Global and regional evolution of short-lived radiatively-active gases and aerosols in the Representative Concentration Pathways
- First Online:
- Cite this article as:
- Lamarque, J., Kyle, G.P., Meinshausen, M. et al. Climatic Change (2011) 109: 191. doi:10.1007/s10584-011-0155-0
- 1.4k Views
In this paper, we discuss the results of 2000–2100 simulations following the emissions associated with the Representative Concentration Pathways (RCPs) with a chemistry-climate model, focusing on the changes in 1) atmospheric composition (troposphere and stratosphere) and 2) associated environmental parameters (such as nitrogen deposition). In particular, we find that tropospheric ozone is projected to decrease (RCP2.6, RCP4.5 and RCP6) or increase (RCP8.5) between 2000 and 2100, with variations in methane a strong contributor to this spread. The associated tropospheric ozone global radiative forcing is shown to be in agreement with the estimate used in the RCPs, except for RCP8.5. Surface ozone in 2100 is projected to change little compared from its 2000 distribution, a much-reduced impact from previous projections based on the A2 high-emission scenario. In addition, globally-averaged stratospheric ozone is projected to recover at or beyond pre-1980 levels. Anthropogenic aerosols are projected to strongly decrease in the 21st century, a reflection of their projected decrease in emissions. Consequently, sulfate deposition is projected to strongly decrease. However, nitrogen deposition is projected to increase over certain regions because of the projected increase in NH3 emissions.
In support of the analysis of future climate change (for instance as part of community efforts such as the CMIP5 exercise (Taylor et al. 2009)), a set of pathways has been generated to define a range of possible future atmospheric composition over the 21st century (van Vuuren et al. 2011a). These pathways, referred to as Representative Concentrations Pathways (RCPs) are derived from estimated emissions computed by a set of Integrated Assessment Models (Masui et al. 2011; Riahi et al. 2011; Thomson et al. 2011; van Vuuren et al. 2011b) and, for each of these pathways, a rich dataset of information regarding emissions, concentrations and land-use has been generated. For long-lived greenhouse gases, these emissions have been converted into concentrations (and consequently radiative forcing estimates) using the Model for the Assessment of Greenhouse-gas Induced Climate Change (MAGICC, Meinshausen et al. 2011). However, because of regional variations in emissions, it is important to perform the calculation of the distribution of short-lived climate agents using a comprehensive three-dimensional chemistry-climate model. It is the results of such calculations that are discussed in this paper. As such, it is a follow-up to Lamarque et al. (2010), where, using the same model, historical emissions (1850–2000) are used and the corresponding computed changes in ozone and aerosols are discussed.
In addition to the changes in atmospheric composition mentioned above, emissions of short-lived gases and aerosols lead to deposition of nitrogen and sulfate compounds that affect the biosphere, on land and in the ocean (Billen et al. 2010; Butchart et al. 2010, and references therein). In particular, specific thresholds for nitrogen deposition above which additional input is detrimental to plants have been used to identify regions where past and future trends may have adverse impacts (Bobbink et al. 1998; Dentener et al. 2006a; Galloway et al. 2008). Conversely, nitrogen deposition may also influence the carbon cycle by stimulating plant growth (Schlesinger 2009). None of those feedbacks are included in our study or in the RCP projections. As such, it will be important to evaluate the response of the climate system (and how it affects radiative forcing estimates) in more comprehensive models with explicit representation of biogeochemistry and the carbon cycle.
The paper is organized as follows: in Section 2, we describe the model used for the chemistry simulations, along with the experimental design. In Section 3, we focus on the analysis of ozone changes, both tropospheric and stratospheric. Section 4 describes the simulated aerosol changes. In both Sections 3 and 4, radiative forcing calculations are included. We analyze nitrogen and sulfate deposition in Section 5. Discussion and conclusions are in Section 6.
2 Model description and experimental design
In order to provide the distributions of future aerosol loadings, deposition rates and ozone concentrations discussed later in this paper, we perform simulations with the same model setup as used for providing the respective historical datasets, as described in Lamarque et al. (2010). We use the global three-dimensional Community Atmosphere Model version 3.5 (Gent et al. 2009) modified to include interactive chemistry to calculate distributions of gases and aerosols in the troposphere and the lower to mid-stratosphere. In order to limit computational cost, this model only solves for the atmospheric and land portions of the climate system, using pre-computed sea-surface temperatures and sea-ice extent as boundary conditions.
The model configuration used in this study includes a horizontal resolution of 1.9° (latitude) by 2.5° (longitude) and 26 hybrid levels, from the surface to ≈ 40 km with a timestep of 30 min; the transient simulations were performed continuously between 2001 and 2100. In order to simulate the evolution of the atmospheric composition over the model vertical range, the chemical mechanism used in this study is formulated to provide an accurate representation of both tropospheric and stratospheric chemistry (Lamarque et al. 2008). Specifically, to successfully simulate the chemistry above 100 hPa, we include a representation of stratospheric chemistry (including polar ozone loss associated with stratospheric clouds) from version 3 of MOZART (MOZART-3; Kinnison et al. 2007). The tropospheric chemistry mechanism has a simplified representation of non-methane hydrocarbon chemistry in addition to standard methane chemistry, extended from Houweling et al. (1998) with the inclusion of isoprene and terpene oxidation and updated to JPL-2006 (Sander et al. 2006). This model has a representation of aerosols based on the work by Tie et al. (2001, 2005); i.e., sulfate aerosol is formed by the oxidation of SO2 in the gas phase (by reaction with the hydroxyl radical) and in the aqueous phase (by reaction with ozone and hydrogen peroxide). Furthermore, the model includes a representation of ammonium nitrate that is dependent on the amount of sulfate present in the air mass following the parameterization of gas/aerosol partitioning by Metzger et al. (2002). Because only the bulk mass is calculated, a lognormal distribution is assumed for all aerosols using different mean radius and geometric standard deviation (Liao et al. 2003). We use a 1.6 days of exponential lifetime for the conversion from hydrophobic to hydrophilic carbonaceous aerosols (organic and black). Natural aerosols (desert dust and sea salt) are implemented following Mahowald et al. (2006a and b), and the sources of these aerosols are derived based on the model-calculated wind speed and surface conditions.
For all the simulations, the initial conditions correspond to the distributions of all chemically active species from the January 1 2001 results of the transient simulation (1850–2000) described in Lamarque et al. (2010). For each RCP one simulation has been performed. Results are presented for all the RCPs: RCP2.6, also known as RCP3PD, RCP4.5, RCP6 and RCP8.5.
At the lower boundary, the time-varying (monthly values) zonal-averaged distributions of CO2, CH4, H2, N2O and all the halocarbons (CFC-11, CFC-12, CFC-113, HCFC-22, H-1211, H-1301, CCl4, CH3CCl3, CH3Cl and CH3Br) are specified following the datasets described in Meinshausen et al. (2011), except for H2 which is kept at a constant 500 ppbv. Emissions from anthropogenic activities and biomass burning (natural and anthropogenically-forced) are taken from the various RCPs. Note that the biomass burning emissions vary amongst RCPs and in time, following changes in land-use; however, no climate feedback with fire frequency is included. Finally, the natural emissions of ozone precursors and of sulfur compounds (from non-eruptive volcanoes, Dentener et al. 2006b) are kept constant (i.e. set at their value in 2000) for the whole duration of the simulations. While reasonable for the historical period (Lathière et al. 2006), this is an assumption that will need to be evaluated in future studies, especially for the case of biogenic volatile organic compounds (VOC) emissions, which can be climate and CO2-dependent (Guenther et al. 2006; Young et al. 2009).
Scenario used in CCSM3 simulation
B1 (4.2 W/m2)
A1B (6.1 W/m2)
A2 (8.0 W/m2)
Total (anthropogenic + shipping + aircraft + biomass burning) global emissions for ozone precursors and aerosols: Tg(species)/year, except NOx emissions expressed as Tg(NO2)/year
In this section, we discuss the results from the simulations described above and focus on the analysis of ozone (tropospheric and stratospheric), both on its distribution of ozone and its associated radiative forcing.
3.1 Analysis of present-day distribution of tropospheric ozone
Overall, the evaluation (along with additional recent studies discussing and evaluating simulations performed with this model, e.g. Sanderson et al. 2008; Shindell et al. 2008; Anenberg et al. 2009; Fiore et al. 2009; Reidmiller et al. 2009; Jonson et al. 2010; Lamarque and Solomon 2010) indicates that the model is quite accurate in its representation of present-day ozone in the troposphere and lower stratosphere. Additional information on present-day and historical ozone trends can be found in Lamarque et al. (2010).
3.2 Change in tropospheric ozone burden
Using the scaling factor of 0.042 W/m2/DU (Ramaswamy et al. 2001) to compute the ozone radiative forcing associated with the previously described change in tropospheric ozone, we find that our estimated globally- and annually-averaged radiative forcing between 1850 and 2000 is in good agreement with the IPCC-AR4 best estimate between 1750 and 2005 (Table 2.12 in Forster and Ramaswamy, 2007). To facilitate the comparison with our simulated change with respect to 1850, we have corrected (in Fig. 4) the 1750-referenced radiative forcing estimate by –0.05 W/m2, as calculated in Meinshausen et al. (2011). In the case of RCP8.5, the additional tropospheric ozone (between 2000 and 2100) is expected to increase the radiative forcing by an additional 0.2 W/m2 by 2100, while the other projections lead to a decrease in the radiative forcing of tropospheric ozone between 0.07 and 0.2 W/m2. It is however clear that the radiative forcing by tropospheric ozone strongly varies regionally (Shindell et al. 2003) but this is not documented here.
It is worth noting that the estimated tropospheric ozone radiative forcing (and hence concentration perturbation) estimated in MAGICC for the RCP8.5 (and to a lower extent RCP4.5) case is noticeably lower than in our model simulations. This lower forcing is likely a combination of the limited applicability of the chemistry parameterizations (Prather and Ehhalt 2001) currently used in MAGICC and the lack of consideration of the increased flux of ozone from the stratosphere to the troposphere (Stevenson et al. 2006; Hegglin and Shepherd 2009), a result of stratospheric ozone increase (Eyring et al. 2010a) and increased stratospheric circulation (Rind et al. 2001; Garcia et al. 2007; Lamarque et al. 2008) in relation to increasing greenhouse gases. A more detailed discussion is provided in the next section. It is however important to note that the overall uncertainty in tropospheric ozone forcing is rather large (Fig. 4).
3.3 Change in stratospheric ozone burden
3.4 Change in zonally-averaged surface ozone
Aerosols are a strong component of the radiative forcing associated with anthropogenic emissions. This is achieved through a combination of direct radiative forcing and cloud-aerosol interactions (indirect effects), as summarized in Forster et al. (2007). In the simulations discussed here, only the direct effects are considered. In the discussion of the results, we therefore focus on the evolution of the distribution of sulfate, black and organic carbon and ammonium nitrate aerosols. Our model consists of a bulk representation (Lamarque et al. 2005a), leading to an external mixing assumption for all radiative calculations. The importance of this latter assumption is discussed below.
4.1 Comparison with present-day observations
In terms of the CAM-chem simulated aerosol optical depth (computed at 550 nm, referenced hereafter as AOD), the annual global average (including natural and anthropogenic sources for all aerosols but ammonia) for present-day is 0.115, in agreement with the satellite-based estimate (valid over the ocean) of 0.10–0.15 by Mishchenko et al. (2007) but lower than recent satellite-based global estimates (Chung et al. 2005; Remer et al. 2008). Our present-day simulated value represents an increase of 0.0315 over the 1850 conditions (0.08). This anthropogenic increase is very much in agreement with the average AeroCom results (Schulz et al. 2006). While the AeroCom reference point is 1750 instead of 1850, the overall AOD associated with anthropogenic emissions in 1850 (and therefore the potential increases between 1750 and 1850) is, however, quite small in our simulation.
4.2 Change in aerosol optical depth and burden
4.3 Direct radiative forcing estimates
Clear sky radiative forcing (W/m2) with respect to 1850. See text for details
Our estimate of the clear-sky direct radiative forcing from aerosols in year 2000 (–0.8 W/m2, with respect to 1850) is similar to the multi-model mean of the AeroCom models (–0.68 ± 0.28) in Schulz et al. (2006). Also, the black carbon radiative forcing computed here for year 2000 (0.1 W/m2) is very comparable to the AeroCom results for fossil-fuel black carbon (i.e., the portion that significantly changed over the historical period, Schulz et al. 2006), which averaged 0.12 ± 0.04 W/m2. It is however on the low end of published estimates (e.g.. Chung and Seinfeld 2005; Bond et al. 2010; Jacobson 2010 and references therein); furthermore, many models are biased low against observations (Koch et al. 2009). It is however important to note that, following Cooke et al. (1999), the Optical Properties of Aerosols and Clouds (OPAC) optics for soot aerosols are used for black carbon. OPAC treats soot as hydrophobic, so that its optic properties are not affected by ambient humidity (Collins et al. 2002). This will lead to an underestimate of the soot radiative forcing. Furthermore, as discussed in Jacobson (2000, 2001) and Chung and Seinfeld (2005), the external mixture assumption will also lead to an underestimate of the overall radiative forcing of black carbon. It will therefore be important to estimate the black carbon radiative forcing in another aerosol model where internal mixture and/or hygroscopicity are considered.
Using the same methodology, we find that, in all cases, the RCP-projected direct radiative effect of aerosols in 2100 (Table 3) is considerably reduced from the 2000 estimates, in agreement with the evolution of their emissions. It is interesting to note that the relative role of ammonium nitrate, especially under the RCP2.6 projection (and to a lesser extent RCP8.5) becomes significantly larger than for the historical period, emphasizing the importance of performing chemistry simulations with adequate representation of the sulfate-dependent ammonium nitrate as described by Adams et al. (2001), Schaap et al. (2004) and Myhre et al. (2006).
4.4 Change in the regional distribution of aerosols
5 Nitrogen and sulfate deposition
Reactive nitrogen, from natural and anthropogenic emissions of nitrogen oxides and ammonia, is deposited over land and ocean through a variety of processes (namely, wet and dry deposition) such that it can potentially act as a fertilizer, with an impact on vegetation in non-agricultural areas (Holland et al. 1997; Bouwman et al. 2002; Holland et al., 2005; Lamarque et al. 2005b; Dentener et al. 2006a; Thornton et al. 2007; Reay et al. 2008). Everything else being equal, these additional N inputs into ecosystems can potentially lead to an increased carbon sink (Reay et al. 2008). Sulfate deposition, on the other hand, has been linked to a decrease in tree growth as a result of acidification (Savva and Berninger 2010).
Nitrogen deposition for present-day conditions: mg(N)/m2/year
Nitrogen deposition 2000
All land (mg(N)/m2/year)
Natural vegetation (mg(N)/m2/year)
Rest of S. Amer.
Nitrogen deposition: mg(N)/m2/year. Bold indicates increase between 2000 and 2100
Rest of S. Amer.
Sulfur deposition: mg(S)/m2/year
Rest of S. Amer.
6 Discussion and conclusions
We have presented and analyzed results from a set of simulation with a comprehensive three-dimensional chemistry-climate model. These simulations focused on the representation of short-lived radiative forcing agents (aerosols and gases) following the emissions from the recently developed Representative Concentration Pathways (RCPs). These projections will be used in a number of research activities that will be assessed in the Intergovernmental Panel on Climate Change (IPCC) 5th Assessment Report.
Using a set of observations representative of present-day conditions, we have shown that the model forced by the 2000 emissions provides a reasonable representation of atmospheric composition. In addition, we have shown that the model response to the 20th century emissions (Lamarque et al. 2010) is in agreement with previously published estimates of ozone and aerosol radiative forcings (except for a low black carbon radiative forcing); note that this can of course only ensure consistency with previous studies and may not reflect the real value of the specific quantity discussed. Based on this positive evaluation, we have documented and discussed projections for ozone (tropospheric and stratospheric), tropospheric aerosols and deposition (nitrogen and sulfate). In particular, we have shown that the tropospheric ozone radiative forcing by 2100 will vary strongly between RCPs, a consequence of the variations in precursor emissions and methane concentrations (Meinshausen et al. 2011). In that calculation, we have also shown that the tropospheric ozone radiative forcing estimate from MAGICC is underestimated for RCP8.5, most likely from a lack of representation of the stratospheric circulation changes under a high-CO2 scenario. The projections of aerosol loading are quite similar amongst RCPs, except for NH3, which acquires in RCP2.6 and RCP8.5 a relatively more prominent role, especially with respect to nitrogen deposition. Projected nitrogen deposition in RCP2.6 and RCP8.5 shows increase in a number of regions around the globe. It is however important to note that these estimates could be very model-dependent due to the inherent difficulties in modeling ammonia chemistry (this system is quite sensitive to the ratio of sulfate and ammonium/ammonia, with ammonium nitrate formation promoted when all sulfate is neutralized) and its deposition, mostly through precipitation processes.
Contrary to projections under the SRES A2x scenario (Prather et al. 2003), the emissions provided by the RCPs lead to surface ozone projections for 2100 that signal improvements over present-day conditions. In particular, at least in a zonal-average sense, surface ozone is projected to worsen little, if not improve, over the 2000 conditions, even in the RCP8.5 projection. It is clear however that more analysis of the specific conditions under which high-pollution events occur is needed to fully identify the projected pollution regimes in 2100.
Beyond the points mentioned above, the application of the RCP emissions as discussed in this paper have by design (e.g., use of prescribed sea-surface temperatures or CO2 and CH4 concentrations) or by lack of process representation (e.g. carbon cycle or wetland emissions) ignored some potentially strong feedbacks between various portions of the Earth system. In particular, the importance of considering SRES-generated sea-surface temperatures remains to be discussed. Similarly, the RCP emission projections provide a new paradigm under which existing simple estimating formulas (such as in MAGICC) for concentrations and/or radiative forcings likely need to be revisited.
The authors would like to thank the three anonymous reviewers, P. Hess and E. Holland for their constructive feedback on previous versions of this paper. A. J. C. and F. V. were funded by the Department of Energy under the SciDAC program. Computing resources were provided by the Climate Simulation Laboratory at NCAR’s Computational and Information Systems Laboratory (CISL), sponsored by the National Science Foundation and other agencies.This research was enabled by CISL compute and storage resources. Bluefire, a 4,064-processor IBM Power6 resource with a peak of 77 TeraFLOPS provided more than 7.5 million computing hours, the GLADE high-speed disk resources provided 0.4 PetaBytes of dedicated disk and CISL's 12-PB HPSS archive provided over 1 PetaByte of storage in support of this research project. The CESM project is supported by the National Science Foundation and the Office of Science (BER) of the U.S. Department of Energy. The National Center for Atmospheric Research is operated by the University Corporation for Atmospheric Research under sponsorship of the National Science Foundation.