IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis Sulphate Aerosol

Atmospheric sulphate aerosol may be considered as consisting of sulphuric acid particles that are partly or totally neutralized by ammonia and that are present as liquid droplets or partly crystallized. Sulphate is formed by aqueous phase reactions within cloud droplets, oxidation of SO2 via gaseous phase reactions with OH, and by condensational growth onto pre-existing particles (e.g., Penner et al., 2001). Emission estimates are summarised by Haywood and Boucher (2000). The main source of sulphate aerosol is via SO2 emissions from fossil fuel burning (about 72%), with a small contribution from biomass burning (about 2%), while natural sources are from dimethyl sulphide emissions by marine phytoplankton (about 19%) and by SO2 emissions from volcanoes (about 7%). Estimates of global SO2 emissions range from 66.8 to 92.4 TgS yr–1 for anthropogenic emissions in the 1990s and from 91.7 to 125.5 TgS yr–1 for total emissions. Emissions of SO2 from 25 countries in Europe were reduced from approximately 18 TgS yr–1 in 1980 to 4 TgS yr–1 in 2002 (Vestreng et al., 2004). In the USA, the emissions were reduced from about 12 to 8 TgS yr–1 in the period 1980 to 2000 (EPA, 2003). However, over the same period SO2 emissions have been increasing significantly from Asia, which is estimated to currently emit 17 TgS yr–1 (Streets et al., 2003), and from developing countries in other regions (e.g., Lefohn et al., 1999; Van Aardenne et al., 2001; Boucher and Pham, 2002). The most recent study (Stern, 2005) suggests a decrease in global anthropogenic emissions from approximately 73 to 54 TgS yr–1 over the period 1980 to 2000, with NH emission falling from 64 to 43 TgS yr–1 and SH emissions increasing from 9 to 11 TgS yr–1. Smith et al. (2004) suggested a more modest decrease in global emissions, by some 10 TgS yr–1 over the same period. The regional shift in the emissions of SO2 from the USA, Europe, Russia, Northern Atlantic Ocean and parts of Africa to Southeast Asia and the Indian and Pacific Ocean areas will lead to subsequent shifts in the pattern of the RF (e.g., Boucher and Pham, 2002; Smith et al., 2004; Pham et al., 2005). The recently used emission scenarios take into account effective injection heights and their regional and seasonal variability (e.g., Dentener et al., 2006).

The optical parameters of sulphate aerosol have been well documented (see Penner et al., 2001 and references therein). Sulphate is essentially an entirely scattering aerosol across the solar spectrum (ωo = 1) but with a small degree of absorption in the near-infrared spectrum. Theoretical and experimental data are available on the relative humidity dependence of the specific extinction coefficient, fRH (e.g., Tang et al., 1995). Measurement campaigns concentrating on industrial pollution, such as the Tropospheric Aerosol Radiative Forcing Experiment (TARFOX; Russell et al., 1999), the Aerosol Characterization Experiment (ACE-2; Raes et al., 2000), INDOEX (Ramanathan et al., 2001b), the Mediterranean Intensive Oxidants Study (MINOS, 2001 campaign), ACE-Asia (2001), Atmospheric Particulate Environment Change Studies (APEX, from 2000 to 2003), the New England Air Quality Study (NEAQS, in 2003) and the Chesapeake Lighthouse and Aircraft Measurements for Satellites (CLAMS; Smith et al., 2005), continue to show that sulphate contributes a significant fraction of the sub-micron aerosol mass, anthropogenic τaer and RF (e.g., Hegg et al., 1997; Russell and Heintzenberg, 2000; Ramanathan et al., 2001b; Magi et al., 2005; Quinn and Bates, 2005). However, sulphate is invariably internally and externally mixed to varying degrees with other compounds such as biomass burning aerosol (e.g., Formenti et al., 2003), fossil fuel black carbon (e.g., Russell and Heintzenberg, 2000), organic carbon (Novakov et al., 1997; Brock et al., 2004), mineral dust (e.g., Huebert et al., 2003) and nitrate aerosol (e.g., Schaap et al., 2004). This results in a composite aerosol state in terms of effective refractive indices, size distributions, physical state, morphology, hygroscopicity and optical properties.

The TAR reported an RF due to sulphate aerosol of –0.40 W m–2 with an uncertainty of a factor of two, based on global modelling studies that were available at that time. Results from model studies since the TAR are summarised in Table 2.4. For models A to L, the RF ranges from approximately –0.21 W m–2 (Takemura et al., 2005) to –0.96 W m–2 (Adams et al., 2001) with a mean of –0.46 W m–2 and a standard deviation of 0.20 W m–2. The range in the RF per unit τaer is substantial due to differing representations of aerosol mixing state, optical properties, cloud, surface reflectance, hygroscopic growth, sub-grid scale effects, radiative transfer codes, etc. (Ramaswamy et al., 2001). Myhre et al. (2004b) performed several sensitivity studies and found that the uncertainty was particularly linked to the hygroscopic growth and that differences in the model relative humidity fields could cause differences of up to 60% in the RF. The RFs from the models M to U participating in the AeroCom project are slightly weaker than those obtained from the other studies, with a mean of approximately –0.35 W m–2 and a standard deviation of 0.15 W m–2; the standard deviation is reduced for the AeroCom models owing to constraints on aerosol emissions, based on updated emission inventories (see Table 2.4). Including the uncertainty in the emissions reported in Haywood and Boucher (2000) increases the standard deviation to 0.2 W m–2. As sulphate aerosol is almost entirely scattering, the surface forcing will be similar or marginally stronger than the RF diagnosed at the TOA. The uncertainty in the RF estimate relative to the mean value remains relatively large compared to the situation for LLGHGs.

The mean and median of the sulphate direct RF from grouping all these studies together are identical at –0.41 W m–2. Disregarding the strongest and weakest direct RF estimates to approximate the 90% confidence interval leads to an estimate of –0.4 ± 0.2 W m–2.

Table 2.4. The direct radiative forcing for sulphate aerosol derived from models published since the TAR and from the AeroCom simulations where different models used identical emissions. Load and aerosol optical depth (τaer ) refer to the anthropogenic sulphate; τaer ant is the fraction of anthropogenic sulphate to total sulphate τaer for present day, NRFM is the normalised RF by mass, and NRF is the normalised RF per unit τaer .

No  Modela  LOAD (mg(SO4) m–2)  τaer (0.55 µm)  τaer ant (%)  RF (W m–2)  NRFM (W g–1)  NRF (W m–2 τaer–1)  Reference 
Published since IPCC, 2001  
A  CCM3  2.23      –0.56  –251    (Kiehl et al., 2000) 
B  GEOSCHEM  1.53  0.018    –0.33  –216  –18  (Martin et al., 2004) 
C  GISS  3.30  0.022    –0.65  –206  –32  (Koch, 2001) 
D  GISS  3.27      –0.96  –293    (Adams et al., 2001) 
E  GISS  2.12      –0.57  –269    (Liao and Seinfeld, 2005) 
F  SPRINTARS  1.55  0.015  72  –0.21  –135  –8  (Takemura et al., 2005) 
G  LMD  2.76      –0.42  –152    (Boucher and Pham., 2002) 
H  LOA  3.03  0.030    –0.41  –135  –14  (Reddy et al., 2005b) 
I  GATORG  3.06      –0.32  –105    (Jacobson, 2001a) 
J  PNNL  5.50  0.042    –0.44  –80  –10  (Ghan et al., 2001) 
K  UIO_CTM  1.79  0.019    –0.37  –207  –19  (Myhre et al., 2004b) 
L  UIO_GCM  2.28      –0.29  –127    (Kirkevag and Iversen, 2002) 
AeroCom: identical emissions used for year 1750 and 2000  
M  UMI  2.64  0.020  58  –0.58  –220  –28  (Liu and Penner, 2002) 
N  UIO_CTM  1.70  0.019  57  –0.35  –208  –19  (Myhre et al., 2003) 
O  LOA  3.64  0.035  64  –0.49  –136  –14  (Reddy and Boucher, 2004) 
P  LSCE  3.01  0.023  59  –0.42  –138  –18  (Schulz et al., 2006) 
Q  ECHAM5-HAM  2.47  0.016  60  –0.46  –186  –29  (Stier et al., 2005) 
R  GISS  1.34  0.006  41  –0.19  –139  –31  (Koch, 2001) 
S  UIO_GCM  1.72  0.012  59  –0.25  –145  –21  (Iversen and Seland, 2002; Kirkevag and Iversen, 2002) 
T  SPRINTARS  1.19  0.013  59  –0.16  –137  –13  (Takemura et al., 2005) 
U  ULAQ  1.62  0.020  42  –0.22  –136  –11  (Pitari et al., 2002) 
Average A to L  2.80  0.024    –0.46  –176  –17    
Average M to U  2.15  0.018  55  –0.35  –161  –20    
Minimum A to U  1.19  0.006  41  –0.96  –293  –32    
Maximum A to U  5.50  0.042  72  –0.16  –72  –8    
Std. dev. A to L  1.18  0.010    0.20  75  9    
Std. dev. M to U  0.83  0.008  8  0.15  34  7      


a CCM3: Community Climate Model; GEOSCHEM: Goddard Earth Observing System-Chemistry; GISS: Goddard Institute for Space Studies; SPRINTARS: Spectral Radiation-Transport Model for Aerosol Species; LMD: Laboratoire de Météorologie Dynamique; LOA: Laboratoire d’Optique Atmospherique; GATORG: Gas, Aerosol, Transport, Radiation, and General circulation model; PNNL: Pacific Northwest National Laboratory; UIO_CTM: University of Oslo CTM; UIO_GCM: University of Oslo GCM; UMI: University of Michigan; LSCE: Laboratoire des Sciences du Climat et de l’Environnement; ECHAM5-HAM: European Centre Hamburg with Hamburg Aerosol Module; ULAQ: University of L’Aquila.