<%BANNER%>

UFIR



Short-term effects of experimentally elevated precipitation and nitrogen on soil fertility and plant growth in a Neotrop...
http://www.esajournals.org/loi/ecsp ( Publisher's URL )
CITATION PDF VIEWER
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/IR00001329/00001
 Material Information
Title: Short-term effects of experimentally elevated precipitation and nitrogen on soil fertility and plant growth in a Neotropical savanna
Series Title: Copeland, S. M. E. M. Bruna, L. V. Barbosa, M. C. Mack, and H. L. Vasconcelos. 2012. Effects of experimentally elevated precipitation and nitrogen on soil fertility and plant growth in a Neotropical savanna. Ecosphere 3(4): 1-20
Physical Description: Journal Article
Creator: Bruna, Emilio
Publisher: Ecological Society of America
Place of Publication: Ecosphere
Publication Date: 19 April 2012
 Notes
Abstract: Increasing nitrogen (N) deposition and changing precipitation patterns in Neotropical savannas could alter plant growth, reproduction, and nutrients by altering soil nutrient and water availability. We examined the potential for simulated N deposition and increased dry season precipitation to have interactive effects on reproduction and growth of two abundant native Cerrado (Brazilian savanna) grasses—Loudetiopsis chrysothrix and Tristachya leiostachya—via feedbacks with soil nutrient status. Plant growth and reproduction responses consistently varied by species. Water addition led to more consistent increases in both growth and reproduction than nitrogen addition and the two treatments did have significant interactive effects. We expected that both treatments would affect plant growth and reproduction via positive effects on soil and plant N. Instead, we found that plant responses were linked to species-specific treatment effects on soil and foliar phosphorus (P). Structural equation models (SEM) confirmed that changes in soil P—rather than changes in soil N or increasing soil acidity—explained plant response to treatments. Our results imply that N deposition and precipitation change could impact Cerrado plant growth and reproduction via subtle effects on plant and soil phosphorus.
Acquisition: Collected for University of Florida's Institutional Repository by the UFIR Self-Submittal tool. Submitted by Emilio Bruna.
Publication Status: Published
 Record Information
Source Institution: University of Florida Institutional Repository
Holding Location: University of Florida
Rights Management: All rights reserved by the submitter.
System ID: IR00001329:00001

Downloads

This item is only available as the following downloads:

Copeland_etal_2012_Ecosphere ( PDF )


Full Text

PAGE 1

Short-termeffectsofelevatedprecipitationandnitrogen onsoilfertilityandplantgrowthinaNeotropicalsavannaSTELLAM.COPELAND,1,5, EMILIOM.BRUNA,1,2LAURAV.BARBOSASILVA,1,4MICHELLEC.MACK,3ANDHERALDOL.VASCONCELOS41DepartmentofWildlifeEcologyandConservation,UniversityofFlorida,P.O.Box110430, Gainesville,Florida32611-0430USA2CenterforLatinAmericanStudies,UniversityofFlorida,P.O.Box115531,Gainesville,Florida32611-0430USA3DepartmentofBiology,UniversityofFlorida,P.O.Box118526,Gainesville,Florida32611-0430USA4InstitutodeBiologia,UniversidadeFederaldeUberl ˆ a ndia,C.P.593,Uberl ˆ a ndia,MinasGerais38400-902Brazil Citation: Copeland,S.M.,E.M.Bruna,L.V.BarbosaSilva,M.C.Mack,andH.L.Vasconcelos.2012.Short-termeffectsof elevatedprecipitationandnitrogenonsoilfertilityandplantgrowthinaNeotropicalsavanna.Ecosphere3(4):31.http:// dx.doi.org/10.1890/ES11-00305.1Abstract.Increasingnitrogen(N)depositionandchangingprecipitationpatternsinNeotropical savannascouldalterplantgrowth,reproduction,andnutrientsbyalteringsoilnutrientandwater availability.WeexaminedthepotentialforsimulatedNdepositionandincreaseddryseasonprecipitation tohaveinteractiveeffectsonreproductionandgrowthoftwoabundantnativeCerrado(Braziliansavanna) grasses— Loudetiopsischrysothrix and Tristachyaleiostachya —viafeedbackswithsoilnutrientstatus.Plant growthandreproductionresponsesconsistentlyvariedbyspecies.Wateradditionledtomoreconsistent increasesinbothgrowthandreproductionthannitrogenadditionandthetwotreatmentsdidhave significantinteractiveeffects.Weexpectedthatbothtreatmentswouldaffectplantgrowthand reproductionviapositiveeffectsonsoilandplantN.Instead,wefoundthatplantresponseswerelinked tospecies-specifictreatmenteffectsonsoilandfoliarphosphorus(P).Structuralequationmodels(SEM) confirmedthatchangesinsoilP—ratherthanchangesinsoilNorincreasingsoilacidity—explainedplant responsetotreatments.OurresultsimplythatNdepositionandprecipitationchangecouldimpact Cerradoplantgrowthandreproductionviasubtleeffectsonplantandsoilphosphorus.Keywords: Brazil;bunchgrass;Cerrado; Loudetiopsischrysothrix ;nitrogenenrichment;phosphorus;precipitation change;savanna;soilfertility; Tristachyaleiostachya Received 4November2011;revisedandaccepted5March2012; published 19April2012.CorrespondingEditor:D.P.C. Peters. Copyright: 2012Copelandetal.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommons AttributionLicense,whichpermitsrestricteduse,distribution,andreproductioninanymedium,providedtheoriginal authorandsourcesarecredited.5Presentaddress:DepartmentofEnvironmentalScienceandPolicy,UniversityofCalifornia,Davis,California95616 USA. E-mail: scopeland@ucdavis.eduINTRODUCTIONAnthropogenicnitrogen(N)additionhasmore thandoubledpre-industrialnitrogeninputsto terrestrialenvironments(reviewedinSchlesinger 2009).IncreasedNdepositionchangessoil acidityandmacronutrientavailability(Vitousek etal.1997,Aberetal.1998)whichaltersplant biomassallocation,phenology,fitness,andcompetitiveinteractions(Clelandetal.2006,Clark andTilman2008,Lauetal.2008).TheeffectsofN depositiononplantgrowthandsoilfertilitycan affectecosystemnetprimaryproduction,plant diversity,andglobalclimateviaeffectsoncarbon v www.esajournals.org1April2012vVolume3(4)vArticle31

PAGE 2

cycling(Vitouseketal.1997,Gruber2008). However,thedirectionandmagnitudeofthe plantresponsestoNdepositioncanvarywidely duetothenutrientrequirementsofdifferent plantfunctionalgroupsandspecies(Craineetal. 2002,Zavaletaetal.2003)aswellasthe limitationsimposedbysoilfertilityandregional climate(Bobbinketal.2010). Inadditiontoincreasingnitrogendeposition, human-causedincreasesinatmosphericCO2are alteringthetimingandabundanceofprecipitationworld-wide(Zhangetal.2007).Plant phenologyandgrowthcanbehighlysensitive tochangesinprecipitation(Fayetal.2003, Zavaletaetal.2003,KochyandWilson2004), particularlyinecosystemswithlongdryseasons wherenitrogenmineralizationandplantnitrogenuptakearesynchronizedwithseasonal precipitationpatterns(Austinetal.2004,Knapp etal.2006,Yahdjianetal.2006).ResultsfromN depositionandprecipitationexperimentssuggest thattheseglobalchangefactorscanhave unpredictableandinteractiveeffectsonplant floweringandgrowth(Zavaletaetal.2003, Clelandetal.2006,Henryetal.2006,Siemann etal.2007). Becausemanytropicalsavannasandgrasslandsexperienceprofounddry-seasons,they maybeparticularlysensitivetochangesinthe timing(Knappetal.2002)orabsoluteamount (PandeyandSingh1992,KochyandWilson 2004)ofprecipitation.Tropicalsavannasmay alsobehighlysensitivetonitrogendeposition becauseplantproductivitytendstoprimarily limitedorco-limitedbyN(Bargeretal.2002, Augustine2003,Sarmientoetal.2006).InNlimitedgrasslandsandsavannas,plantresponse toprecipitationamountandtimingislinkedto theeffectsofrainfallonseasonalNmineralizationpatterns(SeagleandMcNaughton1993, Austinetal.2004,Yahdjianetal.2006)and mass-flowofinorganicNtoplantroots(Borken andMatzner2009).Despiteprojectionsofsharp increasesinNdeposition(Phoenixetal.2006) andpredictionsforprecipitationchange(Christensenetal.2007)intropicalsavannas,few experimentshaveexplicitlytestedhownative plantsandsoilsarelikelytorespondtothese globalchanges(reviewedbyMatsonetal.1999, Bobbinketal.2010).Theuniquecharacteristicsof tropicalsavannascouldleadtoresponsesto globalchangesthatdivergefromthoseobserved intheextensiveexperimentalresultsfromtemperategrasslands(e.g.,FisherandWhitford1995, Carreraetal.2003,Zavaletaetal.2003,Kochy andWilson2004,Clelandetal.2006). Amongtropicalsavannas,theNeotropical BrazilianCerradoisremarkableforitsbiodiversity(10,000plantspecies)andextent(2million km2)(OliveiraandMarquis2002).Cerrado ecosystemfunctionanddiversityisthreatened byland-usechange,invasivegrasses,andurbanization(Ratteretal.1997).Inaddition,increases inurbanfossilfuelcombustionandNfertilizer useareexpectedtomorethandoubleN depositionratesintheCerrado,fromanaverage of5–13kgNha 1yr 1inthe1990sto14–38kg Nha 1yr 1by2050(Bustamanteetal.2006, Phoenixetal.2006).Ndepositionratesofthis magnitudeareassociatedwithplantspeciesloss anddecreasedproductivityinNorthAmerica (ClarkandTilman2008)andEurope(Stevenset al.2004)andhavethepotentialforsimilar negativeimpactsonCerradospeciesandecosystemprocesses(reviewedinBobbinketal.2010). NitrogendepositioncoulddecreasesoilfertilityandplantgrowthintheCerradoandother nutrientpoortropicalecosystemsbydecreasing cationandphosphorusavailability,increasing acidity,andraisinglevelsoftoxicaluminum (Matsonetal.1999).However,thepotentialfor negativeeffectsmaydependuponplantandsoil demandforaddedNgivenextremelynutrient poorCerradosoils.Ndepositionismorelikelyto havenegativeeffectsonCerradoplantspeciesif Pistheco-limitingorprimarynutrientlimiting growthbecausePislikelytobereducedwith increasedsoilacidity.P-limitationinCerrado ecosystemshasbeensuggestedbysomeprevious experimentsandobservationalstudies(Bustamanteetal.2006,Nardotoetal.2006,Kozovitset al.2007).However,thereisalsoevidenceforN limitingorco-limitingplantproductivity—asis oftenthecaseintemperatesavannasand grasslands(Bustamanteetal.2006).Conflicting evidencefornutrientlimitationintheCerrado mayresultfromtheapparentwiderrangeof nutrientuseefficienciesamongNeotropical savannaandrainforestspeciescomparedto temperatespecies(Bustamanteetal.2004,Townsendetal.2007). Herewereporttheresultsofayear-long v www.esajournals.org2April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 3

experimenttestingtheinteractiveeffectsofN depositionandprecipitationchangeonplant biomassallocationandsoilsintheCerradowith twodominantC-4nativegrassspecies,focusing onthelinksbetweensoilandplantresponses (Fig.1).Weexperimentallyaddednitrogenin amountsandratesconsistentwithnitrogen depositionlevelsfortheregionforthecoming century(Phoenixetal.2006).Webasedourwater additiontreatmentonclimatechangemodels whichpredictincreasingdryseasonprecipitation intheCerradoundermoderateclimatechange scenarios,thoughsomemodelsandclimate changescenariosalsopredictdecreasesinprecipitation(Magrinetal.2007). Ourfocalspecieswerechosenbasedontheircodominanceinourstudysystem,suggestingthat anyspecies-specificresponsestoglobalchange factorscouldhaveimportantimplicationsforplant andsoilpropertiesecosystem-wide.Forexample, increasedgrowthofthedominantgrassspecies couldnegativelyaffectwoodyplantrecruitment andaffecttherelativeabundanceofwoodyplants andgrasses,akeyelementofsavannastructure andecosystemdynamics(Gardner2006,Furley 2007).Inaddition,studieswithco-dominantC-4 grassspeciesinNorthAmericanprairies(Silletti andKnapp2001,Swemmeretal.2006,Nippertet al.2009)havedemonstratedthatco-dominant speciesmayresponsedifferentiallytoglobal changesduetotrade-offsinresourceuseand acquisitionstrategies. Ourresearchaddressedtwocentralquestions forpredictingtheeffectsofthecombinationofN depositionandprecipitationchangeonCerrado ecosystems.First,weasked:HowdoNaddition andincreaseddryseasonwateravailabilityaffect thenutrientstatusofinfertileCerradosoils?We expectedthatbothtreatmentscouldincreasesoil N,eitherdirectlywithNaddition,orindirectly withwateradditionviathepositiveeffectsofsoil moistureonplant-availableN.Incontrast,we expectedthatNadditioncouldhaveanegative effectonsoilfertilityifNadditionincreasedsoil acidity,whichcouldfurtherdecreasetheavailabilityofPandessentialcationsinacidic, nutrient-poorCerradosoils.Second,weaddressedthequestion:HowdoNandwater additionaffectgrowthandreproductionofcodominantgrassspecies?Weexpectedthatboth treatmentswouldleadtoincreasedgrowthand reproductionbecauseNanddry-seasonwater arepotentiallylimitinginthisecosystem. Finally,ourapproachallowedustoevaluate whetherplantresponsestoN-depositionanddry seasonwateradditionwereduetothepositive (fertilizing)ornegative(toxicity)plant-soilfeedbacks.Specifically,weexpectedthattreatment effectsonplantgrowthandreproductionwould belinkedtochangesinchangesinsoilessential nutrients,pH,andtoxicaluminumlevelsvia increasesordecreasesinleafsenescence,foliar nutrients,androot:shootratio(Fig.1).We constructedstructuralequationmodelstotestfor supportofpositivevs.negativeplant-soilfeedbacksandtocomparetherelativeimportanceof directversusindirect(soil-mediated)effectsofNdepositionandwateradditiononplantresponses. Fig.1.ConceptualframeworkforpotentialpositiveandnegativeeffectsofNdepositionanddry-season precipitationincreaseonplantsviaeffectsonsoilfertility. v www.esajournals.org3April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 4

MATERIALSANDMETHODSStudysiteandspeciesThisstudywasconductedattheEstac a o Ecolo gicadoPanga,a404hareservelocated40 kmfromUberl ˆ a ndia,MinasGerais,Brazil(19 8 100S,48 8 230W).Monthlyaveragetemperatures rangefrom20–25 8 C,andannualrainfallis approximately1600mmwithanalmostrainless dryseasonfromMayandSeptember(Instituto deGeografia2008).Soilsarehighlyweathered OxisolswithahighclaycontentandlowpH (LatossoloVermelho-Amarelho,Braziliansoil taxonomyEMPRAPA1999,AnionicAcrustoxe, USsoiltaxonomy,SoilSurveyStaff2003).Our studywasconductedin cerradoralo ,avegetation physiognomytypifiedbydensegrasscover interspersedwithsmalltreesandshrubs(Cardosoetal.2009).Thepreserveisprotectedfrom grazingandotheragriculturalactivities,butis subjecttooccasionalanthropogenicfiresoriginatingonadjoiningroads.Themostrecentfirein thestudyareaoccurred2yearsbeforethestartof ourexperiment. OurfocalspeciesweretwonativeC-4perennialbunchgrassesintheTribeArundinellae, Tristachyaleiostachya and Loudetiopsischrysothrix (referredtohereafterbygenus). Tristachya is generallylargerthan Loudetiopsis :theaverage Tristachya genet(individualbunchgrass)is25cm indiameter,whereas Loudetiopsis genetsare10 cminaveragediameter. Tristachya vegetative tillersare90cmtallonaverageand Loudetiopsis vegetativetillersare70cmtallonaverage. Togethertheyaccountfor69 % oftheabovegroundbiomassinthestudyarea( Loudetiopsis 12 % Tristachya 57 % ,allotherspecies 5 % each,E.M.BrunaandH.L.Vasconcelos, unpublisheddata ),withpeakbiomassandfloweringoccurringbetweenFebruaryandApril.Both speciesalsohavebroaddistributions: Tristachya rangesfromsouthernBraziltoParaguayand Loudetiopsis fromeasternBoliviatosouthern BrazilandParaguay(MissouriBotanicalGarden 2009).ExperimentaldesignInMay2008werandomlyselectedN 80 individualbunchgrassesofeachspeciesina150 3 200mareaofhomogenousaspect,slope,and vegetationcover.Theplantswerelocatedalong6 transects50mapart;transectswere150m(5 transects)or50m(1transect)long.Individuals werewithin1/3ofthemediandiameterfor individualsofthatspecies( Tristachya :18.3–36.7 cm, Loudetiopsis 9–18cm,basedonarandom sampleofN 15individualsinthestudyarea) andallplantsusedintheexperimentwereat least2mfromanyotherfocalplants.We establishedplotsof50 3 50cmaroundeach selectedindividualandrandomlyassignedN 20individualsofeachspeciestooneoffour treatments:nitrogenaddition,wateraddition, nitrogenaddition 3 wateraddition,andunmanipulatedcontrols.Wereducedthepotentiallyconfoundingeffectsofabovegroundcompetitionbyclippingallabove-groundbiomass surroundingthefocalindividualbeforethe treatmentsandthroughouttheexperimentat2– 3weekintervals.Allbelow-groundbiomassand leaflitterwereleftundisturbedtolimiteffectson decompositionandroots.TreatmentsWeaddedatotalof25kgha 1yr 1N(2.5g m 1yr 1N)toplantsintheNtreatmentin accordancewithpredictedNdepositionlevels fortheregion(approximately12kgha 1yr 1N, withamaximumofapproximately38kgha 1yr 1Nby2050;Phoenixetal.2006).Thenitrogen wasaddedasammoniumnitrate(NH4NO3)in fourapplicationsofslow-releasecommercial fertilizer(31 % N;Manah,Gu ¨ngeFertilizantesS/ A,Uberaba,MG,Brazil)sprinkledevenlyonsoil surfaceintheplot(June,September,andDecember2008,February2009).Ammoniumnitrate (50 % eachion)isareasonableapproximationof Ndepositionforthissite:from1997–1999Nwet depositionnearUberl ˆ a ndiawas48 % ammonium and38 % nitrate(LilienfeinandWilcke2004). ClimatemodelsfortheCerradoregionpredict bothincreasesanddecreasesinrainfallby2099 undermodestclimatechangescenarios(Christensenetal.2007).Wechosetoexperimentally addprecipitationinourexperimentsbasedon increasingregionalprecipitationtrendsoverthe last40years(Haylocketal.2006).Weadded watertoplantsintheprecipitationaddition treatmentwithdripirrigationatarateoftwo litersper24hours(8mm/day)inthemiddleof thedryseasonof2008(June–August);each treatmentwasdividedbyalternatedryperiods v www.esajournals.org4April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 5

of2and9days.Intotalweadded72mmof water,approximately5 % oftheaverageannual rainfallorseventimestheambient2008dry seasonprecipitation(9.5mmfromJune–August 2008,datafromUberl ˆ a ndia,40kmfromsite, 2003–2008,InstitutodeGeografia2008).Our dailywateradditionratewasasubstantial increaseoveraveragemonthlydryseason precipitationrates(11.0 6 13.9mmrainfall/ month[mean 6 SD]fromJune–August2004– 2008)andwascomparabletohalftheaverage dailyprecipitationrateduringthewetseason (October2003–April2004,13.3 6 4.0mm/day). Asaresult,averagedailyvolumetricwater content(m3/m3)inwateredplotswasapproximatelythreetimesgreaterthaninun-watered plots(seasonmean 0.03,lower95 % CL0.02, upper95 % CL0.04vs.seasonmean 0.01,lower 95 % CL0.01,upper95 % CL0.02,respectively) andaverage,maximum,andminimumdailysoil moistureweresignificantlydifferent(p 0.01) betweenwateredandun-wateredplotsduring thewateradditionperiod(seeAppendix).LightavailabilityBecause,lightlimitationcandecreaseN limitationandshadymicroclimatestendto amelioratewaterlimitationintropicalsavannas (Cruz1997,Ludwigetal.2001),wequantified shadingforeachindividualbymeasuringphotosyntheticallyactiveradiation(PAR)inourplots withan80cmlongquantumlinesensor ceptometer(AccuparLP80,DecagonDevices, PullmanWA)andincorporatedrelativelight availabilityasacovariateinouranalyses.PAR wasquantifiedbetween11amand2pmon March3,2009andwasmeasuredattheheightof thetallestleavesofeachindividualtoquantify shadingbyoverstoryvegetation(primarilytrees andshrubs).Werecordedtheaverageofthree measurementsthroughthecenteroftheplot takenatdifferentorientationsparalleltothe ground.SoilandfoliarnutrientsTomeasureresin-availableN(NH4 and NO3 ),weinstalledmixedbedresinbagsinthe top10cmofsoilsurrounding76ofour experimentalplants(N 32 Loudetiopsis andN 44 Tristachya ).Resinswerechargedwith1M NaClandextractedwith2MKClafter28daysof fieldincubation(February–March2009).Bulk inorganicNpoolswerequantifiedbysampling theupper10cmofsoilinMarch2009(N 50 Loudetiopsis andN 49 Tristachya ).Wemeasured Navailabilitybyextracting10gsamplesoffieldmoistsoilwith2MKClandadjustedforsoil gravimetricwatercontentandbulkdensity. ConcentrationsofNO3 andNH4 insoiland resinextractswereanalyzedcolorimetrically withanAstoriaAutoanalyzer(Astoria-Pacific, Clackamas,OR,USA).Bothionicformswere analyzedbecauseofthepotentialofthetreatmentstohavedifferentialeffectsonthetwosoil ions.Forexample,wateradditioncouldincrease bothmicrobialmineralizationofammoniumand nitrificationofnitrate,whiletheeffectofN additiononsoilNformsmightdependonplant andmicrobeNdemandanduptake. Wequantifiedtheeffectsofourexperimental treatmentsonindicatorsofsoilacidityand toxicity(pHandAl),phosphorus,andessential cations(KandCa)usingsoilsamplescollectedin March2009anddriedat55 8 Cfor48hours(N 10pertreatment 3 speciescombination).We measuredpHindeionizedwater(ratio:1:2.5 soil:H2O).Potassium(K)andphosphorous(P) wereextractedwithMehlich(HCl-H2SO4)solution(Kratio:sample:solution10:1,Psample: solution20:1).Kconcentrationwasdetermined byflameemissionspectrophometry(ModelNo. B462,Micronal,S a oPaulo,SP,Brazil)andP concentrationwithUV/Visiblespectrophometry (Cary50ConcUV-Vis,Varian,PaloAlto,CA, USA).Weextractedaluminum(Al)andcalcium (Ca)in1MKCl(Caratio:100:1,Alratio:10:1 sample:solution).AlconcentrationwasdeterminedbytitrationwithNaOHinthepresenceof bromothymolblue.Caconcentrationswere determinedwithatomicabsorptionspectrophometry(ModelNo.932A,GBCScientific Equipment,Dandenong,VIC,Australia). Wemeasuredtheeffectsofourtreatmentson foliarNandPbysub-sampling20green undamagedleavesfromeachindividual.Change inspecificleafareawasnotquantifiedbecause thismetricwouldhaverequiredremovinga significantproportionofanindividual  sgreen leavespre-treatmentwithpotentialeffectson otherplantresponsestotreatments.Leaveswere washedwithdeionizedH2O,driedfor48hours at60 8 C,andgroundinaplantmill(Marconi v www.esajournals.org5April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 6

Equipamentos,MA048,Piricicaba,SP,Brazil). FoliarNwasdeterminedbyKjeldahldigestion, steamdistillationofthedigestintoH3BO3,and titrationwithH2SO4.WeextractedfoliarPfrom groundtissuewithHNO3andHClO4and determineddigestPconcentrationwithUV/ Visiblespectrophometry(ModelNo.Cary50 ConcUV-Vis,Varian,PaloAlto,CA,USA).Plantreproductiveeffort,growth, andbiomassallocationBunchgrassesreproduceintwodistinctmanners(1)throughseedlingrecruitmentand(2) productionoflateraltiller(Tomlinsonand O  Connor2004).Forourstudywequantified reproductionbasedonmeasuresofreproductive effortandallocationrelatedtoseedlingrecruitmentbymeasuring(1)floweringvs.nonfloweringindividuals,(2)thenumberoffloweringtillersperindividual,(3)thetotalnumberof spikelets,and(4)allocationofspikeletsper floweringtiller.Eachspikeletcorrespondedto onefertilefloretwhichcouldpotentiallyhave developedintooneviableseedandweobserved thattheflowerstructuresofbothspecieswere consistentwithsexualreproduction.However, wedidnotdirectlymeasurethenumberof developedseeds,theirviability,orlossdueto herbivory.Measurementsweremadebothbefore andafterthetreatmentswereapplied(May2008 andMarch2009,respectively). Wemeasuredthediameteraroundthebaseof eachgenet(individualbunchgrass)beforeand10 monthsafterthetreatmentswereappliedand calculatedpercentgrowthbecauseoftheeffectof sizeongrowthrate.Weusedbunchdiameterto quantifygrowthbecausebunchgrassabovegroundbiomassistightlycorrelatedwithdiameterinmanybunchgrassspecies(Nafusetal. 2009)duetobunchgrassgrowththroughtiller production(TomlinsonandO  Connor2004) allowingdiametertobeusedasasurrogate measureofgrowthforgrasseswiththishabit (e.g.,Martyetal.2005).Wealsodidnotobserve anyevidenceofrhizomatousgrowthforanyof theindividualsweexcavated.Forourcontrol individualsofourfocalspecieswefoundthat diameterissignificantlycorrelatedwithtotal abovegroundbiomass( Loudetiopsis :R2 0.23,p 0.001,N 79. Tristachya :R2 0.28,p 0.001,N 80)andtotalmeristems( Loudetiopsis :R2 0.40, p 0.001,N 79. Tristachya :R2 0.45,p 0.001, N 80). Toestimateroot:shootratio,wecollectedall plantsatthepeakofthegrowingseason(4weeks March–April2009,collectiondaterandomizedto eliminatesystematiceffectsofcollectiondateon biomass),separatedtheroots,liveleaves,dead leaves,floweringstems,andfloweringspikes (i.e.,floretsandseeds),anddriedthematerialat 55 8 Cuntilthesamplesreachedconstantweight (2–4days).Werecoveredasmuchoftheroot biomassaspossiblebytrenchingaroundthe perimeteroftheplotandexcavatingtobelowthe depthofthemainrootmass(i.e.,approximately 15cmfor Loudetiopsis and20cmfor Tristachya individuals).Wetestedtheefficacyoftheroot collectionmethodbycollectingfive6cmwide 3 5cmdeepcoresinthesoilremainingintheplot areaafterplantswereremoved,sievedforcoarse rootmass(2mmsieve),anddriedthematerialto constantweightat55 8 C.Rootmassremainingin thesoilaveraged0.002 6 0.001g/cm2(mean 6 SE),whiletheaveragerootmassofourexcavatedgrasseswas62.33 6 38.28gfor Loudetiopsis and399.69 6 255.83gfor Tristachya ,suggesting thatwewereabletorecoverthevastmajorityof thegrassrootsystems.Finally,weestimated reductioninleafsenescencebycountingallgreen leavesontheplantanddividingthenumberby genetareainAugustof2008tocalculategreen leafdensity.Highergreenleafdensitycorrespondedtomoregreenleavesperplantarea (decreasedleafsenescence)duringthedry season.StatisticalanalysesTreatmenteffects .—Totestforspeciesand treatmenteffectsonsoilparameters(bulksoil NH4 ,resinavailableNO3 –,pH,MehlichPand K,KCl-extractedCaandAl)weusedgeneral linearmodelswithnitrogen,water,andspecies asfixedeffects.Treatmentandspecieseffects resinavailableNH4 andbulksoilNO3 –were analyzedwithgeneralizedlinearmodels(gammadistribution).Allplantresponsevariables wereanalyzedbyspecieswithnitrogenand waterasfixedeffects.Thelikelihoodthatan individualwouldflower(response:floweringor non-flowering)inresponsetotreatmentsand separately,toresinavailableNO3 –andNH4 , wereanalyzedwithbinomialmodels.Toexamv www.esajournals.org6April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 7

inetreatmenteffectsonthenumberofflowering tillersperfloweringindividualandnumberof totalspikeletsweusedgeneralizedlinearmodels withnegativebinomialdistributions.General linearmodelswereusedtoevaluatetreatment effectsonthenumberofspikeletspertiller.To testfortheeffectsofourtreatmentsongrowth, weusedanANCOVAwiththepercentdifference indiameterbetweenyearoneandyeartwoasthe responsevariable,nitrogenandwaterasfixed effectsandoriginaldiameterandaveragePARas covariates.Wetestedfortreatmenteffectson senescence(numberofgreenleavesbyarea)with anANCOVAwithoriginaldiameterasacovariate.Wecomparedtheresponseofroot:shoot ratiostoourtreatmentswithagammadistributedgeneralizedlinearmodel.Weanalyzedthe effectsoftreatmentsontotalaboveground biomass,totaldeadleafbiomass,andtotallive leafbiomasswithgenerallinearmodels.We testedfordifferencesinfoliarNandPconcentrationsandN:Pratiosinresponsetotreatments withanANCOVAwithliveleafmassasthe covariatetocontrolforthepossibilitythata changeinleafbiomasscouldhaveaffected nutrientconcentration. Structuralequationmodels .—Weusedstructural equationmodels(SEMs)toevaluate(1)apriori hypothesesforpositiveornegativedirector indirecteffects(viasoils)oftreatmentsonplants (illustratedinFig.1)and(2)anaposteriori modelforhypothesizedrelationshipsbetween treatments,soils,andplantsbasedontheresults ofourunivariatetests(Grace2006).Boththea prioriandaposterioritestswere  confirmatory  usesofstructuralequationmodels.ThisapplicationofSEMsallowedustotestwhetherour experimentaldataconfirmedthehypothesized mechanisticrelationshipsbetweenvariablesin thesystem(GraceandPugesek1998).Structural equationmodelswhichfailtofitthedatacanbe rejectedbasedonp-values(e.g.,p 0.05)while modelswhichcannotberejected(e.g.,p 0.05) areconsideredadequatemodelsforthedata structure(Grace2006).Non-significantp-values forvariablerelationships(paths)inacceptable modelsdonotindicatepoorfitoftheover-all model(Mitchell1992).Forbothtypesofmodels, standardizedcoefficientsarecalculatedtoallow fordirectcomparisonofrelationshipsbetween variablesdespitedifferencesinmeasurement scale(Grace2006). Forourapriorimodelsweconstructedmodels forourhypothesizedrelationshipsbetween treatment,soilandplantcharacteristicsfor negativeandpositivesoil-plantfeedbacks(both positiveandnegativepathsillustratedinFig.1). IfNadditionledtonegativeplant-soilfeedbacks weexpectedthatNadditionwouldincreasesoil acidity,leadingtohigherlevelsoftoxicAl, decreasedP,andultimatelyreducedplant growth(diameterchange)andreproduction (numberoffloweringculms).IfwaterandN additionledtopositivesoil-plantfeedbackswe expectedthatwateradditionwouldincreaseN mineralization(resinNO3andNH4)andreduce leafsenescencewhileNadditionwouldlead increasedavailableN(resinNO3andNH4) leadingtoincreasesinfoliarN:Pandplant growthandreproduction.Weexpectedthat specieswouldstronglyaffectplantresponses giventhepotentialforcoexistingspeciestohave divergentresourceuseandresponsestochanges inNandwateravailability.Thenegativesoilplantfeedbackmodelwasconstructedbyomittingtheprecipitationtreatmentbecausewater additionwasnotexpectedtonegativelyaffect soilfertilityandincludingsoilP,Al,andpHas indicatorsofsoiltoxicity.Thepositivesoil-plant feedbackmodelincludedboththewaterand nitrogentreatments,species,soilN(resinNO3 NH4 ),foliarN:P,numberoffloweringculms, percentdiameterchange,andlight(PAR). WeconstructedouraposterioriSEMbasedon theANOVAresultsfortreatmenteffectsonsoil andplantvariables.Inthismodel,Pwastheonly soilvariableincludedandbothwaterand nitrogentreatmentswereincorporatedintothe model.Modelfitandparametervalueswere calculatedwithmaximumlikelihoodestimation usingthecovariance-variancematrix(Grace 2006).P-valuesformodelfitarebasedonchisquarevalues.Coefficientspresentedarestandardizedvalues.Samplesizewas80forall models. Modelassumptionsandv ariabletransformations .—Forallmodels,thehomogeneityof variancewasevaluatedwithLevene  stestand logtransformationswereappliedwhennecessarytomeettheassumptionofnormallydistributedresidualsforallgenerallinearmodelsand bivariatenormalityforstructuralequationmodv www.esajournals.org7April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 8

els.Foralltransformationsandnon-normal distributions(gamma,binomial,andnegative binomial)wereportback-transformedmeans and95 % confidenceintervals.StructuralequationmodelswereanalyzedinR2.10.1(sem package).Allotheranalyseswereperformedin SAS9.2(SASInstitute,Cary,NC,USA).RESULTSSoilresponsesOverall,wateradditionhadsignificanteffects onsoilphosphorusandmineralized(resins)and inorganicnitrogenpools(bulksoil).Ontheother hand,theonlysignificanteffectofnitrogen additionwasanincreaseinbulknitratepools. Therewasalsolittleindicationofincreasingsoil acidity,aluminumconc entrations,orcation leachingwithaddedN,thoughtherewasa slighttrendtowardsincreasedcalciumand decreasedpotassium(Table1). Wateradditionledtoincreasedresin-available ammoniumandnitrateovera28-dayperiod duringthepeakofthegrowingseason(Table1). Resin-availableammoniumvalues(controls:0.84 6 0.148 l gN[mean 6 SE])tendedtobehigher thannitratevalues(controls:0.632 6 0.196 l gN). Resin-availableNO3 increased69 % (H2Oaddition:1.07 6 0.176 l gN)whileresinNH4 increased71 % (H2Oaddition:1.44 6 0.175 l g N)withaddedwater.Nadditionandspecies identitydidnotsignificantlyaffectresin-availableNofeitherionicform(Table1). IncontrasttoresultsforresinNO3 ,water additionhadamarginallysignificantnegative effectonbulksoilNO3 poolscomparedto controlvalues(control:0.105 6 0.019,H2O:0.103 6 0.019;Fig.2,Table1).SoilpoolsofNO3 almostdoubledwithnitrogenadditionascomparedtocontrolslevels(Naddition:0.206 6 0.038 l gN/g;Fig.2,Table1).However,when waterwasaddedwithnitrogenNO3 concentrationsdecreasedtolevelsbelowcontrolvalues(N andH2Oaddition:0.104 6 0.019 l gN/g;Fig.2) andsimilartovaluesforplotsreceivingonlythe watertreatment,suggestinganinteractiveeffect ofthetwotreatments(p 0.08;Table1).Water andnitrogenadditiondidnotsignificantlyaffect bulksoilNH4 (Table1).BulksoilNH4 values (controls:2.493 6 0.068 l gN/g)wereordersof magnitudegreaterthanNO3 levels(control: 0.105 6 0.019 l gN/g).Speciesidentitydidnot haveasignificanteffectonbulksoilNO3 or NH4 (Fig.2,Table1). Soilphosphorouswassignificantlyaffectedby wateradditionbutthedirectionoftheeffect variedbyfocalplantspecies(Fig.3).Water additionledtoaslightdecreaseinsoilPinsoils associatedwith Loudetiopsis (0.104 6 0.010, control:0.115 6 0.013g/kg)whilesoilPincreased in Tristachya plotsreceivingaddedwater(0.151 6 0.017,controls0.117 6 0.013g/kg).Pincreasedin soilsassociatedwithbothspecieswithN addition(Fig.3, Loudetopsis :0.154 6 0.022, Tristachya :0.138 6 0.019g/kg).For Tristachya thecombinationofNandwateradditionledto greaterincreasesinsoilP—42 % overcontrol values(0.168 6 0.027g/kg)—thanwithNalone. Table1.F-statisticsandp-valuesfortheeffectsoftreatments(nitrogen,water,andtheirinteraction)andspecies identity( Loudetiopsis or Tristachya )onresinammoniumandnitrate,soil-extractableammoniumandnitrate, pH,Al,K,andCa.Variable NitrogenWaterWater 3 NitrogenSpecies FpFpFpFp ResinNH4 , 0.010.984.72 0.03*1.160.280.180.67 ResinNO3 0.010.993.70 0.06 0.830.370.010.93 SoilNH4 0.750.390.210.650.120.730.530.47 SoilNO3 3.42 0.07 3.76 # 0.06 3.03 # 0.08 0.010.94 pH0.380.540.050.830.100.752.240.14 Aluminum1.600.211.090.300.890.350.740.39 Potassium3.00 # 0.09 0.080.770.420.522.490.12 Calcium3.15 0.08 0.090.760.360.550.010.93 Notes: Arrowstodenotethedirectionoftheeffect,eitherincrease( )ordecrease( # )forp 0.10.Degreesoffreedomand samplesize:pH,Al,K,andCa:N 80anddf 1,75.SoilNH4 andNO3 :N 99,df 1,94.ResinNH4 andNO3 :N 76,df 1,71. p 0.10. *p 0.05. v www.esajournals.org8April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 9

However,soilPconcentrationsin Loudetiopsis associatedsoilswiththecombinationofwater andNwerelowerthancontrolvalues(0.115 6 0.016)andsimilartosoilPvaluesforplots receivingonlyaddedwater(Fig.3). Cationssusceptibletoleaching(Ca,K)were mildlyaffectedbyNadditionbutwerenot affectedbywateradditionorspeciesidentity. Indicatorsofsoilacidity(pH)andtoxicity(Al) werenotaffectedanynitrogenorwateraddition orspeciesidentity(Table1).Nadditionledto marginaldecreasesinKavailability(control:8.36 6 0.04,N:8.23 6 0.06g/kg)andmarginal increasesinCa(control:0.0102 6 0.001,N: 0.0157 6 0.004g/kg)butdidnotaffectpH (control:4.97 6 0.05,N:4.92 6 0.04)orAl (control:0.061 6 0.002,N:0.067 6 0.003).Water additionandthecombinedtreatmentdidnot significantlyaffectanyofthefertilityvariables (Table1).PlantresponsesWaterandN-additioneffectsonfoliarnutrientswerehighlyvariableamongthetwospecies. FoliarPandN:Pwasaffectedbytreatmentsbut nosignificantchangeinfoliarNwasobserved withanyspeciesandtreatmentcombination. Loudetiopsis foliarPconcentrationdecreased slightlyforalltreatments,howeveronlythe interactionbetweenwaterandnitrogenapproachedsignificance(p 0.09,Table2).In contrast,therewasatrendtowardsincreased foliarPwith Tristachya foralltreatments—withN orwateradditionleadingto8 % morefoliarP overcontrolsand14 % greaterfoliarPwhereboth Nandwaterwerecombined(Table2).Nitrogen: phosphorusratios(N:P)in Loudetiopsis increased slightlywithalltreatments,thoughonlyN additionhadamarginallysignificanteffect(F 3.74,p 0.06)withN:Pratiosraisingto10.5 6 2.0(mean 6 SD)from9.5 6 1.4controlvalues.In contrast,wateradditionsignificantlydecreased Tristachya N:Pratios(p 0.02,df 1,F 5.87)to 9.0 6 1.3from9.9 6 1.7controlvalues. Therewasasignificantinteractiveeffectofthe Nandwatertreatmentsonthelikelihoodthat Loudetiopsis wouldflower(p 0.03,df 1,F 4.60):95 % oftheplantsreceivingbothtreatments floweredcomparedto80 % withN,65 % with water,and90 % forcontrols(Table3).With Tristachya, onlywateradditionhadaneffect: thepercentoffloweringindividualsincreased Fig.2.Effectoftreatmentsandspeciesonsoilnitrate( l g/kg)(N 98anddf 1,94).Treatmentssignificantat thep 0.10levelareindicatedbydifferentlower-caseletterswiththep-valueindicatedinthefigureand numeratordegreesoffreedomandF-valueinparentheses.Barsrepresent95 % confidenceintervals. v www.esajournals.org9April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 10

from45 % forcontrolsto70 % witheitherwateror waterandnitrogen(Table3).Contraryto expectations,higherresinavailableammonium valueswerecorrelatedwithdecreasingflowering probabilityfor Tristachya (p 0.02,Wald v2 5.44),whileneitherformofresinavailable nitrogenwasasignificantpredictorofflowering likelihoodfor Loudetiopsis.Loudetiopsis individualsthatdidflowerproducedapproximately doublethenumberoffloweringtillerswith waterandthecombinedNandwatertreatments (Table3).Noneofthetreatmentssignificantly affectedthenumberoffloweringtillersproduced by Tristachya individuals(p 0.10),howeverthe meannumberoffloweringtillersisalsoverylow forthisspecies(2.0 6 1.0floweringtillers/ individualincontrols). Fortheplantsthatdidflower,wealsotested whetherthenitrogenandwatertreatments influencedmeasuresoftotalreproductiveeffort—thetotalnumberofspikeletsproducedper floweringtiller—andallocationofspikeletsper Fig.3.Soilphosphorus(MehlichPg/kg)inplotsassociatedwitheachfocalspecies( Loudetiopsis :AIC 131.42, v2 5.75/df 0.16, Tristachya :AIC 114.17, v2 6.18 v2/df 0.17,N 40anddf 1,36forbothspecies). Treatmentssignificantatthep 0.10levelareindicatedbydifferentlower-caseletterswiththep-valueindicated inthefigureandnumeratordegreesoffreedomandF-valueinparentheses.Barsrepresent95 % confidence intervals. Table2.F-statisticsandp-valuesforeffectsoftreatments(nitrogen,water,andtheirinteraction)andcovariatelive leafbiomassonfoliarNandPandN:Pratio.Variable NitrogenWaterWater 3 NitrogenLiveleafbiomass LoudetiopsisTristachyaLoudetiopsisTristachyaLoudetiopsisTristachyaLoudetiopsisTristachya FpFpFpFpFpFpFpFp FoliarN1.620.210.400.530.390.530.690.411.350.250.240.634.77 # 0.03*0.070.79 FoliarP1.780.192.93 0.09 1.170.283.12 0.08 3.01 # 0.09 0.020.881.960.170.210.65 FoliarN:P3.74 0.06 1.570.211.700.205.87 # 0.02*1.180.28 0.010.997.34 # 0.01*0.020.90 Notes: SymbolsareasinTable1.Degreesoffreedomandsamplesize: Loudetiopsis: N 80,df 1,75; Tristachya :N 79,df 1, 74. v www.esajournals.org10April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 11

floweringtiller.In Loudetiopsis ,wateraddition hadasignificantpositiveeffectonthetotal numberofspikelets:wateredplantsproduced 120 % morespikeletsonaveragethancontrol plants(Table3).Nitrogenadditiondidnotaffect totalreproductiveeffortin Loudetiopsis ,butitdid leadtoamarginalincreaseintheallocationof spikeletstoeachtiller(Table3).Therewasno significantdifferenceineithertotalspikeletsper plantorspikeletsperfloweringtillerwith treatmentsfor Tristachya (p 0.10). Aswithreproductivemeasures,growth,allocation,andsenescenceresponsestotreatments variedbetweenthetwofocalspecies.While Loudetiopsis diametergrowthwasnotaffected byeitherwaterorNaddition, Tristachya diameter didincreasesignificantlywithnitrogenaddition (Table4).Forbothspecies,growthdecreased significantlywithplantdiameterandincreased withhigherphotosyntheticactiveradiation (Table4). Loudetiopsis root:shootratiodecreased withtreatments,buttherewerenosignificant treatmenteffectson Tristachya root:shootratio. Nitrogenandwateradditionbothledtodecreasesin Loudetiopsis root:shootratios,leadingtoan additivedecreaseofabout45 % inratiosintheN 3 watertreatmentcomparedtocontrols(Table 4).Wateradditiondidleadtoapproximately28 % higherdensityofgreenleavesinthedryseason for Loudetiopsis butneithertreatmentsignificantlyaffected Tristachya leafsenescence.Theincrease indry-seasongreenleafdensitywasalso significantlycorrelatedwithincreaseddead abovegroundbiomassattheendoftheexperimentfor Loudetiopsis (p 0.04,F 4.34).StructuralequationmodelsStructuralmodelswereusedtotestforsupport oftwoaprioricompetinghypotheses(positive vs.negative)fortheeffectsofNdepositionand dry-seasonprecipitationchangeonplantgrowth andreproductionviachangesinsoilfertilityand plantnutrientstatusorleafsenescenceinthedry season(bothhypothesesrepresentedinpathsin Fig.1).Thepositiveplant-soilfeedbackmodel testedforthepotentialofNdepositionandwater additiontoincreasesoilN(resinnitrateand ammonium)andfoliarnutrients(N:P)andfor wateradditiontoreduceleafsenescencewith positiveeffectsonplantgrowthandreproduction.Thishypothesizedmodelofrelationships betweenvariableswassupportedbythedata (modelcouldnotberejected, v2 16.9,df 24,p 0.86,pathssignificantatp 0.05:water addition resinnitrate,species leafsenescence,floweringculms,foliarN:P,resinammonium foliarN:P).However,therewasno supportforthenegativeplant-soilfeedback modelincorporatingthepotentialforNaddition tohavetoxiceffectsonsoilsviaincreasingsoil acidityandpHandincreasedtoxicAl( v2 66.08, df 18,p 0.01). Wealsoincorporatedtheresultsofthe ANOVAtestsoftreatmenteffectsintoafinala posterioriSEMforthetreatmenteffectsonplant andsoils.Weretainedphosphorusastheonly soilfertilityfactorinthefinalmodelbecause ANOVAtestsconsistentlyfoundsignificant effectsoftreatmentsonplantandsoilphosphorusvariablesbutweakly(N,Ca,K,pH,Al)or equivocallysupported(somesoilNresults significantothersnotsignificant)treatmenteffectsonothersoilvariables.Weincludeddirect Table3.F-statisticsandp-valuesforeffectsoftreatments(nitrogen,water,andtheirinteraction)onreproductive variablesbyspecies.Variable NitrogenWaterWater 3 Nitrogen LoudetiopsisTristachyaLoudetiopsisTristachyaLoudetiopsisTristachya FpFpFpFpFpFp Floweringprobability 1.070.30 0.010.99 0.010.994.99 0.03*4.60 0.03* 0.010.99 Tillers/individual1.410.240.770.385.07 0.03* 0.010.973.85 0.05 0.810.37 Spikelets/individual1.710.200.010.924.77 0.03*0.910.351.160.292.140.15 Spikelets/tiller3.32 0.07 0.420.520.110.741.750.190.880.350.190.67 Notes: SymbolsareasinTable1.Degreesoffreedomandsamplesize:Floweringprobability: Loudetiopsis :N 80,df 1,75; Tristachya :N 79,df 1,74.Tillers/individual: Loudetiopsis :N 66,df 1,62; Tristachya :N 46,df 1,42.Spikelets/individual andspikelets/tiller: Loudetiopsis :N 66,df 1,62; Tristachya :N 43,df 1,39. TeststatisticisWald v2forthebinomialdistribution. v www.esajournals.org11April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 12

pathwaysbetweenourtreatmentsandplant responsevariablesbecauseANOVAresultssuggestedthateffects,suchastheeffectofN additionongrowth,wereuncorrelatedwith measuredtreatmenteffectsonsoilvariables. ThefinalmodelincludingPastheonlysoil variablewasstronglysupportedbythedata( v2 7.87,df 14,p 0.89;Fig.4).Speciesexerted strong,largelysignificant,effects(indicatedby relativelyhighstandardizedcoefficientvalues) onallresponsevariablesbutdiameterchangein thefinalmodel(Fig.4).Incontrast,thewaterand nitrogentreatmentshadstrong,butnotsignificant,effectsondifferentresponsevariables(e.g., NonsoilP,Wonnumberoffloweringculmsand senescence,Fig.4).Lightwasjustifiedasan importantcovariateinplantresponsetothe treatmentsforbothspecieswithastrong significantpositiveeffectonplantdiameter growth(Fig.4).DISCUSSIONOurresultssupportthehypothesisthatnitrogenenrichmentanddryseasonwateradditionat ratesprojectedunderglobalchangescenarios couldalterthegrowthandreproductionof CerradoC-4grasses.However,bothexperimentaltreatmentsinfluencedprimarilysoilandfoliar phosphorous,andnotthelevelsofsoilandfoliar nitrogenaswepredicted.Inaddition,species identitystronglyinfluencedthedirectionand natureofplantandsoilresponsesdespitethe relativesimilarityofourco-dominantstudy species(C-4grassesinthesametribe). Contrarytoourexpectations,ourNaddition treatmentdidnotsignificantlyincreaseresinavailableNandhadonlyaweakpositiveeffect onbulksoilnitrate.However,wedidobserve slightincreasesinresin-availableNofbothion formswiththewateradditiontreatment,a responseweexpectedbasedonincreasesin nitrogenmineralizationandnitrificationwith increaseddryseasonprecipitationintropical savannas(AugustineandMcNaughton2004, Bustamanteetal.2006)andtropicaldryforests (Davidsonetal.1993).Theconflictingeffectsof wateradditiononinorganicN—adecreaseinsoil nitratecoupledwithincreasesinresinammoniumandnitrate—couldbeduetoincreasedplant uptakeofmineralizedNinthedryseason (Bustamanteetal.2006).Ourinabilitytodetect consistentpositiveeffectsofourNadditionon resin-availableN—despitethelowlevelsof availableNinthesoil(mean:2.58,lower95 % CL:1.79,upper95 % CL:3.77 l gNO3 NH4 / g)—couldbeexplainedbysoilorlitterimmobilization(Aberetal.2002),gaseouslosses(Pintoet al.2006),orleachingtodeepersoillayers (Lilienfeinetal.2003).BecauseoursoilN measurementstookplaceoninthewetseason, wecannotexcludethepossibilitythatourN additionmightsignificantlyincreasedsoilN duringotherseasons. TheNadditiontreatmentinthisstudywasnot linkedtostronggrowthorreproductiveresponsesofthefocalC-4bunchgrasses.Onlyaweak growthresponsefor Tristachya wasobserved. However,asafunctionalgroup,C-4grassesare knowntopredicttheeffectsofN-depositionon Table4.F-statisticsandp-valuesforeffectsoftreatments(nitrogen,water,andtheirinteraction)onplantgrowth (covariatesareinitialdiameterandlight–PAR),allocationtorootsvs.shoots,andsenescenceasindicatedby dry-seasonliveleafdensity(covariateisplantdiameter).Variable NitrogenWaterWater 3 Nitrogen LoudetiopsisTristachyaLoudetiopsisTristachyaLoudetiopsisTristachya FpFpFpFpFpFp Diametergrowth 1.580.214.94 0.03*0.010.940.150.690.540.470.030.86 Root:shoot3.06 # 0.08 0.180.674.16 # 0.04*0.020.880.010.91 0.010.99 Dry-seasonliveleafdensity§0.480.490.020.884.80 0.03*0.230.630.320.950.030.87 Notes: SymbolsareasinTable1.Degreesoffreedomandsamplesize:Diametergrowth: Loudetiopsis :N 78,df 1,72; Tristachya :N 79,df 1,73.Root:shoot: Loudetiopsis :N 80,df 1,76; Tristachya :N 79,df 1,75.Dry-seasonliveleaf density:Bothspecies:N 80,df 1,75. Covariates: Loudetiopsis :Initialdiameter:F-value:6.62,p # 0.01; Tristachya :Initialdiameter:F-value:51.93,p # 0.01. PAR: Loudetiopsis :F-value:8.63,p 0.01; Tristachya :F-value:6.79,p " 0.01. §Covariates: Loudetiopsis :Diameter:F-value:6.99,p # 0.01; Tristachya :Diameter:F-value:24.93,p # 0.01. v www.esajournals.org12April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 13

othergrasslandecosystems(Clarketal.2007) andbunchgrassesareknowntorespondtoN addition,particularlythroughtillerrecruitment (TomlinsonandO  Connor2004).Relativelyhigh NuseefficiencyamongC-4grasses(Craineetal. 2002,Reichetal.2003,Hikosaka2004)could explaintheweakresponsetoNadditionifthe traitledtoareduceduptakeratesofaddedN. BothspecieshavelowN:Pratios(control Loudetiopsis :9.43 6 1.16[mean 6 SE], Tristachya :9.74 6 1.17),whichcanbeindicativeofnitrogen limitation(TessierandRaynal2003).However, thelackoffertilizationresponseweobservedis consistentwithrecentanalysesthatsuggestthat N:Pratiosdonotnecessarilypredictabsolute nutrientlimitationintropicalecosystems(secondarytropicalforest:Davidsonetal.2004; tropicalforests:Townsendetal.2007;African savannas:Craineetal.2008). Finally,thelackofresponseobservedcouldbe duetothelowamountofNaddedtosimulateNdepositioninthisstudy(25kgha 1yr 1). However,ourN-additionwasinadditionto ambientN-depositionwhichlikelyexceedsa previousestimateof9.5ha 1yr 1reportedclose tothestudyarea(LilienfeinandWilcke2004). ThoughourNtreatmentislowcomparedtothe amountsaddedinmanyshort-termglobal Fig.4.Finalstructuralequationmodelforrelationshipsbetweenwateraddition(bluearrows),nitrogen addition(darkgreenarrows),species(purplearrows),photosyntheticallyavailableradiation(yellowarrow),soil phosphorus(hollowarrows),andplantresponsevariables(hollowarrows).Plantandsoilresponse(endogenous) variablesaredenotedbydashedoutlinesandpredictor(exogenous)variableswithsolidoutlines.Arrows presentedcorrespondtocovariancevariablesinthemodelwiththicknesscorrespondingtostandardized coefficientvalues.Standardizedcoefficientsarealsoreportedinboxeswithsignificanceindicatedby*p 0.05, p 0.10. v www.esajournals.org13April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 14

changestudies(Henryetal.2006:70kgha 1yr 1,Bradfordetal.2007:50kgha 1yr 1, Vourlitisetal.2007:50kgha 1yr 1)itiswell withintherangeofvaluesforobservedNdepositionimpacts(Fennetal.2003,Stevenset al.2004,Baezetal.2007). Soilacidityandcationconcentrationswerenot assensitivetoourtreatmentsasresinNandsoil P,andthechangesthatwedidobservewerenot correlatedwithnegativetreatmenteffectson plantgrowthandreproduction,suggestingthat negativesoil-feedbacksarenotresponsiblefor theweakresponsesofourfocalspeciestoN additionTheseresultswerehighlightedbythe poorfitofthestructuralequationmodelwhich relatedAl,P,andsoilpHtoplantresponseto treatments.WhileNadditiondecreasedsoilpH andincreasedAlconcentration—indicatorsof soilacidity—thesedifferenceswerestatistically insignificantanduncorrelatedwithadetectable decreaseinsoilP.Theslightincreaseincalcium anddecreaseinpotassiumwiththeNtreatment didnotsupportouroriginalexpectationthat bothnutrientswoulddecreasewithNasaresult ofincreasingacidity.Overall,ourresultsdonot supporttheassertionthatnutrient-poortropical ecosystemsmightexperienceincreasedsoiltoxicityasaresultofshort-termNaddition(Matson etal.1999),orthatlowcationexchangecapacity isrelatedtosensitivitytoNaddition(Clarketal. 2007).However,theseshort-termexperimental resultsdonotprecludethepossibilitythatlongtermNdepositioncouldleadtoNleachingand soilacidity—particularlyiftheNlevelssurpass plantandmicrobecapacitytoimmobilizeN (Aberetal.1998)orifthenegativeeffectson plantsandsoilshavenon-linear,andincreasingly negative,effectsovertime(ClarkandTilman 2008). Incontrasttotheweakrelationshipbetween plantresponsesandsoilnitrogen,soilandplant phosphoruswerealteredbyourtreatments.The significant,thoughdivergent,changesinfoliarP forbothspecies—increasesfor Tristachya ,decreasesfor Loudetiopsis —werelinkedtospeciesspecificeffectsonsoilP.Thesedivergentspecies responsescouldsuggestspeciesdifferencesin phosphorusdemandanduptakerateswith differentglobalchangefactors.Inagrassland globalchangeexperimentinCalifornia,USA, changesinphosphotaseandfoliarPconcentrationsuggestedthatplantdemandforphosphorusdecreasedwithwateradditionandincreased withnitrogenaddition(MengeandField2007). Inthisstudy,thedecreasein Loudetiopsis foliarP andassociatedsoilswithNadditionandwater additionsuggestedincreasedPdemandand limitation.Incontrast,theincreasedphosphorus in Tristachyaassociatedsoilsandincreasedfoliar Pwithbothwaterandnitrogenadditionsuggest thatPlimitationinthisspecieswasreducedby bothtreatments.Additionalexperimentsare neededtoelucidatethespecies-specificmechanisms,suchasincreasedphosphataseactivity (MengeandField2007),changesinplantP uptake,plantcarboninvestmentinmycorrhizae (Treseder2004),ormorecomplexplant-microbe feedbacks(Beveretal.2010)thatmightexplain theseresults. Whiletheresultsofthestructuralequation modelsgenerallysupportpositivesoil-plant feedbacksinourstudysystem,theeffectsof nitrogenandwateradditiondependedonthe focalspeciesandthespecificgrowthorreproductiveresponsemeasured.Whilegrowthwas enhancedbynitrogenbutnotwater,theinverse wasobservedforreproduction.Wateraddition ledtohighfloweringprobabilityandincreased numbersofspikeletsandpotentiallyseednumberfor Loudetiopsis ,whilenitrogenadditionwas negativelycorrelatedwithfloweringfor Tristachya anduncorrelatedwithreproductivetraits in Loudetiopsis .ThisresultcontrastswithpreviousresearchthatstronglylinksNadditionwith higherratesofgrassfloweringinatemperate grassland(Sillettietal.2004). Thesignificantpositiveinteractionofnitrogen andwateronfloweringprobabilityinthisstudy highlightstheimportanceoftestingforthe interactiveeffectsofglobalchangefactorsand indicatesthatnitrogendepositioncouldhave morepositiveeffectsonplantgrowthand reproductionwhencombinedwithincreasing dryseasonrainfallinthisecosystem.Interactive effectsoftreatmentswereobservedformultiple plantresponsesincludingfoliarP,tillernumber in Loudetiopsis,andleafsenescencein Loudetiopsis .However,therewerealsocaseswherethe effectsofthetwotreatmentsappearedtobe simplyadditive(e.g.,decreasedroot:shootratio with Loudetiopsis ). Overall,themostimportantinteractiveeffect v www.esajournals.org14April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 15

observedwasthedifferenceintreatmenteffects onplant-soilfeedbacksbyspecies.WehypothesizedthatthetwoC-4grassspeciesusedinthe experimentwouldrespondsimilarlytothe experimentalmanipulations.However,theresponsesof Tristachya and Loudetiopsis were idiosyncratic,andspeciesidentityexplaineda largerproportionofthevariationinthedatathan ourtreatmentsdidinstructuralequationmodels (Fig.2).Thesespecies-specificresponses,comparabletothoseobservedinotherstudiesofcooccurringC-4grasses(SillettiandKnapp2001, Sillettietal.2004,Fynnetal.2005),alsosuggest particularspeciestraitswhichcouldberesponsibleforthedifferencesinresponsetosimulated globalchangefactors.Themostnotableofthese isdroughttolerance,whichcouldinfluenceleaf senescenceresponsetowateraddition(seealso Swemmeretal.2006).Ofthetwospecies, Tristachya ,whichhadlowerratesofleafsenescence,maybemoretolerantofdroughtsinceits rootsarelargerindiameterandexploitdeeper soillayersthanthoseof Loudetiopsis .Such complementaryrootstructureisknowntoallow co-occurringperennialgrassestoexploitdifferent nutrientandwatersourcesinotherecosystems (FargioneandTilman2005).Futureresearch wouldbenefitfromanexplicitunderstandingof thehydrologicalniches(Arayaetal.2011)for Tristachya and Loudetiopsis ,apotentialmechanismforcoexistenceofthesesavannagrasses, andperhapsapredictoroftheirresponsesto changingprecipitationpatterns. Fireandherbivory—notmanipulatedinour study—couldinteractwithchangingwaterand nutrientregimesandarelikelytoinfluence growthandreproductionofourfocalspecies andotherCerradoplants.Forexample,the reductionintheproportionoffloweringindividualsbetweenyearone(priortofertilization)and yeartwoofthestudy( Tristachya :94–58 % Loudetiopsis :99–83 % )couldhavebeenrelatedtoa fireinthestudyareain2006(Vasconcelosetal. 2009)sinceNeotropicalsavannagrassestendto increasefloweringinresponsetofire(Sarmiento 1992,BaruchandBilbao1999).Insectherbivory, whichwealsodidnotmanipulate,canalsohave profoundeffectsonecosystemnutrientcyclingin theCerrado.Forexample,theeffectsofleafcutterantsonbothplantandsoilN(Sternberget al.2007,Costaetal.2008,Mundimetal.2009)are analogoustotheimpactsthatlargemigrating ungulateshaveonNavailabilityinAfrican savannas(Augustine2003,Holdoetal.2007, Cechetal.2008).Howbothfireandherbivory interactwithglobalchangefactorscouldbe crucialtopredictingsoil-plantfeedbackswith climatechangeandrisingNdepositionratesin theCerradointhecomingdecades. Whileexperimentalprecipitationelicitedmore plantandsoilresponsesthannitrogenaddition inthisstudy,wecannotexcludethepossibility thatambientNdepositioninourstudyarea(9.5 kgha 1yr 1,1997–99,LilienfeinandWilcke 2004)affectedourresultsbyreducingecosystem Nlimitation.Lackofdetailedinformationon currentambientNdepositionratesisacommon limitationoftropicalNenrichmentstudies (Bobbinketal.2010).Futureresearchshould endeavortomeasureandquantifytheeffectsof backgroundNdepositioninNeotropicalsavannasliketheCerradowhichareexperiencing rapidlyrisingNdepositionrates.ConclusionsOurresultssuggestthattheincreasesin nitrogendepositionandprecipitationpredicted fortheCerradoregioncanhaveinteractive positiveeffectsonthegrowthandreproduction oftwodominantgrassspeciesviadecreased nutrientandwaterlimitation.Contrarytoour expectations,theresponsesvariedstrongly amongspeciesandwerelargelyrelatedto feedbacksbetweenplantandsoilphosphorus ratherthannitrogen.Ourresultsdemonstrate strongspecieseffectsonplant-soilfeedbacksand suggestthatspecies-specificresponsestoglobal changescouldexertsignificanteffectsonecosystemproperties.Sucharesultisproblematicfor predictingglobalchangeeffectsinNeotropical savannasgivenextremelyhighplantspecies richnessanddiversestrategiesfornutrient acquisition(Bustamanteetal.2004,Townsend etal.2008).Futureexperiments,conductedover longertime-scalesandwithdifferentCerrado functionalgroups,arenecessarytodetermineif theeffectsobservedinthisstudyarerepresentativeoflong-termecosystemresponsestochanges inprecipitationandchronicNdeposition.We suggestthatstructuralequationmodelscanoffer uniqueperspectivestoanalysisofcomplexglobal changestudies(Clarketal.2007,e.g.,Antoninka v www.esajournals.org15April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 16

etal.2009)becausetheycantestapriorimodels fordirectandindirectinteractionsbetween treatments,soils,andplantsaswellasprovide aposteriorimodelsforcomplexinteractionsthat maymotivatefutureresearch.Finally,ourfindingsemphasizethatglobalchangefactorsshould besimultaneouslymanipulatedatrealisticlevels infutureexperimentsbecauseoftheirpotential tohavecomplexnon-additiveeffectsontropical savannaplantsandsoils.ACKNOWLEDGMENTSWethanktheUniversidadeFederaldeUberl ˆ a ndia forprovidinglogisticalsupportandR.Pacheco,A. NilodaCosta,J.XavierdaSilva,J.DeMarco,J.Schafer, G.Crummer,andtheLaborato riodeAna lisesdeSolos eCalca rios,UniversidadeFederaldeUberl ˆ a ndiafor helpfulfeedbackandassistanceinthefieldandthelab. WeareindebtedtoM.BrennanandJ.ColeeoftheUFIFASStatisticsConsultingUnitandP.Gagnonfor adviceonstatisticaldesignandanalysis.Wealsothank J.Ewel,T.Emam,A.Marklein,andM.Skaerfor helpfulcommentsonthemanuscript.FinancialsupportwasprovidedbygrantsfromtheUSNational ScienceFoundation(DEB-0542287)andtheDavidand LucilePackardFoundationtoEMB,anNSFGraduate ResearchFellowshiptoSC,andaresearchgrantfrom Fapemig(APQ0457.503/07)toHLV.Thedataforthe analysesinthisarticlearearchivedwithDataDryad underrecordnumber10.5061/dryad.dg380p9q.SMC designedthestudy,performedresearch,analyzed data,andwrotethepaper,EMBcontributedtostudy design,methods,andthepaper,LVBSperformed researchandcontributedtomethods,MCMcontributedtostudydesign,methods,andpaper,andHLV contributedtothestudydesignandpaper.LITERATURECITEDAber,J.,W.McDowell,K.Nadelhoffer,A.Magill,G. Berntson,M.Kamakea,S.McNulty,W.Currie,L. Rustad,andI.Fernandez.1998.Nitrogensaturationintemperateforestecosystems—Hypotheses revisited.BioScience48:921–934. Aber,J.D.,S.V.Ollinger,C.T.Driscoll,G.E.Likens, R.T.Holmes,R.J.Freuder,andC.L.Goodale. 2002.Inorganicnitrogenlossesfromaforested ecosysteminresponsetophysical,chemical,biotic, andclimaticperturbations.Ecosystems5:648–658. Antoninka,A.,J.E.Wolf,M.Bowker,A.T.Classen, andN.C.Johnson.2009.Linkingabove-and belowgroundresponsestoglobalchangeatcommunityandecosystemscales.GlobalChange Biology15:914–929. Araya,Y.N.,J.Silvertown,D.J.Gowing,K.J. McConway,H.P.Linder,andG.Midgley.2011.A fundamental,eco-hydrologicalbasisforniche segregationinplantcommunities.NewPhytologist 189:253–258. Augustine,D.J.2003.Spatialheterogeneityinthe herbaceouslayerofasemi-aridsavannaecosystem. PlantEcology167:319–332. Augustine,D.J.,andS.J.McNaughton.2004. Temporalasynchronyinsoilnutrientdynamics andplantproductioninasemiaridecosystem. Ecosystems7:829–840. Austin,A.T.,L.Yahdjian,J.M.Stark,J.Belnap,A. Porporato,U.Norton,D.A.Ravetta,andS.M. Schaeffer.2004.Waterpulsesandbiogeochemical cyclesinaridandsemiaridecosystems.Oecologia 141:221–235. Baez,S.,J.Fargione,D.I.Moore,S.L.Collins,andJ.R. Gosz.2007.Atmosphericnitrogendepositioninthe northernChihuahuandesert:Temporaltrendsand potentialconsequences.JournalofAridEnvironments68:640–651. Barger,N.N.,C.M.D  Antonio,T.Ghneim,K.Brink, andE.Cuevas.2002.Nutrientlimitationtoprimary productivityinasecondarysavannainVenezuela. Biotropica34:493–501. Baruch,Z.,andB.Bilbao.1999.Effectsoffireand defoliationonthelifehistoryofnativeandinvader C-4grassesinaNeotropicalsavanna.Oecologia 119:510–520. Bever,J.D.,I.A.Dickie,E.Facelli,J.M.Facelli,J. Klironomos,M.Moora,M.C.Rillig,W.D.Stock, M.Tibbett,andM.Zobel.2010.Rootingtheoriesof plantcommunityecologyinmicrobialinteractions. TrendsinEcologyandEvolution25:468–478. Bobbink,R.,etal.2010.Globalassessmentofnitrogen depositioneffectsonterrestrialplantdiversity:a synthesis.EcologicalApplications20:30–59. Borken,W.,andE.Matzner.2009.Reappraisalof dryingandwettingeffectsonCandNmineralizationandfluxesinsoils.GlobalChangeBiology 15:808–824. Bradford,M.A.,H.B.Schumacher,S.Catovsky,T. Eggers,J.E.Newingtion,andG.M.Tordoff.2007. Impactsofinvasiveplantspeciesonriparianplant assemblages:interactionswithelevatedatmosphericcarbondioxideandnitrogendeposition.Oecologia152:791–803. Bustamante,M.M.C.,L.A.Martinelli,D.A.Silva,P.B. Camargo,C.A.Klink,T.F.Domingues,andR.V. Santos.2004.15Nnaturalabundanceinwoody plantsandsoilsofcentralBraziliansavannas (cerrado).EcologicalApplications14:200–213. Bustamante,M.M.C.,E.Medina,G.P.Asner,G.B. Nardoto,andD.C.Garcia-Montiel.2006.Nitrogen cyclingintropicalandtemperatesavannas.Biogeochemistry79:209–237. v www.esajournals.org16April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 17

Cardoso,E.M.I.CruzeiroMoreno,E.M.Bruna,and H.L.Vasconcelos.2009.Vegetationdynamicsin theCerrado:18yearsofecologicalsuccessionat Estac a oEcolo gicadoPanga(Uberl ˆ a ndia,MG). CaminhosdeGeografia. Carrera,A.L.,M.B.Bertiller,C.L.Sain,andM.J. Mazzarino.2003.Relationshipbetweenplantnitrogenconservationstrategiesandthedynamicsof soilnitrogeninthearidPatagonianMonte, Argentina.PlantandSoil255:595–604. Cech,P.G.,T.Kuster,P.J.Edwards,andH.O. Venterink.2008.Effectsofherbivory,fireandN-2fixationonnutrientlimitationinahumidAfrican savanna.Ecosystems11:991–1004. Christensen,J.H.,etal.2007.2007:Regionalclimate projections.Pages1–46 in S.Solomon,D.Qin,M. Manning,Z.Chen,M.Marquis,K.B.Averyt,M. Tignor,andH.L.Miller,editors.ClimateChange 2007:ThePhysicalScienceBasis.Contributionof WorkingGroupItotheFourthAssessmentReport oftheIntergovernmentalPanelonClimateChange. CambridgeUniversityPress,Cambridge,UK. Clark,C.M.,E.E.Cleland,S.L.Collins,J.E.Fargione, L.Gough,K.L.Gross,S.C.Pennings,K.N.Suding, andJ.B.Grace.2007.Environmentalandplant communitydeterminantsofspecieslossfollowing nitrogenenrichment.EcologyLetters10:596–607. Clark,C.M.,andD.Tilman.2008.Lossofplantspecies afterchroniclow-leveln itrogendepositionto prairiegrasslands.Nature451:712–715. Cleland,E.E.,N.R.Chiariello,S.R.Loarie,H.A. Mooney,andC.B.Field.2006.Diverseresponsesof phenologytoglobalchangesinagrassland ecosystem.ProceedingsoftheNationalAcademy ofSciencesUSA103:13740–13744. Costa,A.N.,H.L.Vasconcelos,E.H.M.Vieira-Neto, andE.M.Bruna.2008.DoherbivoresexerttopdowneffectsinNeotropicalsavannas?Estimatesof biomassconsumptionbyleaf-cutterants.Journalof VegetationScience19:849–854. Craine,J.M.,C.Morrow,andW.D.Stock.2008. Nutrientconcentrationratiosandco-limitationin SouthAfricangrasslands.NewPhytologist 179:829–836. Craine,J.M.,D.Tilman,D.Wedin,P.Reich,M. Tjoelker,andJ.Knops.2002.Functionaltraits, productivityandeffectsonnitrogencyclingof33 grasslandspecies.FunctionalEcology16:563–574. Cruz,P.1997.Effectofshadeonthegrowthand mineralnutritionofaC-4perennialgrassunder fieldconditions.PlantandSoil188:227–237. Davidson,E.A.,C.J.R.deCarvalho,I.C.G.Vieira, R.D.Figueiredo,P.Moutinho,F.Y.Ishida,M.T.P. dosSantos,J.B.Guerrero,K.Kalif,andR.T.Saba. 2004.Nitrogenandphosphoruslimitationof biomassgrowthinatropicalsecondaryforest. EcologicalApplications14:150–163. Davidson,E.A.,P.A.Matson,P.M.Vitousek,R.Riley, K.Dunkin,G.Garciamendez,andJ.M.Maass. 1993.ProcessesregulatingsoilemissionsofNO andN2Oinaseasonallydrytropicalforest. Ecology74:130–139. EMPRAPA.1999.Sistemabrasileirodeclassific a ode solos.CentroNacionaldePesquisadeSolos, EmbrapaSolos,RiodeJaneiro. Fargione,J.,andD.Tilman.2005.Nichedifferencesin phenologyandrootingdepthpromotecoexistence withadominantC-4bunchgrass.Oecologia 143:598–606. Fay,P.A.,J.D.Carlisle,A.K.Knapp,J.M.Blair,and S.L.Collins.2003.Productivityresponsesto alteredrainfallpatternsinaC-4-dominatedgrassland.Oecologia137:245–251. Fenn,M.E.,etal.2003.Ecologicaleffectsofnitrogen depositioninthewesternUnitedStates.BioScience 53:404–420. Fisher,F.M.,andW.G.Whitford.1995.Field simulationofwetanddryyearsintheChihuahuan Desert:Soilmoisture,N-mineralization,andionexchangeresinbags.BiologyandFertilityofSoils 20:137–146. Furley,P.A.2007.Tropicalsavannasandassociated forests:vegetationandplantecology.Progressin PhysicalGeography31:203–211. Fynn,R.W.S.,C.D.Morris,andK.P.Kirkman.2005. Plantstrategiesandtraittrade-offsinfluencetrends incompetitiveabilityalonggradientsofsoil fertilityanddisturbance.JournalofEcology 93:384–394. Gardner,T.A.2006.Tree-grasscoexistenceinthe Braziliancerrado:demographicconsequencesof environmentalinstability.JournalofBiogeography 33:448–463. Grace,J.B.2006.Structuralequationmodelingand naturalsystems.CambridgeUniversityPress, Cambridge,UK. Grace,J.B.,andB.H.Pugesek.1998.Ontheuseof pathanalysisandrelatedproceduresforthe investigationofecologicalproblems.American Naturalist152:151–159. Gruber,N.G.,andJ.N.Galloway.2008.AnEarthsystemperspectiveoftheglobalnitrogencycle. Nature45:293–296. Haylock,M.R.,etal.2006.Trendsintotalandextreme SouthAmericanrainfallin1960-2000andlinks withseasurfacetemperature.JournalofClimate 19:1490–1512. Henry,H.A.L.,N.R.Chiariello,P.M.Vitousek,H.A. Mooney,andC.B.Field.2006.Interactiveeffectsof fire,elevatedcarbondioxide,nitrogendeposition, andprecipitationonaCaliforniaannualgrassland. Ecosystems9:1066–1075. Hikosaka,K.2004.Interspecificdifferenceinthe photosynthesis-nitrogenr elationship:patterns, v www.esajournals.org17April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 18

physiologicalcauses,andecologicalimportance. JournalofPlantResearch117:481–494. Holdo,R.M.,R.D.Holt,M.B.Coughenour,andM.E. Ritchie.2007.Plantproductivityandsoilnitrogen asafunctionofgrazing,migrationandfireinan Africansavanna.JournalofEcology95:115–128. InstitutodeGeografia.2008.AlturasPluviome tricas MensaisUberl ˆ a ndia,MG2004-2008.Universidade FederaldeUberl ˆ a ndia,Uberl ˆ a ndia,MG. Knapp,A.K.,C.E.Burns,R.W.S.Fynn,K.P. Kirkman,C.D.Morris,andM.D.Smith.2006. Convergenceandcontingencyinproduction-precipitationrelationshipsinNorthAmericanand SouthAfricanC-4grasslands.Oecologia149:456– 464. Knapp,A.K.,P.A.Fay,J.M.Blair,S.L.Collins,M.D. Smith,J.D.Carlisle,C.W.Harper,B.T.Danner, M.S.Lett,andJ.K.McCarron.2002.Rainfall variability,carboncycling,andplantspecies diversityinamesicgrassland.Science298:2202– 2205. Kochy,M.,andS.D.Wilson.2004.Semiaridgrassland responsestoshort-termvariationinwateravailability.PlantEcology174:197–203. Kozovits,A.R.,M.M.C.Bustamante,C.R.Garofalo, S.Bucci,A.C.Franco,G.Goldstein,andF.C. Meinzer.2007.Nutrientresorptionandpatternsof litterproductionanddecompositioninaNeotropicalsavanna.FunctionalEcology21:1034–1043. Lau,J.A.,J.Peiffer,P.B.Reich,andP.Tiffin.2008. Transgenerationaleffectsofglobalenvironmental change:long-termCO2andnitrogentreatments influenceoffspringgrowthresponsetoelevated CO2.Oecologia158:141–150. Lilienfein,J.,andW.Wilcke.2004.Waterandelement inputintonative,agri-andsilviculturalecosystems oftheBraziliansavanna.Biogeochemistry67:183– 212. Lilienfein,J.,W.Wilcke,L.Vilela,M.A.Ayarza,S.D. Lima,andW.Zech.2003.Soilfertilityundernative CerradoandpastureintheBraziliansavanna.Soil ScienceSocietyofAmericaJournal67:1195–1205. Ludwig,F.,H.deKroon,H.H.T.Prins,andF. Berendse.2001.Effectsofnutrientsandshadeon tree-grassinteractionsinanEastAfricansavanna. JournalofVegetationScience12:579–588. Magrin,G.,C.GayGarc a,D.CruzChoque,J.C. Gime nez,A.R.Moreno,G.J.Nagy,C.Nobre,and A.Villamizar.2007.2007:LatinAmerica.Pages 581–615 in M.L.Parry,O.F.Canziani,J.P. Palutikof,P.J.vanderLinden,andC.E.Hanson, editors.ClimateChange2007:Impacts,Adaptation andVulnerability.ContributionofWorkingGroup IItotheFourthAssessmentReportoftheIntergovernmentalPanelonClimateChange.CambridgeUniversityPress,Cambridge,UK. Marty,J.T.,S.K.Collinge,andK.J.Rice.2005. ResponsesofaremnantCalifornianativebunchgrasspopulationtograzing,burningandclimatic variation.PlantEcology181:101–112. Matson,P.A.,W.H.McDowell,A.R.Townsend,and P.M.Vitousek.1999.TheglobalizationofN deposition:ecosystemconsequencesintropical environments.Biogeochemistry46:67–83. Menge,D.N.L.,andC.B.Field.2007.Simulated globalchangesalterphosphorusdemandinannual grassland.GlobalChangeBiology13:2582–2591. MissouriBotanicalGarden.2009.w3-TROPICOS. MissouriBotanicalGarden,St.Louis,Missouri, USA. Mitchell,R.J.1992.Testingevolutionaryandecological hypothesesusingpathanalysisandstructural equationmodeling.FunctionalEcology6:123–129. Mundim,F.M.,A.N.Costa,andH.L.Vasconcelos. 2009.Leafnutrientcontentandhostplantselection byleaf-cutterants, Attalaevigata ,inaNeotropical savanna.EntomologiaExperimentalisEtApplicata 130:47–54. Nafus,A.M.,M.P.McClaran,S.R.Archer,andH.L. Throop.2009.Multispeciesallometricmodels predictgrassbiomassinsemidesertrangeland. RangelandEcologyandManagement62:68–72. Nardoto,G.B.,M.M.D.Bustamante,A.S.Pinto,and C.A.Klink.2006.Nutrientuseefficiencyat ecosystemandspecieslevelinsavannaareasof CentralBrazilandimpactsoffire.Journalof TropicalEcology22:191–201. Nippert,J.B.,P.A.Fay,J.D.Carlisle,A.K.Knapp,and M.D.Smith.2009.Ecophysiologicalresponsesof twodominantgrassestoalteredtemperatureand precipitationregimes.ActaOecologica,InternationalJournalofEcology35:400–408. Oliveira,P.S.,andR.J.Marquis,editors.2002.The cerradosofBrazil:ecologyandnaturalhistoryofa Neotropicalsavanna.ColumbiaUniversityPress, NewYork,NewYork,USA. Pandey,C.B.,andJ.S.Singh.1992.Rainfalland grazingeffectsonnetprimaryproductivityina tropicalsavanna,India.Ecology73:2007–2021. Phoenix,G.K.,etal.2006.Atmosphericnitrogen depositioninworldbiodiversityhotspots:theneed foragreaterglobalperspectiveinassessingN depositionimpacts.GlobalChangeBiology 12:470–476. Pinto,A.D.,M.M.C.Bustamante,M.DaSilva,K.W. Kisselle,M.Brossard,R.Kruger,R.G.Zepp,and R.A.Burke.2006.Effectsofdifferenttreatmentsof pasturerestorationonsoiltracegasemissionsin thecerradosofcentralBrazil.EarthInteractions10. Ratter,J.A.,J.F.Ribeiro,andS.Bridgewater.1997.The Braziliancerradovegetationandthreatstoits biodiversity.AnnalsofBotany80:223–230. Reich,P.B.,C.Buschena,M.G.Tjoelker,K.Wrage,J. Knops,D.Tilman,andJ.L.Machado.2003. v www.esajournals.org18April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 19

Variationingrowthrateandecophysiologyamong 34grasslandandsavannaspeciesundercontrastingNsupply:atestoffunctionalgroupdifferences. NewPhytologist157:617–631. Sarmiento,G.1992.Adaptivestrategiesofperennial grassesinSouth-AmericansavannasJournalof VegetationScience3:325–336. Sarmiento,G.,M.P.daSilva,M.E.Naranjo,andM. Pinillos.2006.Nitrogenandphosphorusaslimiting factorsforgrowthandprimaryproductionina floodedsavannaintheVenezuelanLlanos.Journal ofTropicalEcology22:203–212. Schlesinger,W.H.2009.Onthefateofanthropogenic nitrogen.ProceedingsoftheNationalAcademyof SciencesUSA106:203–208. Seagle,S.W.,andS.J.McNaughton.1993.Simulated effectsofprecipitationandnitrogenonSerengeti grasslandproductivity.Biogeochemistry22:157– 178. Siemann,E.,W.E.Rogers,andJ.B.Grace.2007.Effects ofnutrientloadingandextremerainfalleventson coastaltallgrassprairies:invasionintensity,vegetationresponses,andcarbonandnitrogendistribution.GlobalChangeBiology13:2184–2192. Silletti,A.M.,andA.K.Knapp.2001.Responsesofthe codominantgrasslandspecies Andropogongerardii and Sorghastrumnutans tolong-termmanipulations ofnitrogenandwater.AmericanMidlandNaturalist145:159–167. Silletti,A.M.,A.K.Knapp,andJ.M.Blair.2004. Competitionandcoexistenceingrasslandcodominants:responsestoneighbourremovalandresourceavailability.CanadianJournalofBotany 82:450–460. SoilSurveyStaff.2003.Keystosoiltaxonomy.Ninth edition.USDANaturalResourcesConservation Service. Sternberg,L.D.,M.C.Pinzon,M.Z.Moreira,P. Moutinho,E.I.Rojas,andE.A.Herre.2007.Plants usemacronutrientsaccumulatedinleaf-cuttingant nests.ProceedingsoftheRoyalSocietyB274:315– 321. Stevens,C.J.,N.B.Dise,J.O.Mountford,andD.J. Gowing.2004.Impactofnitrogendepositiononthe speciesrichnessofgrasslands.Science303:1876– 1879. Swemmer,A.M.,A.K.Knapp,andM.D.Smith.2006. GrowthresponsesoftwodominantC4grass speciestoalteredwateravailability.International JournalofPlantSciences167:1001–1010. Tessier,J.T.,andD.J.Raynal.2003.Useofnitrogento phosphorusratiosinplanttissueasanindicatorof nutrientlimitationandnitrogensaturation.Journal ofAppliedEcology40:523–534. Tomlinson,K.W.,andT.G.O  Connor.2004.Controlof tillerrecruitmentinbunchgrasses:unitingphysiologyandecology.FunctionalEcology18:489–496. Townsend,A.R.,G.P.Asner,andC.C.Cleveland. 2008.Thebiogeochemicalheterogeneityoftropical forests.TrendsinEcologyandEvolution23:424– 431. Townsend,A.R.,C.C.Cleveland,G.P.Asner,and M.M.C.Bustamante.2007.Controlsoverfoliar N:Pratiosintropicalrainforests.Ecology88:107– 118. Treseder,K.K.2004.Ameta-analysisofmycorrhizal responsestonitrogen,phosphorus,andatmosphericCO2infieldstudies.NewPhytologist164:347– 355. Vasconcelos,H.L.,R.Pacheco,R.C.Silva,P.B. Vasconcelos,C.T.Lopes,A.N.Costa,andE.M. Bruna.2009.Dynamicsoftheleaf-litterarthropod faunafollowingfireinaNeotropicalwoodland savanna.PlosOne4. Vitousek,P.M.,J.D.Aber,R.W.Howarth,G.E. Likens,P.A.Matson,D.W.Schindler,W.H. Schlesinger,andD.G.Tilman.1997.Human alterationoftheglobalnitrogencycle:Sources andconsequences.EcologicalApplications7:737– 750. Vourlitis,G.L.,S.Pasquini,andG.Zorba.2007.Plant andsoilNresponseofsoutherncaliforniansemiaridshrublandsafter1yearofexperimentalN deposition.Ecosystems10:263–279. Yahdjian,L.,O.Sala,andA.Austin.2006.Differential controlsofwaterinputonlitterdecompositionand nitrogendynamicsinthePatagoniansteppe. Ecosystems9:128–141. Zavaleta,E.S.,M.R.Shaw,N.R.Chiariello,B.D. Thomas,E.E.Cleland,C.B.Field,andH.A. Mooney.2003.Grasslandresponsestothreeyears ofelevatedtemperature,CO2,precipitation,andN deposition.EcologicalMonographs73:585–604. Zhang,X.B.,F.W.Zwiers,G.C.Hegerl,F.H.Lambert, N.P.Gillett,S.Solomon,P.A.Stott,andT.Nozawa. 2007.Detectionofhumaninfluenceontwentiethcenturyprecipitationtrends.Nature448:461–464. v www.esajournals.org19April2012vVolume3(4)vArticle31COPELANDETAL.

PAGE 20

SUPPLEMENTALMATERIALAPPENDIXWemeasuredvolumetricwatercontent(m3/ m3)fromJulytoAugustof2008withSoil MoistureSmartSensors(OnsetComputerCorp., Bourne,MA,USA).Sensorswerelocatedadjacenttoplantsinirrigationandcontroltreatments andrecordedameasurementevery5minutes; theyweremovedtonewplantsapproximately every10days.Duringthestudyperiodthe averagedailyvolumetricwatercontent(m3/m3) inwateredplots(mean0.03,lower95 % CL0.02, upper95 % CL0.04)wasapproximatelythree timesgreaterthaninun-wateredplots(mean 0.01,lower95 % CL0.01,upper95 % CL0.02). Average,maximum,andminimumdailysoil moisturecontentinwateredplotswassignificantlylowerinun-wateredplotsthanwatered plots(p 0.0001forminimum,maximum,and average,F-value:avg.:66.10,max.:105.17,min.: 32.82,generallinearmodelwithdayasarandom effect). v www.esajournals.org20April2012vVolume3(4)vArticle31COPELANDETAL.