiccatastrophescouldbeusedtohelpprovideroughguidelinesforspacingandnumbersofreservesrequired(H.
Possingham,S.
An-delman,S.
Gaines,L.
Botsford,A.
Hastings,andJ.
Lubchenco,unpublishedmanuscript).
SizeThequestionofsizeofmarinereservescancomeintwoforms.
First,ifalargetotalareaisavailableforprotection,howdoesonebestdivide(ornotdivide)thatintoreservesAlternatively,ifaparticularlocationrequiresprotection,howbigshouldareservebeinordertoprovideitAnswerscanbedifferentindif-ferentsituationsandmaydependonthepurposeofthereserve.
Inananalysisof89studiesoffullyprotectedmarinereserves,Halpern(2003)showedthatinvirtuallyeverycase,thecreationofreservesledtoincreasesinabun-dance,biomass,size,anddiversityoforganisms.
Themagnitudeoftheseeffectswasindependentofthesizeofthereserve(thesizerangeexaminedwas0.
002–846km2,withamedianof4.
0km2).
Inotherwords,bothsmallandlargereservesproducedsimilarincreasesinS204CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssueeachofthemeasures.
Thustheaggregatedbiologicalbenetsofreservesincreasedirectlywiththetotalareaprotected,regardlessofhowthisissubdividedintoreserveunits.
Basedonhowmuchbenetisdesired,atotalareaorproportionofhabitatcanbespeciedforprotection(NationalResearchCouncil2000,Rob-ertsandHawkins2000).
Becausemanybiologicalef-fectsofreservesdonotappeartobesizedependent,decisionsabouttheactualsizeanddistributionofre-servesshouldfocusonotherissues.
Beforediscussingtheseissues,itisimportanttoap-preciatethelimitationsofHalpern's(2003)study.
Thetaxonomicrangeofspeciesexaminedinthisstudywasratherlimited,andthequalityofdatapresentedvari-able;henceonlyquitecrudeaggregatemeasuresofperformancewerepossible.
Theseaggregatemeasuresobscuretherangeofspecies-specicresponsestore-servesandtheirpossibledependenceonreservesize.
Furthermore,nomeasuresofecosystemprocesseswereincluded,onlyofspeciespresent,andsuchprocessesmayalsobesensitivetoreservesize.
Exportfunctions.
—Thelargeedge-to-arearatiosofsmallreservesmakethembetteratexportinglarvaeandadults.
Thisisimportantformeetingbothsheryandconservationobjectives:exporttoshinggroundswillhelpsupportyields(Alcala1988,Castillaetal.
1998),whileexporttootherreserveswillpromotelong-termpopulationpersistence.
Fisheryexportisbestservedbysubdivisionofprotectedareasacrossthere-gionofinterest.
Thiswillprovidelocalbenetstosh-eries,throughjuvenileandadultspillover,andmoreregionalbenets,throughgreaterlarvalexport.
Forthisreason,manysmallreservesinanetworkwouldbepreferredoverfewerlargeones.
Viability.
—Smallreserves,however,maynotsup-portpopulationsthatarelargeenoughtopersist,es-peciallyformobilespeciesthatoftencrossreserveboundaries(increasingmobilitydiminishestheeffec-tivesizeofaprotectedarea;KramerandChapman1999).
Ifpopulationscannotsustainthemselves,thereservewillnotserveeithershery,conservation,orotherobjectivesdependentonthosespecies.
Halpern's(2003)conclusionthatevensmallreservesincreaseabundanceanddiversityoforganismsprobablystemsfromtheprotectionofspeciesthatarerecruitedfromelsewhere,i.
e.
,reservepopulationsmaintainedfromothersourcepopulations.
Thisisanimportantpoint:verysmallreserveswillfunctiononlytothedegreethatessentiallinkagestootherhabitatsaremaintained.
Suchlinkageswillinuencethesizeofhabitatfrag-mentsinreservesthatwillbeviable.
Whereahabitatisabundantinaregion,smallfragmentsaremorelikelytobeviablethanifthehabitatisrare,sincelinkagetoothersourcesofrecruitswillbegreaterforabundanthabitats.
Hence,viabilitymustbeviewedinthecontextofhabitatextentanddistribution.
Largerreserveswillprobablybeneededtoprotectrareandfragmentedhab-itatsascomparedtoextensiveandwell-connectedhab-itats.
Ifaparticularhabitatpatchistheonlyrepresen-tativeinalocalareaofthathabitat,itwillbeathigherriskthanapatchwithseveralothersimilarhabitatpatchesinthevicinity.
Viabilityofhabitatsinreserveswillalsodependontheprotectionstatusofothersim-ilarhabitatsnearby.
Wheremanyhabitatpatchesareprotected,theirviabilityinreservesislikelytobegreaterthanwherenootherexamplesofthathabitatareprotected(Roberts2000).
Thereisnosimplecutoffpointinsizewhereahabitatpatchslipsfrombeingviabletoinviable.
Thecriticalareawillbedifferentforeachspeciesthehabitatsup-ports.
Ideally,itwillbepossibletoprotecthabitatpatchesofsufcientsizethatpopulationsofallspeciespresentwillbeviable.
However,undersomecircum-stances,reservesmaybedesignedwithparticularspe-ciesinmind,forexamplespecieslistedasendangered.
Thepatchsizesofhabitatneededforsuchspeciesmightbedeterminedusingpopulationviabilityanalysis(PVA),whichhashadsomesuccessininuencingter-restrialreservedesign.
However,therearefewPVAsformarinespeciesandtheargumentforasingle-spe-ciesapproachtoconservationplanningmaybeevenweakerintheseacomparedtoonland.
Wherethreat-enedspeciesareafocusofconservationaction,otherstepsmaybenecessarytoassurepopulationsarevi-able.
Forexample,ifaspeciesissensitivetoAlleeeffectsandhasfallenbelowcriticalpopulationden-sitiesneededforsuccessfulreproduction,itmaybenecessarytoarticiallyincreasedensitiesinreserves(NationalResearchCouncil2000).
Forexample,thismaybetheonlywaytorestartreproductionbyspecieslikethecriticallyendangeredwhiteabalone(Haliotissorenseni)inCalifornia,whoseglobalpopulationisestimatedatnomorethanafewhundredwidelyscat-teredindividuals(Tegneretal.
1996).
Finally,populationsofsomespecieswillonlybeviableifnetworksofhabitatpatchesareprotectedthatincludevulnerablelifestages(forexample,spawningaggregations;Johannes1998).
Asinglereservecannotencompassthenecessaryrangeofareasnecessarytoensurepersistence(seeSelectioncritieria:Vulnerablelifestages).
Disturbance.
—Smallreservesaremorevulnerabletoperiodicdisturbances,suchasextremelowtidesoralgalblooms,thatcouldwipeoutareservepopulationinasingleevent.
Largerreservescanhelpmitigatethisvulnerability,buttheyalsorisk''puttingallyoureggsintoonebasket.
''Acatastrophiceventcoulddriveen-tirepopulationstoextinctioninasinglelargereserve,whereascreatingsmallerreserveswithinalargenet-workmightavoidsuchafate.
Thispointsoutoneoftheimportanttradeoffsbetweenreservesizeandre-servenumber.
Forexample,WashingtonState(USA)hasthreereservesforrazorclams(Siliquapatula).
ClamsinoneofthemwererecentlywipedoutduetoFebruary2003S205CRITERIAFORSELECTIONOFMARINERESERVESashiftinthecourseofariverwhilethoseinthere-mainingtwowereunaffected(D.
Simons,personalcommunication).
Ifalltheareahadbeenplacedinonereserve,noneoftheclampopulationmightremainpro-tected.
Management.
—Logisticalconcernsalsoweighheavilyindecisionsabouthowbigtomakeareserve.
Thoughmanysmallreservesspreadrisk,theyareoftenmuchhardertoenforcethanafewlargeones,andavastnetworkofsmallreservesmayprovetoocomplexforcompliance.
Largerandfewerreservesoffereasierenforcementbecausetheyallowconcentrationofmon-itoringeffortaswellasmakingiteasierforpeopletorecognizeandadheretoreserveboundaries.
However,largerreservesmaybemoredifcultforpeopletoac-cept,especiallyinintensivelyusedareas.
Thesecon-cernsareimportantsincethedegreetowhichreservessupporthigherabundance,biomass,anddiversityoforganismsisdirectlylinkedtothedegreeofcompliancewithclosuretoshing(Roberts2000).
Giventheaboveconsiderations,weneednotbecomexatedonaparticularreservesize.
Anetworkofmanymoderatelysizedreservespreservesfunctionsoverarelativelylargeareaandappearstobeagoodcompro-mise.
However,somevariationinreservesizeisim-portant.
Inthefewverylargereservessomeareasofalmostpristineecosystemmayberetained,whilesmall-andmedium-sizedreservessuitmanyspecies,spreadoutrisk,andhelpcaptureecosystemvariability.
Halpern's(2003)analysisgivesuscauseforopti-mismthatsmallreservescanbebenecial.
However,thenatureofthedataherevieweddidnotallowhimtoresolvewhetherlargerreserveswouldbenecessarytobenetlargerbodiedormoremobilespecies,butthisislikelytobeso.
Clearly,anetworkoftinyre-serveswillbeutterlyinadequatetoprotectbiodiversityortheprocessesunderpinningit.
Itiscriticaltoviewreservesizeinthecontextofthetotalareacovered.
Themostimportantpointisthatthegreaterthecom-binedareacoveredbyreserves,thegreatertheirab-solutebenets.
Presentevidencesuggeststhatpro-tecting20%ormoreofeachhabitatpresentwillbenecessarytosupportsheryproductionandsafeguardbiodiversityoverthelongterm(NationalResearchCouncil2000,RobertsandHawkins2000).
ConnectivitySeveralformsofconnectivityareimportanttoma-rinereservefunction:(1)theexchangeofoffspring,(2)themovementofjuvenilesandadults,and(3)thetransferofmaterials,suchasorganiccarbon.
Herewedealonlywiththeexchangeofoffspring.
Adultandjuvenilemovementisdiscussedelsewhereinrelationtohabitatheterogeneityandtheprotectionofvulner-ablelifestages.
MovementofmaterialsiscoveredinSelectioncriteria:Ecosystemlinkages.
Exchangeofoffspringbetweenplacesiscriticaltothefunctioningofreserves,especiallyasconditionsoutsidereservesdeteriorate.
Itcanaffectbothsustain-abilityofpopulationsinsideareserveandthedegreetowhichareservecanserveasasourceofrecruitmenttoshinggrounds.
Insitingreservestwoquestionsmustbeasked:(1)Istheresufcientconnectivitytoallowexchangeandreplenishmentwithinthereserveand(2)Willpopulationsinareserveconnectwithoth-ersinunprotectedareasorinotherreservesThean-swerstobothquestionsdependondispersaldistancesfortheorganismsprotected(mostmarinespecieshaveapelagiclarval-dispersalphase),reservesize,andlocaloceanography.
Populationsinisolatedreserveswillonlybeselfsustainingwherethereissignicantlocalretentionofoffspring.
Bycontrast,forsheryenhance-menttooccurtheremustbesubstantialexportofoff-spring.
Perhapstheperfectreservecouldbeviewedasonewherethereissufcientlocalretentiontosustainpopulationsinsidethereserve,butthemajorityofoff-springareexportedandwillreplenishshinggrounds.
Giventhatmostmarinespeciesarehighlyfecund,suchasituationmaynotbeunusual.
Itwouldtakeonlyasmallfractionofthefecunditytoprovidereplenish-ment,andthesurvivalofthisfractionmightbequitehighrelativetothefateoflarvaedispersedfurtheraeld.
Forexample,90%oftherecruitsinareservecouldresultfromretention,buttheserecruitscouldrepresentonly1%oflocalproduction.
Theother99%oflocalproductioncouldbeexportedtoshedareasandtootherreservesifthesearelocatedwithindis-persaldistances.
Becausespecies'dispersalcharacter-isticsvarywidely,theidealdistributionandsizingofreservesforonespeciesmaybeverydifferentfromthatforanother(Roberts1998,ShanksandGrantham2001,Granthametal.
2003).
Atasingle-specieslevel,siteselectionmaysub-stantiallyaffectthedegreeofself-sustainabilityandconnectiontootherareas.
Pastworksuggeststhatareasmaydiffersubstantiallyinthedegreetowhichtheydependonlocalproductionforrecruitment.
OntheCaribbeanislandofSt.
Croix,forexample,Sweareretal.
(1999)foundthattheleewardendoftheislandhadthehighestrecruitmentofblueheadwrasse(Thalas-somabifasciatum),andthatrecruitmentpulseswereretentionevents.
Windwardreefsreceivedlessrecruit-ment,andmostlarvaecamefromelsewhere.
Thusdif-ferencesappeareveninareasincloseproximity,withpopulationsderivingfromamixoflocalandmoredistantsourcesofrecruits.
Inaddition,onehabitatmayaffectanother'srecruitmentbysimplyinterceptingpo-tentialsettlers,creatingdowncurrent''shadows''oflowrecruitmentorlteringsomepotentialsettlersoutofthelarvalpool(Jones1997).
Regardlessofthepur-poseenvisionedforareserve,itisessentialthatnet-workdesignconsiderdispersaldistances(RuckelshausandHays1998)andtheprotectionoflarvalhabitatS206CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssue(CarrandReed1993).
Whilelarvaearenotsheddi-rectly,theycouldbeaffectedbypollutionorhabitatdestruction(Allisonetal.
1998).
Ifrecruitmentisinuencedbylarvalretention,thenreservedesignwillbeaffectedinmanyways.
Wherelocalretentionofoffspringislow,sustainabilityofre-servepopulationswilldependonreplenishmentfromelsewhere,eitherfromotherreservesorfromanyre-productivepopulationsinunprotectedareas(Gerberetal.
2003).
Ifthereserveisexpectedtoexportlarvaetoreplenishshedareas,thenthelocationshouldprefer-ablybeinnonretentionareas.
However,thisisadouble-edgedsword:ifallproductionisexported,theremaybenosourceofrecruitsforthereserveitself(CarrandReed1993).
Itmayalsobelessresilientintheeventofalocalcatastrophethatlowersbiodiversitybecauseofthelackofoutsidesourcesforreplenishment.
Thiswillbecomeacriticalissueifreservesaretobeusedasshery-managementtools(e.
g.
,Bohnsack1996,Laucketal.
1998).
Ifmanagementschemesbecomeoverlydependentonreserves,theremaybeacorrespondingrelaxationoftraditionalmanagementmethods.
Thismightleadtoheavyexploitationofshedareas,whichwouldreducepotentialsourcesofrecruitsforthere-servesthemselves,unlessreservesarecloselynetworked(CarrandReed1993,Allisonetal.
1998).
Forspecieshighlyvulnerabletooverexploitation,reproductivepopulationsmaybecomedepletedinshedareas,andsosourcesofrecruitsarelikelytobefromotherreserves.
Forspeciesthatarenotexploitedoraremoreresilienttoexploitation,therewillprobablystillbesignicantreproductioninunprotectedareas.
Therateandextentofrecoveryofcommunitiesinsidereservesmaydependontheintensityofshingoutsidethereserve.
Wheresourcepopulationsareintensivelyexploited,reservepopulationsmayonlyrecoverslow-ly,ormayfailtoreachthepotentialcarryingcapacityofthehabitatduetorecruitmentlimitation(Roberts2000).
Forexample,inJamaica,fullyprotectedmarinereserveshavefailedtoshowanybuildupoflargersnap-perandgrouperspecies(M.
Watson,personalcom-munication).
Localsourcepopulationshavebeensooverexploitedthattherearenoshlefttoprovidere-cruitmenttoreserves.
Bycontrast,intheBritishVirginIslands,whichhavebeenlessheavilyexploited,re-cruitmentofsnappersandgroupersisseveralordersofmagnitudehigherandpopulationsinreservesestab-lishedthereareexpectedtobuildupquickly.
Manetal.
(1995)providetheoreticalsupportforthesuggestionthattheworsetheconditionsinunprotectedareas,themorereservesarerequired.
VulnerablehabitatsHabitatsespeciallyvulnerabletodisturbancemaydeserveparticularconsiderationforinclusioninre-servenetworks,providedtheycanbeeffectivelypro-tected.
Theseincludehabitatswithlowecologicalre-silience(Folkeetal.
1996),suchascoralreefsandsaltmarshes,andregionallyrareorthreatenedhabitats.
Nonresilienthabitatsareoftencharacterizedandmaintainedbysometypeofbiologicalmatrixorstruc-tureratherthanphysicalstructuresorprocessesalone.
Theecologicalintegrityandfunctioningofmangroveforests,seagrassbeds,saltmarshes,andcoralreefsandtheirassociatedlagoonsalldependonabiologi-callygeneratedmatrix.
Perturbationorlossofthatliv-ingmatrixdestroystheintegrityandfunctioningoftheentireecologicalcommunity.
Recoveryofthebiolog-icalmatrixandhabitatmayrequireyearsfollowingmajordisturbance(Rogers1990,Bythelletal.
1993,Hughes1994,McCoyetal.
1996).
Further,fragmen-tationofthematrixbysmallerscaledisturbancescanerodehabitatfunctioningandreduceresilience(CreedandFilho1999).
Manyofthesenonresilienthabitatsoccurinshallowcoastalareaswherehumandevelop-mentisalsoconcentrated,addingfurtherneedfortheirconservation.
Recoveryofseagrassesandsomemacroalgalma-tricesmaydependonthedamagetorhizomesorhold-fasts.
Experimentallyclearedandtrampledplotsofsea-weedsandmusselsintherockyintertidalrecoveredtoamere30%and25%ofpretreatmentandcontrolvaluesin10yr(Richards1994).
Insomecases,theoriginalbiologicalmatrixmaybereplacedbyspecieswhichdonotprovidethesametypeofmatrix,andthismayper-sistforyears.
Forexample,recoveryofbrownmussel(Pernaperna)bedsmaytakemorethaneightyears,inpartbecausealgaereplacethemusselsfollowingdis-turbanceandmusselbedsarethepreferredsettlementsubstrateforrecruitsofP.
perna(Dyeetal.
1997).
Lossofthebiologicalmatrixmayalsoinuencethesuccessandspreadofinvasivespecies.
InsouthernCaliforniabays,aninvasivemussel(Musculistasen-housia)hadlowergrowthandsurvivalinbedsofnativeeelgrass(Zosteramarina)thanindisturbedhabitatwithnoeelgrass(ReuschandWilliams1999).
Bedsoftheinvasivemusselalsohadnegativeeffectsonrhizomegrowthandvegetativepropagationofnativeeelgrass.
Habitatsthatareregionallyrareorthreatenedmayneedhigherconsiderationinmarine-reservenetworks.
Forexample,coastalwetlandsofsouthernCaliforniahavebeenreducedto10%oftheirformercoverageoverthelast300yr(Zedler1984).
Additionalloss,isolation,andfragmentationofthishabitatmayfurthercompromiseviabilityandecologicalfunctioningonaregionalbasis.
Rockyshoresareecologicallyrich,yetinmanyregionscompriseasmallfractionofthecoasts.
Forsuchhabitats,alargerproportionmayneedtobeincludedinareservenetworktoensureviability.
Rarehabitatsarebytheirnaturevulnerabletochangeandeliminationandmaywellbeirreplaceablewithinare-gion.
VulnerablelifestagesIdentifyinglife-stagetransitionsthatarecriticaltoaspecies'populationdynamicscanbeusefulinchoos-February2003S207CRITERIAFORSELECTIONOFMARINERESERVESingamongsitesforinclusioninareservenetwork.
Wherevulnerablelifestagesareknown,thehabitatsorprocessessupportingthemshouldbeprotected.
Forexample,manyorganismshavelifehistoriesthatin-cludeaggregationspawningorbreeding,migrationbot-tlenecks,orotherhabitat-specicontogenetictransi-tions.
Habitatcharacteristicscanhaveaneffectonsur-vivalprobabilitieswithinalifestage,andtheycanaffectcriticaltransitionprobabilitiesbetweenlifestag-es(M.
RuckelshausandJ.
Dugan,unpublishedman-uscript).
Acommonmethodofidentifyingvulnerablelifestagesistouseamatrixpopulationmodel(Caswell1989)thatcanshowhowachangeineachdemographicparameteraffectspopulationgrowthrate.
M.
Ruckel-shausandJ.
Dugan(unpublishedmanuscript)foundthatfor17speciestheyreviewed,recruitingjuvenilesandindividualsuptosmallerreproductivesizesweremostoftencriticaltopopulationdynamics.
However,forthreespecies,aseaweedandtwopolychaetes,vul-nerablelifestagesvariedamonglocationsandyears.
Ifpotentialreservesitescontainhabitatthatsupportsvulnerablelifestagesforoneormorespecies,thosesitesmaybegivenhigherprioritythanothers.
Under-standingwhichlifestagesaremostcriticaltopopu-lationdynamicscanhelpavoidmisdirectedsiteselec-tion.
Forexample,themosteffectivewaytoenhancepopulationgrowthofloggerheadseaturtles(Carettacaretta)istoenhancejuvenilesurvivorship(Crouseetal.
1987).
Mosteffortsatsea-turtleconservationhavefocusedonprotectingeggsonnestingbeaches.
Crouseetal.
'sanalysissuggestedthiswouldhavelittleeffectonpopulationsize.
Insteadofareserveincludingnest-ingbeaches,oneprotectingsea-grasshabitats,wherejuvenileturtlebycatchinshrimpsheriesishigh,maybemoreproductive.
However,ontheeastcoastofSouthAfrica,wherethereisnoequivalentthreatfromshrimpshing,protectionofnestingsiteshasledtosubstantialpopulationrecovery(Hughes1989).
Useofmatrixpopulationmodelsislimitedbyheavydatarequirements,thedifcultyofkeepingtrackofmarkedindividuals,andtheassumptionthatpopula-tionsareclosed,withlocalrecruitmentdependentonlocalproduction(Caswell1989).
Duetolarvaldis-persal,manymarinespeciesappeartohavemuchmoreopendynamics.
Withoutknowingwhatfractionoflar-valproductionfromoneareacontributestoadultabun-dance,therelativeimportanceofdifferentlife-stagetransitionscannotbepredicted.
Forthevastmajorityofspeciesbeingconsideredforinclusioninareservenetwork,theinformationnec-essarytoperformaformaldemographicanalysiswillnotbeavailable.
WhatguidancecanweofferinthesecasesForone,itisusefultonotethenumberofhab-itatsspeciesoccurinthroughouttheirlifecycle.
Somespecieswillspendtheirentirelifewithinthehabitatscontainedwithinasinglereserve,andothersmayusemultiplehabitatsthroughouttheirlifecycle,includinghabitatsoutsideofareservenetwork.
Forexample,anumberofshandcrustaceanspeciesuseestuariesornearshoresea-grasshabitatsearlyintheirlifehistoryandthenmigrateintodeeperhabitatsoffshoreforfeed-ingandspawning(Wallaceetal.
1984).
Ideally,allimportanthabitatsshouldbenetworkedinreserves.
Somespeciesareparticularlysusceptibletomortal-ityorstresswhentheyareconcentratedinspace,suchasduringmigrationbottlenecksorspawningaggre-gations.
Aclassicexampleistheconcentrationofsquid(Loligovulgarisreynaudii)tomateandlayeggsinwell-denedgroundswheretheybecomeeasytargetsforshing(Saueretal.
1992).
Similarly,shorebirdsundertakinglong-distancemigrationsaggregateinla-goonsandestuariestofeed(Hockeyetal.
1992)andcouldsufferifthesehabitatsbecomedegraded.
Whenpossible,reservesitingshouldincludeconsiderationofthesevulnerablestagesandthehabitatsinwhichtheyoccur(Warneretal.
2000).
Itmaybeimportanttoincludeseveral,linkedlifestagesinreserves,otherwiseprotectionofspawningareascouldredirectshingef-fortintojuvenilenurserygrounds,increasingmortalityratherthanreducingit(Horwoodetal.
1998).
SpeciesorpopulationsofspecialconcernSomespeciesorpopulationshavecharacteristicsthatmakethemofparticularinterestforinclusioninre-serves.
Speciesthatareuniquetoaparticularregion(endemics)areoftenaffordedhigherconservationpri-oritythanspeciesthataremorewidespread(Interna-tionalCouncilforBirdPreservation[ICBP]1992).
Narrowlydistributedspeciesareconsideredtohaveahigherriskofextinctionthanwidespreadspeciesbe-causealocalizedimpactcouldaffecttheirentirerange.
However,thetermendemismisoftenlooselyapplied,andthiscanleadtovaluejudgementsaboutwhichspeciestoinclude.
Forexample,aspeciesmaybefoundonlyinCanadianwatersandsoisendemictothecoun-try,butmightbeverywidespreadwithinthecountry'swaters.
Alternatively,aspeciesmightberestrictedtoafewfjordsalongthenorthwestCanadiancoastbutmightbeaffordedequalvalueinconservationdeci-sions.
Endemismisaspatiallyambiguousconceptandtendstobeusedlooselyatarbitrarilydenedspatialscales.
Asecondproblemwithusingendemism,inthesenseofspeciesthatarerestrictedtoaparticulargeographicorpoliticalunit,isthatthisdenitionmayleadtobiasinevaluatingtheconservationpriorityofasite.
Forexample,placesclosetotheedgeofacountryarelikelytocontainfewerendemicsthanareasclosetothecenter,sincespecies'rangeboundariesatmarginalsitesaremorelikelytooverlapadjoiningcountries.
Thuscen-trallylocatedsitesmightbeassignedarticiallyin-atedconservationvalues.
S208CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssueFIG.
2.
Endemismhotspotsforcoral-reefshesinthetropicalAtlanticbasedonananalysisoftheglobalgeographicrangesof363speciesofsh.
Thelocationsmarkedindicatesiteswheretherewerethreeormorespeciesofrestricted-rangeshespresent,denedasspecieswithrangesizesof800000km2.
ReproducedfromRobertsetal.
(2002),withpermission.
Toovercometheseproblemsandprovidemoreob-jectivity,endemismisnowmoreusuallysubstitutedbytheconceptofrestrictedrange.
Conservationvalueisassignedtospeciesbasedonthesizeoftheirglobalgeographicrange.
ICBP(1992)hasusedabreedingrangeof50000km2orlessasanarbitrarycutoffpointtodenerestricted-rangespeciesonland.
Inthesea,Hawkinsetal.
(2000)haveusedanarbitrarycutoffpointof800000km2,reectingthegenerallylargerrangesizesofmarinespecies.
Inapplyingthiscriterion,thevalueofsitescanbedeterminedaccordingtothenumberofrestricted-rangespeciespresent.
Thiskindofassessmentcanbemadeatavarietyofscales.
Fig.
2shows,forexample,endemismhotspotsforcoral-reefshinthetropicalAtlantic.
Asanalternativetousinganarbitrarycutoffpointforwhatconstitutesarestrictedrange,itispossibletomeasurerangerarity,whichisthereciprocaloftherangesizeofaspecies.
Ifthevalueofrangerarityiscalculatedforeachofthespeciespresentintheareabeingconsideredforreserves,thendifferentsitescanbecomparedbyusingthesumoftherange-rarityvaluesforallofthespeciespresentatanygivensite(seeRobertsetal.
[2002]foranexample).
Toimplementthiscriterion,siteswouldbeselectedbasedoneitherthenumberofrestricted-rangespeciesorthesummedrange-rarityvaluesforthespeciespre-sent.
Thesitethathadthehighestvalueaccordingtooneorotherindexwouldbeselectedrstandsubse-quentsitescouldbeaddedthroughananalysisofcom-plementaritythatwouldensureinclusionofunrepre-sentedspecies.
Relictspecieshavealsobeenidentiedasbeingofspecialconcern.
Relictsarespeciesthathavepersistedinparticularregionsfollowingcontractionorfragmen-tationofpreviouslymoreextensivedistributionsoverevolutionarytimescales.
Thepresenceofsuchspecies,especiallyifthereweremanyofthem,wouldaddvaluetositesbeingconsideredforreserves,duetotheirsci-enticinterest.
Forexample,thetwoknowncoelacanthspecies(LatimeriachalumnaeandL.
menadoensis)intheComoresandIndonesiaappeartobegeographicallyveryrestricted,theirpopulationshavehighlyspecichabitatsfromwhichthereislittlemovement,andtheyarethreatenedbyartisanalsheriesinwhichtheyarebycatch(Erdmann1999,Hissmannetal.
2000).
Theyrepresentidealcandidatesforprotectioninfullypro-tectedmarinereserves.
Anothercategoryofspeciesofspecialconcernin-cludesthosethatarerarewherevertheyoccur,evenifgeographicallywidespread.
Rarityinitselfdoesnotnecessarilymeanthatpopulationviabilityislow(Gas-ton1994),butitcanrenderaspeciesmorevulnerabletohumanimpactsthanifitweremorecommon.
Inmanycases,raritymaybeadirectresultofexploitation.
Whenthisisthecase,fullyprotectedreservescandi-rectlyprotectrarespecies.
Forexample,amongthegianttropicalgroupers,thejewsh(Epinephelusita-jara)hasnowdisappearedfrommostofitsformerrangeintheCaribbean,whileanIndo-Pacicconge-neric,Epinephelustukula,isnowextremelyrare(Mor-risetal.
2000).
TargetinggloballyrarespeciesdeliberatelyexcludesspeciesthataresimplyrarebecausetheyareattheFebruary2003S209CRITERIAFORSELECTIONOFMARINERESERVESedgesoftheirrangesorinotherwisemarginalhabitat.
Conservationconcernforthelatterspeciesismuchlower.
Whilemanycountriesdoattachconservationimportancetolocallyrarespecies,populationsofspe-ciesattheedgesoftheirrangesmaybeveryhardtosustain.
Theirnumberstendtouctuatewidely(LesicaandAllendorf1995)andreservesmaybeunabletosupportviablepopulationsoverthelongterm.
How-ever,theterm''globallyrare''canbeappliedtospeciesthathavepocketsofhigherabundanceinanotherwisesparsedistribution.
SouthernFlorida,forexample,isalaststrongholdforthejewshasitsextensiveareasofmangroveandseagrassprovideexcellentjuvenilehabitat(Bullocketal.
1992).
Thislocalityhasprobablyalwaysbeenacorehabitatforjewsh,evenwhentheyexistedthroughoutmuchoftheCaribbean.
Identifyingspeciesthataregloballyrarerequiresinformationontheirabundancethroughoutthewholeoftheirrange.
Thismayextendfarbeyondthebound-ariesoftheregionbeingconsideredforreserveestab-lishment.
Suchinformationmayalsobehardtocomeby.
Ourknowledgeofmarinespecies,especiallysmall-bodiedorganisms,isgenerallypoorandsomanyspe-ciesthatcouldqualifywillbemissed.
Thisapproachismostapplicabletolargerandwell-knowntaxa.
Itisalsopossibletoconsiderthegeneticdiversityofpopulationsinselectingreservesites(Palumbi2003).
Allelsebeingequal,sitescontainingpopula-tionscharacterizedbygreatergeneticdiversityshouldbegivengreaterpriority.
Intheory,thesepopulationshaveagreaterevolutionarypotential,orabilitytoadapttochangingenvironmentalconditions(LesicaandAl-lendorf1995).
However,thereserve-selectionprocessdevelopedinthispaperisdirectedtowardachievingmultipleobjectivesandwewouldnotrecommendthatthegeneticdiversityofpopulationsofasinglespeciesshoulddictate,orevenheavilyinuence,siteselection.
Knowledgeofgeneticstructureisalwayslikelytobelimitedtoasmallnumberofspecies,whereasselectingreservesitesthatserveentirecommunitiesofspeciesshouldusuallybetheobjective.
ExploitablespeciesHalpern's(2003)reviewofreserveeffectsprovidespowerfulevidenceofthecapacityofreservestoen-hancestocksofexploitedspeciesandsimultaneouslyprotectbiodiversity.
However,forsuchreservestoful-llshery-managementobjectivestheymusteitherac-tuallyprotectpopulationsofexploitablespeciesorbecapableofprotectingthem.
Thelatterdistinctionisimportantsincepopulationsofexploitablespecieshaveoftenbeenseverelydepletedandreservesarebeingconsideredasameansofrebuildingstocks.
Onepos-sibleguidetowhetheranareamightsupportaspeciesiswhetheritdidsointhepast.
Forexample,intheCaribbean,severalspeciesofgrouperinthegenusMyc-teropercawereformerlycommononreefsbuthavebeenvirtuallyextirpatedasaresultofintensiveshing(RobertsandHawkins1999).
Mostreefsarepresum-ablystillcapableofsupportingthesespeciesifsectionsweretobeclosedtoshing.
ThechoiceofreservesinSanJuanCounty,WashingtonState,USA,wasguidedbyshersidentifyingplacesthatwereexcellentshingsitesinthepast(Murray1998).
However,oneproblemwithusingpastabundanceasaguidetothepossibilityforrecoveryisthatifpopulationshavebeeneliminatedoveralargearea,lackofrecruitmentmaylimitrecov-ery(Tegneretal.
1996).
Thepotentialforexploitedspeciestorecovermayalsobelimitedwheredepletionorotherhumanim-pacts,suchasdevelopmentorpollution,havedisruptedshbehaviorormovements.
Forexample,intheCa-ribbean,Nassaugroupers(Epinephelusstriatus)oncespawnedinaggregationsthatnumberedtensofthou-sandsofsh(Sadovy1993).
However,throughoutmuchoftheeasternCaribbeanthoseaggregationshavebeenshedtothepointofdisappearanceandthespe-ciesmayfailtorecoverbecauseremainingindividualscannoteffectivelyreproduceoutsidesuchaggrega-tions.
Thepotentialforrecoveryinsidereservesmayalsobelimitedbymodicationofhabitats.
Forexample,intheNorthSea,habitatsontheDoggerBankweredra-maticallychangedbytheadventoftrawlinginthemid-nineteenthcentury(Alward1932).
Trawlingtrans-formedthehabitatfromsomethingsuitableformanyspeciesofgadoidsh,toahabitatmoresuitableforatshsuchasplaice.
Despitetheforegoingcaveats,theoverridingmes-sagefromempirical(Halpern2003)andtheoreticalwork(Gerberetal.
2003)isthatmarinereservescananddoleadtotherecoveryofexploitedspecies.
EcosystemlinkagesInmanypartsoftheworld,apparentlydisjunctbutjuxtaposedmarineandterrestrialsystemshavebeenshowntobeinterdependent(Newell1984,Suchaneketal.
1985,Mann1988,Dugginsetal.
1989,MannandLazier1991,Bustamanteetal.
1995,PolisandHurd1996,GarmanandMacko1998).
Manytypesoflinkagecanberecognized,butweconcentratehereonthosethathaveatrophicbasisbecauseecosystemfunc-tioningisoftendenedintermsoftheperformanceofafoodweb.
Weonlyconsiderlinkagesthatarerelevanttomarine-reserveselection.
Manylinkagesaremedi-atedbyspeciesorfunctionalgroups(e.
g.
,mangrovelittersuppliessea-grasscommunities,orsubtidalma-croalgaesubsidizesintertidalorterrestrialcommuni-ties).
Consequently,linkagesaredenedhereastheow,orimpediment,ofmaterialsfromonesystemtoanotherthatallows,modies,ormodulatesthefunc-tioningofagivenmarineandcoastalarea.
Perhapsthebestdocumentedmarineandcoastallinkageistheroleofmacrophyte-derivedmatterasaS210CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssueconnectionbetweenthefoodwebsofsubtidal,inter-tidal,andestuarineecosystems(Dugginsetal.
1989,Bustamanteetal.
1995,BustamanteandBranch1996,Soaresetal.
1996,1997,Rodriguez1999).
Suchlink-ages,termed''trophicsubsidies,''canincreasethesec-ondaryandtertiaryproductionofthereceivingeco-system(Dugginsetal.
1989).
Theoperativeuseofthiscriterionfortheselectionofmarinereservesisnotstraightforward,butitsim-plicationsarehighlyrelevant.
Forexample,sessileben-thicfaunaareoftentrophicallydependentonexternalinputsandproductionprocesses.
Hence,inputsofor-ganicmatterviadetritalpathwaysmustbeconsideredwhenselectingreservesthatwillprotectbenthiccom-munities.
Thisisespeciallyimportantasthebenthicbiomassofmostreefs,beaches,lagoons,orestuarineecosystemsisoftendominatedbybenthiclterfeedersorsuspensionanddepositfeeders.
Manystudieshavedemonstratedthattheabundance,growthrates,andstandingstockofsuspensionfeedersaremediatedbytrophicsubsidiesfromadjacentecosystemsorcom-munities(e.
g.
,AlbertandValiela1995,Bustamanteetal.
1995,Slimetal.
1996,DugginsandEckman1997,IrlandiandCrawford1997,Soaresetal.
1997).
Con-sequently,conservationofhabitatsorecosystemsthatcontainsuchimportantgroupsofbasalspecies(sensuMenge1992)willdependontheselectionofsiteswith-inanetworkthatincludethedonorendsofallidentiedlinkages.
Forexample,networksitescontainingkelpforestswillneedtocomplementintertidalrocky-shorereservesthataredependentonthem,andmangroveforestssupplementingsea-grasscommunitiesshouldbeincludedinthereservedesign.
Despitetheirobviousimportance,ecosystemlink-ageshavenotbeenfullyexploredorusedinthede-cision-makingprocessofdesigningandsettingupma-rinereserves.
Again,linkagescannotbeusediniso-lation,buttheiruseasacriterioncanalterormodifythedecision-makingprocesswhenchoicesneedtobemadebetweenrivalsitesbeingconsideredforprotec-tion.
Theyarealsoanimportantconsiderationforthelong-termsuccessofagivenmarinereserve.
Acoastalreserveofsmallislands,designedfortheconservationofparticularfaunalcommunitiesthatincludeinverte-bratesandseabirds,maynotsucceedifthenearshoremarinecommunitiesarenotincluded(PolisandHurd1996,PolisandStrong1996).
EcologicalservicesforhumansEcosystemservicesarebenetsprovidedtohuman-ityandotherlifeonEarthasaresultofthefunctioningofecologicalsystems(Daily1997).
Virtuallyalleco-logicalsystemsprovideservices.
Theprovisionofgoods,suchasshsupplies,isanobviousandvitallyimportantservice.
Intactecologicalsystemsareessen-tialtoprovidetheservicesthathumansuse.
Manyservicesoperateoverlocalscales.
Forex-ample,mangrovesandcoralreefsprotectcoastlinesfromwaveaction.
Wetlandplantsinreedbeds,saltmarshesandmangrovestrapandltersediments.
Filterfeeders,suchasoysters,cleansewaterinbays.
Bio-logicalstructures,forexamplemangroverootsorsea-grassbeds,createcriticalhabitatsforspawning,re-cruitment,orgrowthofbiota.
Healthyecosystemsalsoprovidedesirableplacesforrecreation,inspiration,ed-ucation,andthecreationofnewknowledge.
Someser-vicesoperateoverregionalscales,forexampledetox-icationandsequestrationofpollutants,generationandmaintenanceofbiodiversity,decompositionofdeadbi-ologicalmaterialandcyclingofnutrients.
Yetotherservicesoperateoverglobalscales,forexamplethepartialregulationofclimateandbiogeochemicalcy-cling.
Onlyrecentlyhaveecosystemservicesbeenrecog-nizedasimportantbenets(Daily1997).
Historically,humanshaveprimarilyvaluedthegoodsprovidedbyecologicalsystems,whileserviceshavebeentakenforgranted.
Becauseservicesarelargelyoutsideeconomicsystems,changesinquantityarenotalwaysobvious,unlikegoodswhichincreaseinvaluewhenscarce(Per-manetal.
1996).
TogethergoodsandservicesprovidethelifesupportsystemsforalllifeonEarth.
Ecosystemservicesareaby-productoffunctioningecologicalsys-tems.
Ifthesesystemsaremodied,theservicestheyprovidemaychange.
Thestudyofecosystemservicesisinitsinfancy.
Methodstoquantifythemareunderdevelopmentandthescalesoverwhichservicesareproducedarebeinginvestigated.
Studyoftheeffectsofalterationsinbio-diversityonfunctioningofecologicalsystemsisan-otheractiveareaofresearch.
Despitethelackofspe-cicknowledgeaboutmanycriticalservices,thefactthatmostareirreplaceableandsubjecttoseriousdis-ruptionshouldsignaltheimportanceofpreservingthem.
APPLICATIONOFTHECRITERIAThecriteriadescribedinthispapercanbeappliedinmanydifferentwaystotheproblemofselectingreservesanddesigningnetworks.
Robertsetal.
(2003)havedevelopedamethodofapplyingthefullsetofcriteriainaprocessaimedatevaluatingcandidatere-servesinthecontextoftheircontributiontolargerscalenetworks.
Theyarguethattheperformanceofindivid-ualreserveswillbeenhancedbynetworking,andthatnetworksareessentialtomaintainthelarge-scaleeco-logicalprocessesthatunderpinecosystemfunctioninginthesea.
Airameetal.
(2003)showhowourcriteriacanbeintegratedwithsocialcriteriatochooseloca-tionsforfullyprotectedzoneswithinalargemultiple-useprotectedarea—theChannelIslandsNationalMa-rineSanctuaryinCalifornia.
Possinghametal.
(H.
Pos-singham,S.
Andelman,S.
Gaines,L.
Botsford,A.
Has-February2003S211CRITERIAFORSELECTIONOFMARINERESERVEStings,andJ.
Lubchenco,unpublishedmanuscript)takeanalternatetack,developingthecriteriaintoaseriesofrulesofthumbtosimplifyreserveselection.
Whilealltheseschemesplacebiologicalassessmentaheadofsocioeconomicconcerns,theyallrecognizethatsocio-economicconsiderationswilldominatethenalstagesofsiteselection.
However,useofthecriteriacanhelpfacilitatedecisionsbyrevealingwhethercandidatesitespossessbiologicalattributesthatwillenablethemtomeettheirobjectives.
ACKNOWLEDGMENTSThisiscontributionnumber32fromtheWorkingGroupontheScienceofMarineReservesoftheNationalCenterforEcologicalAnalysisandSynthesis,acenterfundedbyNSF(GrantNo.
DEB-94-21535),theUniversityofCalifornia–SantaBarbara,theCaliforniaResourcesAgency,andtheCal-iforniaEnvironmentalProtectionAgency.
Weareverygrate-fultoJulieHawkinsforherhelpwithpreparingthemanu-script,andtoananonymousreviewerforhelpfulcomments.
LITERATURECITEDAgardy,T.
S.
1997.
Marineprotectedareasandoceancon-servation.
AcademicPress,Dallas,Texas,USA.
Airame,S.
,J.
E.
Dugan,K.
D.
Lafferty,H.
Leslie,D.
A.
McArdle,andR.
R.
Warner.
2003.
Applyingecologicalcriteriatomarinereservedesign:acasestudyfromtheCaliforniaChannelIslands.
EcologicalApplications13:S170–S184.
Albert,M.
,andI.
Valiela.
1995.
Organicaggregatesinde-tritalfoodwebs:incorporationbybatscallopsArgopectenirradians.
MarineEcologyProgressSeries121:117–124.
Alcala,A.
C.
1988.
EffectsofmarinereservesoncoralshabundancesandyieldsofPhilippinecoralreefs.
Ambio17:194–199.
Allison,G.
W.
,S.
D.
Gaines,J.
Lubchenco,andH.
P.
Pos-singham.
2003.
Ensuringpersistenceofmarinereserves:Catastrophesrequireadoptinganinsurancefactor.
Ecolog-icalApplications13:S8–S24.
Allison,G.
W.
,J.
Lubchenco,andM.
H.
Carr.
1998.
Marinereservesarenecessarybutnotsufcientformarinecon-servation.
EcologicalApplications8:S79–S92.
Alward,G.
L.
1932.
TheseasheriesofGreatBritainandIreland.
AlbertGait,Grimsby,UK.
Appeldoorn,R.
S.
,C.
W.
Recksiek,R.
L.
Hill,F.
E.
Pagan,andG.
D.
Dennis.
1997.
Marineprotectedareasandreefshmovements:theroleofhabitatincontrollingontoge-neticmigration.
Pages1917–1922inH.
LessiosandI.
G.
Macintyre,editors.
ProceedingsoftheEighthInternationalCoralReefSymposium.
Volume2.
SmithsonianTropicalResearchInstitute,Balboa,RepublicofPanama.
Ballantine,W.
J.
1997.
Designprinciplesforsystemsof'no-take'marinereserves.
Page4inT.
J.
Pitcher,editor.
Thedesignandmonitoringofmarinereserves.
Volume9,Num-ber1.
FisheriesCentre,UniversityofBritishColumbia,Vancouver,BritishColumbia,Canada.
Barry,J.
P.
,C.
H.
Baxter,R.
D.
Saparin,andS.
E.
Gilman.
1995.
Climate-related,long-termfaunalchangesinaCal-iforniarockyintertidalcommunity.
Science267:672–675.
Bohnsack,J.
A.
1996.
Maintenanceandrecoveryofreefsheryproductivity.
Pages283–313inN.
V.
C.
PoluninandC.
M.
Roberts,editors.
Reefsheries.
ChapmanandHall,London,UK.
Branch,G.
M.
,R.
Bally,B.
A.
Bennett,H.
P.
deDecker,G.
A.
W.
Fromme,C.
W.
Heyle,andJ.
P.
Willis.
1985.
Syn-opsisoftheimpactofarticiallyopeningthemouthoftheBotRiverEstuary:implicationsformanagement.
Trans-actionsoftheRoyalSocietyofSouthAfrica45:465–483.
Branch,G.
M.
,S.
Eekhout,andA.
L.
Bosman.
1990.
Short-termeffectsofthe1988OrangeRiveroodsontheinter-tidalrocky-shorecommunitiesoftheopencoast.
Trans-actionsoftheRoyalSocietyofSouthAfrica47:331–354.
Bullock,L.
H.
,M.
D.
Murphy,M.
F.
Godcharles,andM.
E.
Mitchell.
1992.
Age,growth,andreproductionofJewshEpinephelusitajaraintheEasternGulfofMexico.
FisheryBulletin90:243–249.
Bustamante,R.
H.
,andG.
M.
Branch.
1996.
Thedependenceofintertidalconsumersonkelp-derivedorganicmatteronthewestcoastofSouthAfrica.
JournalofExperimentalMarineBiologyandEcology196:1–28.
Bustamante,R.
H.
,G.
M.
Branch,andS.
Eekhout.
1995.
MaintenanceofanexceptionalintertidalgrazerbiomassinSouthAfrica:subsidybysubtidalkelps.
Ecology76:2314–2329.
Bythell,J.
C.
,E.
H.
Gladfelter,andM.
Bythell.
1993.
ChronicandcatastrophicnaturalmortalityofthreecommonCarib-beanreefcorals.
CoralReefs12:143–152.
Carr,M.
H.
,andD.
C.
Reed.
1993.
Conceptualissuesrelevanttomarineharvestrefuges:examplesfromtemperatereefshes.
CanadianJournalofFisheriesandAquaticScience50:2019–2028.
Carter,J.
,andG.
R.
Sedberry.
1997.
Thedesign,functionanduseofmarinesheryreservesastoolsfortheman-agementandconservationoftheBelizeBarrierReef.
Pages1911–1916inH.
LessiosandI.
G.
Macintyre,editors.
Pro-ceedingsoftheEighthInternationalCoralReefSympo-sium.
Volume2.
SmithsonianTropicalResearchInstitute,Balboa,RepublicofPanama.
Castilla,J.
C.
1988.
Earthquakecausedcoastalupliftanditseffectsonrockyintertidalkelpcommunities.
Science424:440–443.
Castilla,J.
C.
1999.
Coastalmarinecommunities:trendsandperspectivesfromhumanexclusionexperiments.
TrendsinEcologyandEvolution14:280–283.
Castilla,J.
C.
,P.
Manriquez,J.
Alvarado,A.
Rosson,C.
Pino,C.
Espoz,R.
Soto,D.
Oliva,andO.
Defeo.
1998.
Artisanal''Caletas''asunitsofproductionandco-managersofben-thicinvertebratesinChile.
CanadianSpecialPublicationinFisheriesandAquaticScience125:407–413.
Caswell,H.
1989.
Matrixpopulationmodels:construction,analysisandinterpretation.
SinauerandAssociates,Sun-derland,Massachusetts,USA.
Copenhagen,W.
J.
1953.
TheperiodicmortalityofshintheWalvisBayregion;aphenomenonwithintheBenguelaCurrent.
InvestigationalReportNo.
14.
DivisionofSeaFisheriesofSouthAfrica,CapeTown,SouthAfrica.
Creed,J.
C.
,andG.
M.
A.
Filho.
1999.
Disturbanceandrecoveryofthemacrooraofaseagrass(HalodulewrightiiAscherson)meadowintheAbrolhosMarineNationalPark,Brazil:anexperimentalevaluationofanchordamage.
Jour-nalofExperimentalMarineBiologyandEcology235:285–306.
Crouse,D.
T.
,L.
B.
Crowder,andH.
Caswell.
1987.
Astage-basedpopulationmodelforloggerheadseaturtlesandim-plicationsforconservation.
Ecology68:1412–1423.
Daily,G.
C.
1997.
Nature'sservices,societaldependenceonnaturalecosystems.
IslandPress,Washington,D.
C.
,USA.
Day,J.
C.
,andJ.
C.
Roff.
2000.
Planningforrepresentativemarineprotectedareas:aframeworkforCanada'soceans.
WorldWildlifeFund,Toronto,Ontario,Canada.
Dayton,P.
K.
,M.
J.
Tegner,P.
B.
Edwards,andK.
L.
Riser.
1998.
Slidingbaselines,ghostsandreducedexpectationsinkelpforestcommunities.
EcologicalApplications8:309–322.
Dugan,J.
E.
,andG.
E.
Davis.
1993.
Applicationsofmarinerefugiatocoastalsheriesmanagement.
CanadianJournalofFisheriesandAquaticScience50:2029–2042.
Duggins,D.
O.
C.
,andJ.
E.
Eckman.
1997.
IskelpdetritusS212CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssuegoodfoodforsuspensionfeedersMarineBiology128:489–495.
Duggins,D.
O.
C.
,C.
A.
Simenstad,andJ.
A.
Estes.
1989.
Magnicationofsecondaryproductionbykelpdetritusincoastalmarineecosystems.
Science245:170–173.
Dye,A.
H.
,T.
A.
Lasiak,andS.
Gabula.
1997.
Recoveryandrecruitmentofthebrownmussel,Pernaperna(L.
)inTranskei:implicationsformanagement.
SouthAfricanJournalofZoology32:118–123.
Erdmann,M.
V.
1999.
Anaccountoftherstlivingcoela-canthknowntoscientistsfromIndonesianwaters.
Envi-ronmentalBiologyofFishes54:439–443.
Folke,C.
,C.
S.
Holling,andC.
Perrings.
1996.
Biologicaldiversity,ecosystemsandthehumanscale.
EcologicalAp-plications6:1018–1024.
Garman,G.
C.
,andS.
A.
Macko.
1998.
Contributionofmarine-derivedorganicmattertoanAtlanticcoast,fresh-water,tidalstreambyanadromousclupeidsh.
JournaloftheNorthAmericanBenthologicalSociety17:277–285.
Gaston,K.
J.
1994.
Rarity.
ChapmanandHall,London,UK.
Gerber,L.
R.
,L.
W.
Botsford,A.
Hastings,H.
P.
Possingham,S.
D.
Gaines,S.
R.
Palumbi,andS.
Andelman.
2003.
Pop-ulationmodelsformarinereservedesign:aretrospectiveandprospectivesynthesis.
EcologicalApplications13:S47–S64.
Grantham,B.
A.
,G.
L.
Eckert,andA.
L.
Shanks.
2003.
Dispersalpotentialofmarineinvertebratesindiversehab-itats.
EcologicalApplications13:S108–S116.
Hallegraef,G.
M.
1993.
Areviewofharmfulalgalbloomsandtheirapparentglobalincrease.
Phycologia33:79–99.
Halpern,B.
S.
2003.
Theimpactofmarinereserves:DoreservesworkanddoesreservesizematterEcologicalApplications13:S117–S137.
Hawkins,J.
P.
,C.
M.
Roberts,andV.
Clark.
2000.
Thethreat-enedstatusofrestricted-rangecoralreefshspecies.
An-imalConservation3:81–88.
Hissmann,K.
,H.
Fricke,andJ.
Schauer.
2000.
Patternsoftimeandspaceutilisationincoelacanths(Latimeriachal-umnae),determinedbyultrasonictelemetry.
MarineBi-ology136:943–952.
Hockey,P.
A.
R.
,andG.
M.
Branch.
1994.
Conservingma-rinebiodiversityontheAfricancoast:implicationsofaterrestrialperspective.
AquaticConservationandMarineandFreshwaterEcosystems4:345–362.
Hockey,P.
A.
R.
,andG.
M.
Branch.
1997.
Criteria,objectivesandmethodologyforevaluatingmarineprotectedareasinSouthAfrica.
SouthAfricanJournalofMarineScience18:369–383.
Hockey,P.
A.
R.
,R.
A.
Navarro,B.
Kalejta,andC.
R.
Ve-lasquez.
1992.
Theriddleofthesands:whyareshorebirddensitiessohighinsouthernestuaries.
AmericanNaturalist140:961–979.
Hopner,T.
,andB.
Oelschla¨ger.
1997.
Fromthewarningsig-naltothecaseofemergencyhistory,courseandexpla-nationoftheblacktideventofsummer1996.
AktuelleProblemederMeeresumweltS7:11–20.
Horwood,J.
W.
,J.
H.
Nichols,andS.
Milligan.
1998.
Eval-uationofclosedareasforshstockconservation.
JournalofAppliedEcology35:893–903.
Hughes,G.
R.
1989.
Turtles.
Pages230–243inA.
I.
L.
PayneandR.
J.
M.
Crawford,editors.
OceansoflifeoffSouthernAfrica.
Vlaeberg,CapeTown,SouthAfrica.
Hughes,T.
P.
1994.
Catastrophes,phaseshifts,andlarge-scaledegradationofaCaribbeancoralreef.
Science265:1547–1551.
InternationalCouncilforBirdPreservation.
1992.
Puttingbiodiversityonthemap:priorityareasforglobalconser-vation.
InternationalCouncilforBirdPreservation,Cam-bridge,UK.
Irlandi,E.
A.
,andM.
K.
Crawford.
1997.
Habitatlinkages:theeffectofintertidalsaltmarshesandadjacentsubtidalhabitatsonabundance,andgrowthofanestuarinesh.
Oecologia110:222–230.
Johnannes,R.
E.
1998.
Thecasefordata-lessmarineresourcemanagement:examplesfromtropicalnearshorensheries.
TrendsinEcologyandEvolution13:243–246.
Jones,G.
P.
1997.
Relationshipsbetweenrecruitmentandpostrecruitmentprocessesinlagoonalpopulationsoftwocoralreefshes.
JournalofExperimentalMarineBiologyandEcology213:231–246.
Kelleher,G.
,andR.
Kenchington.
1992.
Guidelinesfores-tablishingmarineprotectedareas.
Amarineconservationanddevelopmentreport.
WorldConservationUnion(IUCN),Gland,Switzerland.
Kelleher,G.
,andC.
Recchia.
1998.
Editorial—lessonsfrommarineprotectedareasaroundtheworld.
Parks8(2):1–4.
Kramer,D.
L.
,andM.
R.
Chapman.
1999.
Implicationsofshhomerangesizeandrelocationformarinereservefunc-tion.
EnvironmentalBiologyofFishes55:65–79.
Lafferty,K.
D.
,andA.
M.
Kuris.
1993.
MassmortalityofabaloneHaliotiscracherodiiontheCaliforniaChannelIs-lands:testsofepidemiologicalhypotheses.
MarineBiologyProgressSeries96:239–248.
Lauck,T.
,C.
W.
Clark,M.
Mangel,andG.
R.
Munro.
1998.
Implementingtheprecautionaryprincipleinsheriesman-agementthroughmarinereserves.
EcologicalApplications8:S72–S78.
Lesica,P.
,andF.
W.
Allendorf.
1995.
WhenareperipheralpopulationsvaluableforconservationConservationBi-ology9:753–760.
Love,M.
S.
,J.
E.
Caselle,andW.
VanBuskirk.
1998.
Aseveredeclineinthecommercialpassengershingvesselrocksh(Sebastesspp.
)catchintheSouthernCaliforniaBight,1980–1996.
CaliforniaCooperativeOceanicFish-eriesInvestigationsReports39:180–195.
Man,A.
,R.
Law,andN.
V.
C.
Polunin.
1995.
Roleofmarinereservesinrecruitmenttoreefsheries:ametapopulationmodel.
BiologicalConservation71:197–204.
Mann,K.
H.
1988.
Productionanduseofdetritusinvariousfreshwater,estuarine,andcoastalmarineecosystems.
Lim-nologyandOceanography33:910–930.
Mann,K.
H.
,andJ.
R.
N.
Lazier.
1991.
Dynamicsofmarineecosystems.
BlackwellScientic,Oxford,UK.
Matthews,S.
G.
,andG.
C.
Pitcher.
1996.
WorstrecordedmarinemortalityontheSouthAfricancoast.
Pages89–92inT.
Yasumoto,Y.
Oshima,andY.
Fukuyo,editors.
Harm-fulandtoxicalgalblooms.
IntergovernmentalOceano-graphicCommissionofUNESCO,Paris,France.
McCoy,E.
D.
,H.
R.
Mushinsky,D.
Johnson,andW.
E.
Mes-haka,Jr.
1996.
MangovedamagecausedbyHurricaneAn-drewonthesouthwesterncoastofFlorida.
BulletinofMa-rineScience59:1–8.
Menge,B.
A.
1992.
Communityregulation:underwhatcon-ditionsarebottom-upfactorsimportantonrockyshoresEcology73:755–765.
Miller,R.
J.
,andA.
G.
Colodey.
1983.
WidespreadmassmortalityofthegreenseaurchininNovaScotia,Canada.
MarineBiology73:263–267.
Morris,A.
V.
,C.
M.
Roberts,andJ.
P.
Hawkins.
2000.
Thethreatenedstatusofgroupers(Epinephelinae).
BiodiversityandConservation9:919–942.
Murray,M.
R.
1998.
ThestatusofmarineprotectedareasinPugetSound.
PugetSound/GeorgiaBasinEnvironmentalReportSeries8:103–104.
Nagelkerken,I.
,M.
Dorenbosch,W.
C.
E.
P.
Verberk,E.
C.
delaMoriniere,andG.
vanderVelde.
2000.
Importanceofshallow-waterbiotopesofaCaribbeanbayforjuvenilecoralreefshes:patternsinbiotopeassociation,commu-nitystructureandspatialdistribution.
MarineEcologyPro-gressSeries202:175–192.
February2003S213CRITERIAFORSELECTIONOFMARINERESERVESNationalResearchCouncil.
2000.
Marineprotectedareas:toolsforsustainingoceanecosystems.
NationalAcademyPress,Washington,D.
C.
,USA.
Naturvardsverket.
1995.
Aktionsplanfo¨rbiologiskmang-fald.
(Actionplanonbiologicaldiversity.
)[InSwedish.
]Rapport4463.
Naturvardsverket(SwedishEnvironmentalProtectionAgency),Stockholm,Sweden.
Newell,R.
C.
1984.
Thebiologicalroleofdetritusinthemarineenvironment.
Pages317–344inFlowsofenergyandmaterialsinmarineecosystems,theoryandpractice.
NATOConferenceSeries2,MarineScience.
Volume13.
Plenum,NewYork,NewYork,USA.
Nilsson,P.
1998.
Criteriafortheselectionofmarineprotectedareas.
Report4834.
SwedishEnvironmentalProtectionAgency,Stockholm,Sweden.
NordiskaMinisterradet.
1995.
MarinareservatiNorden,DelI.
(MarineprotectedareasintheNordiccountries,PartI.
)[InSwedish.
]TemaNord553.
NordiskaMinisterradet(Nor-dicCouncilofMinisters),Copenhagen,Denmark.
Ogden,J.
C.
,andE.
H.
Gladfelter.
1983.
Coralreefs,seagrassbeds,andmangroves:theirinteractioninthecoastalzonesoftheCaribbean.
UNESCOReportsinMarineScience23:1–133.
Palumbi,S.
R.
2003.
Populationgenetics,demographiccon-nectivity,andthedesignofmarinereserves.
EcologicalApplications13:S146–S158.
Pauly,D.
,andV.
Christensen.
1995.
Primaryproductionre-quiredtosustainglobalsheries.
Nature374:255–257.
Pauly,D.
,V.
Christensen,J.
Dalsgaard,R.
Froese,andF.
Torres.
1998.
Fishingdownmarinefoodwebs.
Science279:860–863.
Perman,R.
,Y.
Ma,andJ.
McGilvray.
1996.
Naturalresourceandenvironmentaleconomics.
Longman,London,UK.
Pitcher,G.
C.
1999.
HarmfulalgalbloomsoftheBenguelacurrent.
SeaFisheriesResearchInstitute,CapeTown,SouthAfrica.
Polis,G.
A.
,andS.
D.
Hurd.
1996.
Linkingmarineandterrestrialfoodwebs:allochthonousinputfromtheoceansupportshighsecondaryproductivityonsmallislandsandcoastallandcommunities.
AmericanNaturalist147:396–423.
Polis,G.
A.
,andD.
R.
Strong.
1996.
Foodwebcomplexityandcommunitydynamics.
AmericanNaturalist147:813–846.
Reusch,T.
B.
H.
,andS.
L.
Williams.
1999.
Macrophytecanopystructureandthesuccessofaninvasivemarinebivalve.
Oikos84:398–416.
Richards,D.
V.
1994.
EffectsofdisturbanceonpopulationdynamicsofselectedtaxaintherockyintertidalzoneofChannelIslandsNationalPark,California.
Pages45–58inW.
HalvorsonandG.
J.
Maender,editors.
ThefourthCal-iforniaIslandssymposium:updateonthestatusofresourc-es.
SantaBarbaraMuseumofNaturalHistory,SantaBar-bara,California,USA.
Roberts,C.
M.
1995.
Effectsofshingontheecosystemstructureofcoralreefs.
ConservationBiology9:988–995.
Roberts,C.
M.
1996.
Settlementandbeyond:populationreg-ulationandcommunitystructure.
Pages85–112inN.
V.
C.
PoluninandC.
M.
Roberts,editors.
Reefsheries.
Chap-manandHall,London,UK.
Roberts,C.
M.
1997.
Ecologicaladvicefortheglobalsh-eriescrisis.
TrendsinEcologyandEvolution12:35–38.
Roberts,C.
M.
1998.
Sources,sinks,andthedesignofmarinereservenetworks.
Fisheries23:16–19.
Roberts,C.
M.
2000.
Selectingthelocationsofmarinere-serves:optimalityvs.
opportunism.
BulletinofMarineSci-ence66:581–592.
Roberts,C.
M.
,etal.
2003.
Applicationofecologicalcriteriainselectingmarinereservesanddevelopingreservenet-works.
EcologicalApplications.
13:S215–S228.
Roberts,C.
M.
,andJ.
P.
Hawkins.
1999.
Extinctionriskinthesea.
TrendsinEcologyandEvolution14:241–246.
Roberts,C.
M.
,andJ.
P.
Hawkins.
2000.
Fully-protectedmarinereserves:aguide.
WorldWildlifeFund,Washing-ton,D.
C.
andUniversityofYork,York,UK.
Roberts,C.
M.
,D.
J.
McClean,J.
E.
N.
Veron,J.
P.
Hawkins,G.
R.
Allen,D.
E.
McAllister,C.
G.
Mittermeier,F.
W.
Schueler,M.
Spalding,F.
Wells,C.
Vynne,andT.
B.
Warn-er.
2002.
Marinebiodiversityhotspotsandconservationprioritiesfortropicalreefs.
Science295:1280–1284.
Rodriguez,S.
1999.
Subsidiostrocosenambientesmarinos:laimportanciadelasmacroalgaspardasaladerivacomofuenteexogenaderecursosalimentariosparaelerizoTe-trapygusniger(Echinodermata:Echinoidea)enelinter-marealrocosodelacostadeChilecentral.
[InSpanish.
]Dissertation.
PonticaUniversidadCatolicadeChile,San-tiago,Chile.
Roemmich,D.
,andJ.
McGowan.
1995.
ClimaticwarmingandthedeclineofzooplanktonintheCaliforniaCurrent.
Science267:1324–1326.
Rogers,C.
S.
1990.
Responsesofcoralreefsandreefor-ganismstosedimentation.
MarineEcologyProgressSeries62:185–202.
Ruckelshaus,M.
H.
,andC.
G.
Hays.
1998.
Conservationandmanagementofspeciesinthesea.
Pages110–156inP.
L.
FiedlerandP.
M.
Karieva,editors.
Conservationbi-ologyforthecomingdecade.
ChapmanandHall,London,UK.
Sadovy,Y.
1993.
TheNassaugrouper,endangeredorjustunluckyReefEncounter13:10–12.
Sala,E.
,C.
F.
Boudouresque,andM.
Harmelin-Vivien.
1998.
Fishing,trophiccascadesandthestructureofalgalassem-blages:evaluationofanoldbutuntestedparadigm.
Oikos83:425–439.
Salm,R.
,andA.
Price.
1995.
Selectionofmarineprotectedareas.
Pages15–31inS.
Gubbay,editor.
Marineprotectedareas:principlesandtechniquesformanagement.
ChapmanandHall,London,UK.
Sauer,W.
H.
H.
,M.
J.
Smale,andM.
R.
Lipinski.
1992.
Thelocationofspawninggrounds,spawningandschoolingbe-haviourofthesquidLoligovulgarisreynaudii(Cephalop-oda:Myopsida)offtheeasternCapeCoast,SouthAfrica.
MarineBiology114:97–107.
Shanks,A.
L.
,B.
A.
Grantham,andM.
H.
Carr.
2001.
Prop-aguledispersaldistanceandthesizeandspacingofmarinereserves.
EcologicalApplications13:S159–S169.
Slim,F.
J.
,M.
A.
CocheretDeLaMoriniere,andE.
VanDerVelde.
1996.
Tidalexchangeofmacrolitterbetweenaman-grovesforestandadjacentseagrass(GaziBay,Kenya).
NetherlandsJournalofAquaticBiology30:119–128.
Soares,A.
G.
,A.
McLachlan,andT.
A.
Schlater.
1996.
Dis-turbanceeffectsofstrandedkelponpopulationsofthesandybeachbivalveDonaxserra(Roding).
JournalofEx-perimentalMarineBiologyandEcology205:165–186.
Soares,A.
G.
,T.
A.
Schlater,andA.
McLachlan.
1997.
Car-bonandnitrogenexchangebetweensandybeachclams(Donaxserra)andkelpbedsintheBenguelaupwellingregion.
MarineBiology127:657–664.
Suchanek,T.
H.
,S.
H.
Williams,T.
G.
Ogden,D.
K.
Hubbard,andI.
P.
Hill.
1985.
Utilizationofshallow-waterseagrassdetritusbyCaribbeandeep-seamacrofauna:13Cevidence.
DeepSeaResearch32:201–214.
Swearer,S.
E.
,J.
Caselle,D.
Lea,andR.
R.
Warner.
1999.
Larvalretentionandrecruitmentinanislandpopulationofacoral-reefsh.
Nature402:799–802.
Tegner,M.
J.
,L.
V.
Basch,andP.
K.
Dayton.
1996.
Nearextinctionofanexploitedmarineinvertebrate.
TrendsinEcologyandEvolution11:278–280.
Vitousek,P.
M.
,H.
A.
Mooney,J.
Lubchenco,andJ.
M.
S214CALLUMM.
ROBERTSETAL.
EcologicalApplicationsSpecialIssueMelillo.
1997.
HumandominationofEarth'secosystems.
Science277:494–499.
Wallace,J.
H.
,H.
M.
Kok,L.
E.
Beckley,B.
Bennett,S.
J.
M.
Blaber,andA.
K.
Whiteld.
1984.
SouthAfricanes-tuariesandtheirimportancetosheries.
SouthAfricanJournalofMarineScience80:203–207.
Ward,T.
J.
,M.
A.
Vanderklift,A.
O.
Nicholls,andR.
A.
Kenchington.
1999.
Selectingmarinereservesusinghab-itatsandspeciesassemblagesassurrogatesforbiologicaldiversity.
EcologicalApplications9:691–698.
Warner,R.
R.
,S.
E.
Swearer,andJ.
E.
Caselle.
2000.
Larvalaccumulationandretention:implicationsforthedesignofmarinereservesandessentialshhabitat.
BulletinofMa-rineScience66:821–830.
Watling,L.
,andE.
A.
Norse.
1998.
Disturbanceoftheseabedbymobileshinggear:acomparisontoforestclearing.
ConservationBiology12:1180–1197.
Zedler,J.
B.
1984.
TheecologyofsouthernCalifornia'scoastalsaltmarshes:acommunityprole.
FishandWildlifeService,OfceofBiologicalSurveys,81-54.
U.
S.
FishandWildlifeService,Washington,D.
C.
,USA.