Clonal growth buffers the effect of grazing management on the population growth rate of a perennial grassland herb

(1)

Contents lists available atScienceDirect

Flora

j o u r n a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / f l o r a

Clonal growth buffers the effect of grazing management on the population growth rate of a perennial grassland herb

Line Johansen

^∗

, Sølvi Wehn, Knut Anders Hovstad

TheNorwegianInstituteofBioeconomyResearch(NIBIO)Kvithamar,N-7512Stjørdal,Norway

a r t i c l e i n f o

Articlehistory:

Received16June2015

Receivedinrevisedform15March2016 Accepted15April2016

EditedbyHermannHeilmeier Availableonline19April2016

Keywords:

Matrixprojectionmodels LTREanalysis

Growthrate

Life-historycomponents Stagestructure Knautiaarvensis

a b s t r a c t

Grazingisanimportantmanagementactiontoconservebiodiversityinsemi-naturalgrasslandsbutitis importanttounderstandhowgrazinginﬂuencesthelife-historycomponentsandpopulationdynamics ofplantspecies.Inthisstudy,weanalysedeffectsofgrazingintensityandabandonmentonpopulation dynamicsofthesemi-naturalgrasslandspeciesKnautiaarvensiswhichisanimportantnectarsourcefor pollinatingspeciesandanindicatorofbiodiversityinagriculturallandscapes.Werecordedlife-history stage,survival,establishmentofseedlingsandramets, numberofinﬂorescencesandgrazingmarks onpermanentlymarkedindividualsineightpopulationsinmid-Norwayforthreeconsecutiveyears.

Matrixmodellingwasusedtoestimatepopulationgrowthratesandelasticities,andlifeTableresponse experiments(LTREs)wereusedtoassessthecontributionofdifferentlife-historycomponentstothe observedvariationinpopulationgrowthratesbetweendifferentmanagementtreatments.Generalized linearmixedeffectsmodels(GLMMs)wereusedtoinvestigatetheeffectofmanagementonvitalrates andnumberofinflorescencesaswellasdamagetoK.arvensisindividuals.Populationsinabandoned grasslandshadmoreinflorescences,alowerproportionofseedlingsandahigherproportionofflowering rametscomparedtopopulationsingrasslandsunderhighgrazingintensity.Therewerenodifferences inpopulationgrowthratesbetweendifferentgrazingintensities.Fecundityhowever,contributedmore tothegrowthrateingrazedgrasslandscomparedtoabandonedgrasslandswhereclonalregeneration contributedthemost.Survivalofnon-floweringrosettesmadethelargestimpacttooverallgrowthrates.

Ourresultsindicatethatalonglife-spanandclonalgrowthbuffertheeffectofenvironmentalchangein abandonedgrasslandsandthatthereisatrade-offbetweenfertilityandclonalregenerationinK.arvensis populations.

1. Introduction

Semi-natural grasslands harbour a high diversity of plant species(Mariniet al.,2008).However,theareaofsemi-natural grasslandsinEuropehasdecreasedduetoagriculturalintensiﬁca- tion,changedmanagementpracticesandabandonment(Hodgson etal.,2005;PoschlodandWallisDeVries,2002).Semi-naturalgrass- landsthathavebeenabandonedormanagedattoolowintensity tohalt successionalchangewillgraduallybeinvadedbyshrubs andtrees(Wehn,2009).Asaresult,growthconditionsbecomeless favourableforlight-demandingspeciesinparticular(Pykalaetal., 2005),andpopulationsofsuchspeciesmayrapidlydeclineonce

∗Correspondingauthor.

E-mailaddresses:[email protected](L.Johansen),[email protected] (S.Wehn),[email protected](K.A.Hovstad).

grasslandshavebeenabandoned(Endelsetal.,2007b;Hamreetal., 2010).

Extensivegrazing isrecommended asamanagement tool to maintain or improve plantpopulation viabilityin semi-natural grasslands (Meteraet al., 2010; Wrage et al., 2011). Herbivory affects plantabundanceand distribution aswellas planttraits (Louault et al.,2005; Maron and Crone, 2006).Effects of grazing onplantperformanceare both direct andindirect; indirect throughchangingthehabitatqualitybytrampling,reducingcom- petition,additionofnutrientandlitteraccumulation(Brysetal., 2004; Ehrlen etal., 2005) anddirect by damagingtheplantor reducingﬂoweringandseedset(Knight,2004).Hence,different levelsofgrazingintensitycanhavedisparateeffectsonvitalrates andthedemographicbehaviourofplantpopulations(Brysetal., 2004;LennartssonandOostermeijer,2001).Vitalratesmayeven differbetweenpopulationsofasinglespecies,dependingongraz- ingintensityandthetimesinceabandonment(Brysetal.,2004;

JacquemynandBrys,2008).Thelifespanofaspeciesmaycorrelate

http://dx.doi.org/10.1016/j.ﬂora.2016.04.007

4.0/).

(2)

0.650.80

Surviving to t+1

no low high

a)

0.000.100.20

Seedlings

no low high

b)

0.00.30.6

Flowering ramets

no low high

c)

0.01.53.0

Inflorescence

no low high

d)

0.00.20.4

Clonal reproduction

no low high

e)

0.00.40.8

G i

Grazed ramets

no low high

f)

2008 2009 2010 Grazing intensity

Fig.1.Meanandstandarderrorof(a)survivaltothenextyear(t+1),(b)proportionofseedlings,(c)proportionofﬂoweringramets,(d)numberofinﬂorescencesper ramet,(e)proportionofclonalreproduction,and(f)proportionofgrazedrametsinKnautiaarvensispopulationsin2008,2009and2010insemi-naturalgrasslandswithno (abandoned),lowandhighgrazingintensity.

withitspopulationdynamics,andlongevitycanbufferchanging environmentalconditions(Morrisetal.,2008).Long-livedperen- nialplantsareknowntohavemorestablepopulationsizesthan short-livedplants,becauseingeneralsurvivalhaslessvariability thanfecundity,andfecundityisrelativelymoreimportantforpop- ulationdynamicsofshort-livedplantspecies(Garciaetal.,2008).

Becausegrazingaffectsdifferentlife-historystagessimultaneously, itisessentialtointegratemultiplevitalrateswithinasingleanal- ysistofullyunderstandthepopulationdynamicsandviabilityofa species.Instage-structuredpopulationsthiscanbeachievedusing matrixmodels(Caswell,2001).

Theobjectiveof thisstudyistoanalysetheeffectofgrazing intensityandabandonmentonpopulationdynamicsoftheclonal speciesKnautiaarvensis.Wewanttoestimatehowitslife-history components contribute to its population growth rate in semi- naturalgrasslandsunderdifferentland-useregimes.InNorway, K.arvensiscanberegardedasakeystonespeciesupholdingspecies richnessin semi-naturalgrasslandsasit isanimportant nectar sourceformanyspeciesofbutterﬂies,bumblebees,solitarybees and othergroups of pollinatinginsects, e.g. theNorwegian red

listedminingbeeAndrenahattorﬁana(CahenzliandErhardt,2012;

Clausenetal.,2001;FranzenandNilsson,2008;Kålåsetal.,2010;

Totlandetal.,2013).Modelsweredevelopedtoassesshowaban- donmentanddifferentlevelsofgrazingintensitybycattleinﬂuence planttraits,vitalrates,life-historycomponentsandgrowthrate.

Thecontributionofvitalratestothegrowthrateforeachlevelof grazingintensitywasdeterminedtoseewhetherK.arvensishad differentreproductionstrategieswhenexposedtodifferentlevels ofherbivory.

2. Methods 2.1. Studyspecies

Knautiaarvensis(L.)Coult.isaperennial,clonalherbwithagen- eralistpollinationsystemthatgrowsingrasslands,openwoods, onroadvergesandruderalsites.Thespeciesiswidelydistributed inEurope,westAsiaandnorth-westAfrica(LidandLid,2005).It hasasympodial,branchedstockwithleafrosettesandﬂowering stems,ataprootandusuallylateralundergroundrhizomes(Tutin

(3)

0.51.01.52.02.5

Population growth rate 08−09 09−10 08−09 09−10 08−09 09−10 08−09 09−10 08−09 09−10 08−09 09−10 08−09 09−10 08−09 09−1008−09 09−10

no low high

1 1

2 2

3 3

4 4 5 5

6 6

7 7 8 8

Grazing intensity

Transition year

Fig.2. ProjectedpopulationgrowthratesforKnautiaarvensispopulations(1–8)in semi-naturalgrasslandswithno(abandoned),lowandhighgrazingintensityintwo transitionyears(2008–2009and2009–2010).Theerrorbarsarebootstrapped95%

conﬁdenceintervals.

etal.,1976).AccordingtoThompsonetal.(1997)K.arvensishas nopersistentseedbank. Thespecies is abletogerminateunder a widerange of environmentalconditions(Vandvikand Vange, 2003).However,earlyseedlingestablishmenthasbeenfoundto bepoorinclosedgrasslandvegetationbutpositivelyaffectedby disturbancethatcreatestinygapswithbaresoil(Hovstad,2007).

Larsson(2005)foundK.arvensistohaveameanof6.7inflorescences perfertilerametandmeannumberofflowersperinflorescenceto be67.1.Inthepresentstudywemeasuredameanof29seedsper inflorescence.

2.2. Studyarea

The study area is located in mid-Norway (Nord-Trøndelag county)andcoversanareaof20km²(UTM618650E,7070496N) andtheelevationrangesfrom80to180masl.Meanannualtem- peratureis4.7^◦Candmeanannualprecipitationis815mminthe studyarea(normalperiod1961–1990)(NorwegianMeteorological Institute,2014).Thestudyareaisinthemiddleandsouthernboreal vegetationzones,intheslightlyoceanicvegetationsection(Moen, 1999),andthebedrockconsistsmainlyofshaleandsomemica schist(dataprovidedbyGeologicalSurveyofNorwayNGU,2014).

Inthisstudywerefertosemi-naturalgrasslandsasgrasslandswith alongandcontinuousextensivemanagement(grazingormow- ing)wherespecieshavenotbeensownorplanted.Inthestudied grasslandstherewerenocanopiesofwoodyspecies,andthecom- positionsofplantspecieshadnotbeenalteredsigniﬁcantlybyuse offertilizersorherbicides(Norderhaugetal.,2000).

2.3. Studydesignanddatacollection

Allsemi-naturalgrasslandsinthestudyareaweresurveyedin 2008andpresence/absenceofK.arvensiswasrecorded.Fromthis survey,eightsemi-naturalgrasslandsitesofatleast0.2hawere

selected.NeitherpopulationsizenorviabilityofK.arvensiswere usedascriteriafortheselectionofsites.Threeofthegrasslands wereabandoned(8–15yearsago)andfiveweregrazed.Theinten- sityof thegrazingin thegrazed grasslandswasa resultof the farmer’spracticeandnotsetbythestudy.Informationaboutthe managementofthegrasslandswereachievedbysemi-structured interviewswiththefarmers.Grazing tookplacethroughout the grazing season(May–September),but wereregulatedina non- structuralway.Thefarmershadaccesstoseveralgrasslands.Inall grazedsites,livestockweretakenonandoffthegrasslandseveral timesthroughoutthegrazingseasonandsomegrasslandswere preferredandmanagedmoreintensively.Inthepreferredgrass- landsthelivestockweretakenoffthegrasslandswhenswardheight becamelessthan4–8cm.Thelesspreferredgrasslandswerethen usedasoffsetareas.Ingeneral,theoffsetshadahigherstocking rateinlatesummer,comparedtospringandearlysummer.We didnotcontrolthenumberofanimalsperunitareabutbasedon thequalitativedataprovided bytheinterviews,we dividedthe grazedgrasslandsintotwoclassesofgrazingintensities:lowgrazingintensity(n=2;theoffsets)andhighgrazingintensity(n=3;the preferredgrassland).Inthegrasslandsdefinedasunderhighgraz- ingintensity,themeanvegetationheightoveratleast75%ofthe areawas<4–8cm,andingrasslandsdefinedasunderlowgrazing intensitythevegetationheightwas>20cminmid-July.Generally, inthelesspreferredgrassland(underlowgrazingintensity)someof thevegetationwasleftungrazedbytheanimals.Intheabandoned grasslandstherewasnoencroachmentasyet.Allgrasslandswere situatedona hillsidefacingsouth–south-westandhadapproxi- matelysimilarslopesandtherewerenoovergrazingorlargeareas ofbaresoilduetoanimaltrampling.

In 2008,studyplots of1m×1mwereestablishedrandomly withineach oftheselectedsemi-naturalgrasslands.Nineto14 studyplotswithK.arvensisandtwostudyplotswithoutK.arvensis wereestablishedandpermanentlymarkedwithineachgrassland.

StudyplotswithoutK.arvensiswereselectedinordertorecord theestablishmentofnewindividuals.Withineachplot,allram- etsofK.arvensiswerepermanentlymarked.Atotalof72plotsand 1454rametswereincludedinthestudy.Eachsummer,2008–2010, acensusofthestudyspecieswithinallplotswasconductedand foreachrametsurvival,evidenceofgrazing,numberofinﬂores- cencesperramet,andlife-historystagewererecorded.Inaddition, establishmentofseedlingsandclonalrametswererecordedwithin eachplot.Wedistinguishedﬁvelife-historystagesforK.arvensis:

seedlings, non-flowering clonal offspring, flowering clonal offspring,non-floweringrosetteandfloweringrosette.Seedlingswere definedassmallplantsyoungerthanoneyearold.Newindividuals intheplotsweremarkedandincludedinthestudyateachcensus.

2.4. Statisticalanalysis

Theeffectofmanagementandtemporalvariationonthenum- berofinflorescencesper ramet,vitalrates(survivaltothenext year,proportionoffloweringramets,proportionofseedlingsand floweringandnon-floweringclonaloffspring)andproportionof grazedrametsineachplotwereanalysedusingageneralisedlin- earmixed effects model(GLMM). RametID,plotand grassland wereincludedasnestedrandomfactorsinthemodels.Yearand grazingmanagementwereincludedasfixedfactors.Therandom, nestedsamplingstructure wasincludedtoavoidproblemswith pseudoreplication.Theresponsevariablesweremodelledusinga binomialdistribution,exceptfornumberofinflorescenceswherea Poissondistributionwasused.Modelsweremadesequentiallyand reducedbybackwardeliminationofnon-significantfixedeffects (␣definedas0.01;Crawley,2007;Zuur,2009).Theestimatesand statistics presented arefromthefinal reduced modelsfor each responsevariable.

(4)

Seedling

Non- ﬂowering

clonal oﬀspring

Non- ﬂowering

rosee

Flowering rosee Flowering

clonal oﬀspring 0.013

0.003

0.062 0.012 0.051

0.235 0.085

0.040

0.077

0.023 0.048

0.027

0.024

0.033 0.024

0.081

0.055

0.014 0.002

0.087

Fig.3.Life-cycleofKnautiaarvensiscoveringﬁvestages.Thearrowsrepresentyearlytransitionsandelasticityvaluesoftheoverallmeanmatrixaregiven.

Table1

TransitionmatrixmodelforKnautiaarvensiswithfivestageclasses:seedlings,non-floweringclonaloffspring,floweringclonaloffspring,non-floweringrosetteandflowering rosette.Transitionsaregroupedaccordingtosevenlife-historycomponents:clonalreproduction(CR),fecundity(F),survivalofclonaloffspring(SC),seedlingsurvival(SS), stasis(S),retrogression(R)andenteringfloweringstage(EF).“0”indicatesthattherearenotransitionsbetweenthestageclasses.Subscriptnumbersbythematrixelements indicatethetransitionsfromstagejtoiinone-yearintervals(t+1).

yeart

Seedlings Non-ﬂoweringclonaloffspring Floweringclonaloffspring Non-ﬂoweringrosette Floweringrosette

yeart+1 Seedlings 0 0 F13 0 F15

Non-ﬂoweringclonaloffspring 0 CR22 CR23 CR24 CR25

Floweringclonaloffspring 0 CR32 CR33 CR34 CR35

Non-ﬂoweringrosette SS41 SC42 SC43 S44 R45

Floweringrosette SS51 SC52 SC53 EF54 S55

Thedemographic datawereanalysed by constructingstage- classiﬁedpopulationmatrixmodelling(Caswell,2001)witha5×5 projectionmatrixfor eachyear-to-yeartransition(Table1).The transitionmatrixmodelwasintheform:

n_(t+1)=An_(t),

whereAisapopulationprojectionmatrixcontainingthetransition probabilitiesbetweenstagesorstage-specificfecundity,n_(t) is a vectorofstage-classifiedindividualsattimet,andn_(t+1)isthevector ofstage-classifiedindividualsattimet+1.

AllpossibletransitionsaredescribedinTable1.Transitionswere groupedaccordingtosevenlife-historycomponents:clonalreproduction,fecundity,survivalofclonaloffspring,survivalofseedlings, stasis,retrogressionandenteringfloweringstage.Retrogressionis thetransitionfromfloweringtonon-floweringstage,andenter- ingfloweringstageisthetransitionfromnon-floweringrosette tofloweringstage.Sincethespecieshasnopersistentseedbank (Thompsonetal.,1997),noseedstagewasmodelled.Intotal,16 projectionmatriceswereconstructed;oneperpopulationperyear intervalinadditiontoanoverallmeanmatrix.Foreachtransition matrixtheprojectedpopulationgrowthrate()wasestimatedas thedominanteigenvalueofthematrix,inadditiontotheelasticity (proportionalsensitivity)oftheeigenvaluestochangesinthevital

rates.Theelasticityisanestimateoftherelativecontributionof alife-cycletransitiontotheprojectedpopulationgrowthrate(;

Caswell,2001).

Sexualfecunditywascalculatedastheestimatedmeannum- berofinflorescencesperrametmultipliedbyestimatednumberof seedlingsinthepopulationdividedbyinflorescencesinthepopula- tion(Brysetal.,2004).Itwasnotpossibletodeterminefromwhich parentaramet(=clonaloffspring)hademergedwithoutexcavating underground rhizomes.Therefore clonal reproductionwas esti- matedfromthenumbersofclonaloffspringandfloweringclonal offspringinyeart+1dividedbythenumberofpotentialclonal offspringproducersinyeart(Berg,2002).Weassumedthatalllife- historystagesexceptseedlingsproducedclonaloffspringandthat thetransitionsrepresentingclonalreproductionweresimilarforall stages.Whenthestudystarteditwasnotpossibletoknowwhich rametswerenewclonaloffspringthatyear,andhencesurvivalof non-floweringclonesandfloweringclonescouldnotbeestimated.

Thefirstannualtransition(2008–2009)fornon-floweringclones andfloweringcloneswasthereforeestimatedusingthenextannual transition(2009–2010).

AnanalysisofalifeTableresponseexperiment(LTRE)wasused toestimatetherelativeimportanceofgrazinglevel,yearandpop- ulationsonthevariationin(Caswell,2001);anestedfactorial

(5)

designwasused(ElderdandDoak,2006;Endelsetal.,2007a)and aLTRElinearmodelincludinggrazinglevel(i),yearlytransition(j), andpopulationsnestedwithini(l(i)):

_(ij)=^(..)+˛⁽ⁱ⁾+ˇ^(j)+˛ˇ^(ij)+^l(i),

where^(.)istheprojectedpopulationgrowthrateofthemeanref- erencematrix,␣⁽ⁱ⁾and␤^(j)arethemaineffects,␣=grazingintensity effectand␤=yeareffectsatthei^thlevelofgrazingandthej^thlevel ofyearlytransition,␣␤^(ij)istheinteractioneffectand␪^l(i)corre- spondtothenestedeffectsofpopulationlwithingrazingleveli.

Estimatesofthetreatmentseffectsanddecompositionintocon- tributionsfromeach matrixelementweredoneasdescribedin Caswell(2001) and Elderd andDoak (2006). Toinvestigatethe importanceofthedifferentlife-historycomponentstodifferences inpopulationgrowthrate,weseparatelysummedallpositiveand negative contributionsfrom the matrix elementswithin a life- historycomponent.Toestimate95%conﬁdenceintervalsforthe growthratesandLTREeffectsweusedabootstrappingprocedure with3000resamplings(Caswell,2001).Theeffectofmanagement andtemporal variationonthegrowthrates estimatedfromthe matrixmodel,andthevariationingrowthratesamongpopulations werealsoanalysedusingGLMMandthesamebackwardelimina- tionproceduresasdescribedabove.

AllstatisticalanalyseswereperformedinR(RCoreTeam,2013).

Thelinearmodelswereimplementedusingthelmerfunctioninthe lme4package(Batesetal.,2015).Thematrixmodellingprocedures andLTRElinearmodelwereconductedusingthepackagepopbio (StubbenandMilligan,2007).

3. Results

The numbersof inflorescences per ramet and proportion of floweringrametsinK.arvensispopulationswerehighestinaban- donedgrasslands(mean±SEinflorescences:2.88±0.082,p<0.05;

mean±SEﬂoweringramets:0.36±0.008,p<0.05)andlowestin grasslandswithhighgrazingintensity(mean±SEinﬂorescences:

0.45±0.030,p<0.05;mean±SEﬂoweringramets: 0.11±0.005, p<0.05).Theproportionofseedlingswasgreaterinpopulations underhighgrazingintensity(mean±SEseedlings:0.11±0.005, p<0.05), except in 2009. Grazing intensity did not inﬂuence ramet survival but survival waslower in 2009(mean±SEsur- vival: 0.78±0.011, p<0.05) than in 2008 (mean±SE survival:

0.84±0.009,p<0.05).Theeffectofgrazingintensityonthepro- portionofclonalreproduction(floweringandnon-floweringclonal rosettes)andgrazedrametsvariedgreatlybetweenyearsandthere werenooveralltrendsbutsignificantinteractioneffectsbetween yearandgrazingintensity(interactioneffects:p<0.05;Table2;

Fig.1).

Theprojectedpopulationgrowthrate()rangedfrom0.83to 2.13indicatingthatpopulationsbothincreasedanddecreasedin size.Elasticityanalysisoftheoverallmeantransitionmatrixindi- catedthatthesurvivalofnon-ﬂoweringrosetteswasthelife-cycle transitionthathadthelargestimpact(elasticity=0.235)ontheprojectedpopulationgrowthrate(Fig.3).Thepopulationgrowthrate variedsigniﬁcantlyamongyears(␹²=5.35,p=0.02)withthehigh- estgrowthratesin2008–2009.Thedidnotvarywithgrazinglevel (␹²=4.28,p=0.12;Fig.2).Thegrowthratesvariedamongpopula- tionsunderno(␹²=19.74,p<0.01)andhigh(␹²=4.87,p=0.02) grazingintensity,butnotamongpopulationsingrasslandsunder lowgrazingintensity(␹²=0.76,p=0.39).Population1(nograzing intensity)hadthehighestgrowthrate(2.13).K.arvensisdidnot establishinthecontrolplotswherethespecieswasabsentin2008.

The LTRE analysis revealed that the main effect of grazing intensity (mean±SD effect: |0.201±0.103|) contributed more to variation in population growth rate than both year

no low high

Grazing intensity Sum contributions to λ −0.3−0.2−0.10.00.10.20.30.4

Fig.4. Overallgrazingintensityeffect(no,low,high)tovariationinthepopulation growthrate(␭)fromalifeTableresponseexperiments(LTRE)variancedecompositionanalysis.Theerrorbarsarebootstrapped95%conﬁdenceintervals.

(mean±SD effect: |0.003±0.001|) and interaction (mean±SD effect:|0.020±0.0280|)betweenyearandgrazingintensityon. Nograzinghadanetpositiveeffecton␭whilebothlowandhigh grazingintensityhadanegativeeffect(Fig.4).Inabandonedplots (nograzing),sexualreproduction(F;fecundity)waslowcausing anegativeeffectonthegrowthrate(Fig.5).However,thisnega- tiveeffectwascounterbalancedbyamuchlargerpositiveeffectof clonalreproduction(CR)whichinsumcausedthepositiveeffect onpopulationgrowthrates.Ingrazed plots(bothlow andhigh grazing intensity),thenegativeeffectwascausedby lowclonal reproduction(CR).Plotsunderlowgrazingintensityalsoshowed lowsexualreproduction(F).Plotsunderhighgrazingintensityon theotherhand,showedhighsexualreproduction(F),butthiswas counterbalancedbyanevenlowerclonalreproduction(CR).Among populationsundernograzingintensity,theneteffecton␭varied greatlywhiletheneteffectvariedlessamongthegrazedpopula- tions(Fig.S1intheonlineversionatDOI:10.1016/j.ﬂora.2016.04.

007).Thepositivecontributionofclonalreproductiontowaspar- ticularlyclearinoneofthepopulationsundernograzingintensity (Fig.S2intheonlineversionatDOI:10.1016/j.ﬂora.2016.04.007).

4. Discussion

In our study area Knautia arvensis ramets were grazed by cattle.Grazing reducedthenumber ofinﬂorescences perramet and the proportion of ﬂowering ramets withinthe population.

Thiscouldbeduetothecattle’s preferencefor K.arvensis.It is knownthatK.arvensisimprovesNefﬁciencyforruminantsinvitro (Hoffmannetal.,2008),suggestingthatitmightbeapreferredplant amongcattle.Grazingalsoaffectedseedlingestablishmentwithin populations,whichwaslowestinpopulationswithinabandoned grasslands.Thisisnotsurprisingastheabsenceofsmall-scaledis- turbances(e.g.trampling)causesfewgapsonlyinwhichseedlings canestablish(ErikssonandEhrlen,1992).Lowseedlingestablish-

(6)

Table2

Estimatedfixedeffectsstatisticsforlinearmixedmodelswithresponses:survivaltothenextyear(t+1)(survivals),proportionofseedlings(seedlings),proportionof floweringramets(floweringramets,numberofinflorescencesperfloweringramet(inflorescences),proportionofclonalreproduction(clonalreproduction),andproportion ofgrazedramets(grazedramets).Theexplanatoryvariablesstudiedwere:year(2008,2009,2010),grazingintensity(no,low,high)andinteractionsbetweenyearsand grazingintensities.Modelswerefittedwithbackwardselectionandboldteststatisticsindicatep<0.05.

Survivals^a Seedlings Floweringramets Inﬂorescences Clonalreproduction^b Grazedramets^c

Explanatoryvariable Estimate z Estimate z Estimate z Estimate z Estimate z Estimate z

Intercept 1.819 8.303 −6.387 −5.559 −0.605 −1.480 −0.129 0.142 1.396 −6.588 0.262 0.419

Year2009 −0.718 −4.956 1.685 1.395 −0.453 −2.730 0.064 1.502 −1.087 −5.896

Year2010 1.768 1.488 −0.043 −0.265 −0.078 −1.815 0.761 4.146 −0.932 −5.221

Grazinglow −12.769 −0.013 0.507 0.710 −0.618 −4.632 0.358 0.923 −1.461 −1.420

Grazinghigh 4.376 3.648 −2.693 −4.036 −0.864 −3.497 −0.127 −0.398

Year2009:Grazinglow 11.742 0.012 −1.090 −3.608 0.215 3.444 2.344 5.672

Year2009:Grazinghigh −4.457 −3.415 0.293 0.874 0.167 1.096

Year2010:Grazinglow 12.377 −1.919 −1.073 −3.679 0.270 4.420 −0.754 −2.183 1.184 3.061

Year2010:Grazinghigh −2.365 −1.919 0.338 1.080 −0.355 −1.937 −1.204 −4.357

aYear2010notincludedinthemodel.

b Year2008notincludedinthemodel.

c Nograzingnotincludedinthemodel.

SS SC S EF R CR F

no

Sum contributions to λ −0.10−0.050.000.050.100.150.20

SS SC S EF R CR F

low

Life history component

−0.10−0.050.000.050.100.150.20

SS SC S EF R CR F

high

−0.10−0.050.000.050.100.150.20

Fig.5.Grazingintensityeffect(no,low,high)tovariationinthepopulationgrowthrate(␭)fromalifeTableresponseexperiments(LTRE)variancedecompositionanalysis.

Transitionsinthematrixaregroupedaccordingtolife-historycomponents:seedlingsurvival(SS),survivalofclonaloffspring(SC),stasis(S),enteringfloweringstage(EF), retrogression(R),clonalreproduction(CR)andfecundity(F).Retrogressionisdefinedasthetransitionfromfloweringtonon-floweringstage.

mentinabandonedgrasslandshasbeenreportedinseveralstudies (Hamreetal.,2010;Poschlodetal.,2011).

Theeffectofmanagementwasmuchlargerthantheyeareffects ontheprojectedpopulationgrowthrate.However,theclonaltran- sitionratesin 2008–2009wereestimatedfromt+1,and hence therewasnovariationinthesetransitionsbetweenyears.Boththe variationinclonaltransitionsbetweenyearsandtheyeareffectin theLTREanalysisarethereforelikelytobeunderestimated.

Eventhoughgrazingreducedtheproportionofﬂoweringram- etsandnumberofinﬂorescencesperrametandincreasedseedling establishment,thisdidnotaffectpopulationgrowthratessignif- icantly.Alackofeffect ofmanagement onthegrowthratecan indicatethatK.arvensisisnotdependentonmanagedsemi-natural grasslandstopersist.However,ahabitatmodellingstudyshows thateventhoughK.arvensisoccursinseveralvegetationtypesin Norway,semi-naturalgrasslandsarebyfaritsmostcommonhabi-

tat(HovstadandGrenne,2012).Thelargeimpactontheoverall growthratesbysurvivalofnon-ﬂoweringrosettesindicatesthat thelongevityoframetsisimportantforpopulationsurvival.Herba- ceousplantswithlonglife-spansareknowntohavemorestable populationsbecauseforsuchspeciessurvivalislessvariablethan fecundity(Garciaetal.,2008).Populationsoflong-livedorganisms maybeabletobuffernegativeeffectsonvitalrateseveninsmall populationsizes(Kolbetal.,2010).Longevitycanthereforebuffer temporalﬂuctuationsinthepopulationsize.

Inthegrazedgrasslandsinourstudyarea,sexualreproduction contributedthemosttothegrowthratewhereasin abandoned grasslands clonal reproductioncontributed themost. Fecundity andclonalgrowthareknowntobenegativelycorrelatedforclonal plantspecies(Silvertownetal.,1993).Theproportionofseedlings was largest in populations under high grazing intensity, even thoughtheproportionofﬂoweringrametswasleast.Thismeans

(7)

thatthefewﬂoweringrametsinthepopulationsunderhighgraz- ingintensitycontributesigniﬁcantlytothepopulationgrowthrate;

sexualreproductioncontributesmoretopopulationgrowthrate ingrazed grasslandscompared toabandonedgrasslands. Grime (2001)states thatresponsestodisturbanceofteninvolveadap- tiveshiftsinregenerativetraits.Knautiaarvensisisaspecieswith multiplereproductionstrategies,whichisanadvantageinvariable environments(Grime,2001).Severalremnantpopulationsresist extinctionbecauseparts oftheirlife cycle(Eriksson, 1996)and lifehistorystagesotherthanﬂoweringrametsmaintainthepop- ulation.Populationsinabandonedsemi-naturalgrasslandscanbe remnantpopulationsthatsurvivefordecadesandhavehighgrowth ratesduetohighsurvivalandclonalreproduction(Endelsetal., 2007b;Hamreetal.,2010;Johanssonetal.,2011).Theresultsin thisstudyshowthatK.arvensisisabletopersistevenifitshabi- tatchanges,buttheresultsalsoindicatethatthespeciescanform remnantpopulationswithverylowratesofsexualregenerationif semi-naturalgrasslandsareleftwithoutmanagement.

Thethreegrasslands withoutpresent domestic grazerswere allformer haymeadows in which hayharvestingended inthe mid-1980s.Thetwograsslandswiththelowestpopulationgrowth rateswereoccasionallygrazed,andthenunderverylowstocking rates,byhorsesandsheepuntilaboutyear2000.Thethirdgrass- landwiththeexceptionallyhighpopulationgrowthrateswasnot grazedatallbydomesticanimalsafterabandonment.Thesedif- ferencesinmanagementhistorymayexplainthelargedifferences inclonalreproductionamongthegrasslandswithoutgrazingin thisstudy.LindborgandEhrlén(2002)studiedtheperennialherb Primulafarinosaandfoundthatafterabandonmentofgrazingby domesticanimalsthepopulationsexperiencedaperiodofposi- tivegrowthrate.However,bycomparinghistoricaland present distributions they concluded that most populations in habitats wheregrazinghadceased,hadgoneextinct.Theexceptionallyhigh growthrateofK.arvensisinoneofthepopulationsinabandoned grasslandsislikelytolastonlyashorttime.Theincreasedpopu- lationgrowthmightthereforebefollowed byaﬁnal population decreaseas forP.farinosa (Lindborg andEhrlen, 2002).Topre- serveviable populationsofplantspecies inhabitingspecies-rich grasslands,suchas K.arvensis,the effects of differentmanage- mentregimesmustbemonitoredandevaluatedinordertoguide managementdecisions.Itisessentialtounderstandthereproduc- tionandsurvivalprocessinordertoestimateiftherearedelayed responsesinpopulationsize.

5. Conclusions

Knautia arvensis populations tolerate both grazing and short-termabandonmentinsemi-naturalgrasslandandtherewere nodifferencesingrowthratesbetweenpopulationsingrazedand abandonedgrasslands.Alonglife-spanandsurvivalofthespecies can,however,bufferchangingenvironmentalconditions.Inpop- ulations in grazed grasslands,fertility contributed more tothe growthratethaninabandonedgrasslandswhereclonalregenera- tioncontributedthemost,whichindicatesatrade-offbetweenthe life-historycomponents.

Acknowledgements

Thisstudyispartoftheproject“Sauidrift”fundedbytheNorwe- gianResearchCouncil(projectno.208036/010).Thestudyhasalso receivedfundingfromtheEuropeanCommunity’sSeventhFrame- workProgramme(FP7/2007-2013)underthegrantagreementn^◦ FP7-244983(MULTISWARD).Theauthorsthankthelandownersfor grantingthemaccesstotheﬁeldsites,SynnøveNordalGrenne,Liv

Nilsen,LineRosef,andVibeckeMelhuusforﬁeldassistanceandan anonymousreviewerforvaluablecomments.

References

Bates,D.,Maechler,M.,Bolker,B.,Walker,S.,2015.Fittinglinearmixed-effects modelsusinglme4.J.Stat.Softw.67(1),1–48.

Berg,H.,2002.PopulationdynamicsinOxalisacetosella:thesigniﬁcanceofsexual reproductioninaclonal,cleistogamousforestherb.Ecography25,233–243.

Brys,R.,Jacquemyn,H.,Endels,P.,DeBlust,G.,Hermy,M.,2004.Theeffectsof grasslandmanagementonplantperformanceanddemographyinthe perennialherbPrimulaveris.J.Appl.Ecol.41,1080–1091.

Cahenzli,F.,Erhardt,A.,2012.NectarsugarsenhanceﬁtnessinmaleCoenonympha pamphilusbutterﬂiesbyincreasinglongevityorrealizedreproduction.Oikos 121,1417–1423.

Caswell,H.,2001.MatrixPopulationModels:Construction,Analysisand Interpretation,2nded.SinauerAssociatesInc.,USA.

Clausen,H.D.,Holbeck,H.B.,Reddersen,J.,2001.Factorsinﬂuencingabundanceof butterﬂiesandburnetmothsintheuncultivatedhabitatsofanorganicfarmin Denmark.Biol.Conserv.98,167–178.

RCoreTeam,2013.R:ALanguageandEnvironmentforStatisticalComputing.R FoundationforStatisticalComputing,ViennaAustria.

Crawley,M.J.,2007.TheRBook.JohnWileyandSonsLtd.,UnitedKingdom.

Ehrlen,J.,Syrjanen,K.,Leimu,R.,Garcia,M.B.,Lehtila,K.,2005.Landuseand populationgrowthofPrimulaveris:anexperimentaldemographicapproach.J.

Appl.Ecol.42,317–326.

Elderd,B.D.,Doak,D.F.,2006.Comparingthedirectandcommunity-mediated effectsofdisturbanceonplantpopulationdynamics:ﬂooding,herbivoryand Mimulusguttatus.J.Ecol.94,656–669.

Endels,P.,Jacquemyn,H.,Brys,R.,Hermy,M.,2007a.Geneticerosionexplains deviationfromdemographicresponsetodisturbanceandyearvariationin relicpopulationsoftheperennialPrimulavulgaris.J.Ecol.95,960–972.

Endels,P.,Jacquemyn,H.,Brys,R.,Hermy,M.,2007b.Reinstatementoftraditional mowingregimescounteractspopulationsenescenceintherareperennial Primulavulgaris.Appl.Veg.Sci.10,351–360.

Eriksson,O.,Ehrlen,J.,1992.Seedandmicrositelimitationofrecruitmentinplant populations.Oecologia91,360–364.

Eriksson,O.,1996.Regionaldynamicsofplants:areviewofevidenceforremnant, source-sinkandmetapopulations.Oikos77,248–258.

Franzen,M.,Nilsson,S.G.,2008.Howcanwepreserveandrestorespeciesrichness ofpollinatinginsectsonagriculturalland?Ecography31,698–708.

Garcia,M.B.,Pico,F.X.,Ehrlen,J.,2008.Lifespancorrelateswithpopulation dynamicsinperennialherbaceousplants.Am.J.Bot.95,258–262.

GeologicalSurveyofNorwayNGU,2014.Datasets(accessed01.10.14)https://ngu.

no/en/topic/datasets.

Grime,J.P.,2001.Plantstrategies.In:ProcessesandEcosystemProperties.Wiley, Chichester.

Hamre,L.N.,Rydgren,K.,Halvorsen,R.,2010.Theeffectsofmulchingand abandonmentontheviabilityoftheperennialgrasslandspeciesPlantago lanceolata.PlantEcol.211,147–158.

Hodgson,J.G.,Grime,J.P.,Wilson,P.J.,Thompson,K.,Band,S.R.,2005.Theimpacts ofagriculturalchange(1963–2003)onthegrasslandﬂoraofCentralEngland:

processesandprospects.BasicAppl.Ecol.6,107–118.

Hoffmann,E.M.,Seje-Assmann,N.,Becker,K.,2008.Dosestudieson

anti-proteolyticeffectsofamethanolextractfromKnautiaarvensisoninvitro ruminalfermentation.Anim.FeedSci.Technol145,285–301.

Hovstad,K.A.,Grenne,S.N.,2012.Habitatmodellingwithhighresolutioninan agriculturallandscape.In:EuropeanGrasslandFederation24thGeneral Meeting,Lublin,Poland3–7June.GrasslandScienceinEurope17,pp.682–684.

Hovstad,K.A.,2007.Seeddispersalandseedlingestablishmentinsemi-natural grassland.In:PhDThesis.NorwegianUniversityofLifeSciences,Ås.

Jacquemyn,H.,Brys,R.,2008.Effectsofstandageonthedemographyofa temperateforestherbinpost-agriculturalforests.Ecology89,3480–3489.

Johansson,V.A.,Cousins,S.A.O.,Eriksson,O.,2011.Remnantpopulationsandplant functionaltraitsinabandonedsemi-naturalgrasslands.FoliaGeobot.46, 165–179.

Kålås,J.,Viken,Å.,Henriksen,S.,Skjelseth,S.,2010.The2010NorwegianRedList forSpecies.NorwegianBiodiversityInformationCentre,Norway.

Knight,T.M.,2004.Theeffectofherbivoryandpollenlimitationonadeclining populationofTrilliumgrandiﬂorum.Ecol.Appl.14,915–928.

Kolb,A.,Dahlgren,J.P.,Ehrlen,J.,2010.Populationsizeaffectsvitalratesbutnot populationgrowthrateofaperennialplant.Ecology91,3210–3217.

Larsson,M.,2005.Higherpollinatoreffectivenessbyspecialistthangeneralist ﬂower-visitorsofunspecializedKnautiaarvensis(Dipsacaceae).Oecologia146, 394–403.

Lennartsson,T.,Oostermeijer,J.G.B.,2001.Demographicvariationandpopulation viabilityinGentianellacampestris:effectsofgrasslandmanagementand environmentalstochasticity.J.Ecol.89,451–463.

Lid,J.,Lid,D.,2005.NorskFlora.Detnorskesamlaget.

Lindborg,R.,Ehrlen,J.,2002.Evaluatingtheextinctionriskofaperennialherb:

demographicdataversushistoricalrecords.Conserv.Biol.16,683–690.

Louault,F.,Pillar,V.D.,Aufrere,J.,Garnier,E.,Soussana,J.F.,2005.Planttraitsand functionaltypesinresponsetoreduceddisturbanceinasemi-natural grassland.J.Veg.Sci.16,151–160.

(8)

Marini,L.,Fontana,P.,Scotton,M.,Klimek,S.,2008.VascularplantandOrthoptera diversityinrelationtograsslandmanagementandlandscapecompositionin theEuropeanAlps.J.Appl.Ecol.45,361–370.

Maron,J.L.,Crone,E.,2006.Herbivory:effectsonplantabundance,distributionand populationgrowth.Proc.R.Soc.B:Biol.Sci.273,2575–2584.

Metera,E.,Sakowski,T.,Sloniewski,K.,Romanowicz,B.,2010.Grazingasatoolto maintainbiodiversityofgrassland-areview.Anim.Sci.Pap.Rep.28,315–334.

Moen,A.,1999.NationalAtlasofNorway:Vegetation.NorwegianMapping Authority,Norway.

Morris,W.F.,Pﬁster,C.A.,Tuljapurkar,S.,Haridas,C.V.,Boggs,C.L.,Boyce,M.S., Bruna,E.M.,Church,D.R.,Coulson,T.,Doak,D.F.,Forsyth,S.,Gaillard,J.-M., Horvitz,C.C.,Kalisz,S.,Kendall,B.E.,Knight,T.M.,Lee,C.T.,Menges,E.S.,2008.

Longevitycanbufferplantandanimalpopulationsagainstchangingclimatic variability.Ecology89,19–25.

Norderhaug,A.,Ihse,M.,Pedersen,O.,2000.Biotopepatternsandabundanceof meadowplantspeciesinaNorwegianrurallandscape.Landsc.Ecol.15, 201–218.

NorwegianMeteorologicalInstitute,2014.eKlima(accessed01.10.14)http://

www.eklima.met.no.

Poschlod,P.,WallisDeVries,M.F.,2002.Thehistoricalandsocioeconomic perspectiveofcalcareousgrasslands-lessonsfromthedistantandrecentpast.

Biol.Conserv.104,361–376.

Poschlod,P.,Hoffmann,J.,Bernhardt-Roemermann,M.,2011.Effectofgrassland managementontheageandreproductionstructureofHelianthemum nummulariumandLotuscorniculatuspopulations.Preslia83,421–435.

Pykala,J.,Luoto,M.,Heikkinen,R.K.,Kontula,T.,2005.Plantspeciesrichnessand persistenceofrareplantsinabandonedsemi-naturalgrasslandsinnorthern Europe.BasicAppl.Ecol.6,25–33.

Silvertown,J.,Franco,M.,Pisanty,I.,Mendoza,A.,1993.Comparativeplant demography-relativeimportanceoflife-cyclecomponentstotheﬁniterateof increaseinwoodyandherbaceousperennials.J.Ecol.81,465–476.

Stubben,C.J.,Milligan,B.G.,2007.Estimatingandanalyzingdemographicmodels usingthepopbiopackageinR.J.Stat.Softw.22(11),1–23.

Thompson,K.,Bakker,J.P.,Bekker,R.M.,1997.Thesoilseedbanksofnorthwest europe:methodology.In:DensityandLongevity.UniversityPressCambridge, Cambridge.

Totland,Ø.,Hovstad,K.A.,Ødegaard,F.,Åström,J.,2013.StateofKnowledge RegardingInsectPollinationinNorway—TheImportanceoftheComplex InteractionBetweenPlantsandInsects.NorwegianBiodiversityInformation Centre,Trondheim,Norway.

Tutin,T.G.,Heywood,V.H.,Burges,N.A.,Moore,D.M.,Valentine,D.H.,Walters,S.M., Webb,D.A.,1976.FloraEuropaea.UniversityPressU.K,Cambridge.

Vandvik,V.,Vange,V.,2003.GerminationecologyoftheclonalherbKnautia arvensis:Regenerationstrategyandgeographicvariation.J.Veg.Sci.14, 591–600.

Wehn,S.,2009.Amap-basedmethodforexploringresponsestodifferentlevelsof grazingpressureatthelandscapescale.Agric.Ecosyst.Environ.129,177–181.

Wrage,N.,Strodthoff,J.,Cuchillo,H.M.,Isselstein,J.,Kayser,M.,2011.

Phytodiversityoftemperatepermanentgrasslands:ecosystemservicesfor agricultureandlivestockmanagementfordiversityconservation.Biodiv.

Conserv.20,3317–3339.

Zuur,A.F.,2009.MixedEffectsModelsandExtensionsinEcologywithR.Springer, NewYork.