Cork oak physiological responses to manipulated water availability in a Mediterranean woodland

(1)

ContentslistsavailableatScienceDirect

Agricultural

and

Forest

Meteorology

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

Cork

oak

physiological

responses

to

manipulated

water

availability

in

a

Mediterranean

woodland

Cathy

Kurz

Besson

a,∗

_,

_Raquel

_Lobo-do-Vale

b

_,

_Maria

_Lucília

_Rodrigues

b

_,

_Pedro

_Almeida

b

_,

Alastair

Herd

b

_,

_Olga

_Mary

_Grant

c

_,

_Teresa

_Soares

_David

d

_,

_Markus

_Schmidt

e

_,

_Denis

_Otieno

e

_,

Trevor

F. Keenan

f

_,

_Célia

_Gouveia

a

_,

_Catherine

_Mériaux

a

_,

_Maria

_Manuela

_Chaves

b,g

_,

João

S. Pereira

b

a_Centro_de_Geofísica_da_Universidade_de_Lisboa,_Instituto_Dom_Luiz,_Lisboa,_Portugal

b_Instituto_Superior_de_Agronomia,_Universidade_técnica_de_Lisboa,_Tapada_da_Ajuda,_1349-017_Lisboa,_Portugal c_University_College_Dublin,_Belﬁeld,_Dublin_4,_Ireland

d_Instituto_Nacional_de_Recursos_Biológicos,_Quinta_do_Marquês,_Av._da_República,_2784-505_Oeiras,_Portugal e_Universität_Bayreuth,₉₅₄₄₀_Bayreuth,_Germany

f_Harvard_University,₂₂_Divinity_Avenue,_Cambridge,_MA_02138,_USA

g_Instituto_de_Technologia_Química_e_Biológica,_Av._da_República,_2780-157_Oeiras,_Portugal

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received7August2012

Receivedinrevisedform7October2013 Accepted9October2013 Keywords: Quercussuber Throughfallmanipulation Treetranspiration Gasexchange Soilmoisture Precipitationchange

a

b

s

t

r

a

c

t

Thisstudydetailsthephysiologicalresponsesofcorkoak(QuercussuberL.)tomanipulatedwaterinputs.

Treatmentsnamedasdry,ambientandwet,whichreceived80,100and120%oftheannualprecipitation,

respectively,wereappliedtoaMediterraneanwoodlandinsouthernPortugal.Treeecophysiologyand

growthweremonitoredfrom2003to2005.

Theimpactsofthewatermanipulationwereprimarilyobservedintreetranspiration,especially

dur-ingsummerdrought.Rainfallexclusionreducedtheannualstandcanopytranspirationby10%overthe

2-yearstudyperiod,whileirrigationincreaseditby11%.Theaccumulatedtreetranspirationmatched

precipitationinspring2004and2005atthestandlevel,suggestingthatcorkoaktreesrelyon

precip-itationwatersourcesduringthepeakofthegrowingseason.However,duringthesummerdroughts,

groundwaterwasthemainwatersourcefortrees.

Despitethesigniﬁcantdifferencesinsoilwatercontentandtreetranspiration,notreatmenteffects

couldbedetectedinleafwaterpotentialandleafgasexchange,exceptforasingleeventafterspring

irri-gationsintheverydryyear2005.Theseirrigationswereintentionallydelayedtoreducedryspellduration

duringthepeakoftreegrowingseason.Theyresultedinanacutepositivephysiologicalresponseoftrees

fromthewettreatmentoneweekafterthelastirrigationeventleadingtoa32%raiseofstem

diame-terincrementthefollowingmonths.Ourresultssuggestthatinasemi-aridenvironmentprecipitation

changesinspring(amountandtiming)haveastrongerimpactoncorkoakphysiologyandgrowththan

anoverallchangeinthetotalannualprecipitation.

Theextremedroughtof2005hadanegativeimpactontreegrowth.Theannualincrementoftree

trunkdiameterintheambientanddrytreatmentswasreduced,whileitincreasedfortreesfromthewet

treatment.Watershortagealsosigniﬁcantlyreducedleafarea.Thelatterdroppedby10.4%inresponse

totheextremedroughtof2005intreesfromtheambienttreatment.Thereductionwaslesspronounced

intreesofthewettreatment(−7.6%),andmorepronouncedintreesofthedrytreatment(−14.7%).

Corkoakshowedhighresiliencytointer-annualprecipitationvariability.Theannualaccumulatedtree

transpiration,theminimummiddayleafwaterpotentialandtheabsoluteamountofgroundwaterused

Abbreviations:(A),carbonassimilation;(Amax),maximumcarbonassimilation;(CV),crownvolume;(DBH),diameteratbreastheight;(DBHinc),trunkdiameterincrement;

(E),sapﬂux,treetranspiration;(E/DBH),normalizedtreetranspiration;(ET),potentialevapotranspiration;(gs),stomatalconductance;(gsmax),maximumstomatal

conduc-tance;(Gt),wholetreespeciﬁchydraulicconductance;(LAD),leafareadensity;(LAI),leafareaindex;(PAR),photosyntheticallyactiveradiation;(PCA),projectedcrownarea;

(SLA),speciﬁcleafarea;(SWC),soilwatercontent;(T),airtemperature;(VPD),vapourpressuredeﬁcit;(),leafwaterpotential;(md),middayleafwaterpotential;(pd),

predawnleafwaterpotential;(),sapﬂowdrivingforce;(␲),osmoticpotential;(p),leafturgor;(s),leafsolutecontent;(RSC),regressionslopecomparison;(SEM),

standarderrorofthemean.

∗ Correspondingauthorat:InstitutoDomLuiz,CentrodeGeofísicadaUniversidadedeLisboa,FaculdadedeCiências,CampoGrande,Ed.C8,piso3,Sala26,1749-016 Lisboa,Portugal.Tel.:+351217500884;fax:+351217500807.

E-mail addresses: cbbesson@fc.ul.pt, cathybesson@gmail.com (C.K. Besson), raquelvale@isa.utl.pt (R. Lobo-do-Vale), mlrodrigues@isa.utl.pt (M.L. Rodrigues),

pedroccta@yahoo.com(P.Almeida),a.herd@tft-forests.org(A.Herd),olgamgrant@yahoo.com(O.M.Grant),teresasdavid@gmail.com(T.S.David),mwtschmidt@gmx.com

(M.Schmidt),dennis.otieno@uni-bayreuth.de(D.Otieno),tkeenan@oeb.harvard.edu(T.F.Keenan),cmgouveia@fc.ul.pt(C.Gouveia),cameriaux@fc.ul.pt(C.Mériaux),

mchaves@isa.utl.pt(M.M.Chaves),jspereira@isa.utl.pt(J.S.Pereira).

(2)

bytreesappearedunaffectedbytheextremedroughtof2005.Ourstudyshowsthatcorkoakrapidlyand

completelyrecoveredfromtheextremedryyearof2005orfromrainfallexclusion.Ourresultssupport

theeco-hydrologicalequilibriumtheorybywhichplantacquirecomplementaryprotectivemechanisms

tobufferthelargevariabilityinwateravailabilityexperiencedinsemi-aridecosystems.Inoptimizing

theirstructuralbiomassincreaseinresponsetoincreasingdroughtstress,corkoaktreessucceededin

restrictingwaterlossestomaintaintheminimumleafwaterpotentialabovethecriticalthresholdof

xylemembolism,thoughwithnarrowerhydraulicsafetymarginsin2005.

Ourﬁndingshighlightcorkoak’ssensitivitytotheamountandtimingoflatespringprecipitation.This

couldbecriticalasfutureclimatescenariospredictareductionofspringprecipitationaswellasenhanced

severityofdroughtsintheIberianPeninsulabytheendofthe21stcentury.Ininducingwaterstressbefore

theonsetofsummerdroughts,thepredictedspringprecipitationdeclinecoulddrivethespeciescloser

tothethresholdofcatastrophicxylemembolismatthepeakofthedroughtperiod.

1. Introduction

Mediterranean savanna-type evergreen oak woodlands are man-madeecosystems.Theyarecharacterizedbyasparsenative treecover,whichmostlyconsistsofQuercusspecies(mainly ever-green corkoak, Quercus suber L.,and/or holmoak, Quercus ilex ssp. rotundifolia Lam.) and understory vegetationranging from shrubformationstograsslands.InPortugal,corkoaksavanna-type ecosystems-montados-coverabout0.74Mhaandrepresent23%of thenationalforestedarea(ICNF,2013).Thespecieshasa signiﬁ-canteconomicvalue.Itprovides0.7%ofthegrossdomesticproduct, about15,000jobs,andsupplies54%oftheworldwidecork produc-tion(Evangelista,2010).Montadosarealsoecosystemswithahigh conservationvalue,supportinghighlevelsofbiodiversity(Bugalho etal.,2011),andareconsiderednationalheritage.

Accordingto the mostrecent Portuguese forest inventories, montadosshow signsof anincreasing vulnerability.In only 12 years,corkoakmortalityrateandﬁreindicesrosesharply(by450% and200%,respectively)andtreedensitydeclinedby30%onaverage (AFN,2001,2010).Allrecentclimateprojectionsforthe Mediter-raneanbasinagreeinforecastingfuturelongersummerdroughts, withmorefrequentextremeevents,suchasheatwavesorsevere droughts(Barriopedroetal.,2011;CoumouandRahmstorf,2012; GiorgiandLionello,2008;IPCC,2007;Mirandaetal.,2006). Pro-jectionsalso predictthat spring andsummer precipitationwill decrease,byamagnituderangingfrom−20to−40%,depending ontheclimatechangescenarios(Philandrasetal.,2011;Giorgiand Lionello,2008).Montadosaremostlyconcentratedinthesouthern regionofthecountry,whichhasscarcewaterresourcesandalong drysummerseasonusuallycoupledwithhightemperaturesand highradiation(Fariaetal.,1996).Underfrequentlylimitedwater availability,Mediterraneanplantsevolvedphysiologicaland mor-phologicaladaptionstooptimizewateruse(seereviewbySardans and Pe ˜nuelas, 2013).Cork oak have acquiredmany features to copewithwaterstressand,ultimately,avoidxylemcavitationand embolism(Pereiraetal.,2009).Attheleaflevel,stomatalclosure inducedbydroughtstress(Grantetal.,2010;Otienoetal.,2007; Pintoetal.,2012;Vazetal.,2010)restrictswaterlosswhilelimiting therateofCO2assimilation(Chaves,1991;FarquharandSharkey,

1982).BiochemicalconstraintsinthemaximumRubiscoCO2

ﬁx-ationcapacityandthemaximumrateofelectrontransportmay alsotakeplaceunderprolongeddroughtconditions(Lawlorand Cornic,2002;LawlorandTezara,2009;Tezaraetal.,1999)andbe evenmorerelevantunderseveredrought(Botaetal.,2004;Chaves etal.,2003,2009;GrassiandMagnani,2005;Vazetal.,2010).Leaf osmoticadjustmentimproveswateruptakecapacity(Otienoetal., 2007).Attheplantlevel,adeeprootingsystemenhanceswater uptakecapacityviagroundwateraccess(Davidetal.,2007)and hydrauliclift(Kurz-Bessonetal.,2006;Otienoetal.,2006).The reductionof vesseldiameterprovidesfurtherresistanceagainst xylemcavitationwhilethedecreaseofleafareasurface(Aranda

etal.,2005;ChavesandOliveira,2004)avoidsanexcessivewater loss.

ThesedroughtavoidanceadaptationsconferQ.suberwithahigh degreeofresilience(Grantetal.,2010;Vazetal.,2010).However,it remainsuncertainwhethertheseadaptationswillallowplant sur-vivalundermoreextremeconditionsasforecastedforthefuture. Therefore,thereisaneedforfurtherrainfallmanipulation stud-iestobetterpredictthevulnerabilityofMediterraneanecosystems underclimatechange(Beieretal.,2012;Wuetal.,2011).

Althoughcorkoaksseemtobeabletoadapttowaterscarcity insemi-aridambients,thespecieshasneverbeentestedunderthe predictedfutureclimaticconditions.Thisstudyaimedat evalu-atingQ. suber’sphysiological responsestorainfallmanipulation soastoassessthevulnerabilityofMediterraneanSavannah-like ecosystemstofutureprecipitationchanges.Soilwateravailability wasexperimentallymanipulatedinamontadoareainsouthern Portugal,bymeansofrainfallexclusionandaddition.Treatments weredesignedaccordingtothepredictedrangeofprecipitation changes intheMediterranean basin (Giorgi andLionello, 2008; Mirandaetal.,2006).Withincreasingaridity,treeswereexpected toshowprogressivelyhigherstomatalresistance,resultinginlower carbon assimilationrate,especially duringthedrought periods. Theselowercarbonassimilationrateswouldaffectthecarbon bal-anceatthewholeplantlevelleadingtochangedmorphologicaland growthtraits.Someuncertaintiesstillremainonhowhigher sto-matalresistanceunderincreasingwatershortagewouldaffectthe ecosystemwaterbalance,whichwouldexperiencelower transpi-rationratesbuthighergroundwaterdischarges.

Following the evolution of tree water status, sap ﬂux, gas exchange,growthandmorphologicaltraitsovertwoyearsof rain-fallmanipulation,ourresultsshowthatnotalltheexpectations were fulﬁlled. Finally, we discuss the relevance of spring pre-cipitation quantity and distribution to cork oak’s physiological performanceinaparticularlydryyear.

2. Materialandmethods 2.1. Experimentalconditions

ThestudywasconductedinHerdadedaMitra(N38◦31.664,W 8◦01.380,221maltitude)insouthernPortugalfrom2003to2005. TheclimateisMediterraneanmesothermichumid.The experimen-talsiteliesonanacidLitholicnon-HumicsoilderivedfromGneiss witha pHof4–6,ona5%slope.Fromthesurfaceto1mdepth thesoilconsistsof88.9%sand,4.9%silt,and6.3%clay,withalow (5%)waterholdingcapacity.Theexperimentalsite(46m×60m) iscoveredwithQ.suberL.treesplantedin1988,withan under-storymainlycomposedofCistussalviifoliusL.andC.crispusL.and herbaceouswinter–springC3annualplants.Twenty-seven

repre-sentativeQ.subertreeswereselectedformonitoring.Theaverage treedensityontheexperimentalsitewas1997±134treesha−1.

(3)

Table1

Treesizecharacteristicsanddensityatthebeginningoftheexperimentineachtreatment.

TreeNumber Height(m) DBH(cm) LAI PCA(m2₎ _Density_(tree_ha−1₎

Selectedtrees W 9 5.08(0.22) 10.53(0.54) 2.32(0.30) 4.97(0.37) ND A 9 5.39(0.28) 13.15(0.72) 2.00(0.44) 5.51(0.60) ND D 9 5.19(0.23) 13.90(0.97) 4.14(0.59) 4.59(0.43) ND Experimentalarea W 125 2.28(0.16) 3.86(0.40) 0.90(0.12) 2.01(0.17) 2001 A 137 2.23(0.14) 3.93(0.40) 0.84(0.10) 1.97(0.15) 2360 D 103 2.50(0.19) 4.71(0.53) 1.07(0.14) 2.27(0.20) 1724

Treedensity ineach treatmentareawaswithin theconﬁdence

intervalofthemean[1570;2424]oftheentirestand.Tree

den-sity,height,anddiameteratbreastheight(DBH)measuredatthe

beginningoftheexperimentareprovidedinTable1.

FromlateOctober2003onwardsandinaccordancewith predic-tionsofa20%reductioninprecipitation(Mirandaetal.,2006),three treatments(ambient,dryandwet)wereappliedwithinthesite. Rainfallexclusionwasachievedbyplacing0.6mwide×20mlong plasticdrainscovering20%ofthesoilsurfaceandlocatedbetween treerowsunderopencanopyarea.Thedrainswereinstalledon a 5% slope allowing water to ﬂow naturally down to a reser-voirthatwaslaterbeingusedforirrigation.Consideringthat24% oftheexclusionareawascoveredbytreecanopyandthatcork oakinterceptionwas22%ofgrossrainfall(David2000),we cal-culatedthat rainfallinterception slightly reduced the exclusion efﬁciencyfromanexpected20%excludedrainfalltoaneffective 19%.Thecontrol‘ambient’treatmentreceivednaturalrainfallwhile the‘wet’treatmentreceivedextraprecipitation,divertedfromthe ‘dry’treatment.Watercollectedoverthedrytreatmentplotswas redistributedfromthereservoir,ontothewettreatmentplots,via irrigationpipeswithdripemitters.Eachtreatmentwasreplicated overthreeblocksandphysiologicaldatawererecordedinthree treespertreatmentperblock(9blocks).Ablockdesignwasused totakeinto accountpotentialvariation duetotheslope ofthe experimentalsite(10maltitudedifferencewithinthesite).Rainfall exclusionandirrigationstartedattheendofNovember2003.Due totheseveredroughtduring2004/2005,thewettreatmentwas suppliedwithadditionalwaterbetweenJuly2004andJune2005 (126%oftherainfalloccurringintheambienttreatment).Fig.1 showsthetotal amountofwater receivedbyeachtreatmentin 2003,2004and2005comparedtotheaverageofthe30-yearperiod 1951–1980(INMG,1991).

Fig.1.Annualprecipitationforeachtreatmentcomparedtothe30-yearaverage annualprecipitation(1951–1980)(horizontalline).

2.2. Environmentalmonitoring

Precipitation(ARG100rain gauge,EM Ltd.,Sunderland, UK), airhumidity(Fischer431402sensor,K.FischerGmbH,Drebach, Germany),air temperature(T)at2mheight(ThermistorM841, Siemens,Munich,Germany)andphotosyntheticallyactive radia-tion(PAR,LI-190quantumsensor,LicorInc.,Lincoln,NB,USA)were measuredwithadatalogger(DL2e,Delta-TDevicesLtd.,Cambridge, UK)every5min,recording30-minaveragesbasedonthe Coordi-natedUniversalTime/GreenwichMeanTime.Airvapourpressure deﬁcit(VPD)wascalculatedfromtheequationproposedbyWesely (1989).

FromDecember2003toOctober2004,soilwatercontent(SWC) wasmeasuredbyfrequencydomaintechnologyusingechoprobes (ECH20,Decagon,DecagonDevicesInc.,Washington,USA).SWC wascontinuouslymeasuredat0.1,0.2and0.3mdepthinone loca-tionof eachof the9plots. Astheechoprobessuffered froma poorsensorcontactwiththesoilduringthedroughtperiods,the continuousmonitoringofSWCwasreplacedbyweekly measure-mentsusingaproﬁleprobe(PR2,Delta-TDevicesLtd.,Cambridge, UK)fromJanuarytoSeptember2005.ProﬁleprobeSWC measure-mentswereperformedat6depthssimultaneously(0.1,0.2,0.3,0.4, 0.6,1m),on3locationsineachofthe9plots.Gravimetric mea-surementswereperformedperiodicallytocalibratesoilmoisture sensors.

Wewillconsiderherehydrologicalyears(betweenOctoberof thepreviousyearandSeptemberofthecurrentyear),asitreﬂects bettertheavailablewaterfortreefunctioningandgrowth. 2.3. Sapﬂowandhydraulicconductance

Treesapflowwasmeasuredusingthermaldissipationsensor probesof20mmlength(UPGmbH,Landshut,Germany)according totheGraniermethod(Davidetal.,2007;Granier,1985).One sen-sorwasradiallyinsertedinthenorth-facingxylemofeachtreeat breastheight.Sensorswereconnectedtoadatalogger(DL2e, Delta-TDevicesLtd.,Cambridge,UK), readingtemperaturedifferences (T)betweenprobeseveryminuteandrecording30-minaverages. Sapflow wascontinuouslymeasuredon25selectedtreesfrom June2003toOctober2005.Datafrom6treeswerediscardedas theirrecordinghadmorethan65%gapduetorepetitivesensor fail-ures.Missingvaluesintheremainingdatasets,equivalentto29%of totalrecordings,werefilledusinglinearrelationshipsbetween sen-sorsshowinganaveragedeterminationcoefficientof0.77.Thetree transpiration(E)wascalculatedbymultiplyingthesapflux den-sitybytheconductivesapwoodarea,whichwasassumedtobethe 20mm-wideannulusofsapwoodincontactwiththesapflow sen-sor.Adetaileddescriptionofthesensorsandsapflowcalculation canbefoundinDavidetal.(2004).

EexpressedinLday−1wasbestnormalizedtotreediameterat breastheight(E/DBH,expressedinLm−1day−1).E/DBHresultsare presentedas9-dayrunningaverages.Wholetreespeciﬁchydraulic conductance(Gt)wasestimatedaccordingtoDarcy’slaw

(4)

force()(Buccietal.,2005;Cochardetal.,1996).was calcu-latedbysubstractingmidday(md)frompredawn(pd)leafwater

potential.

Treetranspirationwasup-scaledtothestandlevelby calcu-latingthebestdailylinearregressionbetweenthesapﬂuxofthe remaining 19 monitoredtrees and theircorresponding DBHor height.Thedailyrelationshipswerethenappliedoneachofthe479 treescoveringtheexperimentalareaforwhichDBHandheighthad beenmeasuredin2003beforetreatmentimplementation.

2.4. Leafwaterstatusandgasexchange

Dailycoursesofleafwaterpotential()andgasexchangewere performedsimultaneouslyincleardaysthroughouttheyear,and onamonthlybasisduringdroughtperiods(seedatesonTable2). Threematureleavespertreeweresampledforeachmeasurement at0600(only),1000,1300and1600h.

wasmeasuredwitha pressurechamber (PMSInstrument Co.,Corvallis,Oregon,USA).Immediatelyaftermeasuringpredawn leafwaterpotential(pd),discsofeachleafweretaken,placed

inEppendorfs,andplungedintoliquidnitrogenuntilbeingstored at−80◦_C_for_later_{determination}_of_osmotic_potential₍

).After

thawingthesamplesfor10minatroomtemperature,was

mea-suredusingC52chambers(2hfortemperatureandwatervapour equilibration)connectedtoaWescorHR33Tdew-point microvolt-meter(WescorInc.,USA)operatingindewpointmode.Standard NaClsolutionswereusedforeachchambercalibration, generat-ingacalibrationregressionusedtoconvertthe␮voutputtoin

MPa.Leafturgor(p)wasestimatedasthedifferencebetweenpd

and.Leafsolutecontent(s)wascalculatedass=−(V/RT)

whereRistheuniversalgasconstant,Tistheleaftemperatureand Visthetotalvolumeofwaterintheleaftissue.Vwasdeterminedin anothersetofdiscspunchedfromthesameleavesandexpressed onleafdrymassbasis.

Leaf gasexchange measurements wereperformed withtwo cross-calibratedportablephotosynthesissystems(Li-6400;Licor Inc.,Lincoln,NB,USA).Thediurnal changesin CO2 assimilation

(A)andstomatalconductance(gs)wererecordedafter

stabiliza-tion.Environmentalconditions(airtemperatureT,vapourpressure deﬁcitVPD andphotosyntheticphoton ﬂux densityPPFD)were monitoredwiththeLI-6400system.Maximalratesof photosyn-thesis(Amax)andstomatalconductance(gsmax)weredetermined

fromdailycourses.

2.5. Treemorphologicaltraitsandgrowth

Stem diameter was measured at breast height (DBH) using treedendrometers(I-802-D1,UMSGmbH,Munich,Germany).The annual DBHincrement (DBHinc)wascalculated as a percentage

basedonhydrologicalyears(2004and2005).Treeheightswere measuredwithatelescopicheightpole.Sphericallensphotographs (HemiViewSystem-HMV1,Delta-TDevicesLtd.,Cambridge,UK) were alsotaken at dawn in September 2003, March2004 and October 2005 onthe south and north sidesof the 27 selected trees.Crownbasediameter(D),heightandgeometricalshapewere recordedforeachselectedtree.Theseparameterswereusedwithin the Hemiviewsoftware to determinetree leaf areaindex (LAI, one-sideleafareaperprojectedcrownarea),crownvolume(CV) accordingtotreecrowngeometricalshape,andprojectedcrown area(PCA)calculatedas1/4␲D2₍_Berlyn,_1962;_Watson,₁₉₄₇_)._Leaf

areadensity(LAD)wasdeterminedfromtheratiobetweenLAIand CV.Speciﬁcleafarea(SLA)wascalculatedbydividingtheleafarea tothedryweightof6leafdiscs(0.7cmdiameter)collectedfrom3 leavessampledoneachselectedtreeinSeptember2004and2005. _Table

2 Synthesis of ANOVA of the treatment effect on Quercus suber physiological parameters measured on ﬁeld campaigns from 2003 to 2005. Parameter 30-Jun-03 17-Jul-03 20-Aug-03 8-Sep-03 10-Mar-04 2-Jun-04 13-Jul-04 29-Jul-04 2-Sep-04 30-Sep-04 15-Dec-04 3-Feb-05 23-Feb-05 31-Mar-05 1-Jul-05 26-Aug-05 15-Sep-05 13-Oct-05 23-Nov-05 Soil water content SWC − 0.1 m W>=A=>D W>A>D W>A>D W>A>D W>A>D W>A>D W>A>D W>=A=>D W>=A=>D W>=A=>D W>A=D W>=A=>D W>=A=>D ns SWC − 0.2 m ns ns ns ns ns ns ns W>=A=>D ns W>A=D W>A=D W>A=D W>A=D ns SWC − 0.3 m ns ns ns ns ns ns ns ns ns ns ns ns ns ns SWC − 0.4 m ns ns ns W>A=D W>=A=>D W>A=D ns SWC − 0.6 m ns ns W=A=D ns W>A=D ns ns SWC − 1.0 m ns ns W>A=D ns W>A=D ns ns Leaf gas exchange Amax ns ns ns ns ns ns ns ns ns ns W>A=D ns ns ns ns A10:00 ns ns ns ns ns ns ns W>A=D ns ns ns A13:00 ns ns ns ns ns ns ns ns ns ns W>=A=>D ns ns A16:00 ns ns ns ns ns ns W>A=D gsmax ns ns ns ns ns ns ns ns ns ns W>A=D ns ns ns ns gs10:00 ns ns ns ns ns ns ns W>A=D ns ns ns gs13:00 ns ns A>D=W ns ns ns ns ns ns ns W>A=D ns ns gs16:00 ns ns ns ns ns ns W>D=A A /gs 10:00 ns ns ns ns A>=W=>D ns ns A=D>W ns ns ns A /gs 13:00 ns ns ns ns ns ns ns ns ns ns A>D=W ns ns A/gs 16:00 D>=A=>W ns ns ns ns ns ns ns Tree water status p ns ns ns ns ns ns ns ns ns ns ns ns ns W>A=D ns ns ns W>A=D 10:00 ns ns D>A=W ns ns ns ns ns ns 13:00 ns ns ns W>=D=>A ns ns ns ns ns W=D>A ns ns ns W>D=A ns ns ns ns 16:00 ns ns ns W>A=D ns ns ns W>D=A ns ns ns A>=D=>W ns ns ns ns ns ns ns ns ns ns ns ns ns A=D>W s W>=A>=D D=A>W ns ns ns ns ns ns ns ns ns ns ns p ns W>D=A ns ns ns ns ns ns ns ns ns ns ns s ns ns ns ns ns ns ns ns ns W=A>D ns ns Daily E /DBH ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns Gt ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns ns

(5)

Fig.2.Meteorologicalandwaterfluxdataduringtheexperimentalperiod(June2003–December2005).(A)Precipitationandirrigation,(B)dailymeantemperature,(C) vapourpressuredeficit(VPD),(D)maximumphotosyntheticallyactiveradiation(PAR),(E)soilvolumetricwatercontent(SWC)at0.3mdepth(Echo:Echoprobe;Prof: PR1sensor;Grav:Gravimetricalmeasurements),and(F)smoothed(9daysrunningaverage)dailyaveragetreetranspirationforeachtreatmentnormalizedbytheinitial diameteratbreastheight(E/DBH).Theverticallineindicatesthetreatmentbeginning.Thedottedlineindicatestheimpactofthelastirrigationevent.Starsindicatesignificant differencesbetweentreatmentmeans(P<0.05).Greybackgroundindicatessummerdroughtperiods.

2.6. Statisticalanalysisanddataprocessing

Toassesstheeffectofprecipitationchangesontheparameters

measured,aone-waymixedmodelanalysisofvariance(ANOVA)

wasused,consideringtreatmentsasaﬁxedfactorandblocksasa

randomfactor.Whenastatisticallysigniﬁcanteffect(p<0.05)of

thetreatmentwasfound,theposthocTukeyHSDtestwasusedto

identifydifferencesbetweenmeans.Statisticalanalysesand

lin-earregressionﬁtswerecarried outwithSAS9.1(SASInstitute,

Cary,NC,USA).Dataweretransformedwhennecessarytomeet

theassumptionsofparametricanalyses.Statisticalcomparisonsof

regressionslopes(RSC)followedtheanalysisofcovariancemethod

describedbyZar(1999).Time-by-timeANOVAwasperformedon

thederivedmonitoredvariables(E/DBHandSWC)toreduce auto-correlationandemphasizetreatmenteffect(Diggleetal.,2002).In

thetextandﬁgures,meanvaluesarepresentedwithstandarderror ofthemean(SEM).

3. Results

3.1. Environmentaldrivers

Theexperimentcoveredthreehighlycontrastinghydrological years varyingfrom typicalin 2003,tomoderatelydry in 2004, andseverelydryin2005(Fig.1).Annualrainfallamountswere 689mm,607mmand410mmin2003,2004and2005,respectively (Fig.1).In2004,rainfalleventsoccurredmainlyfromOctoberto Mayandweremorefrequentandintenseduringautumn(Fig.2A). 2005wascharacterizedbyunusuallylowprecipitationsbetween NovemberandMarch.Springprecipitationwaslowerin2004than

(6)

Fig.3.Seasonalvariationofthesoilvolumetricwatercontent(SWC)atsixsoildepthsin2005.ValuesareaverageofeachtreatmentandtherespectiveSEM(n=9).Stars indicatesigniﬁcantdifferencesbetweentreatmentmeans(P<0.05).

in2005(51mmand78mm,respectively),whereassummer rain-falldecreasedsubstantiallyoverthetime(50mm,26mmand4mm in2003,2004and2005,respectively).Thelastsigniﬁcantrainfall eventsof2005occurredonMay16th(12mm)and30th(19mm). ThelastirrigationwasintentionallydelayedtoJune22th(13mm). ThemeandailyT,VPDanddailymaximumPARfollowedtypical Mediterraneanseasonaltrends(Miranda etal.,2006),withhigh valuesduringthesummer(Fig.2B,Cand–D).

SWCat0.3mdepth(Fig.2E),wheremostoftheﬁnetreeroots arefound, wasclosetoﬁeldcapacity fromDecember toMarch 2004, averaging 0.23±0.01m3_m−3_. _Compared _with _the

ambi-enttreatment,SWCwashigherinthewettreatment(+11%)and slightlylowerinthedrytreatment(3%).Duringspring(April–June) 2004,SWCdroppedsteeplyinresponsetothelackofrainfall.The soilpermanentwiltingpointwasreachedduringsummer(July, 0.04±0.01m3_m−3₎_with_no_signiﬁcant_difference_between

treat-ments,and persisteduntilOctober.Soilrecoveredﬁeldcapacity duringwinter(December)followingarainyperiod.Fromtheend ofDecember2004tothemiddleofFebruary2005,rainfallevents wererareandweak(<2mm)resultinginlowwinterSWCvalues (overallmeanwas0.13±0.01 m3_m−3_)._At_the_end_of _the

sum-mer2005,thesoilmoisturereachedpermanentwiltingpointagain withSWCaveraging0.05±0.003m3_m−3_._By_the_end_of_October,

SWCwas0.11±0.01m3_m−3_._Differences_in_SWC_at_0.3_m_between

treatmentswerenotsignificant.SWCintheothersoilhorizonsis summarizedinTable2andFig.3.At0.1mdepth,aconsistent signif-icantdifferencebetweenthewetanddrytreatmentswasobserved withahigherSWCinthewetplots,exceptinOctober2005.At0.2m depth,SWCwassignificantlyhigherinthewetplotscomparedto thedryandambientplotsfromFebruarytoSeptember2005.At deeperhorizons(0.4to1mdepth),however,soilbecame signifi-cantlywetterinthewettreatmentfromlateMarch2005untillate August2005.TheincreaseinSWCobservedattheonsetoffallrain eventsin2005diminishedwithincreasingsoildepthandremained verysmallat1mdepth.

3.2. Corkoakphysiologicalresponses

3.2.1. Treetranspirationandhydraulicconductance

E/DBHfollowedtheVPDseasonaltrend(Fig.2C,F)with max-imumvaluesreachedatthebeginningofJune.MaximumE/DBH wasobserved in June 2004,withvalues ranging from96.63to 113.71Lm−1day−1.InJune2005,oneweekafterthelast13mm

irrigationpulse,E/DBHshoweda45%increaseinthewettreatment associatedtoaraiseofsoilwatercontentat0.3mdepth(vertical dashedline,Fig.2EandF).Infallandwinter2003and2004and inearlyspring2004and2005,therewasnosignificantdifference intreetranspirationbetweenthe3treatments(Fig.2F).However, duringthesummerperiodsof2004and2005,treatmenteffects couldbedetectedonseveraloccasions,withsignificantlylower sapfluxinthedrytreatmentthaninthetwootherones. Signif-icanttreatmenteffectsweremorefrequentafterthelast13mm irrigationwithhighervaluesinthewettreatmentcomparedwith thedryone(Fig.2F).

E/DBHwas44%higherinspring2005thaninspring2004.It waslikelyduetothehigheramountofprecipitationinSpring2005 (Fig.2FandTable3).Therewasnosigniﬁcanteffectofthetreatment ontheaccumulatedtranspirationatthetreelevel(Fig.4A).Between May 16thand June 6ththere was hardlyanyE/DBH difference betweentrees fromthedifferenttreatments.Treetranspiration begantodifferbetweentreatments7daysafterthelastheavy rain-fallofMay30th2005,highlightingasubstantialpositiveeffectof thelastirrigationpulsefromthisdateonwards.Thetranspiration rateincreasedby42%fortreesofthewettreatmentanditremained 66%higheronaveragethantheoneoftheambienttreatmentuntil September15th(Fig.5).

Table3

Seasonalstandaccumulatedtranspiration(normalizedtotreedensity×meanDBH) ascomparedtoseasonalaccumulatedrainfallin2004and2005.

Standcumulated transpiration(mm) Rainfall(mm) Groundwater contribution (%) W A D 2004 Fall(OND) 14 13 16 331 0 Winter(JFM) 18 16 19 198 0 Spring(AMJ) 52 53 50 51 0 Summer(JAS) 70 62 45 26 57 Total 154 142 131 607 2005 Fall(OND) 18 19 20 277 0 Winter(JFM) 25 27 28 51 0 Spring(AMJ) 84 80 64 78 0 Summer(JAS) 46 28 23 4 86 Total 174 153 134 410

(7)

Fig.4.(A)Accumulatedtranspirationatthetreelevelinresponsetothewet, ambi-ent,anddrytreatmentscomparedwiththe(B)accumulatedprecipitationin2004 (thinline)and2005(thickline).

Up-scaledatthestandlevel,thecanopytranspirationmost

suf-feredfromrainfallexclusioninthesummer(27%lowerin2004,

and20%in2005thanintheambientarea).Onanannualbasis,

rainfallexclusionreducedthestandcanopytranspirationinthedry

plotsby8%in2004,and13%in2005.Irrigationhadahighpositive

impactinsummer,increasingcanopytranspirationby14%in2004

and63%in2005.Irrigationalsoincreasedtheannualstandcanopy

transpirationby8%in2004and13%in2005.Inspringthe

accumu-latedstandcanopytranspirationwellmatchedtheaccumulated

precipitation.Thetotalaccumulatedstandtranspirationobserved

in2005(134–174mm,Fig.4B)wasunaffectedbythelowerannual

precipitation(410mm)whencompared withthetotal accumu-latedtranspiration(131–154mm)andhigherannualprecipitation (607mm)of2004.Moreover,theabsolutecontributionof ground-watertotreetranspirationoverthe2004and2005summerperiods was35mmand24mm,respectively(Table3),whichoverallwere verysimilarvalues.

3.2.2. Leafwaterrelations

Leafwaterpotentialmeasurementsperformedduringsummer 2003(beforethebeginningofwatermanipulation)showedclose

Fig.5.Meandaily(top)andaccumulatedE/DBH(bottom)fortreesfromthewet (W),ambient(A)anddry(D)plotsinresponseofspringrainfallsandirrigationin 2005.Theverticaldashedlineindicatesthedateatwhichthelastirrigationbegan tohaveaneffectontreetranspiration.

Fig.6.Evolutionof(A)leafpredawnwaterpotential(pd),(B)middaywater

poten-tial(md),(C)osmoticpotential(),(D)turgorpotential(p),(E)solutecontent

(s)and(F)hydraulicconductance(Gt)ofselectedtreesinresponsetothewet,

ambientanddrytreatmentsfrom2003to2005.Valuesareaverageofeach treat-mentwiththerespectiveSEM(n=9).Signiﬁcantdifferencesbetweentreatmentsare indicatedfor*P<0.05,**P<0.01,***P<0.001.Greybackgroundindicatessummer droughtperiods.

similaritybetweentreatments.Aseasonaltrendofcouldbe dis-tinguishedoverthethreeyears,bothatpredawnandmidday.The treesshowedahighleafwaterstatusduringthewinterandspring but wereseverely stressedat theend of thesummerdrought, inSeptember (Fig.6A–B). Signiﬁcantdifferencesbetween treat-mentswereonlyobservedonJuly1st2005(pdandmd)andon

November24th2005(pdonly),wheretreesformthewet

treat-mentshowedahigherwaterstatus(Table2andFig.6A–B).pd

graduallydeclinedthroughoutthesummerandinter-annuallywith particularlylowvaluesin2005attheendofthedroughtperiods. Attheendoftheexperimentafterautumnrainfalls,afastrecovery oftheplantwaterstatuswasobserved,especiallyinthetreesof thewettreatment.Therewasastatisticallysigniﬁcantdifference inminimum pd betweenyears (p<0.001),withoverall means

of−1.40±0.05,−2.21±0.07and−2.88±0.05MPainSeptember 2003,2004and2005,respectively(Fig.6A).

Minimum md, (Fig. 6B) decreased signiﬁcantly from 2003

(−2.62±0.06MPa)to2004(−3.17±0.08MPa)butdidnotshow afurthersigniﬁcantdecreasein2005(_−3.28_±0.06MPa).

Lowpd(−2.47+−0.13MPa)wasmeasuredinFebruary2005

(8)

Fig.7.Evolutionof(A)maximalstomatalconductance(gsmax)and(B)maximalphotosyntheticrate(Amax)ofselectedtreeleavesinresponsetothewet,ambientanddry

treatmentsfrom2003to2005.Signiﬁcantdifferencesbetweentreatmentsareindicatedwhen*P<0.05,**P<0.01.Greybackgroundindicatessummerdroughtperiods.

frozenwaterinleafvessels.However,thehighermd measured

(−1.50+−0.12MPa)afewhoursafterthefreezingnight,whenday temperaturehadreached12◦C,indicatedthatleaveshadquickly recoveredfromthefrostanddidnotsufferirreversibledamage.

Inresponsetoincreasingdroughtamarkeddeclineinwas

observedin 2004and 2005accompaniedby anincrease inthe amountofleafsolutesperunitdrymassindicatingthataseasonal osmoticadjustmenthadoccurredintreesunderthe3treatments (Fig.6CandE).In2004decreasedfromJune(meanrangingfrom

1.07and−1.28MPa)toSeptember(−2.42to−2.51MPa).Asimilar decreaseofwasobservedin2005afterApril.Signiﬁcant

differ-encesinamongtreatmentswereonlydetectedinMarchand

June2004whenthesoilwateravailabilitywasstillhighand favou-ringtreewaterstatus(Table2andFig.6C).Inbothyearsleafturgor decreasedfromJulyonwardsduetotheseasonalreductionofpd

unmatchedbyaproportionaldecreaseof(Fig.6D).Whereas

pwasstillpositivebytheendofthe2004droughtperiod,

tur-gorlosspointwasreachedearlierin2005,inAugust.Differences inpbetweentreatmentswereobservedonJune2nd2004when

thepvalueofwettreatmenttreesbecamehigherthanthatofthe

ambientanddrytreatments,duetolower(Table2;Fig.6Dand

E).

Whilein2004theleafsoluteconcentrationwashigh through-outthedroughtperiod(>0.77mosmolg−1),adecreaseofswas

observedinthemid-summerof2005(<0.45mosmolg−1) proba-blyinresponsetothehighintensityofwaterstressexperienced bythetrees(Fig.6E).Leafsolutecontentdidnotdifferinthethree treatmentsexceptattheendofAugust2005,whenthedry treat-menttreespresentedlowersthanthoseobservedintheother

trees(Table2andFig.6E).Noconsistenttreatmenteffectscould befoundonthetreesapﬂowdrivingforceeveninJuly2005 (Table2).Despitenostatisticallysigniﬁcantdifference,Gtshowed

atrendforhigheraveragevaluesforthewettreatmentduringthe springperiods(Fig.6F).Itsteeplydecreasedto3-foldlowervalues undersummerdryingconditions,butwasunaffectedbytreatments ortheinter-annualvariationindroughtintensity(Fig.6F).

3.2.3. Leafgasexchange

Aseasonaltrendcouldbedistinguishedinleafgasexchange over the study period (Fig. 7A and B), comparable to the sea-sonalvariationof.HighergsmaxandAmaxwereobservedduring

thewinterandearlyspring.Attheend ofthesummerdrought bothparameters werestronglyinhibited.Signiﬁcantdifferences between treatments in gsmax and Amax were only observed on

July1st2005,withwettreatmenttreesshowinghigherleafgas exchangerates(Table2;Fig.7A–B).

Maximal stomatal conductance steeply decreased with the onsetofdrought,fromJuneonwards,in2003and2004,whereas it wasdelayed untilJuly in 2005in thewet treatment, due to springirrigation.Stomatalclosure wassevereattheendof the summerdrought,particularlyin2005.Thelimitationofstomatal conductancewasreversedrapidlyaftertheﬁrstrainsinfall.There wasastatisticallysigniﬁcantdifferenceinminimumgsmaxbetween

years(p<0.001),withoverallmeansof0.081±0.009,0.042±0.003 and0.026_±0.004molm−2s−1inSeptember2003,2004and2005, respectively(Fig.7A).

StomatalclosurelimitedcarbonassimilationasAmaxvariation

wasquiteclosetothevariationobservedingsmax.Comparedto

gsmax,Amaxdeclinedmoreslowlyduringthedroughtstress

devel-opmentbutrecoveredfasterlateron.Signiﬁcantdifferenceswere alsofoundovertheyearsinminimumAmax.Overallmeanswere

6.4±0.5,3.0±0.3and1.8±0.2␮molm−2s−1inSeptember2003, 2004and2005,respectively(Fig.7B).Bytheendofthestudyperiod, gsmaxandAmaxvaluesweresimilarorevenhigherthaninthe

pre-viousrainyseasons(Fig.7AandB).

Noconsistenttrendcouldbeidentiﬁedintheintrinsicwateruse efﬁciencyWUEi,(Amax/gsmax)inresponsetoincreasingdroughtor

treatmenteffect(datanotshown).WUEiwassimilarattheendof summer2003,2004and2005.

3.2.4. Treegrowth

Before treatments were applied, the mean DBH was 11.6±0.3cm withvaluesrangingfrom8.6to21.0.Treeheights variedbetween4.1mand7.2mwithan averageof5.3±0.2m. Inter-annualtrendsoftreesizeandmorphologicalcharacteristics arepresentedin Fig.8.Therewasnomajordifferencebetween treatmentmeansbutthetreeheightincreasewashigherforthe wettreatmenttrees,especially in2005(Fig.8A).Anincreasein DBHwasobservedoverthewholestudyperiod(Fig.8B).However, the growth rate changed over the years. From 2004 to 2005, a decline in DBHinc wasobserved in the ambienttreatment. It

wassigniﬁcantly more pronounced inthe dry treatments(RSC, p<0.05). Inversely, an increase of DBHinc occurred in the wet

treatment (RSC, p<0.001). There was also a highly signiﬁcant positiveimpactofirrigationontreeDBHincin2005.Treesfromthe

wettreatmentshoweda55% and60% higherDBHinc thantrees

fromtheambientanddrytreatments,respectively(Table2).Those signiﬁcantdifferencesinDBHincbecamemostlyevidentafterthe

(9)

Fig.8. (A)Treeheight,(B)incrementofdiameteratbreastheight(DBHinc)expressed

aspercentageofthetotaldiameter,(C)crownvolume(CV),(D)treeprojectedcrown area(PCA),(E)leafareaindex(LAI)expressedasleafm2_per_ground_m2_,_(F)_leaf

areadensity(LAD)expressedasleafm2_per_crown_volume_m3_,_(G)_speciﬁc_leaf_area

(SLA).ValuesareaverageofeachtreatmentandtherespectiveSEM(n=9).Signiﬁcant differencesbetweentreatmentsareindicatedwhen**P<0.01,***P<0.001.

RegressionsofCVandPCAagainsttimeshowsignificantlyhigher slopesfortreesfromthedrytreatmentcomparedwiththoseforthe wetandambienttreatments(RSC,p<0.01)between2003and2005 (Fig.8CandD).Inversely,significantlylowerslopeswerefoundfor thedrytreatmentinLAIfrom2003to2005(Fig.8E,RSC,p<0.001) andLADfrom2004to2005(Fig.8F,RSC,p<0.001).TheSLA(Fig.8G) alsosignificantlydecreasedfrom2004to2005foralltreatments (Pairedt-test,p<0.03).Theregressionslopewassteeperfortrees fromthedrytreatment,andsmootherfortreesofthewet treat-ment.SLAregressionslopesfortreesofthewetanddrytreatments werehighlysignificantlydifferent(RSC,p<0.01).Attheendofthe experiment,SLAalsoshowedsignificantdifferencesbetween treat-ments(p<0.01),withsignificantlyhighervaluesobtainedfortrees ofthewettreatment.

4. Discussion

4.1. Effectoftheinter-annualvariabilityinprecipitation

Theexperimentcoveredthreehighlycontrastinghydrological yearsrangingfromtypicalin2003,tomoderatelydryin2004,and severelydryin2005.Thislasthydrologicalyearwasthedriestinthe

IberianPeninsulaoverthelast140yearsandhadmajoreconomic repercussionsforPortugalandSpain(García-Herreraetal.,2007). TherewerenomajorrainfalleventsbetweenDecember2004and February2005,andprecipitationonlyreached8%ofthelong-term average(García-Herreraet al.,2007).It alsofollowed a moder-atelydryyear(2004),whichemphasizedthesubsequentnegative impactsonwaterresourcesandvegetationgrowth.

Theextremedroughtof2005ledtoa decreaseintheDBHinc

(_−24% in the ambient plot from 2004 to 2005, Fig. 8B) and in the SLA (−17%, Fig. 8G). Similar impacts were observed at theecosystem scalefrom satelliteobservations. These observa-tionsshoweda negativeanomalyof theNormalizedDifference Vegetation Index (NDVI) in the montado areasurrounding our experimentalareafromMarchtoJune2005(−13%ofthemedian averageof1999–2009,datanotshown).Acomparabledropofthe NDVIwasalsoobservedattheregionalscale(Gouveiaetal.,2009). Besides,atthestandscaleBoelmanetal.(2003)showedthatthe increaseinNDVIwasstronglycorrelatedwiththeincreaseinthe abovegroundvegetationbiomassand leafarea. Ourresults con-ﬁrmthoseﬁndingsforanytrends.Thissuggeststhattheregional decreaseinNDVI observedbysatellitebyGouveiaetal.(2009) simplyresultsfromthedecreaseofthevegetationLAIatthelocal scale.

Therewasanoticeablestomatalregulationresponseto sum-merdroughtintensity(Fig.7A)anda66%reductionofthetreesap ﬂowintheambientplot(Fig.2F)attheendofsummer2005 com-paredwiththeendofsummer2004(15/08–15/09).Nonetheless, theaccumulatedannualtreetranspirationdidnotvaryaccordingto theinter-annualvariationsinprecipitationbetween2004and2005 (Fig.4A–C).Pereiraetal.(2007)hadshownsimilarresultsin mea-suringcanopytranspirationthrougheddycovarianceinanearby montado.Limousinetal.(2010)suggestedthattreetranspiration wasregulatedbymechanismsadjustingtotheeco-hydrological equilibriumandmitigatingecosystemchangesinwater availabil-ity.Theauthorsproposedthatleafareaadjustmentwasoneofthe mechanismsinvolvedinoptimizingthebalancebetweenbiomass increaseandwaterloss,aspreviouslyreportedbyEagleson(1982) andRambal(1992).Inourstudy,inter-annualtrendsintheLAI,LAD, SLA,DBH,CVandPCA(Fig.8)indicateaslowdownintheleafbiomass productioninresponsetoprecipitationreductionthatisopposite tothecanopysizeevolution.Thisconclusionissupportedbya sta-bleleaflitterfallamountandanincreasedbranchlitterfallamount from2004to2005(resultsnotshown).Thesemorphologicaltrends showthattrees adjusted towardsa steadyannual accumulated transpirationsupportingthehydro-ecologicalequilibriumtheory. The constant accumulated annual canopy transpiration observed after two contrasting precipitation years (Table 3) stronglysuggeststheuseofgroundwaterbycorkoak.Estimatesof thecontributionofgroundwatertothestandcanopytranspiration during the summer(Table 3) supportthis interpretation.They areinagreementwithformerestimations obtainedfromstable isotopeanalysesatthetreelevel(Davidetal.,2007;Kurz-Besson etal.,2006;Otienoetal.,2007).Deep rootingand groundwater taping provide cork oak a drought avoidance strategy confer-ring trees a high adaptabilityunder scarceand variable water resources. Our results show that groundwater used through deep rooting is to beconsidered as a contributing mechanism towards the eco-hydrological equilibrium in savannah-like oak woodland Mediterranean ecosystems, bufferingchanges in the ecosystemtranspirationregardlessoflargeinter-annualvariations inprecipitation.

There was a larger relative contribution of groundwater to thestandcanopytranspirationinsummer2005(86%)compared with summer 2004 (57%) (Table 3). However, the amount of groundwaterusedbytreeswassimilarin2004and2005.This sug-geststhatthegroundwaterdischargebydeeprootingtapingwas

(10)

littleaffected bytheextremedry year of2004–2005.Allthese resultsraisefurtherquestionsaboutthemechanismsregulating theamountofgroundwaterextractedbyCorkoak’srootsandthe long-termeffectsofincreasingdroughtsonthelevelofsuperﬁcial groundwater.

4.2. Effectsofrainfallmanipulationonphysiologicalresponses andgrowthofcorkoak

AsillustratedbyTable2,therewashardlyanysigniﬁcant treat-menteffectoninstantaneousmeasurementsof,Aorgs,(except oneweekaftertheisolatedlatespringirrigationpulse).Ourresults contrastwiththesigniﬁcanteffectsreportedinrainfallexclusion experiments in Mediterranean woodlands. Alberti et al. (2007) appliedthesamedryandwettreatmentsaswedidinan Arbu-tusunedowoodlandinItaly,andreporteda20%reductionofthe ecosystemgrossprimaryproductivityin2005and35%reduction inecosystemevapotranspirationbetweenthewetanddry treat-ments.Under a precipitation input reduction of 29%, Limousin etal.(2009)reportedacontinuousreductionoftreetranspiration, resultingina23%reductioninannualtranspiration.

In our experiment, the wet, ambient and dry treatments received77%,61%and50%ofthelong-termaverageprecipitationin 2005,respectively.Thustheentireexperimentalplotexperienced severewaterstressregardlessoftheimposedtreatment.Thismight partlyexplaintheweakandirregularimpactofthetreatmentson corkoak’swaterstatusandgasexchangein2005.Nonetheless,tree transpirationdifferedsigniﬁcantlybetweentreesfromthethree treatmentsduringsummerdroughts(Fig.2F).Rainfall exclusion hadaneffectoncanopytranspirationatthestandlevel,aswe mea-sureda20–27%reductionoftranspirationduringsummerinthedry plots,comparedtotheambientones.Incomparison,theannual transpirationonlydecreasedof8to13%(Table3).Bycontrast,the irrigationincreasedthecanopytranspirationby14%and63% dur-ingthe2004and2005summers,respectively,andbyonly8–13% onanannualbasis.

CorkoakSLAreductionfrom2004to2005was10.4%(ambient treatment).Itwaslesspronouncedinthewettreatment(−7.6%) andmorepronouncedinthedrytreatment(−14.7%).Compared withtheotherplots,LAIandLADfromthedryplotsshowed signif-icantoppositetrends(RSC,p<0.001),whichconﬁrmsthetendency forleafareareductionunderwater shortage.Ourresultsarein agreementwiththeLAIreductionpreviouslyobservedbyLimousin etal.(2010).

NosignificanttreatmenteffectwasfoundinCVandPCA.This mightbeexplainedbythenon-limitingsoilmoistureduringthe 3monthsprecedingleafunfoldingin2005(Fig.2Eand3),as pre-viouslysuggestedbyPintoetal.(2012).Thelastisolatedspring irrigation, which occurred after full shoot elongation mayalso explaintheabsenceoftreatmenteffectonPCAandCV.Thisagrees withtheabsenceofsignificanttreatmenteffectonleafstomatal conductanceandcarbonassimilationduringtheshootelongation period.TheconclusionisalsoconsistentwiththeresultsofFotelli etal.(2000),whodidnotfindanyeffectofwaterdeficitsonseedling heightincrementwhilestudyingthephysiologicalresponseof4 oakspeciestowatershortage.

Precipitationinterception bytreecrownsdidnotreducethe exclusionefﬁciencyenoughtoexplaintheweaknessor discontinu-ityofthetreatmenteffectobservedintreesfromthedrytreatment. In ourexperiment, thetreedensity waslowerand theapplied waterregimeshadalowercontrastthanmostoftheother experi-mentsperformedinMediterraneanwoodlands(Albertietal.,2007; Limousinetal.,2009,2010;Missonetal.,2010;Cotrufoetal.,2011). Thesecharacteristicsassociatedwiththeuseofgroundwaterby treesmightexplaintheweakimpactofrainfallexclusioninour study.

Inourexperimentalplot,thelowsoilwaterholdingcapacity(5%, 50mmwaterreservoirbetweenthesurfaceand1mdepth)lowered theabilityoftheecosystemtobufferwaterﬂuctuations,aneffect previouslyproposedbyBeieretal.(2012).Thissoilcharacteristic islikelytoresultinareducedirrigationefﬁciency,especiallywhen irrigationwasappliedimmediatelyafterrainevents.

4.3. Therelevanceofprecipitationandirrigationduringspring Theaccumulatedtreetranspirationmatchedtheprecipitation amount at the stand level in spring 2004 and 2005 (Table 3). Thissuggeststhat standcanopytranspirationand gasexchange inspringaremostlylimitedbytheshallowsoilwateravailability atthepeakofthegrowingseason.Apreviousstudyconductedin thesameexperimentalplotshowedthattreeswererelyingon pre-cipitationandshallowsoilwaterfromfalltospring(Kurz-Besson etal.,2006).AfterlateJune, thesoilwaterpotential( s)inthe

mainrhizospherehorizon(∼−0.3mdepth)rapidlydecreaseduntil −1.5MPa.Then sstartedtodecreaseandﬂuctuateinthedeeper

horizon, suggestingnew rootgrowth.Assummerdrought pro-gressed,treesreliedmoreongroundwaterandredistributedwater fromhydrauliclift(Davidetal.,2007;Otienoetal.,2006; Kurz-Bessonetal.,2006).

Irrigationeventswereintentionallydelayedduringspring2005 toreducedryspelldurationinthewettreatmentatacriticaltime fortheecosystemwateravailability.Atthistimeoftheyear,the high solarradiation coupledwith increasing temperaturesand usuallynot yetlimitedwater availabilityprovidetrees thebest meteorologicalconditionsfor maximumstomatalaperture, car-bonassimilationandtranspiration.Thusalargerimpactofwater manipulationisexpectedduringthisperiod.

Late springirrigation pulses in 2005 ledto a transitorybut significantly betterphysiologicalperformance of trees fromthe wettreatment.Asignificantpositiveimpactwasfoundonmost ofthephysiologicalparametersmeasuredonJuly1st.This posi-tiveeffectcouldnotbedetectedinthemeasurementsperformed thefollowingweeks,despitethesignificanttreatmentdifferences foundinsoilwatercontent(Table2).Treetranspirationresponse tothelastirrigationofJune2005coincidedwiththehighersoil moistureobservedat0.3mdepth(Fig.2EandF,dashedline).On July1st,middayleafwaterpotentialintreeswasalreadyreaching −2MPa,indicatingthatfinerootsinthemainsoilrhizospherelayer (∼0.3–0.4m)wereseverelystressedandprobablydying.Thedelay ofafewdaysinusingirrigatedwaterfromthetop-soilhorizons mightbeduetothenecessityfortreestogrownewfineroots(or repairdamagedfineroots)tobeabletoextractwaterfromthesoil. Thetransientphysiologicalresponseofwettreatmenttreesto springirrigationhadarelevantimpactonstemdiametergrowth resultingina32%increasebetweenJuneandNovember2005,while itonlyreached17%intheambientanddrytreatments.In2004 DBHincreached30%ineachtreatmentoverthesameperiod.This

suggeststhatthespringirrigationin2005compensatedthe neg-ativeimpactoftheextremedryyear2005observedforDBHincon

treesfromthedryandambientplots.Furthermorethedifferences inDBHinc betweentreesfromthewetandtheothertreatments

becamesigniﬁcantfromtheendofspringuntiltheendoffall2005 (datanotshown).Thisrevealsahigh-impactbeneﬁtofspringwater supplementontreestemgrowthandafastintegrationofcarbon assimilatedbyleavestostembiomassafterJuly1st2005(Table2 and Figs.2,5, 6and 8).Asubstantialincrease ofstembiomass inresponsetoirrigationinaMediterraneanforestdominatedby ArbutusunedowasalsoreportedbyCotrufoetal.(2011).

Theconstantexclusionofwaterinthedrytreatment-reducing theprecipitationquantitywithoutchangingrainfall distribution-hadalowimpactonstemgrowth(Table2).Althoughthestudy oftheeffectofprecipitationdistributionwasnotinthescopeof

(11)

thiswork,ourfindingsnotonlyhighlighttheimportanceof pre-cipitationquantityonplantphysiology,but alsoitstiming.Our resultsindicatethatchangesinspringprecipitationmayhave a strongerimpactoncorkoakphysiologyandgrowththanthe over-allchangeintheannualprecipitationinasemi-aridenvironment. SimilarconclusionswereproposedbyBeieretal.(2012).Further experimentsshouldbeperformedtofigureoutwhetherspring pre-cipitationquantity,frequency(dryspelldurationand/ornumberof rainydays),ortiminghasthemaininfluenceoncorkoak physiolog-icalresponses.Thisisofacrucialimportancetobetterpredictcork oakvulnerabilityunderclimatechange,giventhatboth precipi-tationquantityandthenumberofrainydaysinspringdecreased overthelast50years(Gallegoetal.,2011;Mirandaetal.,2006; Paredesetal.,2006), withpotentialconsequences forriverflow andsubsurfacewatersources(Paredesetal.,2006).

4.4. Corkoakdroughtavoidancestrategiesandresilience

Independentlyoftheseverityofdroughts,afullrecoveryof pho-tosyntheticcapacitywasobservedafterfallprecipitation(Fig.7). Stomatalclosureaccompaniedbyadecreaseinleafwater poten-tialisacharacteristicresponseofMediterraneanspeciestoseasonal waterdeﬁcit,whichconfersxericplantswithaconservativewater use(AcherarandRambal,1992;Tenhunenetal.,1981).Asdrought persists,theresultingwaterdeﬁcitsmaycausecavitationinxylem vesselsalongtheroottoleafpathway,andthuscompromisexylem transport(Cochardetal.,1996).Cavitationmaysequentiallylead totheoccurrenceofembolism,impairingwaterdeliverytoliving tissues,whichmightkillthetree(Pereiraetal.,2009).Xylem vul-nerabilitytoembolismisthuscloselylinkedtotheminimummd

atwhichxylemdysfunctionsoccur.Toavoidcatastrophicdamage intheroot-leafhydraulicpathway,leafstomataevolvedtocontrol andrestrainwaterlossessothatplantwaterstatusismaintained abovetheminimummdthresholdofxylemrunawayembolism,

thuspreventingtheplantfromlosingitswatertransportcapability (JonesandSutherland,1991;TyreeandSperry,1988).Thelowest meanvalueofmdmeasuredinourstudyattheendofthesummer

2004and2005(−3.28±0.06MPa)isclosetothewaterpotential causing50%lossinhydraulicconductivity,whichisaproxyofthe specieshydraulicsafetymarginbeforexylemdysfunction(Pinto etal.,2012).

Withpersistentdrought,Q.subershowedamarkeddecreaseof leafosmoticpotentialassociatedtotheraiseofleafsolutecontent, whichsuggeststheoccurrenceofosmoticadjustmentattheleaf level.However,whileleafsolutecontentwaskepthigh through-outthedroughtperiodin2004,adecreaseofswasobservedin

themid-summerof2005.Suchadecreasehasalsobeenreportedby Abrams(1990)andKwonandPallardy(1989).Itwaslikelydueto metabolicimpairmentassociatedwiththeprolongedwaterstress experiencedby thetrees (Chaves et al.,2002; Liu etal., 2011). Nevertheless,theaccumulationofsolutesincellspromotingwater absorptionenabledapositiveleafturgoruptovaluesofpdaround

−2.0MPa(Fig.5).Ourresultsagreewithpreviousstudies show-ingthatseveralMediterraneanevergreenoakspeciesincludingQ. suberpossessthecapacitytoosmoticallyadjust(Arandaetal.,2005; Corcueraetal.,2002;ParkerandPallardy,1988)andreduceleafcell wallelasticityinresponsetodrought(Nardinietal.,1999;Otieno etal.,2006;Pardosetal.,2005).Thesemechanismsactingtogether enhancetreepotentialtoextractwaterfromthedryingsoilsby increasingthesoil-plantwaterpotentialgradient,thusavoidingan excessivedehydrationofthecells.

Leaf surface reduction is a furtherprotective mechanism of dehydrationavoidancerestrainingwaterloss whilemaximizing carbongain,andkeepingminimummd abovethethresholdof

xylemembolism underincreasingdrought intensity.LowerSLA has long been shown to be an adaptive response to drought

(Abrams,1994;Cunninghametal.,1999;Grünzweigetal.,2008) oranindicatorof resourceconditions(Cornelissenetal.,2003). Ramírez-Valienteetal.(2011)foundthatcorkoak’sSLAwas neg-ativelycorrelatedwithannualshootgrowthandpositivelyrelated toannualgrowth.Thisis ingood agreementwiththeopposite trendsweobservedbetweenSLA,CVandPCA(Fig.8).With increas-ing drought severity,leaf areareduction is oftenrelated toan increase in theroot:shootbiomass indicating allocation adjust-mentsfavouringtherootsystem(FernándezandReynolds,2000). Thegrowthofnewrootsindeepersoillayersinresponseto increas-ingdroughtstresshasalsobeenshown bothforcorkoak trees in the same experimentalplot by Otieno et al. (2006) and for desertplantsspecies(FernandezandCaldwell,1975;Peeketal., 2006).

5. Concludingremarks

Ourstudyshowsthatinter-annualdifferencesinprecipitation affectscorkoakannualtrunkgrowth,leafareaandsummer tran-spirationrate.In particularcorkoakphysiologyandgrowthare criticallysensitivetospringprecipitations.Thiswasevidencedby theremarkablecorkoakphysiologicalperformancesfollowinglate springirrigationin2005,whichcompensatedthenegativeimpacts oftheextremedroughtofthehydrologicalyear2004/2005.

Several studies (Knapp et al., 2008; Misson et al., 2010) suggestedthatrainfalldistributionpatterncouldhavelarger phys-iologicalimpacts onoaksthan areduction in precipitation,but our study shows the dominant impact of precipitation in late spring.Thismightweakencorkoakendurancetodrought, espe-cially when theIberian Peninsulahas shown a signiﬁcant 50% (∼40mm)decrease ofprecipitationin Marchover thelastfour decades (Gaetani et al., 2011; Paredes et al., 2006), and when droughteventsarepredictedtooccurmorefrequentlyinApriland MayinthefutureinPortugal(Mirandaetal.,2002).Studying cli-matechangetrendsoverthelast50years,Gallegoetal.(2011) alsoshowedthattheMediterraneanregionsufferedareductionof thenumberofrainydaysinspring.Itis,therefore,notsurprising thatthesouthernIberianPeninsulaisexpectedtobethe Mediter-raneanareamostaffectedbylowergroundwaterrechargesinthe future(Hiscocketal.,2011)andmostthreatenedbydesertiﬁcation (Acácioetal.,2009).

Suchchangesintheamountanddistributionofspringrainfall, andingroundwaterrechargearelikelytoseverelyaffectQ.suberin theIberianPeninsula.Byinducingwaterstressbeforetheonsetof summerdroughts,springprecipitationchangeswillprobablydrive thespeciesclosertothethresholdofcatastrophicxylemembolism byimposingnarrowerhydraulicsafetymarginsatthepeakofthe droughtperiod(Franksetal.,2007;McDowelletal.,2008).Thisis alreadysuggestedbytheevidentsignsofdeclineofthespeciesover thelastdecadeinPortugal(AFN,2001,2010).

Ontheotherhand,ourstudyshowsthatcorkoakrapidlyand completelyrecoveredfromtheextremedryyearof2005or rain-fallexclusion.Treesdidnotexhibitsigniﬁcantsignsofincreased vulnerabilitytothereductionofprecipitationamount.Thisreﬂects howwellthespecieshasadaptedtothehighvariabilityof precip-itationintheMediterraneanclimate.

The20%rainfallexclusionappliedduringtwoconsecutiveyears had little signiﬁcantimpacton corkoak physiology. Thestudy ofMartin-StPauletal.(2013)suggestedthatQuercusilexgrowth (heightandDBH)presentsadelayedresponsetochangesinwater availability.Also,longerrainfallexclusionexperimentshaveshown thatdeleteriouscumulativeeffectssometimesonlyappearafter3 ormoreyearsoftreatmentapplication,orundermoredrastic treat-ments(Brandoetal.,2010;daCostaetal.,2010;Limousinetal., 2010).Therefore,additionallongerexclusionexperimentsshould

(12)

bepromotedtotestforlong-termcumulativeeffectsof precipita-tionchangesoncorkoakspeciesandmontadoecosystems.

Thecorkoaksavannah-typeecosystemshowedahighdegree ofresilienceinresponsetointer-annualprecipitationvariabilityor imposedwatershortage.Ourresultsshowedthatthisresilience partly relies ondeep rootingand theuseof groundwater as a droughtavoidancestrategy.Ourresultsraisenonethelessthe ques-tionoftheevolutionoftheecosystemsustainabilityunderclimate changesifthoseprovokeadecreaseofriverﬂoworanincreaseof anthropogenicpressureonwaterresourcesleadingtoarecurrent declineofgroundwatertablelevelassuggestedbyrecent investi-gations(Estrelaetal.,1996;Noharaetal.,2006;Pac¸oetal.,2009; Paredesetal.,2006).

Acknowledgments

This work was supported by the “Fundação para a Ciência e a Tecnologia” (SFRH/BPD/47021/2008 and Pest-OE/CTE/LA0019/2011-IDL),andwasperformedundertheEuropean MIND-EVK2-CT-2002-00158 (Mediterranean terrestrial ecosys-temsand INcreasingDrought:vulnerability assessment)Project financed bytheEuropean Community.We alsoare very grate-ful toJoão Banza, Elisabete Marianito,Carla Nogueira, for their technicalsupportforfieldinstallation,andtoAndréPestanaand ElsaBreiafortheirtechnicalcontributionwithleafgasexchange measurements.Theauthorsthankthejournalrefereesfor provid-ingconstructivecomments resultingintheimprovementofthe manuscript.

References

Abrams,M.D.,1990.AdaptationsandresponsestodroughtinQuercusspeciesof NorthAmerica.TreePhysiol.7(1–4),227–238.

Abrams,M.D.,1994.Genotypicandphenotypicvariationasstressadaptationsin temperatetreespecies:areviewofseveralcasestudies.TreePhysiol.14(7-8-9), 833–842.

Acácio,V.,Holmgren,M.,Rego,F.,Moreira,F.,Mohren,G.,2009.Aredroughtand wildﬁresturningMediterraneancorkoakforestsintopersistentshrublands. Agrofor.Syst.76(2),389–400.

Acherar,M.,Rambal,S.,1992.ComparativewaterrelationsoffourMediterranean oakspecies.PlantEcol.99–100(1),177–184.

AFN, 2001. Inventário Florestal Nacional: Portugal continental. 3a _revisão:

1995–1998.AutoridadeFlorestalNacional,Lisboa,Portugal.

AFN,2010.5◦_Inventário_Florestal_Nacional_{(2005–2006).}_Relatório_ﬁnal._Autoridade

FlorestalNacional,Lisboa.

Alberti,G.,et al.,2007. Cambiamenti nelregime pluviometricoin ecosistemi mediterranei:ilprogettoMIND.iForest–Biogeosci.or.4(1),460–468.

Aranda,I.,Castro,L.,Pardos,M.,Gil,L.,Pardos,J.A.,2005.Effectsoftheinteraction betweendroughtandshadeonwaterrelations,gasexchangeandmorphological traitsincorkoak(QuercussuberL.)seedlings.ForestEcol.Manag.210(1-3), 117–129.

Barriopedro,D.,Fischer,E.M.,Luterbacher,J.,Trigo,R.M.,García-Herrera,R.,2011.

Thehotsummerof2010:redrawingthetemperaturerecordmapofEurope. Science332(6026),220–224.

Beier,C.,etal.,2012.Precipitationmanipulationexperiments–challengesand rec-ommendationsforthefuture.Ecol.Lett.15(8),899–911.

Berlyn,G.P.,1962.Somesizeandshaperelationshipsbetweentreestemsand crowns.IowaStateJ.Res.37,7–15.

Boelman,N.T.,etal.,2003.ResponseofNDVI,biomass,andecosystemgasexchange tolong-termwarmingandfertilizationinwetsedgetundra.Oecologia135(3), 414–421.

Bota,J.,Medrano,H.,Flexas,J.,2004.IsphotosynthesislimitedbydecreasedRubisco activityandRuBPcontentunderprogressivewaterstress?NewPhytol.162(3), 671–681.

Brando,P.M.,etal.,2010.Seasonaland interannualvariabilityofclimateand vegetationindicesacrosstheAmazon.Proc.Natl.Acad.Sci.U.S.A.107(33), 14685–14690.

Bucci,S.J.,etal.,2005.Mechanismscontributingtoseasonalhomeostasisof min-imum leafwaterpotential andpredawn disequilibriumbetween soiland plantwaterpotentialinNeotropicalsavannatrees.TreesStruct.Funct.19(3), 296–304.

Bugalho,M.N.,Caldeira,M.C.,Pereira,J.S.,Aronson,J.,Pausas,J.G.,2011. Mediter-raneancorkoak savannasrequire human useto sustainbiodiversityand ecosystemservices.Front.Ecol.Environ.9(5),278–286.

Chaves,M.M.,1991.Effectsofwaterdeﬁcitsoncarbonassimilation.J.Exp.Bot.42 (234),1–16.

Chaves,M.M.,Flexas,J.,Pinheiro,C.,2009.Photosynthesisunderdroughtandsalt stress:regulationmechanisms fromwhole planttocell.Ann.Bot.103(4), 551–560.

Chaves,M.M.,Maroco,J.P.,Pereira,J.S.,2003.Understandingplantresponsesto drought–fromgenestothewholeplant.Funct.Plant.Biol.30(3),239–264.

Chaves,M.M.,Oliveira,M.M.,2004.Mechanismsunderlyingplantresiliencetowater deﬁcits:prospectsforwater-savingagriculture.J.Exp.Bot.55(407),2365–2384.

Chaves,M.M.,etal.,2002.Howplantscopewithwaterstressintheﬁeld. Photosyn-thesisandgrowth.Ann.Bot.89,907–916.

Cochard,H.,Breda,N.,Granier,A.,1996.Wholetreehydraulicconductanceand waterlossregulationinQuercusduringdrought:evidenceforstomatalcontrol ofembolism?Ann.For.Sci.53(2/3),197.

Corcuera,L.C.,Camarero,J.J.,Gil-Pelegrín,E.G.-P.,2002.FunctionalgroupsinQuercus speciesderivedfromtheanalysisofpressure–volumecurves.TreesStruct.Funct. 16(7),465–472.

Cornelissen,J.H.C.,etal.,2003.Ahandbookofprotocolsforstandardisedandeasy measurementofplantfunctionaltraitsworldwide.Aust.J.Bot.51(4),335–380.

Cotrufo,M.F.,etal.,2011.Decreasedsummerdroughtaffectsplantproductivity andsoilcarbondynamicsinaMediterraneanwoodland.Biogeosciences8(9), 2729–2739.

Coumou,D.,Rahmstorf,S.,2012.Adecadeofweatherextremes.NatureClim.Change 2(7),491–496.

Cunningham,S.A.,Summerhayes,B.,Westoby,M.,1999.Evolucionarydivergences inleafstructureandchemistry,comparingrainfallandsoilnutrientgradients. Ecol.Monog.69(4),569–588.

daCosta,A.C.L.,etal.,2010.Effectof7yrofexperimentaldroughtonvegetation dynamicsandbiomassstorageofaneasternAmazonianrainforest.New Phy-tologist187(3),579–591.

DavidT.S.,Intercepçãodaprecipitaçãoetranspiraçãoemárvoresisoladasde Quer-cusrotundifoliaLam.DissertaçãodeDoutoramento.UniversidadeTécnicade Lisboa,InstitutoSuperiordeAgronomia,Lisboa,2000,155pp.

David,T.S.,Ferreira,M.I.,Cohen,S.,Pereira,J.S.,David,J.S.,2004.Constraintson tran-spirationfromanevergreenoaktreeinsouthernPortugal.Agr.Forest.Meteorol. 122,193–205.

David,T.S.,etal.,2007.Water-usestrategiesintwoco-occurringMediterranean evergreenoaks:survivingthesummerdrought.TreePhysiol.27(6),793–803.

Diggle,P.J.,Heagerty,P.,Liang,K.-Y.,Zeger,S.L.,2002.AnalysisofLongitudinalData. OxfordStatisticalScienceSeries,2nded.OxfordUniversityPressInc.,NewYork, pp.379.

Eagleson, P.S., 1982. Ecological optimality in water-limited natural soil-vegetation systems: 1.Theoryand hypothesis.WaterResour. Res.18(2), 325–340.

Estrela,T.,Marcuello,C.,Iglesias,A.,1996.WaterResources.ProblemsinSouthern Europe.AnOverviewReport.EuropeanEnvironmentAgency,Copenhagen.

Evangelista,M.,2010.Relatóriodacaracterizaçãodafileiraflorestal2010.Associação paraacompetitividadedaindústriadafileiraflorestal,Portugal,pp.80.

Faria,T.,etal.,1996.Diurnalchangesinphotoprotectivemechanismsinleavesof corkoak(Quercussuber)duringsummer.TreePhysiol.16(1/2),115–123.

Farquhar,G.D.,Sharkey,T.D.,1982.Stomatalconductanceandphotosynthesis.Ann. Rev.PlantPhys.33,317–345.

Fernandez,O.A.,Caldwell,M.M.,1975.Phenologyanddynamicsofrootgrowthof threecoolsemi-desertshrubsunderﬁeldconditions.J.Ecol.63(2),703–714.

Fernández,R.J.,Reynolds,J.F.,2000.Potentialgrowthanddroughttoleranceofeight desertgrasses:lackofatrade-off?Oecologia123(1),90–98.

Fotelli,M.N.,Radoglou,K.M.,Constantinidou,H.-I.A.,2000.Waterstressresponsesof seedlingsoffourMediterraneanoakspecies.TreePhysiol.20(16),1065–1075.

Franks,P.J.,Drake,P.L.,Froend,R.H.,2007.Anisohydricbutisohydrodynamic: sea-sonallyconstantplantwaterpotentialgradientexplainedbyastomatalcontrol mechanismincorporatingvariableplanthydraulicconductance.PlantCell Envi-ron.30(1),19–30.

Gaetani,M.,Baldi,M.,Dalu,G.A.,Maracchi,G.,2011.Jetstreamandrainfall distribu-tionintheMediterraneanregion.Nat.HazardsEarthSyst.Sci.11(9),2469–2481.

Gallego,M.C.,etal.,2011.Trendsinfrequencyindicesofdailyprecipitationoverthe IberianPeninsuladuringthelastcentury.J.Geophys.Res.116(D2),D02109.

García-Herrera,R.,etal.,2007.Theoutstanding2004/05droughtintheIberian Peninsula:associatedatmosphericcirculation.J.Hydrometeorol.8(3),483–498.

Giorgi,F.,Lionello,P.,2008.ClimatechangeprojectionsfortheMediterranean region.Globalplanet.Change63(2-3),90–104.

Gouveia,C.,Trigo,R.M.,DaCamara,C.C.,2009.Droughtandvegetationstress moni-toringinPortugalusingsatellitedata.Nat.HazardsEarthSyst.Sci.9(1),185–195.

Granier,A.,1985.Unenouvelleméthodepourlamesureduﬂuxdesèvebrutedans letroncdesarbres.Ann.For.Sci.42(2),193–200.

Grant,O.M.,etal.,2010.TheimpactofdroughtonleafphysiologyofQuercussuberL. trees:comparisonofanextremedroughteventwithchronicrainfallreduction. J.Exp.Bot.61(15),4361–4371.

Grassi,G.,Magnani,F.,2005.Stomatal,mesophyllconductanceandbiochemical lim-itationstophotosynthesisasaffectedbydroughtandleafontogenyinashand oaktrees.PlantCellEnviron.28(7),834–849.

Grünzweig,J.M.,etal.,2008.Growth,resourcestorage,andadaptationtodrought inCaliforniaandeasternMediterraneanoakseedlings.Can.J.For.Res.38(2), 331–342.

Hiscock,K.,Sparkes,R.,Hodgson,A.,2011.Evaluationoffutureclimatechange impactsonEuropeangroundwaterresources.In:HolgerTreidel,J.L.M.-B., Gur-dak,JasonJ.(Eds.),ClimateChangeEffectsonGroundwaterResources:AGlobal SynthesisofFindingsandRecommendations.IAH–InternationalContributions toHydrogeology.CRCPress,Taylor&FrancisGroup,London,UK,p.414pages.

(13)

ICNF,2013.IFN6-ÁreasdosusosdosoloedasespéciesﬂorestaisdePortugal con-tinental.Resultadospreliminares.InstitutodaConservac¸ãodaNaturezaedas Florestas,Lisboa.

INMG,1991.OClimaemPortugal.NormaisclimatológicasdaregiãodoAlentejoe Algarve,correspondentesa1951–1980.FascículoXLIX,4-4a.InstitutoNacional deMoteorologiaeGeofísica,Portugal,pp.98.

IPCC,2007.FourthAssessmentReport.SummaryforPolicymakers:ClimateChange. ThePhysicalScienceBasis.IPCC.

Jones,H.G.,Sutherland,R.A.,1991.Stomatalcontrolofxylemembolism.PlantCell Environ.14(6),607–612.

Knapp,A.K.,etal.,2008.Consequencesofmoreextremeprecipitationregimesfor terrestrialecosystems.Bioscience58(9),811–821.

Kurz-Besson,C.,etal.,2006.Hydraulicliftincorkoaktreesinasavannah-type Mediterraneanecosystemanditscontributiontothelocalwaterbalance.Plant Soil282(1-2),361–378.

Kwon,K.W.,Pallardy,S.G.,1989.Temporalchangesintissuewaterrelationsof seedlingsofQuercusacutissima,Q.alba,andQ.stellatasubjectedtochronicwater stress.Can.J.For.Res.19(5),622–626.

Lawlor,D.W.,Cornic,G.,2002.Photosyntheticcarbonassimilationandassociated metabolisminrelationtowaterdeﬁcitsinhigherplants.PlantCellEnviron.25 (Part2),275–294.

Lawlor,D.W., Tezara, W., 2009.Causes of decreasedphotosynthetic rate and metaboliccapacityinwater-deﬁcientleafcells:acriticalevaluationof mecha-nismsandintegrationofprocesses.Ann.Bot.103(4),561–579.

Limousin,J.-M.,Longepierre,D.,Huc,R.,Rambal,S.,2010.Changeinhydraulictraits ofMediterraneanQuercusilexsubjectedtolong-termthroughfallexclusion.Tree Physiol.30(8),1026–1036.

Limousin,J.M.,etal.,2009.Long-termtranspirationchangewithrainfalldeclinein aMediterraneanQuercusilexforest.GlobalChangeBiol.15(9),2163–2175.

Liu,C.,etal.,2011.Effectofdroughtonpigments,osmoticadjustmentand antioxi-dantenzymesinsixwoodyplantspeciesinkarsthabitatsofsouthwesternChina. Environ.Exp.Bot.71(2),174–183.

Martin-StPaul,N.K.,Limousin,J.M.,Vogt-Schilb,H.,Rodrıguez-Calcerrada,J., Ram-bal,S.,Longepierre,D.,Misson,L.,2013.Thetemporalresponsetodroughtina Mediterraneanevergreentree:comparingaregionalprecipitationgradientand athroughfallexclusionexperiment.GlobalChangeBiol.19,2413–2426.

McDowell,N.,etal.,2008.Mechanismsofplantsurvivalandmortalityduring drought:whydosomeplantssurvivewhileotherssuccumbtodrought?New Phytol.178(4),719–739.

Miranda,P.,Coelho,F.,Tomé,A.,Valente,M.,2002.20thcenturyPortugueseclimate andclimatescenarios.In:Santos,F.D.,Forbes,K.,Moita,R.(Eds.),ClimateChange inPortugal,ImpactsandAdaptationMeasures–SIAMProject.Gradiva,Lisboa, p.454.

Miranda,P.M.A.,etal.,2006.OclimadePortugalnosséculosXXeXXI.In:Santos,F.D., Miranda,P.M.A.(Eds.),Alterac¸õesclimáticasemPortugalCenários,impactese medidasdeadaptac¸ão.Gradiva,Lisboa,pp.45–113.

Misson,L.,Limousin,J.-M., Rodriguez,R.,Letts,M.G., 2010. Leafphysiological responsestoextremedroughtsinMediterraneanQuercusilexforest.PlantCell Environ.33(11),1898–1910.

Nardini,A.,LoGullo,M.A.,Salleo,S.,1999.Competitivestrategiesforwater availabil-ityintwoMediterraneanQuercusspecies.PlantCellEnviron.22(1),109–116.

Nohara,D.,Kitoh,A., Hosaka,M.,Oki,T., 2006. Impactofclimatechangeon riverdischargeprojectedbymultimodelensemble.J.Hydrometeorol.7(5), 1076–1089.

Otieno,D.O.,etal.,2006.Seasonalvariationsinsoilandplantwaterstatusina QuercussuberL.Stand:rootsasdeterminantsoftreeproductivityandsurvival inthemediterranean-typeecosystem.PlantSoil283(1-2),119–135.

Otieno,D.O.,etal.,2007.RegulationoftranspirationalwaterlossinQuercussuber treesinaMediterranean-typeecosystem.TreePhysiol.27(8),1179–1187.

Pac¸o,T.A.,etal.,2009.EvapotranspirationfromaMediterraneanevergreenoak savannah:theroleoftreesandpasture.J.Hydrol.369(1–2),98–106.

Pardos,M.,Jimenez,M.D.,Aranda,I.,Puertolas,J.,Pardos,J.A.,2005.Waterrelations ofcorkoak(QuercussuberL.)seedlingsinresponsetoshadingandmoderate drought.Ann.For.Sci.62(5),377–384.

Paredes,D.,Trigo,R.M.,Garcia-Herrera,R.,Trigo,I.F.,2006.Understanding precip-itationchangesinIberiainearlyspring:weathertypingandstorm-tracking approaches.J.Hydrometeorol.7(1),101–113.

Parker,W.C.,Pallardy,S.G.,1988.Leafandrootosmoticadjustmentin drought-stressedQuercusalba,Q.macrocarpa,andQ.stellataseedlings.Can.J.For.Res.18 (1),1–5.

Peek,M.S.,etal.,2006.Rootturnoverandrelocationinthesoilproﬁleinresponse toseasonalsoilwatervariationinanaturalstandofUtahjuniper(Juniperus osteosperma).TreePhysiol.26(11),1469–1476.

Pereira,J.S.,Kurz-Besson,C.,Chaves,M.M.,2009.Copingwithdrought.In: Aron-son,J.,Pereira,J.S.,Pausas,J.G.(Eds.),Corkoakwoodlandsontheedge:ecology, adaptivemanagement,andrestoration.PartII.ScientiﬁcBasesforrestoration andmanagement.IslandPress,Washington,DC,p.73-80.

Pereira,J.S.,etal.,2007.Net ecosystemcarbonexchange inthreecontrasting Mediterranean ecosystems –the effect of drought. Biogeosciences 4 (5), 791–802.

Philandras,C.M.,etal.,2011.Longtermprecipitationtrendsandvariabilitywithin theMediterraneanregion.Nat.HazardsEarthSyst.Sci.11(12),3235–3250.

Pinto,C.A.,etal.,2012.Drought-inducedembolismincurrent-yearshootsoftwo Mediterraneanevergreenoaks.ForestEcol.Manag.285(0),1–10.

Rambal,S.,1992.Quercusilexfacingwaterstress:afunctionalequilibrium hypoth-esis.PlantEcol.99–100(1),147–153.

Ramírez-Valiente,J.,Valladares,F.,DelgadoHuertas,A.,Granados,S.,Aranda,I.,2011.

Factorsaffectingcorkoakgrowthunderdryconditions:localadaptationand contrastingadditivegeneticvariancewithinpopulations.TreeGenet.Genomes 7(2),285–295.

Sardans,J.,Pe ˜nuelas,J.,2013.Plant-soilinteractionsinMediterraneanforestand shrublands:impactsofclimaticchange.PlantSoil365(1-2),1–33.

Tenhunen, J.D., Lange, O.L., Braun, M., 1981. Midday stomatal closure in Mediterraneantypesclerophyllsundersimulatedhabitatconditionsinan envi-ronmentalchamber.Oecologia50(1),5–11.

Tezara,W.,Mitchell,V.J.,Driscoll,S.D.,Lawlor,D.W.,1999.Waterstressinhibits plantphotosynthesisbydecreasingcouplingfactorandATP.Nature401(6756), 914–917.

Tyree,M.T.,Sperry,J.S.,1988.Dowoodyplantsoperatenearthepointof cata-strophicxylemdysfunctioncausedbydynamicwaterstress?PlantPhysiol.88 (3),574–580.

Vaz,M.,etal.,2010.Drought-inducedphotosyntheticinhibitionandautumn recov-eryintwoMediterraneanoakspecies(QuercusilexandQuercussuber).Tree Physiol.30(8),946–956.

Watson,D.J.,1947.Comparativephysiologicalstudiesonthegrowthofﬁeldcrops. 1.Variationinnetassimilationrateandleafareabetweenspeciesandvarieties, andwithinandbetweenyears.Ann.Bot.11(41),41–76.

Wesely,M.L.,1989.Parameterizationofsurfaceresistancestogaseousdry deposi-tioninregional-scalenumericalmodels.Atmos.Environ.23(6),1293–1304.

Wu,Z.,Dijkstra,P.,Koch,G.W.,Pe ˜nuelas,J.,Hungate,B.A.,2011.Responsesof ter-restrialecosystemstotemperatureandprecipitationchange:ameta-analysis ofexperimentalmanipulation.GlobalChangeBiol.17,927–942.

Zar,J.H.,1999.BiostatisticalAnalysis,4thed.PrenticeHall,UpperSaddleRiver,New Jersey,998pp.