Impact of urbanization and land-use

Methods),weestimatethattheprobabilityofpopulatingunboundtranslationalstatesuponns(N–H)decay,andhenceproducingmoleculartranslation,is6£1025.Consideringbothfactors(prob-abilityofns(N–H)excitationandprobabilityofdecaytotrans-lation),thesmallmagnitudeoftheyieldmeasuredforpathwayM1isfullyconsistentwithamode-selectivemechanismofinducingmoleculartranslation,whichprevailsatlowertunnellingcurrents.Whenweraisetheexcitationrateinordertobalancethestretch-to-translationdecayratewithanadditionalexcitationofthestretchmode,theenergyaccumulatedinthemoleculeislargeenoughtoactivateanadditionaldecaypathwayproducingmoleculardesorption,whichismediatedbytheumbrellamode(N.L.andJ.I.P.,manuscriptinpreparation).

Moleculardesorptionmediatedbytheds(N–H3)mode,M2,dominateswhenwereducetheelectronenergytoexcluden(N–H)excitations.Thispathwayisdescribedbythreeexcitationsoftheumbrellamodeovertonesinaladder-climbingfashion(Fig.4b).Theammoniagainssufficientvibrationalenergytoovercomethe600-meVadsorptionwell.Theumbrellamodefeaturesasmallercouplingwithtranslationalstatesthanthatfoundforthens(N–H)mode.Moreover,ourcalculationsshowthatds(N–H3)vibrationalstatesabove360meVmaywellpopulatealong-livedtransitionstateconnectedwiththecompleteinversionofthemolecule,whichleadstodesorptionafteranadditionalexcitation.Inthisprocesstheumbrellamodeitselfbecomesthevibrationalstatealongthereactioncoordinate(Fig.4b),aswaspreviouslydeducedfromultravioletandinfraredphotodesorptionexperiments19,20.

Itwouldbeinterestingtoextendthismethodologytomorecomplexprocesses,searchingforstrategiesofcontrollingandenhancingreactivityatsurfacesthatmaybeappliedatthemacro-scopicscale.ThecontrolledenvironmentfurnishedbytheSTMallowsthedetectionofreactionmechanismsinthelimitofverylowyieldandverylowpowerirradiation.Insuchsingle-moleculestudies,weexpectthatmode-selectivestrategieswillbecomeimportantinthediscoveryofreactionpathwaysthatareinaccessiblebyclassical‘thermal’chemistry.A

9.Ho,W.Inducingandviewingbondselectedchemistrywithtunnelingelectrons.Acc.Chem.Res.31,

567–573(1998).

10.Eigler,D.M.,Lutz,C.P.&Rudge,W.E.Anatomicswitchrealizedwiththescanningtunnelling

microscope.Nature352,600–603(1991).

11.Stipe,B.C.,Rezai,M.A.&Ho,W.Couplingofvibrationalexcitationtotherotationalmotionofa

singleadsorbedmolecule.Phys.Rev.Lett.81,1263–1266(1998).

12.Komeda,T.,Kim,Y.,Kawai,M.,Persson,B.N.J.&Ueba,H.Lateralhoppingofmoleculesinducedby

excitationofinternalvibrationmode.Science295,2055–2058(2002).

13.Bartels,L.etal.Atomicscalechemistry:DesorptionofammoniafromCu(111)inducedbytunneling

electrons.Chem.Phys.Lett.313,544–552(1999).

14.Avouris,Ph.Manipulationofmatterattheatomicandmolecularlevel.Acc.Chem.Res.28,95–102

(1995).

15.Stipe,B.C.etal.Single-moleculedissociationbytunnelingelectrons.Phys.Rev.Lett.78,4410–4413

(1997).

16.Hla,S.W.,Bartels,L.,Meyer,G.&Rieder,K.H.Inducingallstepsofachemicalreactionwiththe

scanningtunnelingmicroscopetip:Towardssinglemoleculeengineering.Phys.Rev.Lett.85,2777–2780(2000).

17.Kim,Y.,Komeda,T.&Kawai,M.Single-moleculereactionandcharacterizationbyvibrational

excitation.Phys.Rev.Lett.,126104(2002).

18.Rust,H.-P.,Buisset,J.,Schweizer,E.K.&Cramer,L.Highprecisionmechanicalapproachmechanism

foralowtemperaturescanningtunnelingmicroscope.Rev.Sci.Instrum.68,129–132(1997).

19.Hertel,T.,Wolf,M.&Ertl,G.UVphotostimulateddesorptionofammoniafromCu(111).J.Chem.

Phys.102,3414–3430(1995).

20.Hussla,I.etal.Infrared-laser-inducedphotodesorptionofNH3andND3adsorbedonsingle-crystal

Cu(100)andAgfilm.Phys.Rev.B32,34–3501(1985).

21.Prybyla,J.A.,Heinz,T.F.,Misewich,J.A.,Loy,M.M.T.&Glownia,J.H.Desorptioninducedby

femtosecondlaserpulses.Phys.Rev.Lett.,1537–1540(1990).

22.Salam,G.P.,Persson,M.&Palmer,R.E.Possibilityofcoherentmultipleexcitationinatomtransfer

withascanningtunnelingmicroscope.Phys.Rev.B49,10655–10662(1994).

23.Lorente,N.&Persson,M.Theoryofsinglemoleculevibrationalspectroscopyandmicroscopy.Phys.

Rev.Lett.85,2997–3000(2000).

24.Lorente,N.&Persson,M.Theoreticalaspectsoftunneling-current-inducedbondexcitationand

breakingatsurfaces.FaradayDiscuss.117,277–290(2000).

AcknowledgementsJ.I.P.acknowledgesresearchcontracts‘MarieCurie’(EU)and‘Ramony

´a).N.L.acknowledgessupportfromACIJeunesCajal’(MinisteriodeCienciayTecnologı

Chercheurs,andtheCNRSprogramme‘Nano-ObjetIndividuel’.Allcalculationswereperformed

´rieur(CINES)andtheCentredeattheCentred’InformatiqueNationaldel’EnseignementSupe

´ne´es(CALMIP).CalculMidi-Pyre

CompetinginterestsstatementTheauthorsdeclarethattheyhavenocompetingfinancial

interests.

CorrespondenceandrequestsformaterialsshouldbeaddressedtoJ.I.P.(pascual@icmab.es).

Methods

Toestimatethepotentialbarriersfortranslationanddesorption,weperformedtotal-energycalculationsusingplanewavesandpseudopotentialsinthegeneralizedgradientapproximationofdensityfunctionaltheory.Usingthemethodofrefs23and24,wealsoestimatedchemisorbedammoniafrequencies,theprobabilityofexcitationofeachmodeandtheirlifetime.Thenumericalresultsfortherelevantmodesofthisworkare:thesymmetricandoneantisymmetricstretchmode,ns(N–H)andn1a(N–H),at408meVand422meV(294meVand311meVforND3),bothwithaprobabilityPn¼6£1024ofbeingexcitedperimpingingelectronandalifetimeof4psand8ps,respectively;theumbrellamode,ds(N–H3),at139meV(104meVforND3),40£1024and25ps;thescissorsmodes,da(N–H3),degenerateat200meV(145meVforND3),2£1024and11ps.

Thedecayprobability,Pm,ofasinglens(N–H)excitationdecayingintotheunboundhinderedtranslationlevelspluselectron–holeexcitationisestimatedasthefractionwm.30/w.wm.30isthesumofdampingratesintostatesmabovethebarrierfor

translation,andwisthedampingrateofthens(N–H)mode,viaexcitationofelectron–holepairsandhinderedtranslationalstates.AnupperlimitofthetranslationyieldisthenYR¼Pn£Pm¼3£1028moleculesperelectron.

Received20December2002;accepted8April2003;doi:10.1038/nature019.

1.Jortner,J.,Levine,R.D.&Pullman,B.(eds)ModeSelectiveChemistry(KluwerAcademic,Dordrecht,1991).

2.Dai,H.L.&Ho,W.(eds)LaserSpectroscopyandPhotochemistryonMetalSurfaces(WorldScientific,Singapore,1995).

3.Higgings,J.,Conjusteau,A.,Scoles,G.&Bernasek,S.L.Stateselectivevibrational(2n3)activationofthechemisorptionofmethaneonPt(111).JChem.Phys.114,5277–5283(2001).

4.Potter,E.D.,Herek,J.L.,Pedersen,S.,Liu,Q.&Zewail,A.H.Femtosecondlasercontrolofachemicalreaction.Nature355,66–68(1992).

5.Sinha,A.,Hsiao,M.C.&Crim,F.F.Controllingbimolecularreactions:Modeandbondselectedreactionofwaterwithhydrogenatoms.J.Chem.Phys.94,4928–4935(1991).

6.Bronikowski,M.J.,Simpson,W.R.,Girard,B.&Zare,R.N.Bond-specificchemistry:OD:OHproductratiosforthereactionsHþHOD(100)andHþHOD(001).J.Chem.Phys.95,87–88(1991).

7.Hansma,P.K.(ed.)TunnellingSpectroscopy:Capabilities,Applications,andNewTechniques(Plenum,NewYork,1982).

8.Stipe,B.C.,Rezaei,M.A.&Ho,W.Single-moleculevibrationalspectroscopyandmicroscopy.Science280,1732–1735(1998).

..............................................................

Impactofurbanizationandland-usechangeonclimate

EugeniaKalnay&MingCai

UniversityofMaryland,CollegePark,Maryland20770-2425,USA

.............................................................................................................................................................................

Themostimportantanthropogenicinfluencesonclimatearetheemissionofgreenhousegases1andchangesinlanduse,suchasurbanizationandagriculture2.Butithasbeendifficulttosepar-atethesetwoinfluencesbecausebothtendtoincreasethedailymeansurfacetemperature3,4.Theimpactofurbanizationhasbeenestimatedbycomparingobservationsincitieswiththoseinsurroundingruralareas,buttheresultsdiffersignificantlydependingonwhetherpopulationdata5orsatellitemeasure-mentsofnightlight6–8areusedtoclassifyurbanandruralareas7,8.HereweusethedifferencebetweentrendsinobservedsurfacetemperaturesinthecontinentalUnitedStatesandthecorrespondingtrendsinareconstructionofsurfacetemperaturesdeterminedfromareanalysisofglobalweatheroverthepast50years,whichisinsensitivetosurfaceobservations,toestimatetheimpactofland-usechangesonsurfacewarming.Ourresultssuggestthathalfoftheobserveddecreaseindiurnaltemperaturerangeisduetourbanandotherland-usechanges.Moreover,ourestimateof0.278Cmeansurfacewarmingpercenturydueto

NATURE|VOL423|29MAY2003|www.nature.com/nature

528©2003 Nature PublishingGroup

letterstonature

land-usechangesisatleasttwiceashighaspreviousestimatesbasedonurbanizationalone7,8.

TwomethodsusedintheUStoclassifymeteorologicalstationsintourbanandruralto‘correct’theobservedsurfacetemperaturetrendsforurbanizationeffectsarebasedonpopulationdata5andsatellitemeasurementsofnight-light6–8,respectively,andthecorre-spondingestimatesoftheimpactofurbanizationdifferinmagni-tude(0.06and0.158Cpercentury)7,8.Thefindingthatatmospherictemperaturesasmeasuredbysatellitesandweatherballoonshavesmallerwarmingtrendsthansurfaceobservationshasbeenthesubjectofmuchdiscussion9centredmostlyonthequalityofthedata,butitcouldbepartiallyexplainedbyapredominanceofland-useeffectsovergreenhousewarmingnearthesurface.

WeestimatedtheimpactofurbanizationandotherlandusesonclimatechangebycomparingtrendsobservedbysurfacestationswithsurfacetemperaturesderivedfromtheNCEP-NCAR50-yearReanalysis(NNR)10.IntheNNR(astatisticalcombinationof6-hourforecastsandobservations),surfaceobservationsoftempera-ture,moistureandwindoverlandarenotused11.However,atmosphericverticalsoundingsofwindandtemperature(rawin-sondesandsatellitesoundings)stronglyinfluencetheNNR,andsurfacetemperaturesareestimatedfromtheatmosphericvalues.Asaresult,theNNRshouldnotbesensitivetourbanizationorland-useeffects,althoughitwillshowclimatechangestotheextentthattheyaffecttheobservationsabovethesurface.

AsindicatedbyFig.1andmanyotherstudies,theNNRcaptureswellsurfacetemperaturevariationscausedbyatmosphericstorms,advectionofwarm/coldair,andvariationsinthefrequencyortrackofmajorstorms.Incontrasttotheactualsurfaceobservations,wefindnostatisticallysignificantdifferenceintheNNRestimationofurbanandruralstationtrends(seeMethods).Theseargumentssuggestthatwecouldattributethedifferencesbetweenmonthlyorannuallyaveragedsurface-temperaturetrendsderivedfromobser-vationsandfromtheNNRprimarilytourbanizationandotherchangesinlanduse.

Wecomparethedailymaximumandminimumtemperaturesof1,982surfacestationslocatedbelow500minthe48contiguousUnitedStates,andthedailysurfacemaximumandminimumtemperaturesona2.58gaussiangridfromtheNNRinterpolatedtothestationlocations,bothfortheperiod1950–1999.Wecomputetemperatureanomalieswithrespecttothe50-yearmeanannualcycleforeachsiteandeachdataset.Trendsarecomputedaschangesindecadalaveragesintheanomaliestoreducerandomerrors.TheNNR(1948tothepresent)hasbeenconstructedwithamodelanddataassimilationsystemkeptunchanged,butitisaffectedbychangesintheobservingsystems,especiallytheintroductionofthesatelliteobservingsystemin1979.Therefore,inthecompu-tationoftrendsweexcludechangesfromthedecadeofthe1970stothe1980s.

Figure1comparestimeseriesof50yearsofmonthlymeantemperatureanomaliesforBaltimore,alargecityinMaryland,includingtheaveragedecadaldifferencebetweenobservationsandtheNNR.Thereisgoodagreementintheinterannualvariability,withacorrelationofover0.9,butalsoagrowingtrendinthedifferencebetweenthesurfaceobservationsandNNR,increasingto1.48Cduringthe1990s,adifferenceweattributetourbanizationandothersurfacechangesthatdonotaffecttheNNR.Asimilaranalysisonallsurfacestations(SupplementaryFig.1)indicatesa50-yearcorrelationofabout0.9everywhereexceptinmountainousregions,whereitisbetween0.4and0.7,whichiswhyweonlyincludestationslocatedbelow500m.Thecorrelationisalsolowerinthewestcoast,possiblyowingtotheproximityofmountainsortolowdatadensityinthePacificOcean.Decadaltrendscanbelocallydominatedbyinterannualanddecadalvariabilityofthetemperatureduetoanomaliesinthecirculationratherthantolandusechange—effectsthatareexcludedbytakingthedifferencesbetweensurfaceandNNRtemperatures.

NATURE|VOL423|29MAY2003|www.nature.com/nature

Thedecadaltrendaveragedoverthetwoseparate20-yearperiods(1980–1999,and1960–1979)iscomputedforeverystationandaveragedinboxesof0.58latitudeby0.58longitude,withanoverallaveragecomputedoveralltheboxes.Figures2and3showtheaveragetrendsfortheobservationsandfromtheNNR,andalsothedifferencebetweenthesetwotrends,whichisatleastpartiallyattributabletochangesinuseofthelandsurface.

Themaximumtemperature(Fig.2)showsawarmingtrendintheobservationsintheeasternandwesternUSandacoolingtrendintheMidwest,withaslightlynegativeoverallaverageof20.0178Cperdecade.TheNNRissimilarbutsmoother,withanaverageofþ0.0088Cperdecade.ThedifferencebetweentheobservedandNNRtrendsissomewhatnegativeinmostofthecountryeastoftheRockies,butisstronglypositiveinCaliforniaandtoalesserextent,inOregonandWashington,withanaveragedifferenceof20.0258Cperdecade.

Theminimumtemperature(Fig.3)observationsshowamuch

Figure1ComparisonofmonthlymeanstationandNNRsurfacetemperatureanomalieswithrespecttotheirannualcyclesforthecityofBaltimore,Maryland,USA.Tob,observedmonthlymeantemperaturein8C,showninred.Tan,analysedmonthlymeantemperaturein8C,showninblue.Fivedecades(1950to1999)areshownforcomparison.

529

©2003 Nature PublishingGroup

letterstonature

strongerpositivetrendinmostofthecountry,withanaverageofþ0.1938Cperdecade.IntheNNR,theminimumtemperatureincreaseseverywhereexceptintheMidwestandCalifornia,withanaverageofþ0.1138Cperdecade.ThedifferenceinminimumtemperaturetrendsbetweenobservedandNNRvaluesispositiveinmostofthecountry,especiallyinCalifornia,withanaverageof0.0808Cperdecade(40%oftheobservedtrend).

SupplementaryFig.2showsthetrendinthedifferencebetweenmaximumandminimumtemperaturesordiurnaltemperaturerange(DTR).IntheobservationstheDTRtrendisstronglynegativeinmostofthecountrywithanaveragedecreaseof20.2108Cperdecade.TheNNRalsoshowsageneraldecreaseofDTR,withanationalaverageof20.1058CperdecadesothatabouthalfofthedecreaseinDTRcouldbeattributabletosurfacechanges.

Thedailymeantemperatureobservationtrends(SupplementaryFig.3)obtainedastheaverageofthemaximumandminimumtemperaturesshowanincreaseinmostofthecountry,withanaveragetrendofþ0.0888Cperdecade.TheNNRtrendshaveanaverageofþ0.0618Cperdecade.Ofthetwo“urbancorrection”estimates7,8,ourestimateof0.278Cpercenturyattributabletolanduseisclosertotheestimatebasedonthenight-lighturbaneffect(seecentrebottomofplate3Cinref.7)thantotheestimatebasedonpopulationdensity(seecentrebottomofplate3Binref.7).Itshouldbenotedthatourobserveddailymeantemperaturetrends

a

4845423936333027

–0.017 °C

10.50.30.20.10.050–0.05–0.1–0.2–0.3–0.5–1

(SupplementaryFig.3a)aredifferentfromprevious50-yearor100-yeartrendestimates(seeplates3Aand7Ainref.7)becauseinourcomputationswedidnotinclude(1)thedecadaltrendscorre-spondingtothe1980s–1970sandespeciallythe1960s–1950s,and(2)urbanandnon-urbandataadjustments.Thenon-urbanadjust-mentstendtobestronglypositiveexceptovertheRockies(seeplate3Binref.7),sothatifwehadaddedthemtotherawobservations,ourestimateoftheland-useimpactonthemeantemperaturetrendswouldhavebeengeographicallysimilarbutlarger.

AlthoughitisnotpossibledefinitivelytoattributethedifferencesbetweentheobservationandtheNNRtemperaturetrendssolelytolanduse,includingurbanization,agricultureandirrigation,ourresultsarecompatiblewithsuchaninterpretation.Thewell-known‘urbanheatisland’effectactuallytakesplaceatnight,whenbuild-ingsandstreetsreleasethesolarheatingabsorbedduringtheday.Atthetimeofthemaximumtemperaturetheurbaneffectisoneofslightcooling,owingtoshading,aerosols,andtothermalinertiadifferencesbetweencityandcountrythatarenotcurrentlywellunderstood12.

Theeffectofagriculturaldevelopment,increasingevaporationduringtheday,wouldalsotendtodecreasethemaximumtem-perature:irrigationwouldincreasetheheatcapacityofthesoil,thusincreasingtheminimumtemperature.Therefore,bothurbaniz-ationandagricultureeffectscouldbeconsistentwiththegeneralincreaseintheminimumtemperatureandslightdecreaseinthemaximumtemperature,andcontributetothereductioninthediurnaltemperaturerangeshowninourestimateseastoftheRockies(Figs2cand3c).

Thisimpliesthatthecomparisonofurbanandruralstationswithoutincludingagriculturaleffectswouldunderestimatethetotal

a

48454239

Maximum temperature (°C)36333027

+0.193 °C

10.50.30.20.10.050–0.05–0.1–0.2–0.3–0.5–1

b

4845Latitude (°N)423936333027

+0.008 °C

10.50.30.20.10.050–0.05–0.1–0.2–0.3–0.5–1

b

4845Latitude (°N)423936333027

+0.113 °C

10.50.30.20.10.050–0.05–0.1–0.2–0.3–0.5–1

Maximum temperature (°C)c

4845423936333027

120115110105100959085

Longitude (°W)

80

75

70–0.025 °C

10.50.30.20.10.050–0.05–0.1–0.2–0.3–0.5–1

c

4845423936333027

120115110105100959085

Longitude (°W)

80

75

70+0.080 °C

10.50.30.20.10.050–0.05–0.1–0.2–0.3–0.5–1

Figure2DecadaltrendofthemaximumtemperatureaveragedforeveryUSstationbelowanelevationof500m.Eachvalue(in8Cperdecade)wascalculatedfromtheaverageofthe‘1990sminusthe1980s’andthe‘1970sminusthe1960s’maximumtemperatures.Thestationvaluesaredisplayedasaveragesinboxesof0.58latitudeby0.58longitude.Blankboxesindicatethatnoneofthe1982stationsiswithintheboxes,andthenationalaverageistheaverageoftheseboxes.Theaveragevalueofthetrendisindicatedineachpanel.a,Station(observed)maximumtemperaturetrends.b,NNR(analysed)maximumtemperaturetrends.c,‘Observedminusanalysed’maximumtemperaturetrends.

530

Figure3DecadaltrendoftheminimumtemperatureaveragedforeveryUSstationbelow500m.LegendasforFig.2butfortheminimumtemperatures.

NATURE|VOL423|29MAY2003|www.nature.com/nature

©2003 Nature PublishingGroup

letterstonature

impactofland-usechanges.Morestudiesareneeded,includingacomparisonofgeographicaldistributionofNNRtrendswithotherupper-airobservations,suchasrawinsondesandsatellites,amoreprecisedefinitionoftheurbanandruralobservingstations,andtheimpactofotherhumanactivitiessuchascontrailsandaerosolsthatcanalsoreducethediurnaltemperaturerange13.

Ourmethodcanincorporateupdatedobservationsastheybecomeavailable,canbeappliedtolandstationsthroughouttheworld,toothervariablessuchashumidityandwinds,detectseasonaltrends,andsignalchangesinstationlocationsthatareotherwisedifficulttoidentify.A

9.

climatologynetworkbasedonsatellite-designatedlanduse/landcover.J.Clim.12(5),1344–1348

(1999).

NRCBoardonAtmosphereSciencesandClimateCommitteePanel,JohnM.Wallace,chair,pages1–71.ReconcilingObservationsofGlobalTemperatureChange(NationalResearchCouncil,WashingtonDC,2000).

Kalnay,E.etal.TheNCEP/NCAR40-yearreanalysisproject.Bull.Am.Meteorol.Soc.77,437–431(1996).

Kistler,R.etal.TheNCEP/NCAR50-yearreanalysis:MonthlymeansCD-ROManddocumentation.Bull.Am.Meteorol.Soc.82,247–267(2000).

Runnalls,K.E.&Oke,T.R.Dynamicsandcontrolsofthenear-surfaceheatislandofVancouver,BC.Phys.Geogr.21,283–304(2000).

Travis,D.J.,Carleton,A.M.&Lauritsen,R.G.Contrailsreducedailytemperaturerange.Nature418,601(2002).

10.11.12.13.

Methods

Data

Forthesurfaceobservations,weusethedailysurfacemaximumandminimum

uncorrectedsurfacestationtemperaturesfromtheNationalClimateDataCenter(NCDC)‘CooperativeSummaryoftheDay’datasetoverthe48contiguousstatesoftheUnitedStatesfor1950–1999.FortheNNR,weusetheglobaldailysurfacemaximumandminimumtemperaturesgriddedon2.58gaussianboxes,alsofortheperiod1950–1999.

SupplementaryInformationaccompaniesthepaperonwww.nature.com/nature.

AcknowledgementsThisstudywaspartiallysupportedbyaRiskPredictionInitiativegrant.WearegratefultoA.Senserini,whoperformedmostofthecomputations,toJ.E.JanowiakandW.Ebisuzakiwhoprovidedthedata,andtoR.Murnane,T.Oke,J.Hansen,E.Rassmusson,R.PielkeSr,T.vonderHaarandZ.Lifordiscussions.

CompetinginterestsstatementTheauthorsdeclarethattheyhavenocompetingfinancialinterests.

CorrespondenceandrequestsformaterialsshouldbeaddressedtoE.K.(ekalnay@atmos.umd.edu).

Analysis

WeinterpolatelinearlythegriddedNNRdatatoeachobservationalsite,andonlyconsiderthesitesthathaveatotalofatleast480(whole)monthsofobservations.Inaddition,becausetheNNRhassurfaceheightsdifferentfromthoseofthereallocations,andextrapolationsundergroundcanintroduceerrorsoverwhelmingthesignaloftherealtrends(SupplementaryFig.2),inthecomputationofthetrendsweonlyconsidersiteswithelevationslowerthan500m.Thereare1,982USsurfacestationssatisfyingthesetwoconditions.Weobtainmonthlymeansbyaveragingdailydata;dailymeantemperaturesareobtainedbyaveragingmaximumandminimumtemperatures,anddailytemperaturerangesbysubtractingtheminimumfromthemaximumtemperature.

BecausetheNNRcanhavesystematicdifferenceswithobservations,especiallynearthesurface,owingtodeficienciesinthemodelforecastorthemethodofassimilation,weremovethe50-yearmonthlymeanannualcycleforeachsitefromboththeobservationsandtheNNR.Wearethuscomparinganomalieswithrespecttothe50-yearmeanannualcycle.Intheresultswepresentbothcomparisonsofthe50-yeartimeseriesandtrends.Thetrendsarecomputedaschangesindecadalaveragesinordertoreducerandomerrors.Weonlyconsidertwodecadaltrends:thedecade1990–1999minus1980–19,and1970–1979minus1960–1969.Wedonotincludeinthetrendsthedifferencebetweenthedecades1960–1969and1950–1959,becausetheobservingsystemduringthe1950swas

considerablylessreliablethaninlaterdecades,anditunderwentsignificantschedulingchangesduring1958(ref.11).

Inaddition,wehavetoaddresschangesintheobservingsystems,especiallytheintroductionofthesatelliteobservingsystem(ofwhichthemostimportantistheTIROS-NOperationalVerticalSounder,TOVS)startingin1979.ThesetwomajorchangesarethemainreasonwhytrendsintheNNRneedtobecarefullyestimated.Wethereforedonotincludethechanges1980–19minus1970–1979.Thetwodecadalchangesthatwekeepcorrespondtothe1990sminus1980s(20yearswithsatellitedata),and1970sminus1960s(20yearsessentiallywithoutsatellitedata).Thus,whenweaveragethemweobtaindecadaltrendsfromtwoindependentandlargelyhomogeneous20-yearperiods.

Wecomparedthe1990sversus1980strendof775stationsclassifiedasurbanversus167stationsclassifiedasrural.Themeansurfacetemperatureincreasedby0.318Cfortheurbanstationsand0.138Cfortheruralstations,withstandarddeviationsofabout0.58Ceach.Thedifferencebetweenurbanandruralwarming,0.188C,issignificantata99%levelofsignificance.Thetrendsforthereanalysisstationestimatesare0.268Cforurbanand0.258Cforrural,withstandarddeviationsofabout0.228C,andthedifference0.018Cbetweenurbanandruralisinsignificant,showingthattheNNRisinsensitivetosurfaceeffects.

Inthetimeserieswecomputethe1950–1959averagetemperaturedifferencebetweentheNNRandthesurfacestationateachstationandsubtractitfromtheNNR.Thisforcesthetwotimeseriestohavethesame10-yeartimeaverageduringthe1950sandisdonefordisplaybutdoesnotaffectthecomputationofthetrendsorcorrelations.

Received18December2002;accepted23April2003;doi:10.1038/nature01675.

1.IPCCClimateChange2001:TheScientificBasis(CambridgeUniv.Press,Cambridge,UK,2001).2.Pielke,R.A.Sretal.Theinfluenceofland-usechangeandlandscapedynamicsontheclimatesystem:

Relevancetoclimate-changepolicybeyondtheradiativeeffectsofgreenhousegases.Phil.Trans.R.Soc.Lond.A360,1–15(2002).

3.Gallo,K.P.&Owen,T.W.Satellite-basedadjustmentsforurbanheatislandtemperaturebias.J.Appl.

Meteorol.38,806–813(1999).

4.Owen,T.W.,Gallo,K.P.,Elvidge,C.D.&Baugh,K.E.UsingDMSP-OLSlightfrequencydatato

categorizeurbanenvironmentsassociatedwithUSclimateobservingstations.Int.J.RemoteSensing19,3451–3456(1998).

5.Easterling,D.R.etal.Maximumandminimumtemperaturetrendsfortheglobe.Science277,3–

367(1997).

6.Gallo,K.P.,Easterling,D.R.&Peterson,T.C.Theinfluenceoflanduse/landcoveronclimatological

valuesofthediurnaltemperaturerange.J.Clim.9(11),2941–2944(1996).

7.Hansen,J.E.etal.AcloserlookatUnitedStatesandglobalsurfacetemperaturechange.J.Geophys.

Res.106,23947–23963(2001).

8.Gallo,K.P.,Owen,T.W.,Easterling,D.R.&Jamason,P.F.TemperaturetrendsoftheUShistorical

..............................................................

FieldsportsandconservationintheUnitedKingdom

T.E.E.Oldfield,R.J.Smith,S.R.Harrop&N.Leader-Williams

DurrellInstituteofConservationandEcology,UniversityofKent,Canterbury,KentCT27NS,UK

.............................................................................................................................................................................

Manynaturalhabitatsexistonprivatelyownedlandoutsideprotectedareas1,butfewgovernmentscanaffordtoenforceorsubsidizeconservationofthisbiodiversity.Eveninsomedevel-opedcountries,conservationsubsidyschemeshaveonlyachievedlimitedsuccess2–4.Fortunately,somelandownersmaybewillingtoacceptmanagementcostsinreturnforotherbenefits5,althoughthisremainscontroversialwhenitinvolvesthekillingofcharismaticspecies.Forexample,participantsinBritishfieldsports,suchasfoxhuntingandgame-birdshooting,mayvolun-tarilyconserveimportanthabitatsthatarerequiredbyquarryspecies6–8.HerewereportresultsfromamultidisciplinarystudythataddressedthisissuebyfocusingonthreesitesacrosscentralEngland.Wefoundthatlandownersparticipatinginfieldsportsmaintainedthemostestablishedwoodlandandplantedmorenewwoodlandandhedgerowsthanthosewhodidnot,despitetheequalavailabilityofsubsidies.Therefore,voluntaryhabitatmanagementappearstobeimportantforbiodiversityconserva-tioninBritain.CurrentdebatesonthefutureoffieldsportsinBritain,andsimilaractivitiesglobally,maybenefitfromcon-sideringtheirutilityasincentivestoconserveadditionalhabitatonprivateland.

Privatelandownersplayanincreasinglyimportantroleinbio-diversityconservation1.Thisisespeciallyimportantwherehabitatsformisolatedremnantsinanagriculturalmatrix,anditispoliticallydifficulttoestablishlargeprotectedareas9.ThisistypifiedbythesituationinBritain,wherefarmlandcovers76%ofthecountryandincreasesinagriculturalefficiencyhavecausedgreatdeclinesinbiodiversity7,10,11.TheBritishgovernmenthasrespondedbyintro-ducinglegislationtoprotectimportanthabitatsandspeciesonpublicandprivateland12–14,aswellasestablishingsubsidyschemes11,15.However,conservationlegislationremainsunpopular

531

NATURE|VOL423|29MAY2003|www.nature.com/nature©2003 Nature PublishingGroup

letterstonature

19.Otwinowski,M.&Minor,W.ProcessingofX-raydiffractiondatacollectedinoscillationmode.MethodsEnzymol.276,307–326(1997).

20.Terwilliger,T.C.&Berendzen,J.AutomatedMADandMIRstructuresolution.ActaCrystallogr.D55,849–861(1999).

21.Terwilliger,T.C.Maximumlikelihooddensitymodification.ActaCrystallogr.D56,965–972(2000).22.Jones,T.,Zou,J.-Y.,Cowan,S.&Kjeldgaard,M.Improvedmethodsforbuildingproteinmodelsinelectrondensitymapsandthelocationoferrorsinthesemodels.ActaCrystallogr.A47,110–119(1991).ˆnger,A.T.etal.Crystallography&N.M.R.system:anewsoftwaresuiteformacromolecular23.Bru

structuredetermination.ActaCrystallogr.D54,905–921(1998).

..............................................................

corrigenda

SupplementaryInformationaccompaniesthepaperonwww.nature.com/nature.

AcknowledgementsWearegratefultoJ.R.H.Tameforacriticalreadingofthemanuscript.WethankZ.Noforprovidingsildenafilcitrate;D.-K.Kimforprovidingvardenafil;D.K.Shinfordiscussionandfigures;andH.-S.LeeandG.-H.KimfortheirassistanceatthePohangLightSource(PLS),beamline6B.ExperimentsatPLSweresupported,inpart,bytheMinistryofScienceandTechnology(MOST)ofKoreaandPOSCO.WealsothankS.Y.P’sgroupfortheirassistanceatSpring-8forhigh-resolutiondata.ThisworkwassupportedpartiallybyagrantfromtheNationalResearchLaboratoryProgramandtheCenterforBiologicalModulatorsofthe21cFrontierR&DProgram,subsidizedMOST.ThisworkwasalsosupportedpartlybyYuyuInc.andKT&GCo.Ltd..

CompetinginterestsstatementTheauthorsdeclarethattheyhavenocompetingfinancialinterests.

CorrespondenceandrequestsformaterialsshouldbeaddressedtoS.G.R.

(sgro@crystalgenomics.com)orJ.M.C.(jmcho@crystalgenomics.com).Coordinatesforthesildenafil,tadalafilandvardenafilcomplexstructurehavebeendepositedintheProteinDataBankunderaccessioncodes1UDT,1UDUand1UHO,respectively.

StructureofthereplicativehelicaseoftheoncoproteinSV40largetumourantigen

DaweiLi,RuiZhao,WayneLilyestrom,DahaiGai,RongguangZhang,JamesA.DeCaprio,EllenFanning,AndrzejJoachimiak,GerdaSzakonyi&XiaojiangS.Chen

Nature423,512–518(2003).

.............................................................................................................................................................................

ThenameofA.J.wasmisspeltintheauthorlistandshouldbeAndrzejJoachimiak.Also,heisintheBiosciencesDivisionofSBC(andnotattheAdvancedPhotonSource,aspublished).A

..............................................................

corrigendum

Impactofurbanizationand

..............................................................land-usechangeonclimate

erratum

E.Kalnay&M.Cai

Theexpressiondomainof

PHANTASTICAdeterminesleafletplacementincompoundleaves

MinsungKim,SheilaMcCormick,MarjaTimmermans&NeelimaSinha

Nature423,528–531(2003).

.............................................................................................................................................................................

Nature424,438–443(2003)..............................................................................................................................................................................

IntheMethodssectiononpage443ofthisLetter,alinewasomitted.Thesentenceshouldread:“Wedeterminedantibodypurityandspecificitybywesternblotanalysisonproteinextractspreparedfromwild-typeandrs2mutantapices.”A

Whencalculatingarealaverages,whichinvolveweightinggridded

datawiththecosineofthelatitudeofhalf-degreegridboxes,wedividedthesumbythetotalnumberofgridsbutomittedtodividethesumalsobytheaveragecosinelatitudeofthedomain,whichis0.786.Thiserroraffectsonlythearea-averagedvalues,notthemapsorthestationvalues,ortherelativedifferencesbetweenstationandreanalysisvalues.Asaresult,theaveragenumbersonthemapsshouldbedividedbythisfactor.Astheproportionsquotedremainthesame,thiserrordoesnotaffectourconclusions,exceptthatthevaluesofarea-averagedtrendshavetobemultipliedby1.272.Thecorrectedestimateofthetrendindailymeantemperatureduetolandusechangesis0.358Cpercentury.A

102©2003 Nature PublishingGroupNATURE|VOL425|4SEPTEMBER2003|www.nature.com/nature