VegetationResponseafterRemovalofthe InvasiveCarpobrotusHybridComplex in Andalucía,Spain
Jaraandreu,EsperanzaManzano-Piedras, ignasiBartomeus,ElíasD.DanaandMontserratVilà
AbStRACt
Weevaluatedtheecologicalsuccess of themanualremovalof Carpobrotusspecies,aputativehybridcomplexof aSouth Africanperennialmat-formingplant,bycomparingtreated,noninvaded,andinvadedplotsacrosscoastalAndalucíain southern Spain.Asameasureofthemanagementeffectiveness,we quantifiedthedensityofCarpobrotusseedlingsand resproutsintreatedplotsoneyearaftertreatment.Responseoftheplantcommunitytoremovalwasassessedbycom- paringnativespeciesrichness,cover,diversity,andspeciescompositionamongtreatments.Removalgreatlyreducedto agreatextentCarpobrotusdensity.However,successfulcontrolwillrequirerepeatedhand-pullingtreatments.Treated plotshadasignificantincreaseinspecies richness,especiallyannualplants,comparedtoinvadedplots,butbothhad thesamenativeplantcoveranddiversity.Wefoundsimilarspeciescompositionbetweenremovalandnoninvaded plots, indicatingthatrevegetationisnotnecessary.Long-termmonitoring isnecessarytodeterminewhethertheseobserved patternsofcommunityresponsearetransientorstablethroughsuccession.
Keywords:alienplant,coastaldunes,management,plantdiversity,Spain
iven the potential negative impact of non-native plants
onnativespecies,ecosystems,and
humanhealthandeconomies(Mack etal.2000,McKinneyandLockwood
1999,McNeely2001,Pimenteletal.
2005),controlofinvasivespecieshas becomeanimportantchallengefor landmanagers,aswellasacommon componentofrestoration efforts (Zavaleta etal.2001,Smithetal.
2006).Sinceremovinginvasivespecies requirestremendoustimeandeffort, thepotentialcostsandbenefitsof managinginvadersshouldbeassessed. Suchevaluationsneedtoincludemea- suresofplantcommunityresponse, notjustsuch factorsasfundingand degreeofcommunitycommitment.
TheinvasiveSouthAfricansuccu- lentgenuscarpobrotus(Aizoaceae)is anexampleofanon-nativeplantthat
oftendominatesplantcommunities inmanyMediterraneanregionsofthe world(D’AntonioandMahall1991, D’Antonio1993,D’Antonioetal.
1993,Vilàetal.2006).carpobrotus speciesarecrawling,mat-forming,suc- culentchamephytes(plantswithbuds neargroundlevel,Raunkier1977) easilyrecognized by theirsucculent, finger-shaped,and triangular-section leaves(D’Antonio1990,D’Antonio
1993). Introgressive hybridization isverycommon(Vilàetal.1998), occurringthroughoutcoastalCalifor- niabetweenthenon-nativehottentot fig(carpobrotusedulis)and the puta- tivenativeseafig(c.chilensis),lead- ingtoahighabundanceofinvasive
hybridmorphotypes thatcompete aggressivelywithnativeplantcoastal species(D’Antonio1990,Albertet al.1997, VilàandD’Antonio1998).
carpobrotus formslarge matson coastalrocks,cliffs,andsanddunes owingtoitsprofuse clonalgrowthand long-distancedispersalbyvertebrates (D’Antonio1990,D’Antonio1993, Travesetetal.2008).Thefleshyfruits bearalargenumber,oftenoverathou- sand,ofsmallseeds(Bartomeus and Vilà2009)thatareeatenandwidely dispersedbyseveralmammals such asrabbits(D’Antonio1990)andrats (Bourgeoisetal.2005).carpobrotus hasalong-livedseedbankthatcan remainviableinthesoilforatleasttwo years(D’Antonio1990).carpobrotus isabletogrowfrommultipleaxes, rootingwherenodescontactthesoil, andspreadsradiallyatrates ashigh asonemeterperyear(Wisuraand Glen1993).Itcompetesaggressively withnativeplantspecies,achieving highratesofspacecolonization,which
InSpain,carpobrotusspecies,known
suppressesthegrowthandestablish-
.
locallyasuñadegato,maybehybrids betweenhottentotfigandSally-my- handsome(c. acinaciformis).
mentofotherplants(D’Antonioand
Mahall1991,Albert1995,Suehsetal.
2004,Vilàetal.2006).Furthermore, italsointeractsindirectlywithnative
vegetationbyalteringsoilchemistry
(ConserandConnor2009).
Itwasfirstintroducedasanorna- mentalplantintoEuropearoundthe
17thcenturyattheLeydenBotanical
Garden,theNetherlands,andsince thenithasbeencultivatedinother Europeanbotanicalgardens(Fournier
1952).However, theprogressive expansionandnaturalizationinthe MediterraneanBasinstartedinthe beginningofthe20thcentury(Sanz- Elorzaetal.2004).Nowadays,itis considered invasiveinEurope(Alba- nia,France,southernUK,Portugal,
Italy,Greece,Montenegro,andthe CanariesandotherMediterranean islands),northern Africa,southern Australia,NewZealand,andUSA (CaliforniaandFlorida) (Sanz-Elorza etal.2004).
InSpain,itwasintroducedinten-
tionallyforgardening, landscaping, anddunestabilizationinthebegin- ningofthe20thcentury,owingtoits
Figure1.sixexperimentalsiteswhereCarpobrotuswasremovedandmonitoredinandalucía (southernspain): a)paradorMazagón(huelva); b)puntacamarinal(cádiz);c)torrelapeña (cádiz);d)artola1and2(Málaga);ande)puntaentinas(almería).
fastclonalgrowth,lowwaterrequire-
ments,andshowy,largeflowers(Sanz- Elorzaetal.2004).Owingtothelarge extentofinvadedareasincoastalcom- munities,itisoneofthemostcostly invasivespeciesinSpain (Andreuet al.2009).Weconducted aregional surveytotesttheshort-termeffec- tivenessofcarpobrotus removaland nativevegetationresponseincoastal sanddunesacrossAndalucía(southern Spain).Inparticular,weaddressedthe followingtwoquestions:1)Hascar- pobrotusbeensuccessfullycontrolled inourstudysitesoneyearaftertreat- ment?2)Arenativespeciesrichness,
cover,diversity, andcomposition aftercarpobrotus removalsimilarto referencecommunities?
StudySitesExperimentalDesign
The studywasconducted insixcoastal localitiesinfourprovincesofAnda- lucía where carpobrotushadbeen removedtheyearbeforeaspartof theAndalusianPlanforControlof InvasiveSpecies(OrtegaAlegreand Ceballos2006)(Figure1).Overall,
300hahadbeentreatedbyhand- pullingcarpobrotus. Atotalof400
Tofplantmaterialwastransferred tocompostareas.Thisplantisread- ilyclonedbyrootingstemfragments thatcontain justonenode;thusitwas crucialtofullyremoveall individuals andalsoanyburiedstems(D’Antonio
1990).
Theselocalitiesprovidedarepre- sentativesampleoftheentireAndalu- siancoastandthetypicalhabitattypes wherecarpobrotusinvadesworldwide. Thevegetationinthestudysitesis typicalMediterranean coastalshrub- landsdominatedbychamephytes.The mainspeciesareEuropean beachgrass (ammophilaarenaria,Poaceae),mastic tree(Pistacialentiscus,Anacardiaceae), salviacistus(cistussalviifolius,Cista- ceae),silene nicaeensis (Caryophylla- ceae),sweetalyssum(lobularia mari- tima,Brassicaceae),silverseastock (Malcolmia littorea,Brassicaceae), Echium gaditanum(Boraginaceae),and anacyclusradiatus(Asteraceae).The climateisMediterranean,withwarm drysummersandmild,wetwinters.
Inordertodetermineeffectiveness andvegetationresponsestoremoval
treatments,weestablishedthefol- lowingthreetreatments:1)control plots(n=18)wherecarpobrotuswas notpresentandtherewasnohistory ofinvasion andvegetationremoval management;2)invadedplots(n=
13)withhighcarpobrotuscover(circa
70%)andnohistoryofremoval;and
3)treatedplots(n=46)wherecar- pobrotus wasremoved.Thedistance betweenplotswithinalocalityranged from10mto50m.Thisapproach allowedusto distinguishchanges caused byremoval,sinceobserving treatmentplotsovertimemaynot allow differentiation oftreatment effectsfromchangesduetonatural fluxes(Swabetal.2008).Table1pro- videsdetailsforallsixlocalitiesabout location,habitat,mostcommonplant species,andnumberandsizeofplots foreachtreatment.
Thelocation,number,andsizeof experimentalplotsweredecidedwith theaid oflandmanagersoneyearafter carpobrotus removal.Owingtothe idiosyncrasiesoftheremoval treat-
ments,causedmainlybydifferences inthesizeandspatialpositionof carpobrotus patchesandvegetation
table1.siteinformationforallsixlocalitiesinanexperimenttocontrolCarpobrotusincoastalvegetationof andalucía,spain.Raunkierlife-formofeachspeciesisindicated inparenthesis:p= phanerophyte,t= therophyte, c= chamephyte,G= Geophyte, h= hemicryptophyte.
Number ofplots(size,inm)
Locality(province) / habitat / Nativespecies / Invaded / Noninvaded
(control) / treated
PuntaCamarinal
(Cádiz) / Rockshores,sea-cliffs, andstabledunes / Anacyclusclavatus(T) Armeriapungens(H) Cyperuscapitatus(G) Euphorbiaterracina(C) / 5(10× 10) / 0 / 5(10× 10)
Torrelapeña
(Cádiz) / Shiftingandstable dunes / Ammophilaarenaria(H) Lotuscreticus(C) Malcolmialittorea(C) Medicagolittoralis(T) / 5(5× 5) / 0 / 5(5× 5)
Artola1 (Málaga) / Shiftingandstable dunes / Helichrysumstoechas(C)
Lotuscreticus(C) Ononisnatrix (C) Pistacialentiscus(P) / 0 / 5(5× 5) / 5(5× 5)
Artola2 (Málaga) / Shiftingdunes / Cynodondactylon(H) Lotuscreticus(C) Medicagolittoralis(T)
Pancratiummaritimum(G) / 3(5× 5) / 3(5× 5) / 3(5× 5)
Mazagón
(Huelva) / Stabledunes / Cistussalviifolius(P) Medicagolittoralis(T) Paronychiaargentea(C) Rumextingitanus(C) / 0 / 5(5× 5) / 10(2.5× 2.5)
PuntaEntinas
(Almería) / Stabledunes / Cyperuscapitatus(G) Helichrysumstoechas(C) Phagnalonsaxatile(C) Reichardiatingitana(T) / 0 / 5(2× 10) / 5(2× 10)
structure,plotsizeswerenotidentical betweensites.Plotsizesrangedfrom
2.5×2.5m(6.25m2)to10×10m
(100m2).
Asameasureofthemanagement effectiveness,density ofcarpobro- tusseedlingsorresprouts(hereafter recruits)wasdeterminedwithineach plotasthenumberofrecruitsper squaremeter.carpobrotusrecruitswere classifieddepending ontheirnumber
ofleavesin4differentcategoriesbased onincreasingbranchingpatterns(<10 leaves, 10–24 leaves, 25–49 leaves, and>50leaves).Recruitswithfewer than10leaveswereprobablyseed- lings,andthosewithmorethan50 leaveswereprobablyremnantsleftin placeunintentionallywhenremoving thespecies.
Vegetationresponsetocarpobrotus removalwasmeasuredbythepoint- interceptmethod.Plantsurveyswere carriedoutat20cmintervalsalong
theperimeteroftheexperimentalplot
byrecordingallplantspeciescontact- ingan imaginaryverticallineateach intervalpoint.Ineachplot,weidenti- fiedallspeciesat300–1,000points, dependingonthesizeoftheplot.This methodprovidedarecordofspecies compositionandabundanceineach plot,ameasureofplantcover,and, indirectly,anestimationofspecies richnessanddiversity.
Mosttaxawereidentifiedtothespe- cieslevel,withgrasses(Poaceae)being pooled,andthenlabeledasnativeor non-nativeandassignedtooneof thefiveRaunkier(1977)plantlife- forms.Thisplantclassificationsystem isbasedonthepositionofperennating budsinrelationtothesoilsurface: chamephytes,geophytes(survivalvia underground food-storage organs suchasrhizomes,tubers,orbulbs), hemicryptophytes(perennatingbuds atgroundlevelandaerialshootsdie- back),phanerophytes(perennating budsorshootapicesonaerialshoots)
andtherophytes(survivalasseeds—
annualorephemeralplants).
We calculatedtherelativecoverof individualplantspeciesasthepro- portionofcontactsofeachspecies relativetothetotalnumberof plant contactspertransect,followedbythe relativecoverofnon-nativeplantsand eachRaunkierlife-form.Totalnative vegetationcoverwascalculatedasthe
totalnumberofcontactsofallnative speciesinrelationtototalnumberof contactsincludingbareground,car-
pobrotus andothernon-nativespecies, ifpresent.Wemeasuredthespecies richnessofnativesandnon-natives. Nativediversitywascalculatedusing theShannon–Wienerdiversityindex (H),whichissensitivetorarespecies. Alloftheseresponsevariableswere comparedbetweencontrol,invaded, andtreatedplots.
Inordertodetermine whether, withinasite, treatedplotsexhib- itedthesamespeciescomposition
ascontrolplots,wecalculatedthe Sorensen SimilarityIndex(s).This indexranges between1(samespe- ciescomposition) and0(mostvaried speciescomposition).
DataAnalysis
Asdatadidnotfitparametrictest assumptions,differencesintheden- sityofthedifferentleafcategoriesof carpobrotus recruitswereanalyzed withKruskal-Wallistests.Wealso performed amultiplecomparison testafterKruskal-Wallistoassessdif- ferencesinleafcategoriesusingthe pgirmesspackageandtheKruskalmc functionofR(vers.2.6.2,RFoun- dationforStatistical Computing, Vienna,Austria).
Plotsizesdifferedbetweensites.
Figure2.Mean(±se)Carpobrotusrecruit densityoneyearafterremovalfromsixcoastalsites inandalucía,spain,grouped bythenumber ofleavesperrecruit. Differentlettersindicate significantlydifferentvalues(H=8.74,df=3,p=0.032).
Sincespeciesrichnessisdependent onthenumberofspecimens counted and,therefore,onsamplesize,we transformedobservedspeciesnum- berstoexpectedvaluesforagiven samplingsizebyrarefaction curves (Sanders1968,Hurlbert1971,Gotelli andColwell2001)usingtherarefy functionintheveganpackageofR. Rarefactioncurvesstandardizedthe samplingineachoftheplotstothe minimum samplesize(n =300), whichpermitsustouserarefieddata tomakespeciesrichnessanddiver- sitycomparisons(GotelliandColwell
2001).Allanalyseswerewithrarefied data.
Someresponsevariablescouldnot benormalizedwithdatatransforma- tion.Therefore,differencesinnative speciescover,richness,diversity,and
functionalgroupcoverbetweentreat- ments(invaded,control,andtreated) weretestedwithageneralizedlinear mixedmodel(GLMM)withaPois- sondistributionoferrorsandalogit
linkfunction(Crawley2002).The logitlinkfunctionensuresthatallthe fittedvaluesarepositive,whilePoisson errorsarerecommendedtodealwith integer(count)variables,whichare oftenright-skewedandhavevariances thatareequaltotheirmeans(Crawley
2002).Treatmentwasincludedasthe
fixedfactor,andsitewasincludedas arandomfactortoaccountforspa- tialautocorrelation.Modelswererun usingtheglmmPQLfunctionofthe MASSpackageinR.Wealsocalcu- latedthepowerofouranalysis(β)to assesstheprobabilitiesofTypeIIerror, givenoursmallsamplesize.Inorderto test if carpobrotusremovalincreased colonizationbyothernon-nativespe- cies,wecomparednon-native species coveramongtreated,control,and invaded plotsusingtheKruskal-Wallis test.Meanvalues±standarderrorsare giventhroughout thetext.
ResultsandDiscussion
Effectivenessofcarpobrotus
Removal
Recruitdensityintreatedplotsacross sitesaveraged0.13±0.09recruitsper squaremeter.Noreestablishmentof carpobrotusoccurredin52%ofthe treatedplots.Oneofthesampledsites (PuntaCamarinal) accountedformost oftheobservedcarpobrotus recruits (63%).Recruitswith fewerthanten leavesweresignificantlymoreabun- dantthanrecruitswithatleast10 leaves(Figure2).
Low densities of carpobrotus
recruits one yearafter treatment
indicated thatshort-termmanage- menthadconsiderablyreducedcar- pobrotus presence,althoughithadnot eradicatedthespecies.Recruitswith fewerthan10leaves,probablyseed- lings, werethemostabundant,which suggesttheimportanceoftheseed bank inthereestablishmentcapacity ofcarpobrotus.
Native PlantSpeciesCover, Richness,andDiversity
TheGLMMmodelrevealedsignifi- cantlyhighervaluesofrarefiedspecies richnessintreatedplots(7.20±0.40) thanininvadedplots(6.64±0.29; Figure3a),indicatingthatcarpobro- tusmayhaveanimpactonspecies richnessby replacingnativespeciesin thecommunitiesitinvades(Brandon etal.2004,HejdaandPyšek2006, HulmeandBremner2006).However, wefoundnosignificantdifferencesin totalnativespeciescover(Figure3b) anddiversity(Figure3c).Theseresults areconsistentwithothercasestudies (OgdenandRejmánek2005,Vidra etal.2007,Swabetal.2008,Pavlovic atal.2009)andprobablyaredueto alowabundanceofnewrecruitsand theshort-termscaleofourstudy.Nei- therdidwefindsignificantdifferences betweentreatedandcontrolplotswith respecttoplantcover,richness,and
Figure3.BoxplotsofnativeplantresponsetoCarpobrotusremovalininvaded,control,and treated plotsinsixcoastalsitesinandalucía,spain,measuredbya)rarefied speciesrichness; b)total nativespeciescover;andc)shannon-Wienerspeciesdiversity. theboxitselfcontains
50%ofthedata(75th percentile indicated bytheupperedge,themedianbythecenterline,
andthe25thpercentile bytheloweredge),with outliersasopencirclesandmaximum/minimum valueattheterminus oftheverticalline.
diversity,indicatingthat regeneration aftercarpobrotus removalresultsin coastaldunecommunitiessimilarto referencenativecommunities.None- theless,wehavetobecautiouswhen interpretingtheseresults,mostlyin thecaseoftotalnativespeciescover betweentreatedandinvadedplots (p=
0.07);thelimitedpowerofourstatis- ticalanalysis(β=0.67)couldprevent usfromdetectingpossiblesignificant differencesamongtreatments.
Native SpeciesComposition
Despitethelackofchangesintotal nativespeciescoveranddiversity,there weresomechangesinspecies compo- sition.TheSorensenSimilarityIndex betweencontrolandtreatedplotswas, onaverage,0.77±0.034,whichpro- videsadditionalsupportfortheidea thatregenerationaftercarpobrotus removalresultsincoastaldunecom- munitiessimilartoreferencenative communities.
InFigure4,percentcoverofthedif- ferentRaunkierfunctionalgroupshave beencomparedbetweentreatedand controlplots(Figure4a)andbetween treatedandinvadedplots(Figure4b). Onlytwoofthefivefunctionalgroups respondedsignificantlytocarpobrotus removal. Coveroftherophyteswas significantlygreaterintreatedplots thanincontrol and invadedplots.This observedincrease inannualplants suggeststhatthecoastaldunesthat weretreatedareinanearlysucces- sionalstage.Otherstudieshaveshown responsesofannualplantsfollowing removalofinvasivespecies(McCarthy
1997,CarlsonandGorchov2004, CrimminsandMcPherson 2008), whichincreaseslight,soiltempera- ture,andresourceavailability,favor- ingthegerminationofspeciesinthe seedbank,suchasannuals(D’Antonio andMeyerson 2002).However,cover ofchamephytes(excludingcarpobro- tus)waslowerintreatedplotsthanin control plots,andnosignificantdif- ferenceswerefoundbetweentreated
andinvadedplots(Figure4a,also seeonlineappendixatuwpress.wisc
.edu/journals/journals/er_suppl.html).
Thiscanbeexplained bythefactthat chamephytesgrowmore slowlythan therophytesandneedmoretimeto reestablish.Asnosignificantdiffer- encesinotherfunctionalgroupswere foundbetweentreatedandcontrol plots,weexpectthatnaturalcom- munitydynamicswill leadthemto becomemorematurecommunities withamorehomogeneousrelative coverofdifferentlifeforms.
Withouttakingintoaccountgrami- noids,atotalof107specieswere foundinourplots.Ofthese,63were neverfoundininvadedplots,27of whichappearedonlyintreatedplots and11onlyincontrolplots.Another
43speciesoutofthe107werenot presentincontrolplots,andonly11
specieswereneverfoundintreated plots,ofwhich7werepresentonlyin controlplots.
Therelativecoverofparticular speciesdifferedbetweentreatments (Figure5).Forexample, water medick(Medicagolittoralis),cype- ruscapitatus, buckhorn plantain (Plantagocoronopus),andwhitebut- tons(anacyclusclavatus)werepoorly represented incontrolplotsand appearedveryfrequentlyintreated plots(Figure5a).
Onthecontrary,cretatrefoil(lotus creticus),curry plant(Helichrysum italicum),Helichrysum stoechas,and Rumex tingitanus havehigherrela- tivecoverinthecontrolthaninthe treatedplots(Figure5a).Infact,nei- therHelichrysum waseverfoundin invadedplots,probablybecausethey areassociatedwithstableandundis-
Figure4.Raunkierlife-formcoverasafunctionoftreated plotsfora)controlandb)invaded plots.proximitytothelineofunityindicateslackofdifference fromthetreated plotsforthe life-formgroup.
turbedsites.Therewerealsodiffer- encesbetweeninvadedandtreated plots.Forinstance,Malcolmialittorea, Geraldtoncarnationweed(Euphorbia terracina),andEngelsgras(armeria pungens) weremorerepresentedin invaded thantreatedplots,while cyperuscapitatuswasasabundantin invadedasintreatedplots(Figure5b).
Onlyfourof the107specieswere non-native.ThesewereAmericancen- tury plant(agaveamericana),Cape weed(arctothecacalendula),salthelio- trope (Heliotropiumcurassavicum),
andBermudabuttercup(oxalispes- caprae).Nosignificantdifferenceswere foundbetweenthetotalcoverofnon- nativespeciesintreated(4.78±3.32), control(4.08±2.25),andinvaded (8±6.33)plots(Kruskal-WallisH=
0.95,df=2,p=0.620). Manystud- ieshave documentedanincreasein undesirableinvasivespeciesfollow- ingdisturbances(BurkeandGrime
1996,Pickartetal.1998,Zavaletaet al.2001,MasonandFrench2007, CrimminsandMcPherson 2008). Suchsecondaryinvasionsfollowing
controleffortscanbeproblematicfor ecologicalrestoration(Hartmanand McCarthy2004,HulmeandBremner
2006).Thelackofsuch findingsinour short-termstudyis,therefore,encour- agingfromacommunitymanagement perspective.
Conclusionsand
ManagementImplications
Wheneradicatingcarpobrotus itis importanttoremoveanyremnants, asanyremainsleftin placesoon
Figure5.Individualnativespeciesrelativecoverasafunctionoftreated plotsfora)control andb)invadedplots.Raunkierlife-formcategoryisindicated inparentheses:t=therophyte, c=chamephyte, G=Geophyte, h=hemicryptophyte.
therefore, ourfindingsforthesecom- munitiesshould notyetberegarded asdefinitive, sincethemanaged sitesarestillinanearlysuccessional stage. Althoughrepeatedsampling isnecessaryto determinewhether anyobservedpatternofcommunity responseistransientorstable(Sax and Brown2000),thesefindingscan beappliedtoachievecost-effective removalstrategies that accomplish overallrestorationgoals.
Overall,ourresultsrevealedthat removalofcarpobrotusisnotfacilitat- inginvasionbynon-nativesandthat recoveryof nativespeciesishigh.This suggeststhatifseedsofnativespecies arepresent,naturalreestablishmentis possible.Naturalregenerationwould becheaperinthesecoastaldunecom-
munitiesthanseedingaftercarpobro- tusremoval.Thusalthoughplanting desirednativespecies isapotential schemetofacilitatethenativerecov- eryofacommunity,itisanexpensive methodandwedonotconsiderit necessaryin ourcase.
Acknowledgments
Wethank NúriaGassóandCorinaBasnou forstatisticaladviceandusefulcomments. WealsothankA.J. Pickartandtwoanony- mousreviewersforhelpfulcommentson apreviousversionofthemanuscript.This studyhasbeenpartiallyfundedbythe ConsejeríadeMedioAmbientedeAnda- lucía(NET852690), the6thFramework ProgrammeoftheEuropeanCommission’s ALARMproject(GOCE-CT-2003-506675; see
becomeanactivefocusofregenera-
tion(Fragaetal.2006),whichwas demonstratedbytheverylow densities oflargerecruits,probablyresprouts, oneyearafterpulling(Figure2).Our researchrevealedthathand-pulling greatlyreducedcarpobrotus;however, successfulcontrolwilllikelyrequire perseveranceandacommitmentto long-termplanning,implementation, andmonitoring(Pickartetal.1998, ManchesterandBullock2000).More- over,regionaleradicationwouldbe neededinordertopreventnewinva- sions fromneighboring populations (PickartandSawyer1998).
Our findingssuggestthatnative
speciescouldeasily establishafter carpobrotusremoval, particularly annualplants.However,withjust
oneyearofgrowth,thesespeciesare notabletooccupyallbareground (Díazet al.2003).Inaddition,com- parisonsbetweentreatedandcontrol plotsshowedthatmanagementhas resultedincoastalduneswithveg- etation similarto referencenative communities.
Somestudiessuggestthatnative speciesrecoveryafternon-nativespe- ciesremovalrequiresseveralyears.This couldalsobetrueforcarpobrotusand,
al.2005),andtheMONTES projectof
CONSOLIDER(CSD2008-00040).
References
Albert,M.E.1995. Morphologicalvaria- tionandhabitatassociationwithinthe carpobrotusspeciescomplexin coastal California.MSthesis.Universityof California,Berkeley.
Albert,M.E.,C.M.D’AntonioandK.A.
Schierenbeck.1997. Hybridization andintrogressionin carpobrotusspp. (Aizoaceae)in California:I. Morpho- logicalevidence.americanJournalof Botany84:896–904.
Andreu,J.,M.VilàandP.E.Hulme.
2009. Anassessmentofstakeholder
perceptionsandmanagementofnox- iousalienplantsin Spain.Environ- mentalManagement43:1244–1255.
Bartomeus,I. andM.Vilà.2009. Breed- ingsystemandpollenlimitation
oftwosupergeneralistalienplants invadingMediterraneanshrublands. australianJournalofBotany57:1–8.
Bourgeois,K.,C.M.Suehs,E.Vidaland
F.Médail.2005. Invasionalmeltdown potential:Facilitationbetween intro- ducedplantsandmammalsonFrench Mediterraneanislands.Ecoscience12:
248–256.
Brandon,A.L.,D.J. GibsonandB.A.
Middleton.2004. Mechanismsfor dominancein anearlysuccessionalold fieldbytheinvasivenon-native lespe- dezacuneata(Dum.Cours.)G.Don. Biologicalinvasions6:483–493.
Burke,M.J.W.andJ.P.Grime.1996. An experimentalstudyofplantcommu- nityinvasibility.Ecology77:776–790.
Carlson,A.M.andD.L. Gorchov.2004.
Effectsofherbicideontheinvasive biennialalliariapetiolata(garlicmus- tard)andinitialresponsesofnative plantsin asouthwesternOhioforest. Restoration Ecology12:559–567.
Conser,C.andE.F.Connor.2009. Assess- ingtheresidualeffectsofcarpo-
brotusedulisinvasion, implications forrestoration.Biologicalinvasions
11:349–358.
Crawley,M.J.2002. statisticalcomputing: anintroduction toDataanalysisUsing s-Plus.ChichesterUK:Wiley.
Crimmins,T.M.andG.R. McPher-
son.2008.Vegetationandseedbank responsetoEragrostislehmanniana removalin semi-desertcommunities. WeedResearch48:542–551.
D’Antonio,C.M.1990. Seedproduc- tionanddispersalin thenon-native, invasivesucculentcarpobrotus edulis (Aizoaceae)in coastalstrandcommu- nitiesofcentralCalifornia. Journalof appliedEcology27:693–702.
. 1993. Mechanismscontrollinginva- sion ofcoastalplantcommunitiesby thealiensucculentcarpobrotusedulis. Ecology74:83–95.
D’Antonio,C.M.andB.E.Mahall.
1991. Rootprofilesandcompetition betweentheinvasive,exoticperennial, carpobrotusedulis,andtwonative shrubspeciesin Californiacoastal scrub.americanJournalofBotany
78:885–894.
D’Antonio,C.M.andL.A.Meyerson.
2002. Exoticplantspeciesasproblems andsolutionsin ecologicalrestoration:
Asynthesis.Restoration Ecology10:
703–713.
D’Antonio,C.M.,D.C.OdionandC.M.Tyler.1993. Invasionofmari-time chaparralbytheintroducedsuc- culentcarpobrotusedulis.oecologia
95:14–21.
Díaz,S.,A.J.Symstad,F.S.Chapin, D.A.
WardleandL.F.Huenneke.2003. Functionaldiversityrevealedby removalexperiments.trends in Ecology
Evolution18:140–146.
Fournier,P.1952. Dicotylédones.Vol.2 of Floreillustréedes JardinsetdesParcs: arbres,arbustes etFleursdePleine terre.Paris:P.Lechevalier.
Fraga,P.,I. Estaun,J.Olives,G.Da Cunha,A.Alarconetal.2006. Erad- icationofcarpobrotus(L.)N.E.Br.
in Minorca.IUCNSpeciesSurvival
Commission casestudyreport. Gotelli,N.J.andR.K.Colwell.2001.
Quantifyingbiodiversity:Procedures andpitfallsin themeasurementand comparisonofspeciesrichness.Ecology letters4:379–391.
Hartman,K.M.andB.C.McCarthy.
2004. Restorationofaforestunder- storyaftertheremovalofaninva- siveshrub,Amurhoneysuckle
(loniceramaackii).RestorationEcology
12:154–165.
Hejda,M.andP.Pyšek.2006.What is theimpactofimpatiens glandulifera onspeciesdiversityofinvadedripar-
ianvegetation? Biologicalconservation
132:143–152.
Hulme,P.E.andE.T.Bremner.2006.
Assessingtheimpactofimpatiens glan- duliferaonriparianhabitats:Parti- tioningdiversitycomponentsfollow- ingspeciesremoval.Journalofapplied Ecology43:43–50.
Hurlbert,S.H. 1971. Nonconceptofspe- ciesdiversity:Acritiqueandalterna- tiveparameters.Ecology52:577–586.
Mack,R.N.,D.Simberloff,W.M.Lons- dale,H.Evans,M.CloutandF.A. Bazzaz.2000. Bioticinvasions: Causes,epidemiology,globalcon- sequencesandcontrol.Ecological applications10:689–710.
Manchester,S.J.andJ.M.Bullock.2000.
The impactsofnon-nativespecieson UKbiodiversityandtheeffectiveness ofcontrol.JournalofappliedEcology
37:845–864.
Mason,T.J.andK.French.2007. Man- agementregimesforaplantinvader differentiallyimpactresidentcom- munities.Biologicalconservation
136:246–259.
McCarthy,B.C.1997. Responseofa forestunderstorycommunityto experimentalremovalofaninvasive nonindigenousplant(alliaria petio- lata,Brassicaceae).Pages117–130 in J.O.LukenandJ.W.Thieret(eds), assessment andManagement ofPlant invasions. NewYork:Springer-Verlag.
McKinney,M.L.andJ.L.Lockwood.
1999. Biotichomogenization: Afew winnersreplacingmanylosersin the nextmassextinction.trends in Ecology
Evolution14:450–453. McNeely,J.2001. Invasivespecies:A
costlycatastrophefornativebiodi- versity.landUseandWaterResources Research1(2):1–10.
Ogden,J.A.E.andM.Rejmánek.2005.
Recoveryofnativeplantcommuni- tiesafterthecontrolofadominant invasiveplantspecies,Foeniculumvul- gare:Implicationsformanagement. Biologicalconservation125:427–439.
OrtegaAlegre,F.andG.Ceballos.2006.
Control despeciesexóticasinvasoras: Actuaciones.Medioambiente54:30–
39. ambiente/contenidoExterno/Pub_ revistama/revista_ma54/ma54_32
.html
Pavlovic,N.B.,S.A.Leicht-Young,K.J.
FrohnappleandR.Grundel.2009. EffectofremovalofHesperismatrona- lis(dame’srocket)onspeciescoverof forestunderstoryvegetationin NW Indiana.americanMidlandnaturalist
161:165–176.
Pickart,A.J., L.M.MillerandT.E.Due- bendorfer.1998.Yellowbushlupine invasionin northern California coastaldunes:I. Ecologicalimpacts andmanualrestorationtechniques. Restoration Ecology6:59–68.
Pickart,A.J.andJ.O.Sawyer.1998. Ecol- ogyandRestorationofnorthern cal- iforniacoastalDunes.Sacramento: CaliforniaNativePlantSociety.
Pimentel,D.,R.ZunigaandD.Morri- son.2005. Updateontheenviron- mentalandeconomiccostsassoci- atedwithalien-invasivespeciesin the United States. EcologicalEconomics
52:273–288.
Raunkier,C.1977. Biologicaltypeswith referencetotheadaptationofplantsto survivetheunfavorableseason.Chap- ter1 in F.N.Egerton(ed),lifeForms
ofPlantsandstatisticalPlantEcology. Reprintofthe1934 Englishtransla- tionfromtheoriginal1904 article(in Danish).NewYork:ArnoPress.
Sanders,H.L.1968. Marinebenthicdiver- sity:Acomparative study.american naturalist102:243–282.
Sanz-Elorza,M.,E.D.Dana andE.
Sobrino.2004. Atlasdelasplan-tas alóctonasinvasorasenEspaña. Madrid:DirecciónGeneralparala Biodiversidad, MinisteriodeMedio Ambiente.
Sax,D.F.andJ.H.Brown.2000. Thepar- adoxofinvasion.globalEcologyBiogeography9:363–371.
Settele,J.,V.Hammen,P.Hulme,U.Karlson,S.Klotzetal.2005. ALARM:AssessingLArge-scaleenvi- ronmentalRisksforbiodiversitywith testedMethods.gaia14:69–72.
Smith,R.G., B.D.Maxwell,F.D.
MenalledandL.J.Rew.2006. Les- sonsfromagriculturemayimprove themanagementofinvasiveplantsin wildland systems.Frontiersin Ecology andtheEnvironment4:428–434.
Suehs,C.M.,F.MédailandL.Affre.2004.
Invasiondynamicsoftwoaliencarpo- brotus(Aizoaceae)taxaonaMediter- raneanisland:I. Geneticdiversityand introgression.Heredity92:31–40.
Swab,R.M., L.ZhangandW.J.Mitsch.
2008. Effectofhydrologicrestora- tionandloniceramaackiiremoval onherbaceousunderstoryvegeta-
tionin abottomland hardwoodforest.
Restoration Ecology16:453–463.
Traveset,A.,G.Brundu,L.Carta,
I. Mprezetou,P.Lambdonetal.2008. Consistentperformanceofinvasive specieswithinandamongislandsof theMediterraneanbasin.Biological invasions10:847–858.
Vidra,R.L.,T.H.ShearandJ.M.Stucky.
2007. Effectsofvegetationremoval onnativeunderstoryrecoveryin an exotic-richurbanforest.Journalofthe torreyBotanicalsociety134:410–419.
Vilà,M.andC.M.D’Antonio.1998.
Hybridvigorforclonalgrowthin carpobrotus(Aizoaceae)in costal California.Ecologicalapplications
8:1196–1205.
Vilà,M.,M.Tessier,C.M.Suehs,
G.Brundu,L.Cartaetal.2006. Local andregionalassessmentoftheimpacts ofplantinvadersonvegetationstruc- tureandsoilpropertiesofMediterra- neanislands.JournalofBiogeography
33:853–861.
Vilà,M.,E.WeberandC.M.D’Antonio.
1998. Floweringandmatingsystemin hybridizingcarpobrotus(Aizoaceae)in coastalCalifornia.canadianJournalof Botany76:1165–1169.
Wisura,W.andH.F.Glen.1993. The SouthAfricanspeciesofcarpobro- tus(Mesembryanthema-Aizoaceae). contributions totheBolusHerbarium
15:76–107.
Zavaleta,E.S.,R.J.HobbsandH.A.
Mooney.2001.Viewinginvasive
speciesremovalin awhole-ecosystem context.trends in EcologyEvolution
16:454–459.
Jaraandreu,centreforEcological Research andForestryapplications,Uni- versitatautònomadeBarcelona,E-08193
Bellaterra,Barcelona,catalonia,spain,
34-954232340,Fax:34-954621125,
EsperanzaManzano-Piedras, Estación BiológicadeDoñana(EBD-csic),avda. américoVespucio,s/n,isladelacartuja,
41092 sevilla,spain
ignasiBartomeus,centreforEcological ResearchandForestryapplications,Uni- versitatautònomadeBarcelona,E-08193
Bellaterra,Barcelona,catalonia,spain.
ElíasD.Dana,Programaandaluzparael controldeEspeciesinvasoras,EgMasa– consejeríadeMedioambiente,Junta de andalucía,avda.américoVespucio,5,
isladelacartuja, 41092 sevilla,spain
MontserratVilà,EstaciónBiológicade Doñana(EBD-csic),avda.américo Vespucio,s/n,isladelacartuja, 41092 sevilla,spain