acp-7-1193-2007 | N Ox | Human Impact On The Environment

of 20

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
20 pages
0 downs
  Atmos. Chem. Phys., 7, 1193–1212, © Author(s) 2007. This work is licensedunder a Creative Commons License. AtmosphericChemistryand Physics The influence of African air pollution on regional and globaltropospheric ozone A. M. Aghedo 1,2 , M. G. Schultz 3 , and S. Rast 11 Max Planck Institute for Meteorology, Hamburg, Germany 2 International Max Planck Research School on Earth System Modelling, Hamburg, Germany 3 ICG-II, Research Centre, J¨ulich, GermanyReceived: 3 May 2006 – Published in Atmos. Chem. Phys. Discuss.: 4 July 2006Revised: 19 January 2007 – Accepted: 15 February 2007 – Published: 22 February 2007 Abstract.  We investigate the influence of African biomassburning, biogenic, lightning and anthropogenic emissions onthe tropospheric ozone over Africa and globally using a cou-pled global chemistry climate model. Our model studies in-dicate that surface ozone concentration may rise by up to50ppbv in the burning region during the biomass burningseasons. Biogenic emissions yield between 5–30ppbv in-crease in the near surface ozone concentration over tropicalAfrica. The impact of lightning on surface ozone is negli-gible, while anthropogenic emissions yield a maximum of 7ppbv increase in the annual-mean surface ozone concentra-tion over Nigeria, South Africa and Egypt. Our results showthat biogenic emissions are the most important African emis-sion source affecting total tropospheric ozone. The influenceof each of the African emissions on the global troposphericozone burden (TOB) of 384Tg yields about 9.5Tg, 19.6Tg,9.0Tg and 4.7Tg for biomass burning, biogenic, lightningand anthropogenic emissions emitted in Africa respectively.The impact of each of these emission categories on AfricanTOB of 33Tg is 2.5Tg, 4.1Tg, 1.75Tg and 0.89Tg respec-tively, which together represents about 28% of the total TOBcalculated over Africa. Our model calculations also sug-gest that more than 70% of the tropospheric ozone producedby each of the African emissions is found outside the con-tinent, thus exerting a noticeable influence on a large partof the tropical troposphere. Apart from the Atlantic and In-dian Ocean, Latin America experiences the largest impact of African emissions, followed by Oceania, the Middle East,Southeast and south-central Asia, northern North America(i.e. the United States and Canada), Europe and north-centralAsia, for all the emission categories. Correspondence to:  A. M. Aghedo( 1 Introduction Air pollution emitted in Africa comes from four sources,mainly biomass burning, natural emission from vegetationand soil, lightning NO x  emissions, and other anthropogenicsources – such as emissions related to the combustion of fossil-fuel for energy, industrial, transport and domestic uses.Africa contributes a significant amount to the global emis-sions from the first three categories, while emissions fromfossil fuel combustion are important only on the regionalscale. Emissions of trace species (e.g. CO, NO x , volatile or-ganic compounds (VOCs)) significantly affect troposphericchemistry and lead to the formation of tropospheric ozone,which influences the radiative forcing (e.g. Naik et al., 2005;Dentener et al., 2005).African biomass burning activities, generally categorisedas savanna, forest and agricultural waste burning, are drivenby the “slash and burn” agricultural practices that take placeduring the dry seasons – late November to early Marchin the northern hemisphere (NH), and July to October inthe southern hemispheric (SH) part of Africa (Marenco etal., 1990). African biomass burning contributes about 40% to trace species emitted by global biomass burning activi-ties (Crutzen et al., 1979; Andreae, 1991; Helas et al., 1995, Schultz et al., 2007 1 ), and exerts a large influence on tro-pospheric chemistry (Crutzen and Andreae, 1990; Andreae,1993; Helas et al., 1992, 1995; Marufu et al., 2000). Marufu et al. (2000) used a global chemistry transport model to test the sensitivity of tropospheric ozone over Africa to emissionsfrom biomass burning from all over the world and found thatabout 16% of the 26Tg tropospheric ozone burden (TOB)over Africa is due to these global biomass burning emissions. 1 Schultz, M. G., Heil, A., Hoelzemann, J. H., Spessa, A.,Thonicke, K., Goldammer, J., Held, A. C., and Pereira, J. M.:Global Emissions from Wildland Fires in 1960 to 2000, Global Bio-geochem. Cycles, in review, 2007. Published by Copernicus GmbH on behalf of the European Geosciences Union.  1194 A. M. Aghedo et al.: African air pollution influence on tropospheric ozoneLightning produces NO x , mostly in the middle to up-per troposphere (Ridley et al., 1996; Pickering et al., 1996; Lamarque et al., 1996; Pickering et al., 1998; DeCaria et al., 2000), where it has a longer lifetime and greater ozoneproduction potential than in the lower troposphere (Liu etal., 1987; Pickering et al., 1990). For example, DeCaria et al. (2005) calculated a photochemical ozone enhancement of about 10ppbv 24h after a storm observed during theStratosphere-Troposphere Experiment: Radiation, Aerosolsand Ozone (STERAO-A) using a 3-D cloud-scale chem-istry transport model (CTM). Lightning activity is maxi-mum in the tropics. However, tropical thunderstorms are theleast well characterized, therefore, the uncertainty of trop-ical lightning NO x  is particularly large. Generally, the to-tal contribution of lightning to the global NO x  budget ishighly uncertain. Estimates range from 2–20Tg ( N )  peryear (Lawrence et al., 1995; Price et al., 1997; Huntrieser et al., 2002; Labrador et al., 2005). However, several stud-ies published after 2000 have suggested an estimate closerto the lower limit. Huntrieser et al. (2002) estimated about 3Tg ( N ) yr − 1 from lightning NO x  based on detailed anal-ysis of airborne NO x  measurements of European thunder-storms. Tie et al. (2001) and Martin et al. (2002) found that using a global lightning emission value of 7Tg ( N ) yr − 1 and6Tg ( N ) yr − 1 respectively, theirmodelsimulationsshowrea-sonable agreement with airborne observations of reactive ni-trogen species.Vegetation emits a wide range of VOCs (Kesselmeier andStaudt, 1999). Among these biogenic VOCs, isoprene is one of the most important (e.g., Fehsenfeld et al., 1992; Guenther et al., 1995), followed by terpenes and methanol (CH 3 OH).Estimates of global isoprene emissions vary between 200Tgand 600Tg (Kesselmeier and Staudt, 1999; Guenther et al., 2006, and references therein), of which Africa contributesabout a fifth. Biogenic VOCs can have a significant impacton tropospheric chemistry as soon as they are released intothe air, because of their high reactivity. They lead to the pro-duction (or destruction) of ozone in high (or low) NO x  con-ditions. Wang and Shallcross (2000) found an increase in surface ozone concentration of about 4ppbv over the oceanand about 8–12ppbv over the mid-latitude land areas whenisoprene emissions were included in their 3-D model simula-tions. Using biogenic CH 3 OH emissions of 117Tg ( C ) yr − 1 ,Tie et al. (2003) calculated an increase of about 3–4% in the tropical ozone at 300hPa.In this study we performed multi-year sensitivity calcula-tions with a newly developed global chemistry climate modelusingthebiomassburning, biogenicandanthropogenicemis-sions employed in the recent IPCC-ACCENT simulationsin preparation for the fourth assessment report (e.g. seeStevenson et al., 2006; Shindell et al., 2006). The light-ning NO x  emissions are calculated interactively within ourmodel. These sensitivity studies are used to investigate theregional and global influence of each of the African emis-sions on primary and secondary tropospheric trace speciesconcentrations. In particular, we focuson tropospheric ozoneproducedfromthephotochemicalreactionsinvolvingprecur-sors emitted in Africa. Tropospheric ozone is a greenhousegas (Wang et al., 1980; Hansen et al., 2002), high ozone con- centrations in the air affect human health (e.g. Peden, 2001;Desqueyroux et al., 2002; Mortimer et al., 2002) and damagevegetation, including agricultural crops (e.g. Mauzerall andWang, 2001; Oksanen and Holopainen, 2001).A brief description of our model and the setup of the sim-ulation experiments are given in Sects. 2 and 3 respectively. An evaluation of ECHAM5-MOZ is presented in Sect. 4.The results of the sensitivity experiments are discussed inSect. 5. Conclusions and a summary are given in Sect. 6. 2 The global chemistry climate model ECHAM5-MOZ The full description of ECHAM5-MOZ and its sensitivityto the use of different emission inventories can be foundin Rast et al. (2007) 2 . The 3-D global chemistry climatemodel ECHAM5-MOZ is part of the Max Planck Insti-tute, Hamburg Earth System Model (ESM) and consists of the 3-D global general circulation model (GCM) ECHAM5(Roeckner et al., 2003) and the 3-D global CTM, MOZART2 (Horowitz et al., 2003), with modified parameterizations of dry and wet deposition, surface ultraviolet (UV) albedo, andlightning NO x  production. The ECHAM5-MOZ model em-ploys a consistent link of the chemistry calculation withthe parameterisation of the dynamics and the physics of theECHAM5 model.2.1 Atmospheric dynamicsThe dynamical core of ECHAM5 solves prognostic equa-tions for vorticity, divergence, logarithm of surface pres-sure and temperature expressed in spectral coefficients. Thevertical axis uses a hybrid terrain-following sigma-pressurecoordinate system (Simmons and Burridge, 1981). Themodel uses a semi-implicit leapfrog time integration scheme(Robert et al., 1972; Robert, 1981, 1982) with a special time filter (Asselin, 1972). Details of the physical parameteri-sations including radiation, surface processes, gravity wavedrag, convection, stratiform cloud formation, orbit varia-tions, and subgrid scale orography can be found in Roeck-ner et al. (2003). ECHAM5 (and thus ECHAM5-MOZ) can be run in various horizontal resolutions such as T42( ∼ 2.8 ◦ × 2.8 ◦ ), T63 ( ∼ 1.9 ◦ × 1.9 ◦ ) and T106 ( ∼ 1 ◦ × 1 ◦ ) us-ing 19 or 31  σ  -hybrid vertical levels. It can also be runas a coupled ocean-atmosphere model or in an atmosphere-only mode. In this study, we have constrained sea surface 2 Rast, S., Schultz, M. G., Aghedo, A. M., Diehl, T., Rhodin, A.,Schmidt, H., Stier, P., Ganzeveld, L. and Walters, S.: Sensitivity of a chemistry climate model to changes in emissions and the drivingmeteorology, in preparation, 2007. Atmos. Chem. Phys., 7, 1193–1212, 2007   A. M. Aghedo et al.: African air pollution influence on tropospheric ozone 1195temperatures (SST) and sea ice (SIC) by output from cou-pled ocean-atmosphere model simulations performed in theframework of the fourth IPCC assessment report (Roeck-ner et al., 2006) in an Atmospheric Model IntercomparisonProject 2 (AMIP2, Gates et al., 1999) set-up. The effect of  varying model resolutions on the simulated climate is de-scribed in Roeckner et al. (2006).2.2 Tracer transport and depositionTracers in ECHAM5-MOZ undergo advective and convec-tive transport, vertical diffusion, dry and wet deposition, andchemical reactions in the atmosphere. The advection of trac-ers is based on a mass conserving flux-form semi Lagrangiantransport scheme (Lin and Rood, 1996) on a Gaussian grid(Arakawa C-grid, Mesinger and Arakawa (1976)). Convec- tive transport is parameterized according to the mass-flux al-gorithm of  Tiedtke (1989) with modifications proposed byNordeng (1994). ECHAM5-MOZ extends the vertical diffu- sion equations of ECHAM5 to include the net flux of trac-ers at the earth’s surface (e.g. emission and dry deposition).The dry deposition is formulated according to the schemeof  Ganzeveld (2001). The wet deposition is based on the scheme of  Stier et al. (2005), with modifications for below- cloud scavenging for HNO 3  (Seinfeld and Pandis, 1998, ,page 1003). This dynamical wet deposition scheme takesinto consideration the solubility of the tracers and the pos-sibility of the release of trace gases into the atmosphere byre-evaporation of precipitation. The sensitivity of transportof tracers in ECHAM5 to model resolution, forcing meteo-rology and their chemical lifetime is discussed in Aghedo etal. (2007) 3 .2.3 Chemistry schemeThe ECHAM5-MOZ model uses the MOZART2 tropo-spheric chemistry scheme, consisting of 63 transportedspecies and 168 chemical reactions. The details of the chem-ical species, reactions, kinetic equations and the chemistrysolveraredescribedinHorowitzetal.(2003). Asintheorigi-nal MOZART model, the ECHAM5-MOZ chemical reactionscheme is flexible due to the MOZART2 preprocessor whichproduces machine dependent optimized (e.g. vectorized andparallelized) code for a specific set of user-defined reactions.An implicit Euler method is applied for the integration of the kinetic nonlinear differential equations for most of thespecies.2.4 Lightning emissionsECHAM5-MOZ includes interactive lightning NO x  emis-sions according to the parameterisation of  Grewe et al. 3 Aghedo, A. M., Schultz, M. G., and Rast, S.: Sensitivity of tracer transport to model resolution and forcing data in the generalcirculation model ECHAM5, in preparation, 2007. (2001). The lightning frequency is calculated as a functionof the mean updraught velocity in a convective column. Themean updraught velocity is resolution dependent, because itdepends on the size of the grid boxes. Therefore, the param-eterisation contains one freely adjustable global factor thataccounts for this grid-box dependency. The NO x  emissionsare proportional to the calculated flash frequency and are dis-tributed vertically in the atmosphere using C-shaped profilesfor tropical and extratropical continental and marine cloudsas described in Pickering et al. (1998). This parameterisa-tion yields global lightning emissions of about 2.7Tg ( N )/ yrin ECHAM5-MOZ. Over Africa, total lightning emissionsare about 0.7Tg ( N )/ yr. A visual comparison of the spa-tial and seasonal distribution of our lightning flashes withthat of Lightning Imaging Sensor (LIS) data (Christian et al.,1989, 1992 available at query/distributions.html) shows good agreement.2.5 Biogenic emissionsIn this study, we use the prescribed biogenic emissions fromthe recent IPCC-ACCENT Photocomp 2030 intercompari-son experiment (Stevenson et al., 2006). The biogenic emis- sions are 756Tg ( C ) yr − 1 , 68Tg ( C ) yr − 1 , and 8Tg ( N ) yr − 1 for non-methane VOCs (NMVOC), CO and NO x  respec-tively. Isoprene, terpenes and methanol account for about68%, 17% and 11% of the biogenic NMVOC respectively.As an alternative to prescribed globally-gridded biogenicNMVOC emissions, ECHAM5-MOZ offers the option of anonline calculation of biogenic NMVOC emissions accordingto the Model of Emissions of Gases and Aerosols from Na-ture (MEGAN) (Guenther et al., 2006).2.6 Other emissionsThe ECHAM5-MOZ model needs gridded emission data foremissions that are not calculated interactively. Emissionssuch as biomass burning (all open fires including savanna,forest, and agricultural burning), aircraft, ocean and anthro-pogenic emissions (such as fossil-fuel combustion by thedomestic, transport and industrial sectors) are prescribed tothe ECHAM5-MOZ model as monthly-mean globally grid-ded files. They are injected into the model at various modelheights.With the exception of the lightning emissions, which arecalculated from the interactive lightning parameterisationin ECHAM5-MOZ as discussed in Sect. 2.4, all emissionsused for this study are identical to those used in the IPCC-ACCENT experiment (Stevenson et al., 2006). These datasets are a combination of emission inventories of the Insti-tute for Applied System Analysis (IIASA), the Global Emis-sions Inventory Activity (GEIA), the Global Fire EmissionsDatabase (GFED) version 1 (Randerson et al., 2005) andthe Emission Database for Global Atmospheric Research(EDGAR) version 3.2 (Olivier et al., 1999) Atmos. Chem. Phys., 7, 1193–1212, 2007  1196 A. M. Aghedo et al.: African air pollution influence on tropospheric ozone Table 1.  Global trace gas emissions by source used in this studyand contribution from the African continent to the respective globaltotal in parenthesis.CO NMVOC NO x Source (Tg ( C )/ yr) (Tg ( C )/ yr) (Tg ( N )/ yr)Industrial 201 (16%) 66 (14%) 28.0 (5%)Biomass burning 217 (43%) 19 (42%) 10.0 (46%)Biogenic 69 (20%) 756 (25%) 8.0 (30%)Lightning – – 2.7 (26%)Aircraft – – 0.7 (4%)Ocean 9 (–) 4 (–) –All sources 496 (28%) 845 (24%) 49.4 (18%)Marufu et al. (2000) 482 (23%) 534 (23%) 40.0 (18 %) Anthropogenic CO, NO x  and NMVOC emissions suchas domestic, industrial, road transport, off-road and power-plantsfossil-fuel combustion and gasflaring areascalculatedby the IIASA global version of the Regional Air PollutionInformation and Simulation (RAINS) model (Amann et al.,1999) for the year 2000. The international shipping CO, NO x and NMHC emissions are based on the EDGAR3.2 globalemission inventory (Olivier et al., 1999), while aircraft NO x emissions are specified according to the IPCC special reporton Aviation and the Global Atmosphere (IPCC, 1999).The biomass burning emissions, which include savanna,forest and deforestation fires, and agricultural waste burn-ing are from the GFED version 1 (Randerson et al., 2005)database available at For the simula-tions described in this paper, we use the 1997–2002 averagedata. Ocean CO and soil CO, H 2  and NO x  emissions arefrom the GEIA database (see Horowitz et al., 2003). The biogenic VOC emissions from vegetation are based on theglobal model of natural VOC contributed to the GEIA activ-ity by Guenther et al. (1995).Table 1 lists the global CO, NO x  and NMVOC emissionsby source used in this study, and the amount contributedby African emissions are included as the percentages of theglobal emissions. It also contain the comparison of our emis-sions to those of  Marufu et al. (2000). Our biogenic emis- sions are about twice as high as those of  Marufu et al. (2000),while others are comparable. 3 Model simulations The model experiments follow the general setup for theIPCC-ACCENT intercomparison study (Stevenson et al.,2006). Theexperimentswererunforpresent-dayclimateandemissions. The climate conditions (sea surface temperaturesand sea ice fields) were taken from six consecutive years of coupledocean-atmospheresimulationsperformedattheMaxPlanck Institute for Meteorology, Hamburg. Present-dayconstant concentrations of 1760ppbv, 367ppm and 316ppbvwere maintained for CH 4 , CO 2  and N 2 O respectively.ECHAM5 model climate simulations give better results athigher spatial resolutions (Roeckner et al., 2006). Specifi- cally, T42L19 and T63L31 have a particularly good balancebetween computational costs and quality of the results. How-ever, Aghedo et al. (2007) 3 show that the simulated transportof tracers in T42L31 is significantly different from that inT42L19, but comparable to that in the computationally moreexpensive T63L31 resolution. Each experiment in this studywas therefore performed in the T42L31 resolution for 5 years(1997–2001) after a spin-up of 6 months.We performed one reference experiment and 4 sensitiv-ity experiments. The reference experiment includes all theemissions, while in each of the sensitivity experiments, weswitch off one of the following emission categories over theAfrican continent: biomass burning, biogenic, lightning, oranthropogenic emissions, respectively. The differences be-tween the reference and the sensitivity experiments thereforeshow the impact of each of the African emissions.We are aware that setting an emission source to zero af-fects the lifetime of other trace species in the troposphere.Nevertheless, this approach provides a relatively uncom-plicated method in assessing the potential impact of thesedifferent emission types. Also, the method has the ad-vantage that the combined effect of different species (e.g.CO, NO x  and NMVOC) from the same emission category(e.g. biomass burning) on the overall tropospheric chem-istry can be assessed. The methane lifetime (average of 7.1years based on 150ppbv ozone-threshold tropopause) showsa small increase of about 13, 35 and 50 days in the experi-ment without anthropogenic, biomass burning and lightningemissions respectively, when compared to the reference ex-periment, while in the experiment where biogenic emissionsare switched off, it decreases by only 32 days. 4 Model Evaluation An extensive evaluation of the ECHAM5-MOZ model is de-scribed in Rast et al. (2007) 2 . In the ACCENT-IPCC sce-nario studies, ECHAM5-MOZ showed a high bias of about20% in the global tropospheric ozone production and losscompared to the mean of all participating models (Steven-son et al., 2006). In terms of global dry deposition, it has alow bias of about 5%. Methane and CH 3 CCl 3  troposphericlifetimes are at the lower end of the currently accepted esti-mations (Prinn et al., 1995; Ehhalt et al., 2001). The surface ozone concentration is known to have a high bias in heav-ily polluted areas, such as industrial centres and large cityagglomerations but also in the Mediterranean basin.In this section we evaluate the ECHAM5-MOZ modelover Africa and its surroundings, by comparing model-Atmos. Chem. Phys., 7, 1193–1212, 2007 
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks