Amparo Carrillo-Gavila N Josepmaria Espelta
Establishmentconstraintsofanalienandanativeconifer indifferent habitats
Amparo Carrillo-Gavila´n• JosepMaria Espelta •
MontserratVila`
Abstract Alien plants are subjected to different bioticandenvironmental barriersthatlimittheir establishmentsuccessinthe introducedrange.Pseud- otsugamenziesii(Douglasfir),anativeconiferfrom Northwest America,isconsidered oneofthemost invasiveforestryconifersinEurope.However,littleis knownabouttheecologicalfiltersthatconstrainplant establishmentatearlylife-cyclestagesanddifferences inhabitatinvasibilitytothisspecies.Weconducted field experimentstocomparetheestablishment potential(i.e.post-dispersal seedremoval,seedger- mination,seedlingsurvivalandgrowth)ofDouglasfir inbeechforests,holm-oakforestsandheathlands;and compared itwiththetaxonomicallyclosenative conifer Abiesalba(Silverfir).Douglasfirseedswere moreremovedthanSilverfir inholm-oakandin heathlands.Inallhabitats, seedgerminationwas significantlyhigherforDouglasfircomparedtothatof Silverfir and,seedlingmortalitywasextremelyhigh inbothspeciesduetosoildisturbance bywildboars anddroughtstress.Douglasfirmortalitywasonly
A.Carrillo-Gavila´n() M.Vila`
Estacio´nBiolo´gicadeDon˜ana(EBD-CSIC), Avda.Ame´ricoVespucios/n,IsladelaCartuja,
41092Sevilla,Spain
e-mail:
J.M.Espelta
CentredeRecercaEcolo`gicaiAplicacionsForestals
(CREAF),Edifici C,CampusdeBellaterra(UAB),
08193CerdanyoladelValle`s,Barcelona,Spain
lowerthanSilverfir inbeechforests.However, speciesdidnotdifferinseedlinggrowth.Overall, theprobability ofinvasionsuccessofDouglas fir decreased alongthesequentialstagesofplantestab- lishmentinallhabitats. Onlyhighseedgermination ratesofDouglasfirwouldpredictitshighinvasive capacitybutthese advantagesarecounterbalanced by highseedling mortality.Resultsshowedamismatch betweeninvasibilityandcurrentpatternofDouglasfir invasion inthestudyarea.Therefore,futureresearch focused onseedproductionandondifferentcompo- nentsofbioticresistanceisrecommendedtoeluci- datewhichprocesses arefavoringitsestablishment success.
Keywords Abiesalba Germination Invasibility
Pseudotsugamenziesii Seedremoval
Seedling survival
Introduction
Alienplantsaresubjectedtodifferent bioticand environmentalbarriersthatlimittheirestablishment successintheintroduced range(Lonsdale1999). Manyofthesebarriershaveaneffectonearlylife- cyclestages.Forexample,bylimitingthesupply of seedsduetoseedconsumption bynativefauna (McCay andMcCay2009), oralso,bycontrolling seedling establishment either caused by unfavor- able climatic conditions (Lambrinos 2002) or by
competitionwithnativeplants(Dietzetal.1999).Asin nativeplants,theseecologicalbarriers varyspatially andtemporally(Schupp 1988;Vila`andLloret2000; Travesetetal.2003).Forinstance,post-dispersalseed removal variesamong habitats,whichcanbemore intenseinareasfarfromalienconiferplantationsthan inothersclosetothem(Nun˜ezetal.2008).
Therearestudiesthathavecompareddifferencesin theestablishment traitsbetweenalienandnative congenersin the introducedrange(Shafrothet al.
1995; Lambrinos2002; FerrerasandGaletto 2010)to
betterunderstandtheinvasionpotentialofthealien species(i.e.invasivenesss),andtodeterminediffer- encesinhabitatsusceptibilitytoinvasion (i.e.invasi- bility)(PysˇekandRichardson2007).Forthispurpose, experiments arethebestapproachtodisentanglethe mechanismsofinvasion (Soletal.2008).Further- more, these experimentsneedtolinkthecausal relationships ofinvasion,fromseeddispersalto seedling establishment,intheintroduced range(Vila` etal.2006;Vila`andD’Antonio1998)andcompare themtothatofnativespecies.
Severalconiferspeciesareinvasive inmany regionsoftheworld,mainlyintheSouthernHemi- sphere(Richardson andRejma´nek2004).Therehas beenafairamount ofresearchonseedandseedling traitsconcerningconiferinvasiveness(Grotkoppetal.
2002;RichardsonandRejma´nek2004),butdiffer-
encesintheecologicalbarriersthatconstrain their establishment havebeenlessexplored,withthe exceptionofbioticresistancethroughherbivory by nativefauna(Lombarderoetal.2008;Nun˜ezetal.
2008; Carrillo-Gavila´n et al. 2010). Pseudotsuga
menziesiiMirb.Franco,theDouglasfir,isconsidered oneofthemostinvasiveforestryconiferspeciesofthe world(RichardsonandRejma´nek2004).Douglasfiris reportedasinvasiveinareasclosetoplantationsin NewZealand(Kay1994), South Africa(Richardson andHiggins1998),ArgentinaandChile(Simberloff etal.2010).Forinstance,Douglasfir invasionis facilitatedbydeerbrowsingonnativeplants(Relva etal.2009) andbythepresence ofbelow ground mutualisms (Nun˜ez et al. 2009) in Argentina. In Europe,itisnaturalizedinseveralcountries(Carrillo- Gavila´nandVila`2010).Accordingtopreviousstudies (RichardsonandBond1991;Sarasolaetal.2006),the mostsuitablehabitatstobeinvadedbythissortofalien coniferswouldbegrasslandsandshrublandsfollowed byopen forestsand,finally,byclosedforests.
ToassesshowDouglasfir establishmentismodu- latedbyitsinteractionwiththehostcommunity, we conductedfield manipulationexperimentsduringthe earlylife-cyclestages(i.e.post-dispersalseedremoval, seedsurvival,seedgermination,seedlingsurvivaland growth)andcomparedthemwithataxonomicallyclose nativeconiferAbiesalba (Silverfir)asabaseline species.Ourmainobjectiveswere:(a)toknowwhich habitats,closetoDouglasfir plantations,havethe highestinvasibility, (b)toassessthemainecological barrierslimitingDouglasfir establishment,(c)to determinehowdifferent these patternsarecompared toataxonomicallyclose,coexistingnativeconifer species.GiventhecurrentwidedistributionofDouglas fir plantationsinEurope,thisstudycancontributeto habitatriskassessmenttoinvasionwhenthisspeciesis introducedforafforestation.
Materialsandmethods
Studyarea
ThestudyareaislocatedintheMontsenyNaturalPark inBarcelona-Spain(latitude41°420–41°520N,longi- tude 2°160–2°330E), a 40,000ha mountainous area declaredaUNESCO BiosphereReservein1978. Montsenyencompassesawideclimaticgradient,from aMediterranean toasubalpineclimate,andconse- quentlyalargephytogeographicrange(Pen˜uelasand Boada2003).Atmediumandhighaltitudes(800–1,700 m.a.s.l.),themeanannualrainfallis1,148mmand mean annual temperature is 8.7°C (means for the
2007/10 period,CanLleonartmeteorological station). Thisareahoststhemostextensivesouthern European distributionareasofFagussylvatica(beech)forestsand thewesterndistributionofCallunavulgarisheathlands (Bolo`sandVigo1990).Also,MontaneQuercusilex (holm-oak)forestsoccuratthelowestaltitudinalranges (\800m.a.s.l.),withameanannualrainfallof943mm andmeanannualtemperatureof11.7°C(meansforthe
2007/10period,Fontmartinameteorologicalstation).
Theexperimentswereconductedinthefollowing habitattypes:beechforests,holm-oak forestsand heathlands(n=4sites/habitat).Wechose thesetypes ofhabitatbecauseoftheircloseproximitytoDouglasfir plantations andthereforepotentiallyvulnerableto invasion.Siteswere0.2–3kmapart.Thesehabitats differinvegetationstructureandlightavailability,and
seed remover community (Torre and Arrizabalaga
2009).Ineachhabitat,3plotsof10910mwere randomlychosen tomeasurethediameteratbreast height(DBH, cm)inallindividualtrees,andlight availabilitywasmeasured atthesoilsurfaceasphoto- syntheticallyactive radiation (PAR,lmols-1 m-2) usingaportablelightmeter(Sunfleck Ceptometer, Decagon,Pullman,USA).Themainseedremoverinall three habitats is Apodemussylvaticus,followedby A.flavicollis.Otherlessfrequentseedremoversare Myodesglareolus(beechandholm-oakforests),Mus spretus(holm-oak),Microtusagrestis(heathlands)and Glisglis(beech)intheMontseny (I.Torre,personal communication).
Studyspecies
Pseudotsugamenziesiivar.menziesii(Mirb.)Franco (Pinaceae)(Douglasfir, hereafter)isaconiferof NorthwesternAmericanoriginandaninvasivespecies inmanypartsoftheworldwhereithaspropagated fromplantations(Richardson andRejma´nek2004). ThefirstpublishedrecordsofDouglasfirintroductions inEuropewereintheCzechRepublic (1842), Germany (1900),Denmark andtheU.K(1940) (Carrillo-Gavila´nandVila`2010).InMontsenyNat- uralPark,itwasintroduced duringthe1950sand throughoutthefollowing decades,whenmarginal croplandsweretransformed totreeplantationsof mainlyfastgrowing alienconifers(Boada2000). Douglasfirwasoneofthemostplantedspeciesin Montsenywith230scatteredplantations (Boadaand Broncano2003)withatotalareaof250ha.Atpresent, Douglas firisconsideredanaturalizedspecies(sensu Pysˇeketal.2004;Hulme2011)inadjacentheathlands butnotinbeechorholm-oak forestsinMontseny, whicharealsoadjoinedtoplantations(Broncanoetal.
2005).According toBroncano etal.(2005),an incipientexpansionofDouglasfirisobservedinone ofthe230plantations,whichapproximatelycoversan areaof22 ha.Therefore,observationalsurveyssug- gestthatinvasibilitymightbelowintheseforesttypes.
AbiesalbaP.Mill(Silverfir,hereafter)isanative coniferinMontsenywhere itrepresentsthesouthern- mostEuropeandistribution siteforthisspecies.In Montseny,thepresence ofAbiesalbaisarelictsince thelastglaciations.According tohistorical records thisspecies wasmoreabundant inthepast,but intensive exploitation up to the beginning of the
twentiethcenturyreducedandfragmenteditsdistri- bution (Llobet1990). Inthepresentthereareonly threerelict naturalpopulationsofSilverfiroccupying atotalof28.5hasurroundedbybeechforestsand heathlands(BoadaandBroncano 2003),where recruitmentofseedlings isobservedintreefallgaps andareaswithbaresoil(Carrillo-Gavila´n,personal observation). AlthoughothernativePinaceaespecies exist,eitherinnaturalwoodlands orinplantations, such as Pinus pinaster, P. nigra spp salzmanii, P.pinea,P.sylvestrisandP.halepensis(Boadaand Broncano2003),wechosetocompareA.albawith Douglasfirbecausebothspeciesaregroupedinthe subfamilyAbietoideae.Furthermore, accordingto Wangetal.(2000)andBesendorferetal.(2005),the genusAbies ismorecloselyrelatedtoLarixand Pseudotsugathanto PiceaandPinus.Forthisreason, theyhavesimilar dispersalmechanisms,reproductive phenology and climatic requirements (Lo´pez- Gonza´lez2002). Forinstance,Douglas andSilverfir seedsareprimarilydispersed bywindandgravityin autumnwhileinsecond stage,dispersal couldbe promotedbymice,chipmunksorsquirrels(Hemstrom etal.1987;Wolf2003).
Toconductourstudy,wepurchasedseedsfromthe Intersemillas nursery,Spain( las.es)becausefieldcollectionattemptsinprevious yearsprovidedfewviableseeds.However,thiswasof littleconcernsinceourfocuswastocompare the potentialforseedlingestablishment betweenthetwo speciesunderarangeofdifferentecologicalcondi- tionsratherthandocumenting actuallocalratesof establishment.
Foreachspecies,20seedswererandomlychosento measure freshseedmass.Coatstrength wasalso estimatedfollowingRogerdson’s(1998)protocolwith aChatillonUniversalForceTester(Amtek/Chatillon, Largo,Florida,USA).
Seedremoval
Totestdifferences inpost-dispersal seedremoval (seedremoval,hereafter)inAutumn2008, wedelim- iteda1009100mplotateachsitewherewe placed five50mlongtransects.Ineachtransect,weplaced
20seedsofeachspeciesrandomlychosenfromapool ofseeds. Eachseedwasgluedtoapieceofnylon fishingline,tiedtoawirestake(Schupp1988)and placedonthegroundsurfacekeepingadistanceof
approximately3mbetweeneachother.Atotalof
1,200seeds(50seeds92species93habitats94 sites)weresurveyedforseedremovalafter2,6,13,20 and27days.Sinceconiferseedsaremoresuccessfully dispersedthroughtheirprimarydispersal—wind and gravity—(Hemstrometal.1987;Wolf2003),andless sothroughsecondarydispersal(Ordo´n˜ezandRetana
2004),weconsidered eithermissingseedsorthe presenceofseedcoatremnantsasevidenceofseed removal.Wefinishedtheexperimentwhenaftertwo consecutivevisitstheseedremovalcurves didnot vary.Afterthatdayallremainingseedswereremoved fromthefield.
Weconductedtwoanalyses.First,wetestedfor
differencesinpercentageof removedseedsattheend oftheexperimentamonghabitatsandbetweenspecies usingaGeneralizedMixedLinealanalysisonSAS macroGLIMMIX (Littelletal.2006)applyinga binomial error and a logit link function. Habitat (n=3),species(n=2)andhabitat9speciesinter- action were considered as fixedeffects. Site and site9habitatinteractionwereconsideredasrandom factors.Second,asinmanyotherstudies(e.g.Aerts etal.2006), seedremovaltimewasanalyzedusing a timetoeventapproach (AltmanandBland1998).In ourstudytheeventisseedremoval,whichislinkedto escape frompredation andtherefore,thisdatais conventionallycalledseedsurvivaltime(Aertsetal.
2006).Differencesinmeanseedsurvivaltimeamong habitatsand betweenspecieswerecomparedwiththe Gehan-Wilcoxon test(PykeandThompson1986). ThesoftwareSPSS13.0wasusedforsurvivalanalysis (SPSSInc.,Chicago, Illinois,USA).
Seedgermination
InNovember2008weburied10metallicbagsof
10910cmand0.2cmfinemeshperspecies,ata
2cmdepth ineachsite.Eachbagcontained10seeds perspecies.Bagsrestedflat,sothatallseedswerein directcontact with thesoil.Intotal2,400 seeds were buried(10seeds910bags92species93habi- tats94sites).InJuly2009,240daysaftersowing, weretrievedthebags, andseedswere checkedinthe laboratory.Theywereconsideredtohavegerminated whentheradicleand/orthecotyledonhademerged fromtheseedcoat.
Differencesinthepercentageofgerminatedseeds ineachbagamonghabitats andbetweenspecieswere
analyzedusingSASmacroGLIMMIX applyinga binomialerrorandlogitlinkfunction.Habitat(n=3) andspecies(n=2)wereconsideredasfixedfactors while site (n=4)and site9habitatinteractionwere consideredasrandom factors.
Seedling survivalandgrowth
InFebruary2009,wetransplanted1yearoldseed- lingsgrownatanursery(Centre deJardineria Sils, site,andthirtyseedlingsperspecieswererandomly planted3mapartfromeachotheralongtransects. In total,720seedlingswereplanted(30seedlings92 species93 habitats94 sites)inthisway.Seedlings wereexamined1monthafter beingplantedandthose thathaddiedfrom transplant shock wereexcluded fromtheexperiment.Seedling survival wasrecorded everymonth untilJune2010.Plantheight wasmea- suredbeforesummerdrought(July2009andJune
2010, i.e. 5 and 16months after transplanting, respectively)andafterthefirstautumnrains(October
2009,i.e.8monthsaftertransplanting).
Seedlingsurvival timewascalculatedthrough Kaplan–Meier estimationsanddifferencesinmean survivaltimeamong habitats andbetweenspecies werecomparedwiththeGehan-Wilcoxon test(Pyke andThompson1986).
Differences inseedlingrelativegrowthrate(i.e. relativeincreaseinheightofsurviving seedlings betweentwoconsecutivemonitoringtimes)werealso analyzedwiththemacroGLIMMIX withthesame fixedandrandomfactorsasmentionedintheseed germinationanalysis above,butapplyinganormal errorandidentity linkfunctioninstead.Duetohigh seedling mortalityintheholm-oakstandsatthe beginning oftheexperiment(seeresults),weper- formedcomparisons ofseedlinggrowthbetween beechforests andheathlandsat5monthsaftertrans- plantation,andonlybetween specieswithinbeech forestsat 10monthsandat the endofthe experiment.
Probability oftransitionfromseedtoseedling
Inordertolinkthedifferent stagesoftheearlylife- cycle,weestimatedtheprobabilityofaseedbecoming anestablished seedling,foreachhabitattype,inboth coniferspecies(i.e.probabilityofrecruitment).The modelassumedthattheprobabilityofaseedachieving
recruitmentcanbeestimatedastheproduct ofthe previouselemental transitionprobabilities(Herrera etal.1994;Vila`andD’Antonio 1998).Weassumed thatthismodelisindependent ofpropagulepressure (i.e.seedproduction); thesequential recruitment stagesconsidered wereseedsurvival,seedgermina- tionandseedlingsurvival.
Results
Habitatandconifer speciescharacterization
Diameter atbreastheightoftreeswassignificantly biggerinbeechthaninholm-oakforests(16.7±0.8 and 9.4±0.36cm, respectively; t test: t=9.3, df=627,P\0.0001).Heathlandsweredominated byshrubbyshapedJuniperuscommunis, whereas beechandholm-oak forestshadalowunderstory cover.Lightavailability(PAR,lmols-1 m-2) at soil surfacewassignificantly differentamonghabitats (Kruskal–Wallistest; v2 =118.7,P\0.0001),being muchhigherinheathlands(928±28)comparedto holm-oak(51±11)andbeechforests(53.8±13).
Onthe otherhand,thetwoconiferspeciesdifferin seedmass(ttest:t=12.04,df=38,P\0.0001)and seed coat thickness (t test: t=12.18, df=38, P\0.0001).SeedmassislargerinSilverfir(0.03±
0.001g)thaninDouglasfir(0.01±0.0006g).Seed coat strength is also greater in Silver fir(16.7±
0.97g)thaninDouglas fir(4.1±0.33g). Seedremoval
Overall,seedremovalwasnotveryintenseinallthree
habitats.Attheendoftheexperiment,ofthe1,200
Fig.1 (a)Post-dispersalseedremoval(LSmeans?SEM), (b)seedsurvivaltime(mean?SE)and(c)seedgermination (LSmeans?SEM)forAbiesalbaandPseudotsuga menziesiiin threehabitattypes.Asterisksindicatesignificantdifferences betweenspecies(*P\0.05or**P\0.005)withinhabitats
holm-oak forestsandheathlands. Also,seedsurvival timeofSilverfirwaslongerthanDouglasfir(Fig.1b). Referringtodifferenceswithinhabitats,seedsurvival timewassignificantlyshorterinDouglasfirthanin
2
seeds deployed per species, only 20.5% were
Silverfirinheathlands(v
2
=8.17,P=0.004)and
removed.Thereweresignificantdifferencesbetween
inholm-oakforests(v
=9.38,P=0.002)butnotin
species(F1,9 =6.56,P=0.03)butnotamonghab- itats(F1,6 =0.84,P=0.47).Therewasa significant habitat9speciesinteraction(F2,9 =4.15, P=0.05) indicatingthatDouglasfirseedswereremovedmore frequentlythanSilverfir seedsinholm-oakand heathlands,whereas inbeechforeststhisdifference wasnotsignificant(Fig.1a).
There were differences in seed survival time among habitats (v2 =16.48, df=2, P=0.0002) andbetweenspecies(v2 =9.19,P=0.002),being significantly longer in beech forests, followed by
beechforests (v2 =1.07,P=0.3)(Figs.1b,2).
Seedgermination
Thereweresignificantdifferencesingerminationrates betweenspecies(F1,8 =19.05,P=0.002),yetnone werefoundamonghabitats(F2, 6 =0.22,P=0.80). Also,habitat9speciesinteractionwasnotsignificant (F2, 8 =0.07,P=0.93).Inallhabitats,thepercent- ageofseedgerminationwassignificantlyhigherin DouglasfirthaninSilverfir(Fig.1c).
Fig.2 Seedsurvivalcurves(mean±SE)ofAbiesalbaand Pseudotsugamenziesiiin(a)beechforests,(b) holm-oakforests and,(c)heathlands
Seedling survivalandgrowth
Seedlingsurvivalwasverylowinholm-oakforests duetosoildisturbancebynativefauna.Inthishabitat, wild boars pull up more than 50% of seedlings
3months aftertransplantation, andbytheendofthe experimentallseedlingshaddied.
Inheathlandsandbeechforests drought stress seemed tobethemainbarrier againstseedling survival.Inheathlands,only4%ofseedlingssurvived afterthesummer, while,inbeechforests,seedling survivalwerelessthan40%(Fig.3).
Fig.3 SeedlingsurvivalcurvesofAbiesalbaandPseudotsuga menziesiiin (a) beech forests, (b) holm-oak forests and, (c)heathlands
Overall, attheendoftheexperiment,seedling survivaltimewassignificantly differentamonghab- itats(v2 =319.8,df=2,P\0.0001).Survivaltime waslongerinbeechforests andshorterinholm-oak forests (Fig.4). There were differences between species(v2 =25.4,P\0.0001)withSilverfirseed- lingshavinglongersurvivaltimethanDouglas fir. Likewise, these differences between species were alsoobservedwithinhabitats(beech:v2 =8.77,P=
0.003;holm-oak:v2 =11.6,P=0.0006;andheath-
lands:v2 =6.05,P=0.013)(Fig.4).
Duetohighseedling mortality,seedling relative growthratecouldnotbeanalyzedinholm-oakforests. Overall, 5 months after transplantation, relative growthratewasverylowinbothspecies(Douglas fir:0.009±0.01;andSilverfir:0.022±0.01cm). Therewerenosignificantdifferencesbetweenspecies
Fig.4 Seedling survivaltime(mean?SE)ofAbiesalbaand Pseudotsuga menziesiiinthreehabitattypes.Asterisksindi- cate significant differences between species (*P\0.05,
**P\0.005)withinhabitats
(F1, 468 =0.75,P=0.38)orbetweenbeechforests and heathlands (F1, 6 =0.4, P=0.55). In beech forests, relativegrowthratewasnotsignificantly differentbetween species, both8monthsaftertrans- plantation(F1,234 =0.68,P=0.41)andattheendof theexperiment(F1, 237 =2.36,P=0.12).
Transitionfromseedtoseedling
Overall,theprobability ofoneseedbecomingan established seedlingwasverylowforbothconifer species(Fig.5).Inbeechforests,approximately4%of Douglasfirseedsand5%ofSilverfirseedsofinitial seedsetsbecomeestablishedseedlings(Fig.5a).In heathlands,Douglasfirshowedahigherprobabilityof survivalthan Silverfir(0.4 and 0%, respectively)due toahigher germinationrate,while inholm-oak forests, survivalwasnullforbothspecies.
Discussion
Theprobabilityofinvasion successofDouglas fir decreasedalongthesequentialstagesofestablishment in allhabitats,showingthe bestperformancein beech forests.ThehighergerminationcapacityofDouglasfir seemedtopredictabetterestablishment success comparedtoSilverfir.However,pressuresbynative faunaanddrought stressarepotentialmechanisms behindseedlingmortalitylimitingtheirestablishment ofbothconiferspecies.
Fig.5 Probabilityofsurvivalalongeachestablishmentstageof Abiesalba andPseudotsuga menziesiiin(a) beechforests, (b)holm-oakforests and,(c)heathlands
InDouglasfir,seedremovalwasnotveryintensein anyhabitatcomparedtoitsnativerange(Huggard andArsenault 2009)orwithotherconiferspecies (Borchertetal.2003;Peters etal.2004;Carrillo- Gavila´netal.2010).Seedremovalwassimilaramong habitats despiteofthefactthatthethreehabitattypes analyzedwere locatedatdifferentaltitudes.These resultsdonotmatchwithpreviousstudies,whichhave observed thatsmallmammalabundancedecreases with altitudinal gradients (Torre and Arrizabalaga
2009).Thelowseedremovalratesmightbedueto seedsatiation(Stowe etal.2000)astheexperiment coincidedwithseeddispersalofmanynativeplant
species.Moreover,seedremoversmighthaveahigher preferenceeithertowardsholm-oakacorns(Go´mez
2004;Espeltaetal.2009)orbeechnuts,whichare largerthanfirseeds.
Douglasfirgerminationrateswerehighanddidnot
differbetweenhabitatsdespitelargedifferences in vegetationstructureandlightavailability.Thisisin accordancewithotherfieldstudieswhereDouglas fir seedlingemergencewasveryhomogeneous under differentmicrosites(DunneandParker1999).Inspite ofthis,ourseedgerminationratesshould beconsid- eredwithcaution.Sincetheycomefromanursery,our seedsamplemightbecausinganoverestimationof thisparameterattributabletoahigherqualityofseeds. Nonetheless,seed germinationwas higherinDouglas thaninSilverfir.Silverfir’s sensitivitytoother environmentalstressessuchasdrought(Gradecˇkiand Posˇtenjak2001), mightbethecauseofitslow germination rates.Inthiscontext,arecentreviewby PysˇekandRichardson(2007)revealsthatalienspecies aregenerallyreportedtogerminateearlier,betterand inawiderrangeofconditionsthannativespecies.
Sincemanyseedlingsdisappeared inholm-oak forests duetosoildisturbancebywildboars,native fauna mightbeapotential mechanismlimitingthe establishment ofseedlingsinthisforesttype.In contrast, inbeechforests andinheathlands,seedling mortalityduetosummerdroughtstress(Alpertetal.
2000)wasaprominentecologicalfilterforbothnative
andalienplantspecies ashasbeen reported inother Mediterranean regions(DunneandParker1999;Rey andAlca´ntara2000;Dome`nechandVila`2006).
Thehigherseedling survivalinbeechforests compared tootherhabitatsmightsuggestabetter establishmentsuccessofDouglasfirinthishabitat, andthereforeahigherinvasibilityincomparisontothe otherhabitattypes.However,Douglasfirisconsidered apioneerspeciesthatprospersinpartialshade,but onceestablished itrequiresstronglightavailability (HermannandLavender 1999).Aswithmostalien conifers,non-disturbed forestsareresistantcom- munitiestoinvasion(Richardson andBond1991; Sarasolaetal.2006).Simberloffetal.(2002)observed thatDouglasfirdidnotestablishitselfintoclosedforest plantations.Therefore,wewouldexpectinvasibilityto decreaseinbeechforestsinsubsequentstages(Caccia andBallare´1998),exceptinnaturalgaps(Spiesand Franklin1989)orifforestsareintensivelymanaged.If disturbanceoccursandresourcesfluctuate(Davisetal.
2000) thisspecies mightnaturalize,evenifcurrently thereisnoevidenceofseedling establishmentinthis foresttype.
Conifertraitssuchasashortjuvenileperiod,ashort intervalbetween largeseedcrops,asmallseedmass (Rejma´nekandRichardson1996;Richardson and Rejma´nek 2004)andfastrelativegrowthrate (Grotkopp etal.2002)arecorrelatedwithhigh invasion success. Differences in seedling survival timebetween conifer speciesmightbeexplainedby theseed-seedlingconflict(Schupp1995),forwhich smallseededspecieshavebetterdispersalcapacitybut lowseedlingsurvivalduetolowerinvestment in reserves thanlargeseededspecies. Smallseeded speciessuchas Douglasfirtendtoproducelargeseed cropstocounterbalancethelowsurvival probability fromseedtoadulthood,whichwouldbeexpectdueto smallseedmass(MolesandWestoby 2004).
According toGreeneandJohnson(1999),seed productionintreesisinverselyproportionaltoseed mass.LargeDouglasfir seedcropsintheUKare producedevery4–6years(Wilan1985).Hermannand Lavender (1999)estimatethatseedproductionis about2.2kg/ha,whichvarieswidelyamongyears; Gashwiler (1969)estimates annualmeanseeddis- persaltobearound3x105seeds/hainthenorthwestof theUnitedStates.Whileseedproduction ofSilverfir seemstobelowerthanDouglasfir, italsovaries highlyamong years.Forexample,seedproduction variedmarkedlyintwoconsecutiveyearsfrom59.6to
118.8seeds/m2 intheFrenchAlps(Sagnardetal.
2007).
Inconclusion,ourresultsshowedthattheproba- bilityofestablishmentofbothfirspeciesdecreased alongtheearlylife-cyclestagesinallhabitats.Soil disturbancebynativefaunainholm-oak forestsand droughtstressin heathlandsmightbethemechanisms controllingseedlingestablishment.Douglasfir per- formed betterinbeechforests whereseedsurvival, seedgerminationandseedlingsurvivalwerehigher thanintheotherhabitats.Onlythehighergermination capacityofDouglasfir seemedtopredictabetter establishmentsuccesscomparedtoSilverfir. Cur- rently,Douglasfirisonlynaturalizedinheathlands (Broncano etal.2005).Thisindicatesthatthereisa mismatchbetweenthedegreeof invasionobservedin MontsenybyDouglasfir andourinvasibilityassess- ment.ThismismatchmightpredictthattheDouglasfir establishmentobservedinheathlandsisaconsequence
ofa higherseedrainfromadjacentplantationswhich, coupled withitshighgerminationcapacity,might predictitsinvasionsuccess.Thisisinaccordancewith theimportancethatpropagulepressurehasoninva- sionsuccess(Lockwoodetal.2005;Colauttietal.
2006,butseeNun˜ezetal.2011).Manyconiferspecies thatescapefrom plantationsandbecomeinvasiveare often thosethathavebeencultivatedthemost widely andforthelongest time(Krˇiva´nek etal.2006). However, arecentstudyrevealsthatpropagule pressurein80-yearoldwoody species’ plantationsis notthekeytopredictinvasion(Nun˜ezetal.2011).In thissense,climateandbioticresistanceseemtobetwo ofthemostimportantcomponents topredictthe invasionsuccess ofconiferspeciesintheintroduced range (Haugo 2010; Essl et al. 2011; Nun˜ez and Medley2011).Therefore,wehighlighttheneedfor futureresearch tofocusonseedproductionandon othercomponentsofbioticresistance,thatwerenot exploredinthisstudy(Nun˜ezandMedley2011),to determinetheinvasivenessofDouglasfir andthe invasiblityofdifferenthabitatstothisconifer.
Acknowledgments WeareverygratefultoEPe´rez-Garc´ıa,J Andreu,VMeco,andMHernanzforfieldassistance,andto all personnelatFontmartina andCanCasadesintheMontseny NaturalParkfortheirassistanceduringfieldwork.Researchwas partiallyfundedbytheSpanishMinisterio deCienciae Innovacio´nprojects:Consolider-IngenioMONTES(CSD2008-
00040),COMPROPIN(AGL2010-18724),andCSIC(project
200840I153).
References
AertsR,MaesW,NovemberE,NegussieA,HermyM,MuysB (2006)RestoringdryAfromontane forestusingbirdand nurseplanteffects: directsowingofOleaeuropaeassp cuspidataseeds.ForEcolManag230:23–31
AlpertP,BoneE,HolzapfelC(2000)Invasiveness,invasibility
andtheroleofenvironmental stressinthespreadofnon- nativeplants.Perspect PlantEcolEvolSyst3:52–66
AltmanDG,BlandJM(1998)Statisticsnotes—timetoevent
(survival)data.BrMedJ317:468–469
BesendorferV,Krajacˇic´-SokolI,Jelenic´S,PuizinaJ,LinarecJ, SvibenT(2005)Twoclassesof5SrDNAunitarraysofthe silver fir,Abiesalba Mill.: structure, localization and
evolution.Theor ApplGenet110:730–741
BoadaM(2000)ElMontseny. Fiftyyearsoflandscapeevolu- tion.Publicationsdel’AbadiadeMontserrat,Barcelona
BoadaM,BroncanoM(2003)Estudisobreladistribucio´d’es-
pe`ciesforestalsal.lo`ctonesalParcNaturaldelMontseny. Evide`ncies delanaturalitzacio´del’avetdeDouglas (Pse- udotsugamenziesii)alParcNaturaldelMontseny.ICTA-
InstitutdeCie`ncieiTecnologiaAmbientals,Universitat
Auto`nomadeBarcelona,Barcelona
Bolo`sO,VigoJ(1990)FloraDelsPa¨ısosCatalans.Barcino
Press,Barcelona
BorchertM,JohnsonM,SchreinerD,VanderWallS(2003) Earlypost-fireseeddispersal,seedlingestablishmentand seedling mortalityofPinuscoulteri(D.Don)incentral coastalCalifornia,USA.PlantEcol168:207–220
BroncanoM,Vila`M,BoadaM(2005)EvidenceofPseudotsuga
menziesii naturalizationinmontaneMediterraneanforests. ForEcolManag 211:257–263
CacciaFD,Ballare´CL(1998)Effectsoftreecover,understory vegetation, and litter on regeneration of Douglas-fir
(Pseudotsugamenziesii)insouthwesternArgentina.CanJ ForRes28:683–692
Carrillo-Gavila´nMA,Vila`M(2010)Littleevidenceofinvasion byalienconifers inEurope.Divers Distrib16:203–213
Carrillo-Gavila´nMA,HadrienL,Vila`M(2010)Comparing seedremoval of16pinespeciesdiffering ininvasiveness.
BiolInvasions12:2233–2242
ColauttiR,Grigorovich I,MacIsaacH(2006)Propagulepres- sure:anullmodel forbiologicalinvasions.BiolInvasions
8:1023–1037
DavisMA,GrimeJP,ThompsonK(2000)Fluctuatingresources in plant communities:a general theoryof invasibility. JEcol88:528–534
DietzH,SteinleinT,UllmannI(1999)Establishmentofthe invasive perennialherbBunias orientalisL.:anexperi- mentalapproach.ActaOecol20:621–632. doi:10.1016/ S1146-609X(99)00104-6
Dome`nechR,Vila`M(2006)Theroleofsuccessionalstage,
vegetationtypeandsoildisturbanceonCortaderia sello- anainvasion.JVegSci17:591–598
DunneJA,ParkerVT(1999)Species-mediatedsoilmoisture availabilityandpatchyestablishment ofPseudotsuga menziesiiinchaparral.Oecologia119:36–45
EspeltaJM,BonalR,Sa´nchez-HumanesB(2009)Pre-dispersal
acornpredationinmixedoakforests:interspecificdiffer- encesaredriven bytheinterplayamong seedphenology, seedsizeandpredator size.JEcol97:1416–1423
EsslF,MangT,Dullinger S,MoserD,HulmePE(2011) Macroecological driversofalienconifernaturalizations Worldwide.Ecography34:1–9
FerrerasAE,GalettoL(2010)Fromseedproductiontoseedling establishment:importantstepsinaninvasiveprocess.Acta Oecol36:211–218
GashwilerJS(1969)Seedfallofthreeconifersinwest-central
Oregon.ForSci15:290–295
Go´mez JM (2004) Bigger is not always better: conflicting selective pressuresonseedsizein Quercusilex. Evolution
58:71–80
Gradecˇki M,PosˇtenjakK(2001) Relationbetween laboratory andnurserygerminationofsilverfirseed(Abiesalba
Mill.).In:Matic´S,KrpanAPB,GracˇanJ(eds)Znanostu
potrajnomgospodarenju hrvatskimsˇumama:znastvena knjiga.SˇumarskifakultetSveucˇilisˇtau Zagrebu,Sˇumarski institute,Zagreb, pp191–200
Greene DF,Johnson EA(1999) Estimatingthemeanannual seedproductionoftrees.Ecology 75:642–647
GrotkoppE,Rejma´nekM,RostTL(2002)Towardacausal explanationofplantinvasiveness:seedlinggrowthand
life-historystrategiesof29pine(Pinus)species.AmNat
159:396–419
HaugoRD(2010)Causesandconsequencesofconiferinvasion intoPacificNorthwestgrasslands.PhD,Thesis.University ofWashington
Hemstrom MA,LoganSE,PavlatW(1987)Plantassociation andmanagementguide:WillametteNationalForest. R6- Ecol257-B-86.U.S.Department ofAgriculture, Forest Service, PacificNorthwestRegion,Portland,p312
Hermann RK,Lavender DP(1999)Pseudotsugamenziesii (Mirb.)Franco. In:BurnsR,HonkalaBH(eds)Silvics of NorthAmerica, vol1.Conifers. USDAForestService. Washington,D.C.,pp540–557
Herrera CM, Jordano P, Lo´pez-Soria L, Amat JA (1994)
Recruitmentofamass-fruiting, bird-dispersed tree: bridgingfrugivore activityandseedlingestablishment. EcolMonogr 6:315–344
Huggard DJ,ArsenaultA(2009)Coniferseedpredationin harvestedandburneddryDouglasfir forestsinsouthern BritishColumbia.CanJForRes39:1548–1556
HulmePE(2011)Addressingthethreattobiodiversityfrom botanicgardens.TrendsEcolEvol26:168–174
Kay M(1994)Biologicalcontrolforinvasivetreespecies.NZ For39:35–37
Kˇriva´nekM,PysˇekP,JarosˇikV(2006)Plantinghistoryand propagule pressureaspredictorsofinvasion bywoody speciesinatemperateregion.ConservBiol20:1487–1498
LambrinosJG(2002)ThevariableinvasivesuccessofCorta-
deriaspecies.Inacomplexlandscape.Ecology83:518–529
LittellRC,MillikenGA,StroupWW,Wolfinger RD,Schan- benberberOD(2006)SASformixedmodels,2ndedn.SAS PressSeries,Cary
LLobetS(1990)ElmediilavidaalMontseny.Museude
Granollers
LockwoodJL,CasseyP,Blackburn T(2005)Theroleof propagulepressureinexplainingspeciesinvasions.Trends EcolEvol20:223–228
LombarderoMJ,Va´zquez-MejutoP,Ayres MP(2008)Roleof plantenemiesintheforestryofindigenous vs.nonindige-
nouspines.EcolAppl18:1171–1181
LonsdaleWM(1999)Globalpatternsofplantinvasionsandthe conceptofinvasibility.Ecology80:1522–1536
Lo´pez-Gonza´lezG(2002)Gu´ıadelosa´rbolesyarbustosdela
Pen´ınsulaIbe´rica y Baleares.EdicionesMundi-Prensa, Madrid
McCayTS,McCayDH(2009)Processesregulatingtheinva-
sionofEuropeanbuckthorn(Rhamnuscathartica)inthree habitatsofthenortheasternUnitedStates.BiolInvasions
11:1835–1844
Moles AT,WestobyM(2004)Seedlingsurvivalandseedsize:
assynthesis oftheliterature.JEcol92:372–383
Nun˜ezMA,MedleyKA(2011)Pineinvasions:climatepredicts invasionsuccess;somethingelsepredictsfailure.Divers Distrib.doi:10.1111/j.1472-4642.2011.00772.x
Nun˜ezMA,SimberloffD,RelvaMA(2008)Seedpredationsas abarriertoalienconiferinvasions.BiolInvasions8:1389–
1398
Nun˜ezMA,HortonTR,SimberloffD(2009)Lackofbelow- groundmutualismshindersPinaceaeinvasions.Ecology
90:2352–2359
Nun˜ezMA,MorettiA,SimberloffD(2011)Propagulepressure hypothesis notsupported byan80-yearexperimenton woody species invasion. Oikos120:1311–1316
Ordo´n˜ez JL, Retana J (2004) Early reduction of post-fire recruitmentofPinusnigrabypost-dispersalseedpredation
indifferenttime-since-firehabitats.Ecography27:449–
458
Pen˜uelasJ,BoadaM(2003)Aglobalchange-inducedbiome shiftintheMontsenymountains(NESpain).GlobChange Biol9:131–140
PetersSH,Macdonald SE,BoutinS,MosesRA(2004)Post- dispersalseedpredationofwhitespruceincutblocksinthe borealmixedwoods:ashort-termexperimentalstudy.Can JForRes34:907–915
PykeDA,Thompson JN(1986)Statisticalanalysisofsurvival andremoval rateexperiments.Ecology 67:240–245
PysˇekP,RichardsonDM(2007)Traitsassociatedwithinva-
sivenessinalienplants:where dowestand? In:Nentwing W(ed)Biologicalinvasions, ecologicalstudies193. Springer,Berlin, pp97–126
PysˇekP,RichardsonDM,Rejma´nekM,WebsterG,Williamson
M,KirschnerJ(2004)Alienplantsinchecklistsandfloras: towardsbettercommunicationbetweentaxonomistsand ecologists.Taxon53:131–143
Rejma´nekM,RichardsonDM(1996)Whatattributesmake
someplantspeciesmoreinvasive?Ecology77:1655–1661
RelvaMA,Nun˜ezMA,SimberloffD(2009)Introduceddeer reducenativeplantcoverandfacilitateinvasionofnon- nativetreespecies:evidenceforinvasionalmeltdown.Biol Inv.doi:10.1007/s10530-009-9623-0
ReyPJ,Alca´ntaraJM(2000)Recruitmentdynamicsofafleshy-
fruitedplant(Oleaeuropaea):connectingpatternsofseed dispersal toseedlingestablishment.JEcol88:622–633
RichardsonDM,BondWJ(1991)Determinantsofplantdis- tribution: evidence from pine invasions. Am Nat
137:639–668
Richardson DM,HigginsSI(1998)Pinesasinvadersinthe SouthernHemisphere.In:RichardsonDM(ed)Ecology andbiogeographyofPinus.CambridgeUniversityPress, Cambridge,pp450–470
RichardsonDM, Rejma´nekM (2004) Conifers as invasive
aliens:aglobalsurveyandpredictiveframework.Divers
Distrib 10:321–331
RodgersonL(1998)Mechanical defenceinseedsadaptedfor antdispersal. Ecology79:1669–1677
SagnardF,PichotC,DreyfusP,JordanoP,FadyB(2007)
Modellingseeddispersaltopredictseedlingrecruitment:
recolonizationdynamicsinaplantationforest.EcolModell
203:464–474
Sarasola MM,RuschVE,Schlichter TM,GhersaCM(2006) Invasio´ndecon´ıferasforestalesena´reasdeestepaybos- quesdecipre´sdelacordilleraenlaRegio´nAndinoPa- tago´nica.AustEcol16:143–156
SchuppEW(1988)Seedandseedlingpredationintheforest
understoryandintreefallgaps.Oikos51:71–78
SchuppEW(1995)Seed-seedlingconflicts,habitatchoice,and patterns ofplantrecruitment.AmJBot82:399–409
ShafrothPB,AubleGT,ScottML(1995)Germination and establishment ofthenativeplainscottonwood (Populus deltoides Marshall subsp. Monilifera) and the exotic
Russian-olive(ElaeagnusangustigoliaL.).ConservBiol
9:1169–1175
SimberloffD,RelvaMA,Nun˜ezM(2002)Gringosenelbos- que:introduced treeinvasioninanativeNothofagus/Aus- trocedrusforest. BiolInvasions4:35–53
SimberloffD,Nun˜ezMA,LedgardNJ, PauchardA,Richadson DM,Sarasola M,wanWilgenBW,ZalbaSM,ZenniRD, BustamanteR,Pen˜aE,ZillerSR(2010)Spreadandimpact ofintroduced conifersinSouthAmerica:lessonsfrom other southernhemisphereregions.AustEcol35:489–504
SolD,Vila`M,Ku¨hnI(2008)Thecomparativeanalysisof
historical alien introductions. Biol Invasions 10:1119–
1129
SpiesSE,Franklin J(1989)Gapcharacteristicsandvegetation responseinconiferousforestsofthePacificNorthwest. Ecology70:543–545
StoweKA,MarquisRJ,HochwenderCG,SimmsEL(2000)The
evolutionaryecologyoftolerancetoconsumerdamage. AnnuRevEcolEvolSyst31:565–595
TorreI,ArrizabalagaA(2009)Speciesrichnessandabundance
of small mammals along an elevational gradient of a
Mediterraneanmountain.VieMilieu59:203–212
TravesetA,GuliasJ,RieraN,MusM(2003)Transition prob- abilitiesfrompollinationtoestablishment inararedioe- ciousshrubspecies(Rhamnus ludovici-salvatoris)intwo habitats.JEcol91:427–437
Vila`M,D’AntonioCM(1998)FitnessofinvasiveCarpobrotus
(Aizoaceae) hybrids in coastal California. Ecoscience
5:191–199
Vila`M,LloretF(2000)Seeddynamicsofthemastseeding tussockgrassAmpelosdesmosmauritanica inMediterra-
neanshrublands.JEcol88:479–491
Vila`M,BartomeusI,GimenoI,TravesetA,MoraguesE(2006) DemographyoftheinvasivegeophyteOxalispes-caprae
acrossaMediterraneanisland.AnnBot97:1055–1062
WangX-Q,TankDC,SangT(2000)Phylogenyanddivergence timesinPinaceae:evidencefromthreegenomes.MolBiol Evol17:773–781
WilanRL(1985)Aguidetoforestseedhandlingwithspecial
referencetothetropics.FAOForestryPaper,no.20/2. FAO,Rome(Italy)
WolfH(2003)EUFORGENtechnicalguidelinesforgenetic
conservationanduseforsilverfir(Abiesalba).Interna- tionalPlantGeneticResourcesInstitute, Rome