Ozone — the persistent menace: interactions with the N cycle and climate change
David Simpson
1,2, Almut Arneth
3, Gina Mills
4, Sverre Solberg
5and Johan Uddling
6Troposphericozoneisinvolvedinacomplexwebofinteractions withotheratmosphericgasesandparticles,andthrough ecosysteminteractionswiththeN-cycleandclimatechange.
Ozoneitselfisagreenhousegas,causingwarming,and reductionsinbiomassandcarbonsequestrationcausedby ozoneprovideafurtherindirectwarmingeffect.Ozonealsohas coolingeffects,however,forexample,throughimpactson aerosolsanddiffuseradiation.Ecosystemsarebothasourceof ozoneprecursors(especiallyofhydrocarbons,butalso nitrogenoxides),andasinkthroughdepositionprocesses.
Theinteractionswithvegetation,atmosphericchemistryand aerosolsarecomplex,andonlypartiallyunderstood.Levels andpatternsofglobalexposuretoozonemaychange dramaticallyoverthenext50years,impactingglobalwarming, airquality,globalfoodproductionandecosystemfunction.
Addresses
1EMEPMSC-W,NorwegianMeteorologicalInstitute,Oslo,Norway
2Dept.Earth&SpaceSciences,ChalmersUniversityofTechnology, Gothenburg,Sweden
3KarlsruheInstituteofTechnology,InstituteofMeteorologyandClimate Research/AtmosphericEnvironmentalResearch,82467Garmisch- Partenkirchen,Germany
4CentreforEcologyandHydrology,EnvironmentCentreWales,Deiniol Road,Bangor,GwyneddLL572UW,UK
5NorwegianInstituteforAirResearch,P.O.Box100,2027Kjeller, Norway
6Dept.BiologicalandEnvironmentalSciences,Universityof Gothenburg,P.O.Box461,SE-40530,Sweden
Correspondingauthor:Simpson,David([email protected])
CurrentOpinioninEnvironmentalSustainability2014,9–10:9–19 ThisreviewcomesfromathemedissueonSystemdynamicsand sustainability
EditedbyCarolienKroeze,WimdeVriesandSybilSeitzinger ForacompleteoverviewseetheIssueandtheEditorial Received10March2014;Accepted16July2014 Availableonline5thAugust2014
http://dx.doi.org/10.1016/j.cosust.2014.07.008
1877-3435/#2014TheAuthors.PublishedbyElsevierB.V.Thisisan openaccessarticleundertheCCBYlicense(http://creativecommons.
org/licenses/by/3.0/).
Introduction
Tropospheric ozone (O3) is unique among the gases whichcontributetoglobalwarming(GW),inthataswell as beingthethirdmostimportantanthropogenicgreen- house gas [1], it causes major health problems (both
directlyandthroughproductsofozone-relatedreactions), andalsohasstronginteractionswithvegetationandhence the carbon andnitrogen cycles [2,3,4]. Measurements andmodelsbothsuggestthatozonehasbeen increasing asaresultofanthropogenicemissions.Indeed,thetitleof this paperreflectstheidentificationoflong-range trans- ported ozoneasa‘mountingmenace’in theearly1980s [5],whichstillpersists.Futuretrendsinozonearehighly uncertain. Levels and patterns of global exposure to ozone are likely to change dramatically over the next 50 years, impacting GW, air quality, global food pro- ductionand ecosystemfunction[6].
The range of issues to be discussed in this paper is sketched outin Figure1.Acompletepicture wouldbe farmorecomplex,butbelowwerefertorelevantreview articleswhichcovereachtopicinmoredetail.Theitalic lettersinthesectionheadingsbelowrefertothepathways indicatedin Figure1.
Atmospheric chemistry (Figure1a,b)
Although produced naturally in the stratosphere, O3in thetroposphereismainlyproducedfromchemicalreac- tionsinvolvingorganicprecursors(CH4andnon-methane volatile organic carbon, NMVOC), CO and nitrogen oxides (NOx,=NO+NO2). The biggestsourceof NOx
emissions is from fossil-fuel combustion, but emissions fromlightning,biomassburningandsoil-microbesarealso significant [but highly uncertain; 7,8,9]. Emissions of biogenicNMVOC(BVOC)aresignificantlygreaterthan anthropogenicNMVOC;thissourceisdiscussedbelow.
Chemical processes, frequently enhanced by anthropo- genic emissions, account for over 90% of ozone pro- duction, and almost 80% of ozone loss (Table 1).
Figure 2 illustrates some of themain reactions in con- nectiontoreactivenitrogen(Nr)species,aswellasnoting thedryandwetdepositingcompounds.Thischemistryis complexinthatmanyNrspeciesactasbothsourcesand sinks of O3 and other oxidants (see e.g. [10], or more descriptivesummariesin[3]).Inparticular,NOisadirect sinkofO3closetosources,butwithsufficientlyhighNOx
levels, O3formationisenhanced downwind.Ozoneisa productof photo-chemistry,but alsothemainsourceof thekeyOHradicalwhichcontrolsthelifetimeof many traces gases, the most important among these for GW being methane. At high NOx levels ozone production is sensitive to NMVOC compounds emitted from
anthropogenic(AVOC) orbiogenic (BVOC) sources.As indicatedinFigure2,highO3andhenceOHalsospeeds theconversionofslowlydepositingprecursorspeciesNO andNO2tocompoundswhicharemorequicklyremoved bydryandwetdeposition,notablyHNO3andparticulate nitrates.Otherimportantproductsincludeperoxy-acetyl nitrate,PAN,whichisverystableatlowtemperature,but which can dissociate into O3-forming NO2 and peroxy radicals(RO2)inwarmerregions:allowing,forexample, emissionsofBVOCinNorthAmericatohavesignificant impactsonO3inEurope [11].
Products of ozone-induced reactions include inorganic particles(e.g.nitrate,ammonium,Figure2)andsecond- aryorganicaerosol,SOA.Thecomplexityincomposition, mechanisms and impacts of SOA formation has been stressedin recentreviews[12,13].
Importantly, both O3 and SOA formationare processes wherethecontributionfromBVOC(mostlyisopreneand, forSOA,monoterpenes) candominateover combustion VOC sources, as seen in numerous modelling [e.g.
14,13] or observational studies using 14C and other source-apportionmenttechniques[e.g.15].
Radiativeforcing, aerosols (Figure1b0,c,d) Thedirectradiativeforcing(RF)potentialofO3(pathd), ca.400mWm2from1750to2010[7],isof near-equal magnitude to that of methane. Ozone also causes an indirect warming throughthe impact of O3on primary productivityasdiscussed inthenextsection.
Productsof ozone chemistry have a number of cooling effects, however. Scattering aerosols from Nr or SOA generally reduce RF (path b0) [4,13,16]. Myhre etal. [17]estimated meandirectRFover theindustrial era of 80mWm2 (range 20–120) for nitrate, and 60mWm2 (range 10–210) from SOA, althoughsuch estimates(especially fromSOA)arefraughtwithuncer- tainty,anddonotincludefeedbackswithBSOA-induced cloud albedo change such as those highlighted in
Figure1
Aerosol
CLIMATE
(b’)
(c) (b)
(a)
(d)
(e) (f)
(k) (Q)
(j)
(i)
(h) (g)
CH4
ECOSYSTEM O3
OH
NOx, VOC, CH4, CO NO2
NO PAN
Nr
Soil-NO
CO2
BVOC HNO2
Current Opinion in Environmental Sustainability
Overviewofozone–chemistry–climateinteractions.Mainprocesses whicharediscussedfurtherinthetextare(a)changesinCH4lifetime,(b) generationofaerosol,(c)aerosoleffectsecosystemsthroughradiation changes,(d)directeffectofozoneonclimatewarning,(e)indirecteffect ofphyto-toxicozonethroughbiomassandstomatalchanges,(f)impact ofNrdepositiononecosystemgrowth,(g)impactofstomatalchanges onwaterbudget.BVOCemissionsareaffectedbyCO2increases(h)and biomasschanges(i),aswellasO3itself(j),withBVOCaffectingozone chemistry(j).SoilNOemissions(k)alsochange,inturnbeingaffectedby depositionofreactiveNitrogen,Nr(f).Atmosphericchemistryamong oxidantssuchasO3andOHandvariousNrandotherprecursorspecies (Q)islooselyindicatedanddiscussed.
Table1
Tropospheric ozone budget from ACCMIP comparison [9].
Fifteen models used for burden, six for other terms, data representyear2000.Wrepresentsonestandarddeviation
Burden(Tg) 33723
Transportfromstratosphere(Tg/year) 47796 Chemicalproduction—troposphere
(Tg/year)
4877853
Chemicalloss(Tg/year) 4260645
Deposition(Tg/year) 1094264
Lifetime(days) 23.4
Figure2
Aerosol
+O3 hv
Δ +RO2
OH O3
hv M hv
Kp
+
O3 hv
NO NO2
RH
NH3
HNO3 PAN
OH
RO2
SOA
SOA(p) (g)
NO−3 NH4NO3
N2O5
NO3
NO2
OH,O3,NO3
OH,O3,NO3
Current Opinion in Environmental Sustainability
Overviewofsomeimportantnitrogenreactionsinthe(polluted) troposphere.Thegreenandbluearrowsindicateddryandwet deposition.Emittedcompoundsaregiveninwhitecircles,andozonein red.
Paasonenetal.[16].Further,althoughBSOAismainly associatedwith‘natural’VOCprecursors,BSOAloadings havelikelychangedoverthelastcenturytimeasaresult ofchangesinozone(seeOzonetrendssection)andother factors[18].Suchassessmentsarecomplicated,however, bytheinfluenceofCO2andevenozoneitselfonBVOC emissionrates,seebelow.
Ozone alsoimpactsblack-carbon(BC) aerosol,another key air-quality and (warming) RF component [19].
Increases in O3 increase the rate at which oxidised compounds coat (or ‘age’) BC.Such aged BC is much morereadilywet-depositedthanfreshhydrophobicBC;
faster aging would give lower residence times in the atmosphere [20], hence reduced RF. Aerosols also impact ecosystems in a number of ways (c) that can affect growth and hence CO2 uptake beyond, for example, direct Nr-fertilisation. Aerosols reduce total radiationreachingthesurface,butincreasethefraction ofdiffuseradiationrelativetodirect.Mercadoetal.[21]
estimatedthatvariationsinthediffusefraction,associ- ated largely with ‘global dimming’ enhanced the land carbonsinkbyapproximatelyone-quarterbetween1960 and1999 [see also4,20].
Ozone impactsonprimaryproductivity (Figure1e)
Ozoneisconsideredtobemoredamagingtovegetation thananyotherairpollutant[6],withsignificanteffects on the growth of trees, semi-natural vegetation, and several importantcrops, including wheat,soybean and rice [6,23,24]. Globally, ozone is estimated to account for yield losses of between 3% and 20% for crops[25],andtoreducebiomassproductionofnorthern hemisphereforesttreesbyca.7%atcurrentozonelevels [26].
Reducedphotosynthesisimpliesreduceduptakeofozone andCO2;allowingmoreofbothtoremainintheatmos- phere, enhancing RF. This indirect warming effect of ozone may contribute as much warming as the direct radiative effect of O3 itself [2] and for NOx and VOC emissions, ozone impacts on the carbon cycle are the dominant contributor to changes in global surface temperature [22].
It should be noted though that all estimates of these indirecteffectsofO3arebuiltuponanumberofuncertain assumptions. For example, Kvalevag and Myhre [27]
suggest that inclusion of N-limitation effects on plant growthwouldreducethenegativeeffectofO3oncarbon uptake by a factor of four, and RF by a factor of six compared to earlier studies. This study may however haveunderestimatedozoneeffectsas itdidnotaccount for the important effect of ozone on leaf-senescence/
shedding.
Phyto-toxicozonemetric,PODY
WithinthescopeoftheLRTAPConvention,1theInter- national Cooperative Programme onEffects of Air Pol- lutiononNaturalVegetationandCrops(ICPVegetation) hasbeeninstrumentalindevelopingozoneriskmethod- ology for Europe. In the last decade, a new metric for assessing cumulative ozone uptake through stomata, PODY, (Phyto-toxic Ozone Dose over threshold Ynmolem2s1)hasbeendevelopedbyICPVegetation [28–30](Figure 3).PODYtakesintoaccounttheinstan- taneouseffects of climaticfactors(temperature,humid- ity,light,soilmoisture)andplantfactors(growthstage)on theamountofozonethatistakenupbytheplant.Unlike earliermetricswhichwerebaseduponO3concentration rather than uptake,PODYtypically haslower valuesin hot, dry conditions (reflecting stomatal closure) whilst oftenhavingrelativelyhighvaluesincentralandnorthern climates that are highly conducive to stomatal uptake, leadingtoamoreevenmapofozone-riskacrossEurope thangivenbyconcentration-basedapproaches[31].This isalsomore consistentwithfieldevidence [23].
Forests
Although peat-wetlands accumulate tremendous amountsofCovermillenia[4],forestecosystemshave thegreatestCsinkcapacityovertime-scalesofdecadesto centuries [32].Therefore we here focus specifically on evidence ofozone effectsonforestproductivity.
Severalmethodshavebeenusedtodetermineeffectsof ozoneonforests,withthemostcommonbeingopen-top chambers(OTCs,usuallyca. 3m diameterandca.2.5–
3mhigh)inwhichjuveniletrees(910years)areexposed to controlled concentrations of ozone, usually under ample water supply. Deciduous trees are found to be more responsive to ozone than conifers within these systems[e.g. 29](Figure 3).The challengehasbeen to relateeffectsdetectedinjuveniletreesgrowinginanon- competitive OTC environment to effects in real forest stands.Untilnow,therehavebeenonlytwoecologically realisticfree-airO3enrichmentexperimentsinforests.In the largest of those, the so called Rhinelander Aspen FACE experiment in Wisconsin, stands with northern hardwoodtreespecieswereexposedto50%elevatedO3
and/orCO2concentrationsover11years[33].Attheend of the experiment, total tree biomass and ecosystem carbon content werereduced by 16% and 9%, respect- ively, in elevated O3. Negative effects on productivity diminished towardstheendoftheexperiment,possibly becauseofalteredtreecommunitycompositioninfavour of O3 tolerant genotypes [34,33]. There was no evi- denceofelevatedCO2modifyingproductivityresponses toelevatedO3[33].Reductionsinbiomassproduction perunitPODYwereofsimilarmagnitudeinthisfree-air
1The Convention on Long-range Transboundary Air Pollution, www.unece.org/env/lrtap.
O3 enrichment experiment (ca. 1% per mmole O3m2year1 POD1.6; biomass data in [33], POD1.6
datain [35]) as in thejuvenile beech and birch exper- iments of Karlsson et al. [29] (1.2% per mmole O3m2year1).
Inanotherfree-airO3experimentina50-yearto70-year oldmixedbeechandspruceforestinsouthernGermany, fivetreesofeachspecieswereexposedtoexperimentally doubledO3concentrationsduring eightyears. Account- ingforapretreatmentdifferenceinproductivitybetween theelevatedO3plotandtheneighbouringcontrolplot,it wasconcludedthatelevatedO3stronglydecreasedstem volumegrowthinbeech (44%) butnotinspruce[36].
Expressed per unit POD1, the negative O3 effect on maturebeechstemvolumeincrementswerelargerthan biomassreductionsfoundintheOTCexperimentswith juvenilebeechandbirchexperimentsasusedinLRTAP [30].
Another,thus farpoorly explored,approach to estimate O3impactsonforestproductivityistoapplymultivariate statisticalmethodstodisentangletheeffectsofO3from thoseofotherenvironmentalvariables[37].Otherstudies havedetectedshort-term effectsofelevated O3oneco- system CO2 fluxes as measured with eddy covariance (EC)techniques[38].Indeed,thelargenetworkofsites measuringfluxesbyECoffersagreatpotentialforstand scaleO3impact estimationusing multi-variateanalysis.
However,carefulconsiderationofexposureandresponse indicesand theirtemporal integration isneeded, given
thecumulativeimpactsofO3exposureonphotosynthesis andstomatalconductance[e.g.39,40].
Stomatalsensitivity
RisingCO2concentrations arelikelytoreduce stomatal conductance(gs)andhavebeenexpectedtoreduceozone impacts by restricting stomatal uptake of ozone [6].
However,there is agrowingbody of evidence thatthe picture is more complex in a future environment with multiplestressfactors.Chronicozoneexposurehasbeen found to reduce stomatal sensitivity to environmental stimuli[e.g.41],leadingtoeitherslowerresponsiveness orenhancedopeninginseveralspeciesandlowerdrought resistance[42].Thisphenomenonhasbeenmeasuredin the field too; elevated O3 caused progressive loss of stomatal control over summertime transpiration in the Aspen FACE experiment [40]. Further, Sun et al. [40]
attributedasignificantproportionofspatialandtemporal variation in late-season streamflow across six forested watershedsto O3effectsontranspiration.
This evidence, together with new results showing that ozoneexposurecanuncouplethecriticallyimportantleaf processesofstomatalconductanceandphotosynthesisin thefield[e.g.38],isleadingtoare-thinkoverhowozone effectsinafuturechangingclimateshouldbemodelled.
Finally,onecommonfallacyinconnectionwithgsisworth amention;namelythatchangesings(atleastweightedby leaf-area)giveproportionalchangesinevapotranspiration orotherfluxes.Generally,therelationshipFlux=gsD
Figure3
0 0.0
POD1, mmol m-2
Birch Beech
y = 1.00–0.011 *POD1 r2 = 0.64
p < 0.001
Relative total bimoass
0.2 0.4 0.6 0.8 1.0 1.2
20 40 60 0
0.0
POD1, mmol m-2 y = 1.00–0.0024 *POD1 r2 = 0.55
p < 0.001
Relative total bimoass
0.7 0.8 0.9 1.0 1.1
10 20 30 40
Current Opinion in Environmental Sustainability
TherelationshipbetweentherelativetotalbiomassandPOD1forsunlitleavesof(a)birch(Betulapendula)andbeech(Fagussylvatica)basedondata fromFinland,SwedenandSwitzerland,and(b)Norwayspruce(Piceaabies)basedondatafromFrance,SwedenandSwitzerland.Thedashedlines indicatethe95%-confidenceintervals;notethedifferentstartingpointoftheY-axisforNorwayspruce.Fromtheso-called‘MappingManual’(http://
www.icpvegetation.ceh.uk/manuals/mapping_manual.html);thesedataunderliethecriticallevelssummarisedinMillsetal.[30].
(where Dissome drivingforce, e.g. humidity deficitor concentrationdifference)isonlytrueifthedriverDisnot affected by the flux, for example when near-canopy humidity levels are not affected by the changes in gs
for thevegetationunder consideration. This point, and indeed links between gs, water-vapour, and large-scale meteorology, is discussed in detail in Jarvis and McNaughton [43]. For ozone, thenear-canopy O3con- centrationdrivingtheflux(here,Disnear-canopyminus intercellularO3,thelatterusuallyassumedtobezero)is itself a function of the ozone-uptake, with higher gs
leading to lower near-canopy O3, a classical negative feedback. For ozone, accounting for non-stomatal con- ductances isalsocritical[44].
LinkstoNsequestration
Ozone-induced reductions in C-sequestration imply changesinN-sequestrationalso.C/Nratiosinvegetation are reasonably well known (ca. 25–50, [8]). However, ozone impactson tree foliage alter many below-ground processes involvedinN cycling,including finerootpro- duction, mycorrhizal formation, nutrient acquisition by rootsandsoilrespiration.Forexample,intheAspenFACE experimentsdescribedabove,ozonetreatmentgenerally decreased theN mass (g(N)m2) of leaf litterthereby reducing N availability for microbialdecomposition and subsequently wholetreeNuptake[e.g.33,45,andrefs therein].Conversely,depositionofNr(Figure1f)impacts C-sequestration, although therelationship ismore com- plexthanasimplefertilisationeffect[4,46].
Ozone also has more subtle effects such as changing speciesdiversity.
Biogenic emissions (Figure1h–k)
Globally, emissions of BVOC far exceedanthropogenic VOC emissions [47,48]. BVOC emissions play an
importantroleforozoneproduction[10]andforsecond- ary organic aerosol [14,13,16,18]. Although there is some, possibly ‘illusory’, consensus on global emission rates of isoprene [47], emission estimates over smaller regionsvarywidely (Figure4).Therolesof BVOCand climate for future O3 and SOA formation are unclear.
Climatechangemaywellincreasefoliageinmanyareas, especially intheborealand temperateregions[e.g. 49].
This,anddirecttemperatureeffects,mightbeexpected to promote increasesin BVOC emissions infuture,and indeed many studies have thereby estimated notably increased emissions of BVOC, thus enhancing tropo- sphericO3formation andSOAformation.
However,anumberofstudieshavereportedthathigher CO2 levels will reduce BVOC emission rates [e.g.
48,50]. Arneth et al. [51,52] suggested that including theinhibitionofCO2onisoprenemetabolismcounteracts the warming/CO2 fertilisation effect and keeps BVOC emissionsnearcurrentlevelsforlongtimescalesintothe future.Otherstudieshaveshowndifferentoveralleffects, however; large uncertainties arise fromboth the ‘CO2– BVOC’ algorithm that is used, and from assumptions about how changes in climate and CO2 concentration interact with vegetation growth [e.g. 53]. Calculations indicateasignificant andregionallyvery heterogeneous effectontroposphericozoneattheendofthe21stcentury [54]. The experimental basis for such predictions is at presenttoolimitedtodrawfirmconclusions;thesignof changesinBVOCandhenceBSOAinfutureawaitsnew studies.
Other responses are also complex. For example, some BVOC species seemto playarole in reducingO3con- centrationsinvegetationcanopies[e.g.55],thusprotect- ingvegetationfromthetoxiceffectsofO3[48].Itmight therefore be speculated that BVOC emissions would
Figure4
DEHM EMEP SILAM MATCH
D N O S A J J M A M F J 0 500 1000 1500 2000 2500
Gg/month
Current Opinion in Environmental Sustainability
Anuncertaininput.Isopreneemissionestimates(GgC5H10/month)fromfourchemicaltransportmodels.DataarefortheEuropeandomain,using regionalclimatemodelmeteorologyfor2000–2009.FromLangneretal.[71].
increasewithincreasingO3.However,bothincreasesand decreases have been found [56]. Land use change, in particularinthetropics,canalsosignificantlyaffectlocal andindeedglobalO3andSOAlevels[52,57].
LoretoandFares[48]havereviewedmanyotherinter- actions(e.g.drought)ofawiderangeofBVOC;theystate that‘longer-termand fieldstudies arestill missing,and are deeply needed, to assess whether acclimation to highertemperatureswillalsoaffect futureBVOCemis- sions’.Thissentimentcouldbeappliedtomanyaspects ofBVOC emission.
Finally,bothNr-depositionandecosystemchangesmight affectsoilNO(andC2O)emissions(k),withfeedbacksto O3production[58].An interesting new developmentis the recognition that GW might substantially enhance NH3 emission rates, and hence Nr-deposition, above currentforecasts[59,60].ThecomplexitiesofC–Ninter- actions and soil–NO emissions are discussed elsewhere [4,61,8].
Ozonetrends
‘Baseline’trends
Owing to its lifetime in the atmosphere (ca. 23 days, Table1,[9])theconcentrationsandlong-termtrendsof ozonearethenetresultofahemispheric‘baseline’level andmore local/regional effects. Recentstudiesof base- lineozone[e.g.62,63,64]paintaratherconsistentpicture ofaroughdoublingofO3fromthe1950sinallsitesinall seasonsupto abouttheyear2000followedbyadecade withnogrowthorevenreductionsinO3atsomesitesin someseasons,particularlyinsummer.(Databefore1950 showmuchlowerlevelsthaninthe1950s,butthesedata areofuncertainqualityand generality[7].)
Loganetal.[62]showedthatatleastsomeofthetrends reportedintheliteraturecouldbeascribed toproblems withinstrumentation, orwereinconsistent in someway withotherdata.DatafromthreeAlpinesitesweredeter- mined to provide the most reliable trend data over Europe,withmeantrendsof 6.5–10ppbfor 1978–1989, 2.4–4.5ppb in the 1990s. From 2000 onwards, ozone decreasedby4ppbduringthesummermonths,butwith no significant trends in other seasons. The German mountain station Hohenpeisenberg [63] shows similar features. Recent studies also indicate a change in the meanseasonalcycleofthebaselineO3withtheseasonal maximumbeingshiftedfromsummertospringinrecent years [65,64].This could have important consequences fortheozone/vegetationinteractionsdiscussed above.
Europeantrends
In contrast to the consistent picture for the baseline studies,theresultsaremoremixedforsurfacemonitoring stationsinEurope.Owingtothesubstantialreductionin Europeanemissionsduringthelasttwodecades(31%for
NOx,46% for NMVOC[66]),adeclinein O3levelsis expected,butformanypartsofthecontinentthisisnot seen. Colette et al. [67] found very good agreement betweenobserved(Airbasedata)andmodelledmonthly NOxlevelsfortheperiod1998–2007,but nosystematic trendsinO3.Wilsonetal.[68]foundsignificantincreases inO3measurements (158sites, 1996–2005)for the5th- percentilesand95th-percentiles(p5,p95)ofhourlydata foraroundhalfthesites,buttheresultsweresubstantially influencedbyindividualyearsliketheheatwaveanomaly in summer 2003. Sicard et al. [69] found significant reductions in various O3 parameters at Mediterranean sitesfor2000–2010formostanalysedregions,particularly when looking at rural sites. Using rural background EMEPdata over 1990–2010,Tørseth et al. [66] found a decrease in the highest levels (and a corresponding increasein thelow percentiles)intheUK,Netherlands and some other sites, but no trends in Switzerland or Austria.For discussionof otherstudies,see[66].
Itisunclearwhetherthelackoftrendscanbeexplained byotherphysicalprocessescounteractingtheinfluenceof theprecursoremissionsorifitissimplyaproblemwith the‘signal:noise’ratio.Thelatterwouldindicatethatthe effectofthe reducedprecursorsis maskedbythelarge inter-annual variations in O3, caused by, for example, meteorology, or biomass burning events. One likely reason for the differences between studies is that the selectionoftimeperiodisdecisiveforthetrendestimates [70,62].Thus,trendassessments becomeuncertain for networks with significantdifferences in the monitoring historyfor thevarioussubregions.Inaddition,thetrend estimatesaredeterminedbythechoiceofO3parameter (percentiles,meanvalues,etc.)andthemethodsapplied (e.g.linearorquadratic).Akeymessageseemstobethat thetimeseriesneedto bemuchlongerthan10years in order to distinguish a significant long term trend from inter-annual variability. Secondly, significant trends are mostly seenin the highest (p>95)and lowest (p<5) percentilesoftheO3concentrationdistributionandnotin meanvalues.
Inorderto illustratetherelationshipbetween trendsin differentpercentiles,Figure5showsthechangesinthe mean annual percentiles of O3 from the decade 1990–
1999to2000–2009forEMEPsites.Resultsareshownfor someNordic,north-westEurope(GreatBritain,Ireland, Netherlands),andcentralEuropeansitesseparately.The results indicate significant regional differences within Europewithstrongreductionsinthehighestpercentiles (p95)forthenorth-westEuropesites,variableresults fortheNordicsitesandverysmallchangesforthecentral Europeansites.
Futureozone
Althoughozone may haveimportant effects onclimate changeasdiscussedabove,recentmodelstudiessuggest
Figure5
0.1 1 5 10 25 50 75 90 95 99 99.9 0.
1 1 5 10 25 50 75 90 95 99 99.9 0.
1 1 5 10 25 50 75 90 95 99 99.9
−20
−15
−10
−5 0 5
Δppb/decade
AT02 Illmitz AT32 Sulzberg AT41 Haunsberg AT42 Heidenreichstein AT43 Forsthof CH03 Tänikon
NO01 Birkenes SE11Vavihill SE12 Aspvreten SE32 Norra–Kvill
GB13 Yarner Wood GB36 Harwell IE31 Mace Head NL09 Kollumerwaard
Current Opinion in Environmental Sustainability
Thechangeinmeanannualpercentiles(ofhourlyozonedata)fromthedecade1990–1999tothedecade2000–2009,thatis,Px(2000s)-Px(1990s), wherexrangesfrom0.1to99.9,forselectedEuropeansites.Dataandsitesfrom[66],withadata-capturerequirementof75%completenessof hourlydataineachyear.
Figure6
Years
O3 Chage (ppb) O3 Chage (ppb)
Years
Europe South Asia
–62000 –4 –2 0 2
–2 0 2 4 6 8 10 12
4 6 8
2010 2020 2030 2040 2050 2000 2010 2020 2030 2040 2050
Current Opinion in Environmental Sustainability
Anuncertainfutureforozone.PlotsshowestimatesoffuturesurfaceozoneinEuropeandSouthAsia.ThegreenareashowstherangeofO3predicted fromtheIPCC4thAssessmentReport(SRESscenariosA2,A1B,B2,B1),andtheyellowareagivestheupdatedrangeusingtheIPCC5thAR (RCP8.5,6.0,4.5,2.6).FigureredrawnfromWildetal.[75].
low or modest impact of climate change on future ozone and/or Nr-deposition [71,72,60]. The possibility remains however that future climate may be more extremethanusedinthesestudies,whichcouldchange O3dramatically.Theyear2003providesaclearexample, with severe ozone episodes and widespread drought in Central Europe [73]. Using regional climate simu- lations,Beniston[74]concludedthatfor‘manypurposes the 2003 event can be used as an analogue of future summersincomingdecadesinclimateimpactsandpolicy studies’.
Regardlessofclimate,thedevelopmentofozoneinfuture iscriticallydependentuponemissionchanges.Figure 6 illustrates this with estimates presented by Wild et al.
[75],inwhichtheresultsof14globalchemicaltransport modelswereparameterisedsothatsurfaceozonecouldbe estimated from emissions of NOx, CH4and other pre- cursors.Thenewerandmore stringent‘RCP’emissions scenariosproducemuchsmallerincreasesinO3thanthe older‘SRES’ estimates.About75% of the5ppbdiffer- ence between the outlying RCP 2.6 and RCP 8.5 scenarios couldbeattributed to differencesinmethane abundance.Thereisclearlyplentyofscopeforemission controltochangefutureozone.
Discussion andconclusions
Ozone is clearly involved with the N-cycles and C- cycles in a complex, and only partially understood way. Gas-phase atmospheric chemistry is reasonably wellunderstoodinprincipal,butemissionsofespecially naturalVOCandNOprecursorsareveryuncertain.The responseofsuchemissionstoclimatechangeisunclear evenwithregardtothesignofthechange.Changesin stratospheric–tropospheric exchange of O3 may also affect future ozone, but uncertainties are again large [e.g.9].
OzoneimpactsonvegetationandhenceNandCseques- trationare alsodifficulttoquantify,especiallyforforest ecosystems which are not amenable to small-scale and short-termexperiments.Thereisaclearneed tounder- stand how ozone acts within the mix of climate, other pollutant, and biotic stresses (e.g. insect pests, fungal diseases)thatoccurnowandaremorelikelyinthefuture withinnaturalorman-managedecosystems.Manyofthe issuesaddressedabovepointtotheneedforbetterlong- term monitoring data (e.g. of fluxes) in order to help untanglethecomplexweb ofinteractions.
ModellingoftheeffectsofO3onvegetationisdependent on improvements in the dose–response algorithms. A major challenge now is to take the PODY approach to thenextstage, incorporatingeffects ofmultiplestresses and climate changeas wellas the growingevidence of effectsofozoneonstomatalfunctioningandthecoupling withphotosynthesis [see6,andrefs.therein].
Theimportanceof ozone asashort-lived climategasis receiving increasing attention, and mitigation of ozone throughprecursorcontrolisseenasapromisingstrategy tohelpmitigateclimatewarming[3,19].Somemeasures arecomplexhowever,withforexampleemissioncontrol ofNOxlikelytoleadtowarmingintheshortterm(ca.20 years)butcoolinginthelongerterm[22].Manystudies stressthebenefitsofCH4controlonaglobalscale,since emissionsreductionsarebeneficialformostenvironmen- talissues.
Acknowledgements
ThisstudybuildsuponsupportfromtheEUFP7projectsECLAIRE (#282910)adPEGASOS(#265148),EMEPunderUNECE,theSwedish ResearchprojectsBECCandMERGE,andICPVegetationsupportedby Defra,UNECEandNERC.
Referencesand recommendedreading
Papersofparticularinterest,publishedwithintheperiodofreview, havebeenhighlightedas:
ofspecialinterest ofoutstandinginterest
1. IPCC,ClimateChange2013:ThePhysicalScienceBasis.
ContributionofworkingGroupItotheFifthAssessmentReportof theIPCC,www.ipcc.ch.InternationalPanelonClimateChange;
2013:.www.ipcc.ch.
2. SitchS,CoxPM,CollinsWJ,HuntingfordC:Indirectradiative forcingofclimatechangethroughozoneeffectsontheland- carbonsink.Nature2007,448:791-795.
3. RoyalSociety:Ground-LevelOzoneinthe21stCentury:Future Trends,ImpactsandPolicyImplications;Vol.PolicyDocument15/
08. London:TheRoyalSociety;2008,. 4.
ArnethA,HarrisonSP,ZaehleS,TsigaridisK,MenonS,BartleinPJ etal.:Terrestrialbiogeochemicalfeedbacksintheclimate system.NatGeosci2010,3(8):525-532http://dx.doi.org/10.1038/
ngeo905.
Sumarisesalargenumberofterrestrialbiogechemicalfeedbacks,with newestimatesofRFtowardstheendofthe21stcenturyanddiscussion ofbothN-cyclesandC-cycles.
5. GrennfeltP,SchjoldagerJ:Photochemicaloxidantsinthe troposphere:amountingmenace.Ambio1984,13:61-67.
6.
AinsworthEA,YendrekCR,SitchS,CollinsWJ,EmbersonLD:The effectsoftroposphericozoneonnetprimaryproductivityand implicationsforclimatechange.AnnRevPlantBiol2012, 63:637-661http://dx.doi.org/10.1146/annurev-arplant-042110- 103829.
Amoredetailedreviewofmanyofthesubjectscoveredinthisoverview, coveringtheliteratureuptoabout2010.
7. StevensonDS,YoungPJ,NaikV,LamarqueJF,ShindellDT, VoulgarakisAetal.:Troposphericozonechanges,radiative forcingandattributiontoemissionsintheAtmospheric ChemistryandClimateModelIntercomparisonProject (ACCMIP).AtmosChemPhys2013,13(6):3063-3085http://
dx.doi.org/10.5194/acp-13-3063-2013http://www.atmos-chem- phys.net/13/3063/2013/.
8.
FowlerD,CoyleM,SkibaU,SuttonMA,CapeJN,ReisSetal.:The globalnitrogencycleinthetwenty-firstcentury.PhilosTrans RoySocB:BiolSci2013,368(1621)http://dx.doi.org/10.1098/
rstb.2013.0164.
Arecentoverviewofthenitrogencycle,includinggooddiscussionsof sources,sinks andtrends. Nicelyillustrated, withup-to-datebudget estimates.
9.
YoungPJ,ArchibaldAT,BowmanKW,LamarqueJF,NaikV, StevensonDSetal.:Pre-industrialtoend21stcentury projectionsoftroposphericozonefromtheAtmospheric ChemistryandClimateModelIntercomparisonProject (ACCMIP).AtmosChemPhys2013,13(4):2063-2090http://
