1Key Laboratory for Semi-Arid Climate Change of the Ministry of Education, College of Atmosphere Science,
Lanzhou University, Lanzhou 7300002Laboratory for Climate Studies of China Meteorological Administration, National Climate Center,
China Meteorological Administration, Beijing 1000813Chinese Academy of Meteorological Sciences, Beijing 100081
4Department of Atmospheric Sciences, University of Washington, Seattle, 98195 USA
(Received 14 September 2015; revised 17 March 2016; accepted 11 April 2016)
ABSTRACT
This study simulates the effective radiative forcing (ERF) of tropospheric ozone from 1850 to 2013 and its effects onglobal climate using an aerosol–climate coupled model, BCC AGCM2.0.1 CUACE/Aero, in combination with OMI (OzoneMonitoring Instrument) satellite ozone data. According to the OMI observations, the global annual mean tropospheric col-umn ozone (TCO) was 33.9 DU in 2013, and the largest TCO was distributed in the belts between 30◦N and 45◦N and atapproximately 30◦S; the annual mean TCO was higher in the Northern Hemisphere than that in the Southern Hemisphere;and in boreal summer and autumn, the global mean TCO was higher than in winter and spring. The simulated ERF dueto the change in tropospheric ozone concentration from 1850 to 2013 was 0.46 W m−2, thereby causing an increase in theglobal annual mean surface temperature by 0.36◦C, and precipitation by 0.02 mm d−1 (the increase of surface temperaturehad a significance level above 95%). The surface temperature was increased more obviously over the high latitudes in bothhemispheres, with the maximum exceeding 1.4◦C in Siberia. There were opposite changes in precipitation near the equator,with an increase of 0.5 mm d−1 near the Hawaiian Islands and a decrease of about −0.6 mm d−1 near the middle of the IndianOcean.
Citation: Xie, B., H. Zhang, Z. L. Wang, S. Y. Zhao, and Q. Fu, 2016: A modeling study of effective radiative forcing andclimate response due to tropospheric ozone. Adv. Atmos. Sci., 33(7), 819–828, doi: 10.1007/s00376-016-5193-0.
1. Introduction
Since the early 1970s, ozone has been receiving increas-ing attention as its concentrations are increasingly influencedby human activities. As a radiatively active gas, ozone hasan effect on longwave and shortwave radiation in both thestratosphere and troposphere. Stratospheric ozone has strongabsorption in solar UV radiation, leading to a surface coolingof the Earth, while tropospheric ozone causes a greenhouseeffect by interacting with terrestrial longwave radiation. Anychanges in the concentration of ozone can cause radiativeforcing (RF) and therefore lead to climate change (Lacis etal., 1990; Forster et al., 2007).
The concentration of tropospheric ozone has significantlyincreased since the industrial revolution. The main sourcesof tropospheric ozone include the downward transport of
ozone from the stratosphere (Stohl et al., 2003; Hsu andPrather, 2009) and the photochemical oxidation of precur-sors. Methane (CH4), carbon monoxide (CO), and non-CH4volatile organic compounds in the presence of nitrogen ox-ides (NOx) are the main anthropogenic precursors of ozone(Crutzen, 1974; Derwent et al., 1996; Zhang et al., 2004).When the emissions of NOx and CO double in India, tropo-spheric ozone can be increased up to 21 ppbv, with a relativechange of more than 10% in the upper troposphere (Liu etal., 2003). Tropospheric ozone is depleted by several chem-ical reactions (Crutzen, 1974). Increases in absolute humid-ity (driven by global warming), changes in the ozone distri-bution itself, and changes of hydroxyl and peroxyl radicals,have influenced the rapid consumption of tropospheric ozoneby chemical processes (Johnson et al., 2001; Stevenson et al.,2006; Isaksen et al., 2009). Tropospheric ozone can also becleared by dry deposition, and it has an adverse effect on thephotosynthesis of plants (Fowler et al., 2009). Anthropogenicprecursor emissions are probably the main driver of changes
820 ERF AND CLIMATE RESPONSE DUE TO TROPOSPHERIC OZONE VOLUME 33
in tropospheric ozone, and have increased dramatically sincethe industrial era (Lamarque et al., 2010).
The global distribution of ozone in the atmosphere can beretrieved by satellite remote sensing, with early observationalozone data having been obtained from the Total Ozone Map-ping Spectrometer (TOMS) and the Solar Backscatter Ultra-violet (SBUV) sensor (Heath et al., 1975). More recently, theOzone Monitoring Instrument (OMI) onboard the Aura satel-lite has offered improvements by monitoring the UV/visiblecontinuous spectrum with high spatial resolution (Miyazakiet al., 2012). Inversion methods have also been developed, in-cluding the continuously improving TOMS inversion methodand a combination of TOMS and Differential Optical Absorp-tion Spectroscopy (Anton et al., 2009). To obtain the updatedRF due to tropospheric ozone, the ozone observational dataof the OMI in 2013 are utilized in this work.
IPCC AR5 reported that the radiative forcing (RF) atthe tropopause was 0.40 ± 0.2 W m−2 due to the changesin tropospheric ozone since the industrial revolution. Tro-pospheric ozone is an important short-lived greenhouse gas,and controlling the emissions of its precursors can effectivelyslow down global warming (Zhang et al., 2014a). Skeie et al.(2011) estimated that tropospheric ozone increased by about11.4 DU from pre-industrial times to 2010. Shindell et al.(2006) simulated the impact of tropospheric ozone on globalmean surface temperature, and found that the increase in tro-pospheric ozone from pre-industrial times to 2000 could havecontributed approximately 0.11◦C to global warming. Changet al. (2009) simulated the climate responses to the directRF due to tropospheric ozone in eastern China during 1951–2000, and found that the ozone forcing led to changes inthe mean surface air temperature and precipitation in easternChina by +0.43◦C and −0.08 mm d−1, respectively.
The most recent estimation only reported RFs due totropospheric ozone during 1850–2000 (Conley et al., 2013;Lamarque et al., 2013; Stevenson et al., 2013, Stordal etal., 2003), and very few studies have estimated the updatedRF due to ozone from the pre-industrial era to the year2013 and its climate effect. In this study, the updated tro-pospheric ozone observational data of 2013 were inputtedinto BCC AGCM2.0.1, which had previously been coupledwith the China Meteorological Administration Unified Atmo-spheric Chemistry Environment for Aerosols (CUACE/Aero)model (Wang et al., 2014; Zhang et al., 2014b), to simulatethe effective RF (ERF) due to tropospheric ozone and its im-pact on the global climate. As defined in IPCC AR5, ERF isthe net downward radiative flux change at the TOA allowingatmospheric temperatures, clouds and water vapor to adjust,but with surface or sea surface temperature and sea ice un-changed. Rapid tropospheric adjustments can influence theflux perturbations, leading to differences in long-term climatechange due to the forcing agents. So, allowing the rapid ad-justments in the troposphere can provide a relatively accuratecharacterization of the forcing due to tropospheric constituentchanges.
In section 2, the ozone data, model and ERF calculationmethod are briefly introduced. In section 3, the distribution
of tropospheric ozone concentration and the ERF due to tro-pospheric ozone are analyzed, and then the climate effect oftropospheric ozone is discussed. A summary and discussionof the study is presented in section 4.
2. Data, model, and scheme description
2.1. Satellite data
The ozone profile data observed by the OMI in 2013 wereused in this study. The OMI is one of the instruments onboardthe Aura spacecraft (data available at: http://aura.gsfc.nasa.gov/), which was launched by NASA on 15 July 2004, as partof the EOS. The OMI is a nadir-scanning instrument, and itcan measure column ozone in the UV (270–314 nm and 306–380 nm) and visible (350–500 nm) wavelengths. OMI has aglobal coverage (except polar night latitudes) at a spatial res-olution of 13× 24 km, and a swath width of approximately2600 km. According to the studies of Buchard et al. (2008)and Veefkind et al. (2006), the relative uncertainty of OMIdata is about 3%. The ozone profile data are given in termsof the layer-columns of ozone in DU for an 18-layered atmo-sphere (about 2.5 km for each layer) (Miyazaki et al., 2012).NCEP daily mean tropopause data for 2013 (http://www.esrl.noaa.gov/) were also used in this study, for the calculation oftropospheric column ozone (TCO).
2.2. Models
The aerosol–climate model BCC AGCM2.0.1 CUACE/Aero developed by Zhang et al. (2012a), Zhang et al. (2014b),and Wang et al. (2014), was used in this study. The model hasjoined AeroCom Phase II (Myhre et al., 2013a), and it hasbeen used to study the RFs of aerosols and the subsequenteffects on climate (e.g., Zhang et al., 2012b; Wang et al.,2013a, 2013b, 2014, 2015; Zhao et al., 2015). BCC AGCMis a general circulation model developed by the Beijing Cli-mate Center of the China Meteorological Administration, andits main features have been described by Wu et al. (2010).The CUACE/Aero aerosol model, developed by the Insti-tute of Atmospheric Composition of the Chinese Academy ofMeteorological Sciences, was coupled with BCC AGCM byZhang et al. (2012b). The model employs a horizontal T42spectral resolution (approximately 2.8◦ × 2.8◦) and verticalhybrid δ-pressure coordinates, including 26 layers, with thetop located at about 2.9 hPa. Recently, several revisions havebeen made to improve the physics of the model. First, sev-eral schemes were incorporated into BCC AGCM (Zhang etal., 2014b), which included the cloud overlap scheme of theMonte Carlo independent column approximation (Pincus etal., 2003) and the radiation scheme BCC RAD (Beijing Cli-mate Center Radiation Transfer Mode) developed by Zhanget al. (2003, 2006a, 2006b). These schemes improved the ac-curacy of the sub-grid cloud structure and its radiative trans-fer process (Zhang et al., 2014b). Second, the methods ofNakajima et al. (2000) and Zhang et al. (2014a) were adoptedto calculate the optical properties of water and ice clouds.More details about this coupled system can be found in Zhang
JULY 2016 XIE ET AL. 821
et al. (2012b) and Wang et al. (2014). Xin et al. (2013) esti-mated the sensitivity and general performance of the climatemodel, with their results showing the transient climate re-sponse (TCR) of the BCC model to be 1.94◦C (the mean TCRwas found to be 2.0◦C for 16 climate models participating inCMIP5), and the model overestimating the global warmingbut underestimating the warming amplitude over China in theearly 21st century. Furthermore, BCC AGCM’s participationin CMIP5 has shown reasonable performance in simulatingimportant meteorological factors (Zhou et al., 2014a, 2014b).
2.3. Experimental designFour experiments were performed in this study, named
EXP1, EXP2, EXP3 and EXP4 (EXP1 and EXP2 were usedto calculate ERF; EXP3 and EXP4 were for calculating theclimate response; experiment configurations shown in Ta-ble 1). In EXP1, we used tropospheric ozone for the year1850, given in Lamarque et al. (2010); and the stratosphericozone was from the OMI observational data for 2013. Thetropopause in 1850 was defined as the height where the ozoneconcentration was 150 ppb, as used in Young et al. (2012).In EXP2, everything was kept the same as in EXP1 exceptthat the tropospheric ozone data from the OMI in 2013 wereused. Each experiment was run for 15 years with prescribedSSTs. The last 10 years of results were analyzed to calculatethe ERF due to the change in tropospheric ozone from 1850to 2013. According to previous research (Kristjansson et al.,2005), after a period of adjustment (generally five years fora model with prescribed SST), the global mean surface tem-perature will basically reach equilibrium (with small fluctu-ations). So, in this study, we just used the last 10 years ofresults of the 15-year runs of EXP1 and EXP2 to calculatethe ERF due to the change in tropospheric ozone, as follows:
ERF = ΔFEXP2−ΔFEXP1 ,
where ΔF was the net radiation flux (the difference betweenincoming and outgoing radiative flux, both shortwave andlongwave) at the top of the model (for there is little differ-ence in net radiation flux between the top of the model andthe TOA). The ΔFEXP1 (ΔFEXP2) was the net radiation fluxof EXP1 (EXP2).
The same tropospheric ozone data in EXP1/EXP2 wereused in EXP3/EXP4. However, a coupled slab ocean modelwas used to replace the prescribed SST, in order to fully con-
sider the climate response. EXP3 and EXP4 were run for70 model years. In order to allow the global mean surfacetemperature to roughly reach equilibrium, the coupled slabocean model needed 30 years of adjustment (Kristjansson etal., 2005). So, we just used the last 40 years of results of the70-year runs of EXP3 and EXP4 to facilitate discussion onthe climate response due to the change in tropospheric ozone.
The t-test was used on the model results to estimate theirstatistical significance. More specifically, the t-test used wasthe test for a difference between two sample means, as fol-lows:
t =X1−X2√
(n1−1)S 21 + (n2−1)S 2
2
n1+n2−2
(1n1+
1n2
) ,
where X and S was the average and variance of the sample,and n was sample size.
3. Results
3.1. Distribution of tropospheric ozoneFigure 1 shows the seasonal mean distribution of TCO in
2013 from the OMI observation. The distribution of TCOvaried both spatially and temporally. The largest TCO wasmainly distributed in the belt between 30◦N and 45◦N, wherethe main emissions sources of anthropogenic ozone precur-sors were located. Additionally, another belt of peak TCOwas located at approximately 30◦S. Both belts of peak TCOin the NH and SH were due to some stratospheric ozonebrought into the troposphere by the Brewer–Dobson circula-tion and latitudinal variations in the tropopause height (Lin etal., 2013). The area of high TCO in the SH was uniformlydistributed in different seasons, with a maximum near thesouthern Indian Ocean. There were two low-TCO areas overthe Tibetan Plateau and western North America. The TCO“trough” in the Tibetan Plateau is due to the increasing heightof the tropopause over this region and dilution by low-ozoneair from tropical areas (Zhou et al., 1995; Liu et al., 2010).The TCO was relatively low over the equator and polar ar-eas, with a minimum in the Antarctic. The global mean TCOwas higher in JJA (June–July–August) and SON (September–October–November), and lower in DJF (December–January–February) and MAM (March–April–May) (Table 2). The
Table 1. Experimental design.
Number Ozone data Sea temperature Running time Purpose
EXP1 Tropospheric ozone data in 18501 Prescribed SST3 15 years ERF calculationEXP2 Tropospheric ozone data in 20132 Prescribed SST 15 years ERF calculationEXP3 Tropospheric ozone data in 1850 Slab ocean model 70 years Climate response calculationEXP4 Tropospheric ozone data in 2013 Slab ocean model 70 years Climate response calculation
1The tropospheric ozone in 1850, from Lamarque et al. (2010), and the stratospheric ozone in 2013, from the OMI observational data. The tropopause wasdefined as the height where the ozone concentration was 150 ppb, as used in Young et al. (2012).2The tropospheric and stratospheric ozone was from the OMI observations in 2013.3Monthly mean prescribed SST observational dataset (Hurrell et al., 2008).
822 ERF AND CLIMATE RESPONSE DUE TO TROPOSPHERIC OZONE VOLUME 33
Fig. 1. Seasonal mean distribution of TCO in 2013 from OMI observations: (a) DJF; (b) MAM; (c) JJA; (d) SON (units: DU).
largest TCO (hemispheric mean: 40.74 DU) was recorded inMAM in the NH. The seasonal changes in the high-TCO ar-eas in the SH were not as dramatic as in the NH, but there wasalso an obvious seasonal difference in the SH, with a maxi-mum in SON (hemispheric mean: 36.94 DU). In China, theseasonal fluctuation of TCO was more notable, and the con-centration reached a maximum in MAM (43.64 DU) beforefalling to a minimum in SON (30.98 DU). The temporal andspatial distributions of TCO were possibly a result of ozonegenerated by lightning discharges, fossil fuels and biomassburning, soil emissions, and seasonal transport due to plane-tary circulation (Ziemke et al., 2011).
The global annual mean TCOs were 15.82 DU and 33.90DU in 1850 and 2013, respectively [the TCO in 1850 givenby Lamarque et al. (2010) and the TCO in 2013 from the OMIobservation). The spatial distributions of TCO in 1850 and2013 were similar to some extent (Fig. 2). The TCO increasedsignificantly in the midlatitudes (about 25 to 35 DU), whereindustrialized regions were located. The increase in TCO
Table 2. Seasonal mean TCO averaged over the globe, NH, SH andChina in 2013 (units: DU).
was less than 10 DU over high latitudes and less than 5 DUin the Antarctic. The TCO increased in the Tibetan Plateauand north of the Yangtze River in China (except northeasternChina). Compared with other areas, the change in TCO wasmore apparent in China (increased by 20.99 DU in China andthe global mean change in TCO was 18.06 DU). Gao et al.(2009) showed that spatiotemporal variations in troposphericozone concentrations over East Asia were not only affectedby the seasonal variation in solar intensity and photochemicalactivity [the most important contributor to ozone seasonalityover Northeast Asia, as shown in Kim and Lee (2010)], butalso influenced by the monsoons.
3.2. The ERF of tropospheric ozoneThe increased emissions of ozone precursors into the at-
mosphere during the industrial era has led to a change in theozone concentration and subsequently their ERF. However,for the effect of instant adjustment, the distributions of thechange in tropospheric ozone and the consequent ERF didnot agree very well. As shown in Fig. 3c, positive ERF oc-curred over the low latitudes of both hemispheres, such asMexico, South Asia and the west coast of southern Africa.The ERF due to change in tropospheric ozone was negativeover the midlatitudes of northern continents, such as northernAfrica, Mongolia and especially northeastern China (approx-imately −5 W m−2) (see Fig. 3c). The shortwave ERF of tro-pospheric ozone was remarkably negative in eastern China,South Asia, and the Pacific region near the equator. How-ever, in Southeast Asia and Siberia, the shortwave ERF wasmore notably positive. The most significantly negative and
JULY 2016 XIE ET AL. 823
Fig. 2. Global distribution of (a) annual mean TCO in 1850 and (b) the difference in TCO between 1850 and 2013 (units:DU). The TCO in 1850 is that given in Lamarque et al. (2010), while the TCO in 2013 is from the OMI observations.
Fig. 3. The distribution of annual mean (a) shortwave ERF, (b) longwave ERF and (c) total ERF (units: W m−2). Thedots represent statistical significance according to the t-test at the � 95% confidence level.
positive shortwave ERFs were in Southeast China (approxi-mately −11 W m−2) and New Guinea (approximately 10 Wm−2), respectively (see Fig. 3a). The distribution of long-wave ERF was opposite to that of shortwave ERF, with themost significantly negative and positive longwave ERF in thePhilippine basin (approximately −5 W m−2) and Arabian Sea(approximately 5.5 W m−2), respectively (see Fig. 3b). AsIPCC AR5 reported, the difference between the ERF and RFfor tropospheric ozone is likely to be small compared to theuncertainty in the RF (Shindell et al., 2013). The global an-nual mean total ERF was 0.46 W m−2 in this study, and thebest estimate of tropospheric ozone RF reported in IPCC AR5
was 0.40 [0.20–0.60] W m−2. A comparison of our resultsand other model results is shown in Table 3.
Table 3. Comparison of ERF in this work and the simulated RFin other works due to changes in tropospheric ozone from the pre-industrial period to the present (Units: W m−2).
References ERF RF Present
This study 0.46 - 2013Shindell et al. (2013) - 0.33 [0.31–0.35] 2010Søvde et al. (2011) - 0.38 2000Skeie et al. (2011) - 0.41 [0.21–0.61] 2010
824 ERF AND CLIMATE RESPONSE DUE TO TROPOSPHERIC OZONE VOLUME 33
3.3. Climate response of tropospheric ozone
3.3.1. Temperature
Tropospheric ozone is an important short-lived green-house gas. The ERF of tropospheric ozone is generally pos-itive, and leads to a warming effect on the near-surface cli-mate. The simulations (EXP3 and EXP4) showed that the in-crease in tropospheric ozone since the industrial revolutioncaused an increase of 0.36◦C in the global annual mean sur-face temperature, with a significance level above 95% on theglobal scale. As shown in Fig. 4a, the surface temperature
was increased over most of the world, with a slight negativechange in Alaska, in the southeastern and northwestern Pa-cific basin, and especially the Norwegian Sea (approximately−0.2◦C). The warming over the middle and high latitudes ofthe boreal hemisphere was prominent, with a maximum ex-ceeding 1.4◦C, and it was also remarkable in the AntarcticCircle (approximately 1.0◦C). Figure 4b shows the responseof the surface net radiation flux (SNRF) to the change in tro-pospheric ozone. The distribution of SNRF was very consis-tent with surface temperature over the ocean areas, and therewere significant increases in SNRF over the eastern Pacific
Fig. 4. Annual mean distributions of simulated differences in (a) surface temperature (units: ◦C), (b) SNRF (units: Wm−2), (c) low cloud cover (units: %), (d) high cloud cover (units: %), and (e) zonal-mean total atmospheric merid-ional heat transport (units: K m s−1) depicted by VVVT , where VVV is the meridional velocity (units: m s−1) and T is theatmospheric temperature (units: K), and (f) wind field at 850 hPa (units: m s−1) between EXP4 and EXP3. The dotsrepresent statistical significance according to the t-test at the � 95% confidence level.
JULY 2016 XIE ET AL. 825
Ocean near the equator, the west bank of southern Africa, theeastern sea near Japan and near the Antarctic Circle (with asignificance level above 95% in these areas). The SNRF in-creased by more than 3.0 W m−2 in all the regions above, andthe increase was particularly notable over the Tibetan Plateauwhere the increase reached approximately 8.0 W m−2. Theincreases in SNRF were also significant over the equatorialoceans, with a maximum increase of 3.52 W m−2. The SNRFdecreased significantly over most of the Indian Ocean, SouthPacific Ocean, west of the Ural Mountains, and the high lati-tudes of the SH (the decrease in the South Pacific Ocean hada significance level above 95%). The cooling in the North Pa-cific near the equator was most likely caused by the low-levelcloud cover (below 680 hPa), which increased by around2.0% (Fig. 4c) (with a significance level above 95%). Theincrease in low cloud resulted in a decrease in SNRF, therebyleading to a cooling effect at the surface. The increase inhigh-level cloud (above 440 hPa) can cause an increase insurface temperature. Therefore, the marked increase in highcloud might be the reason for the warming in the west nearGreenland. Local temperature changes may not necessarilybe explained by local processes (surface net radiation flux andcloud changes); they can be strongly influenced by changesin ambient heat transport. It should be noted that the surfacetemperature decreased in areas such as the southeastern Pa-
cific basin and North Atlantic near the North Pole, but thetotal ERF was positive in these areas (Fig. 3c). This inconsis-tency might have been caused by the effect of air flow (Fig.4f); there was cold advection from high latitudes in those ar-eas. As shown in Fig. 4f, warm advection mainly caused theanomalous increasing of temperature in East Siberia, thoughthe total ERF and SNRF there were both negative. Figure 4eshows the simulated annual mean difference in zonal-meantotal atmospheric heat transport due to changes in the TCO.There was an increase in the total heat transport from highlatitudes of the NH to regions near the North Pole. This in-crease also appeared at 30◦S, and might have led to surfacetemperature increasing over Australia. Finally, the changesin total atmospheric heat transfer fed back to the changes insurface temperature.
3.3.2. Evaporation, clouds and precipitation
The increase in tropospheric ozone resulted in a warm-ing effect on the atmosphere by positive ERF at the TOA,thereby causing an increase in surface evaporation (Fig. 5d).The spatial distributions of changes in surface evaporationand surface radiation flux were similar (Figs. 4b and 5d).The evaporation was dramatically increased at the sea sur-face over low and middle latitudes in both hemispheres, es-pecially the Northwest and East Pacific (with a significance
Fig. 5. Simulated annual mean differences in (a) zonally averaged relative humidity (units: %), (b) cloud fraction (units:%), (c) precipitation (units: %), and (d) surface evaporation (units: %) between EXP4 and EXP3. The dots representstatistical significance according to the t-test at the � 95% confidence level.
826 ERF AND CLIMATE RESPONSE DUE TO TROPOSPHERIC OZONE VOLUME 33
level above 95%), where evaporation was increased by morethan 0.15 mm d−1. The surface evaporation was significantlyweakened with the declining surface radiation flux in mostareas.
The increase in tropospheric ozone caused the amount ofhigh, medium and low cloud to be reduced by 0.17%. Inaddition to surface evaporation, air flow convergence and di-vergence, changes in relative humidity and aerosol–cloud in-teractions also contribute to the changes in cloud cover. Thus,the changes in cloud cover and surface evaporation were notvery consistent. Surface evaporation increased notably in thePacific near 30◦S, Yellow Sea and the Sea of Japan (Fig. 5d),with a significance level above 95%. However, the total cloudcover (mainly low cloud cover; Fig. 4c) decreased (with asignificance level above 95%) due to the air divergence in theabove regions (Fig. 4f).
The vertical distribution of cloud was significantly influ-enced by relative humidity. As shown in Fig. 5a, the sim-ulated annual mean differences in zonal mean relative hu-midity between EXP4 and EXP3 were consistent with thatin zonal mean cloud fraction (Fig. 5b). The relative humiditywas clearly reduced (maximum close to −1.0%) throughoutthe troposphere of the midlatitude NH, upper troposphere ofthe tropics in both hemispheres, and in most of the tropo-sphere of the midlatitude SH, with a significance level above95%. This might have been due to the decrease in surfaceevaporation and the air divergence. Because of the decreasein relative humidity, cloud cover was also reduced by about0.2% to 0.8% in these areas (with a significance level above95% in all the areas mentioned before), and led to a clearincrease in SNRF, except in the Antarctic region. In con-trast, the relative humidity and cloud cover increased remark-ably throughout the troposphere near 60◦S, the lower tro-posphere in the tropics, and the stratosphere in both hemi-spheres. There were significant increases of cloud cover inthe lower troposphere over the high latitudes of both hemi-spheres, and these might have been due to the air convergenceover these regions.
The global distribution of changes in precipitation wassimilar to that of the changes in cloud cover. Near the equa-tor, the changes in precipitation were notable, with a signif-icance level above 95%. The precipitation and total cloudcover (mainly low cloud cover) increased significantly (over0.5 mm d−1 and 4%) in the Hawaiian Islands, Philippinesand Somalia. The precipitation and cloud cover were sharplyreduced (over −0.6 mm d−1 and −4%) in the central Indianbasin and Palau Islands. The total cloud cover and precipita-tion increased in the high latitudes of both hemispheres. Theprecipitation decreased in most of China, except the south-eastern area.
4. Conclusions
This work studied the ERF and climate impact due tothe tropospheric ozone concentration change since 1850 byusing a coupling aerosol climate model, BCC AGCM2.0.1
CUACE/Aero, and OMI ozone concentration data in 2013.By analyzing the ozone data, we found that there were sig-nificant regional differences in TCO. The TCO was higherin the midlatitudes of the NH, mainly due to more emissionsof ozone precursors from industry in this region. The maxi-mum TCO was mainly located in the banded areas between30◦N to 45◦N and around 30◦S. The TCO was high in JJAand SON, and low in MAM and DJF. The maximum TCOwas 36.58 DU in JJA, and the minimum was 30.46 DU inDJF. The TCO “trough” located over the Tibetan Plateau oc-curred in all seasons. The global mean TCO in 2013 (33.9DU) was nearly double the value in 1850 (15.82 DU).
The distribution of the simulated ERF due to the ozoneconcentration change since 1850 was dissimilar to that ofTCO for the reason of instant adjustment. The annual meanERF was 0.46 W m−2. A relatively positive ERF occurred inthe low latitudes of both hemispheres, and a negative one inthe mid-high latitudes of the NH.
Tropospheric ozone, as an important short-lived green-house gas, has caused an average rise of 0.36◦C in globaltemperature since 1850. The warming in the middle and highlatitudes of both hemispheres was noticeable, and could ex-ceed 0.8◦C. The simulated annual mean difference in globalsurface evaporation and precipitation were both 0.02 mm d−1.Because of the change in tropospheric ozone concentrations,there were remarkable increases in the cloud cover near 60◦S,the North Pole and the South Pacific, but the cloud cover de-creased sharply near 40◦N. The changes in the global distri-bution of precipitation and cloud cover were similar, espe-cially near the equator. Precipitation increased significantlyover the oceans from 10◦N to 30◦N, but was reduced (byabout 0.6 mm d−1) over the central Indian basin and southof Guam.
Acknowledgements. This work was supported by the NationalNatural Science Foundation of China (Grant No. 41575002).
REFERENCES
Anton, M., M. Lopez, J. M. Vilaplana, M. Kroon, R. McPeters, M.Banon, and A. Serrano, 2009: Validation of OMI-TOMS andOMI-DOAS total ozone column using five Brewer spectrora-diometers at the Iberian Peninsula. J. Geophys. Res.: Atmos.,114, D14307.
Buchard, V., C. Brogniez, F. Auriol, B. Bonnel, J. Lenoble, A.Tanskanen, B. Bojkov, and P. Veefkind, 2008: Comparison ofOMI ozone and UV irradiance data with ground-based mea-surements at two French sites. Atmos. Chem. Phys., 8, 4517–4528.
Chang, W.-Y., H. Liao, and H.-J. Wang, 2009: Climate responsesto direct radiative forcing of anthropogenic aerosols, tropo-spheric ozone, and long-lived greenhouse gases in easternChina over 1951–2000. Adv. Atmos. Sci., 26, 748–762, doi:10.1007/s00376-009-9032-4.
Conley, A. J., J. F. Lamarque, F. Vitt, W. D. Collins, and J. Kiehl,2013: PORT, a CESM tool for the diagnosis of radiative forc-ing. Geoscientific Model Development, 6, 469–476.
Crutzen, P. J., 1974: Photochemical reactions initiated by and in-
JULY 2016 XIE ET AL. 827
fluencing ozone in unpolluted tropospheric air. Tellus, 26, 47–57.
Derwent, R. G., M. E. Jenkin, and S. M. Saunders, 1996: Photo-chemical ozone creation potentials for a large number of re-active hydrocarbons under European conditions. Atmos. Env-iron., 30, 181–199.
Forster, P., and Coauthors, 2007: Changes in atmospheric con-stituents and in radiative forcing. Chapter 2. Climate Change2007: The Physical Science Basis. Contribution of WorkingGroup I to the Fourth Assessment Report of the Intergovern-mental Panel on Climate Change, Intergovernmental Panel onClimate Change, Ed., Cambridge University Press.
Fowler, D., and Coauthors, 2009: Atmospheric compositionchange: Ecosystems-Atmosphere interactions. Atmos. Envi-ron., 43, 5193–5267.
Gao, L. J., M. G. Zhang, and Z. W. Han, 2009: Model analy-sis of seasonal variations in tropospheric ozone and carbonmonoxide over East Asia. Adv. Atmos. Sci., 26, 312–318, doi:10.1007/s00376-009-0312-9.
Heath, D. F., A. J. Krueger, H. A. Roeder, and B. D. Henderson,1975: The solar backscatter ultraviolet and total ozone map-ping spectrometer (SBUV/TOMS) for NIMBUS G. OpticalEngineering, 14, 323–331.
Hsu, J., and M. J. Prather, 2009: Stratospheric variability andtropospheric ozone. J. Geophys. Res.: Atmos., 114(D6), doi:10.1029/2008JD010942.
Hurrell, J. W., J. J. Hack, D. Shea, J. M. Caron, and J. Rosinski,2008: A new sea surface temperature and sea ice boundarydataset for the community atmosphere model. J. Climate, 21,5145–5153.
Isaksen, I. S. A., and Coauthors, 2009: Atmospheric composi-tion change: Climate–Chemistry interactions. Atmos. Envi-ron., 43, 5138–5192.
Johnson, C. E., D. S. Stevenson, W. J. Collins, and R. G. Derwent,2001: Role of climate feedback on methane and ozone studiedwith a coupled ocean-atmosphere-chemistry model. Geophys.Res. Lett., 28, 1723–1726.
Kim, J. H., and H. Lee, 2010: What causes the springtime tropo-spheric ozone maximum over Northeast Asia? Adv. Atmos.Sci., 27, 543–551, doi: 10.1007/s00376-009-9098-z.
Kristjansson, J. E., T. Iversen, A. Kirkevåg, Ø. Seland, and J. De-bernard, 2005: Response of the climate system to aerosol di-rect and indirect forcing: Role of cloud feedbacks. J. Geo-phys. Res.: Atmos., 110(D24), doi: 10.1029/ 2005JD006299.
Lacis, A. A., D. J. Wuebbles, and J. A. Logan, 1990: Radiativeforcing of climate by changes in the vertical distribution ofozone. J. Geophys. Res.: Atmos., 95, 9971–9981.
Lamarque, J. F., and Coauthors, 2010: Historical (1850–2000)gridded anthropogenic and biomass burning emissions of re-active gases and aerosols: methodology and application. At-mospheric Chemistry and Physics, 10, 7017–7039.
Lamarque, J. F., and Coauthors, 2013: The Atmospheric Chem-istry and Climate Model Intercomparison Project (ACCMIP):Overview and description of models, simulations and climatediagnostics. Geoscientific Model Development, 6, 179–206.
Lin, L., Q. Fu, H. Zhang, J. Su, Q. Yang, and Z. Sun, 2013: Upwardmass fluxes in tropical upper troposphere and lower strato-sphere derived from radiative transfer calculations. Journal ofQuantitative Spectroscopy and Radiative Transfer, 117, 114–122.
Liu, C. X., Y. Liu, Z. N. Cai, S. T. Gao, J. C. Bian, X. Liu, andK. Chance, 2010: Dynamic formation of extreme ozone min-
imum events over the Tibetan Plateau during northern win-ters 1987–2001. J. Geophys. Res.: Atmos., 115(D18), doi:10.1029/2009JD013130.
Liu, Y., W.-L. Li, X.-J. Zhou, I.-S.-A. Isaksen, J.-K. Sundet,and J.-H.-He, 2003: The possible influences of the in-creasing anthropogenic emissions in India on troposphericozone and OH. Adv. Atmos. Sci., 20, 968–977, doi: 10.1007/BF02915520.
Miyazaki, K., H. J. Eskes, K. Sudo, M. Takigawa, M. van Weeleand K. F. Boersma, 2012: Simultaneous assimilation of satel-lite NO2, O3, CO, and HNO3 data for the analysis of tropo-spheric chemical composition and emissions. Atmos. Chem.Phys., 12, 9545–9579.
Myhre, G., and Coauthors, 2013a: Radiative forcing of the di-rect aerosol effect from AeroCom Phase II simulations. At-mospheric Chemistry and Physics, 13, 1853–1877.
Myhre G., and Coauthors, 2013b: Anthropogenic and natural ra-diative forcing. Climate Change 2013: The Physical ScienceBasis. Contribution of Working Group I to the Fifth As-sessment Report of the Intergovernmental Panel on ClimateChange, IPCC, Ed., Cambridge University Press.
Nakajima, T., A. Higurashi, K. Kawamoto, and J. Penner, 2000:A possible correlation between satellite-derived cloud andaerosol microphysical parameters. Geophys. Res. Lett., 28:1171–1174.
Pincus, R., H. W. Barker, and J. J. Morcrette, 2003: A fast, flex-ible, approximate technique for computing radiative trans-fer in inhomogeneous cloud fields. J. Geophys. Res.: Atmos.,108(D13), doi: 10.1029/2002JD003322.
Shindell, D., G. Faluvegi, A. Lacis, J. Hansen, R. Ruedy, and E.Aguilar, 2006: Role of tropospheric ozone increases in 20th-century climate change. J. Geophys. Res., 111, D08302.
Shindell, D., and Coauthors, 2013: Attribution of historical ozoneforcing to anthropogenic emissions. Nature Climate Change,3, 567–570.
Skeie, R. B., T. K. Berntsen, G. Myhre, K. Tanaka, M. M.Kvalevåg, and C. R. Hoyle, 2011: Anthropogenic radiativeforcing time series from pre-industrial times until 2010. At-mospheric Chemistry and Physics, 11, 11 827–11 857.
Søvde, O. A., C. R. Hoyle, G. Myhre, and I. S. A. Isaksen, 2011:The HNO3 forming branch of the HO2 + NO reaction: Pre-industrial-to-present trends in atmospheric species and radia-tive forcings. Atmospheric Chemistry and Physics, 11, 8929–8943.
Stevenson, D. S., and Coauthors, 2006: Multimodel ensemble sim-ulations of present-day and near-future tropospheric ozone. J.Geophys. Res.: Atmos. (1984–2012), 111(D8), D08301.
Stevenson, D. S., and Coauthors, 2013: Tropospheric ozonechanges, radiative forcing and attribution to emissions in theAtmospheric Chemistry and Climate Model IntercomparisonProject (ACCMIP). Atmos. Chem. Phys., 13, 3063–3085.
Stohl, A., and Coauthors, 2003: Stratosphere-troposphere ex-change: A review, and what we have learned from staccato. J.Geophys. Res.: Atmos. (1984–2012), 108(D12), doi: 10.1029/2002JD002490.
Stordal, F., and Coauthors, 2003: Climate impact of troposphericozone changes. Ozone-Climate Interactions, I. S. A. Isak-sen, Ed., European Commission Air Pollution research reportNo.81.
Veefkind, J. P., J. F. de Haan, E. J. Brinksma, M. Kroon, and P. F.Levelt, 2006: Total ozone from the ozone monitoring instru-ment (OMI) using the DOAS technique. IEEE Trans. Geosci.
828 ERF AND CLIMATE RESPONSE DUE TO TROPOSPHERIC OZONE VOLUME 33
Remote Sens., 44(5), 1239–1244.Wang, Z.-L., H. Zhang, X.-W. Jing, and X.-D. Wei, 2013a: Effect
of non-spherical dust aerosol on its direct radiative forcing.Atmos. Res., 120–121, 112–126.
Wang, Z.-L., H. Zhang, J.-N. Li, X.-W. Jing, and P. Lu, 2013b:Radiative forcing and climate response due to the presence ofblack carbon in cloud droplets. J. Geophys. Res.: Atmos., 118,3662–3675.
Wang, Z. L., H. Zhang , and X. Y. Zhang, 2014: Black carbonreduction will weaken the aerosol net cooling effect. Atmos.Chem. Phys. Discuss, 14, 33 117–33 141, doi: 10.5194/acpd-14-33117-2014.
Wang, Z. L., H. Zhang, and X. Y. Zhang, 2015: Simultaneous re-ductions in emissions of black carbon and co-emitted specieswill weaken the aerosol net cooling effect, Atmos. Chem.Phys., 15(7), 3671–3685.
Wu, T. W., and Coauthors, 2010: The Beijing Climate Center at-mospheric general circulation model: description and its per-formance for the present-day climate. Climate Dyn., 34, 123–147.
Xin, X. G., T. W. Wu, J. L. Li, Z. Z. Wang, W. P. Li, and F. H. Wu,2013: How well does BCC CSM1.1 reproduce the 20th cen-tury climate change over China? Atmospheric and OceanicScience Letters, 6(1), 21–26.
Young, P. J., and Coauthors, 2012: Pre-industrial to end 21st cen-tury projections of tropospheric ozone from the AtmosphericChemistry and Climate Model Intercomparison Project (AC-CMIP). Atmospheric Chemistry and Physics Discussions, 12,21 615–21 677.
Zhang, H., T. Nakajima, G. Y. Shi, T. Suzuki, and R. Imasu, 2003:An optimal approach to overlapping bands with correlated kdistribution method and its application to radiative calcula-tions. J. Geophys. Res.: Atmos. (1984–2012), 108(D20), doi:10.1029/2002JD003358.
Zhang, M. G., Y. F. Xu, I. Uno, and H. Akimoto, 2004: A numeri-cal study of tropospheric ozone in the springtime in East Asia.Adv. Atmos. Sci., 21, 163–170, doi: 10.1007/BF02915702.
Zhang, H., G. Y. Shi, T. Nakajima, and T. Suzuki, 2006a: The ef-fects of the choice of the k-interval number on radiative calcu-lations. Journal of Quantitative Spectroscopy and Radiative
Transfer, 98, 31–43.Zhang, H., T. Suzuki, T. Nakajima, G. Y. Shi, X. Y. Zhang, and Y.
Liu, 2006b: Effects of band division on radiative calculations.Optical Engineering, 45, 016002.
Zhang, H., Z. Shen, X.-D. Wei, M. Zhang, and Z. Li, 2012a: Com-parison of optical properties of nitrate and sulfate aerosol andthe direct radiative forcing due to nitrate in china. Atmos. Res.,113, 113–125.
Zhang, H., and Coauthors, 2012b: Simulation of direct radiativeforcing of aerosols and their effects on East Asian climateusing an interactive AGCM-aerosol coupled system. ClimateDyn., 38, 1675–1693.
Zhang, H., Q. Chen, B. Xie, and S. Y. Zhao, 2014a: PM2.5 andtropospheric ozone in china and pollutant emission control-ling integrated analyses. Progressus Inquisitiones de Muta-tione Climatis, 10, 289–296. (in Chinese)
Zhang, H., X.-W. Jing, and J. Li, 2014b: Application andevaluation of a new radiation code under McICA schemein BCC AGCM2.0.1. Geoscientific Model Development, 7,737–754.
Zhao, S.Y., H. Zhang, S. Feng, and Q. Fu, 2015: Simulating directeffects of dust aerosol on arid and semi-arid regions usingan aerosol climate model system. Int. J. Climatol., 35, doi:10.1002/joc.4093.
Zhou, X. J., C. Luo, W. L. Li, and J. E. Shi, 1995: Total ozonechanges in China and low center over the Tibetan plateau.Chinese Science Bulletin, 40, 1396–1398. (in Chinese)
Zhou, T. J., L. W. Zou, B. Wu, C. X. Jin, F. F. Song, X. L. Chen,and L. X. Zhang, 2014a: Development of earth/climate sys-tem models in China: A review from the Coupled Model In-tercomparison Project perspective. Journal of MeteorologicalResearch, 28(5), 762–779.
Zhou, T. J., and Coauthors, 2014b: Chinese contribution to CMIP5:An overview of five Chinese models’ performances. Journalof Meteorological Research, 28(4), 481–509.
Ziemke, J. R., S. Chandra, G. J. Labow, P. K. Bhartia, L. Froide-vaux, and J. C. Witte, 2011: A global climatology of tropo-spheric and stratospheric ozone derived from aura OMI andMLS measurements. Atmos. Chem. Phys., 11, 9237–9251.
