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

Mineral dust from anthropogenic sources originates mainly from agricultural practices (harvesting, ploughing, overgrazing), changes in surface water (e.g., Caspian and Aral Sea, Owens Lake) and industrial practices (e.g., cement production, transport) (Prospero et al., 2002). The TAR reported that the RF due to anthropogenic mineral dust lies in the range of +0.4 to –0.6 W m–2, and did not assign a best estimate because of the difficulties in determining the anthropogenic contribution to the total dust loading and the uncertainties in the optical properties of dust and in evaluating the competing shortwave and longwave radiative effects. For the sign and magnitude of the mineral dust RF, the most important factor for the shortwave RF is the single scattering albedo whereas the longwave RF is dependent on the vertical profile of the dust.

Tegen and Fung (1995) estimated the anthropogenic contribution to mineral dust to be 30 to 50% of the total dust burden in the atmosphere. Tegen et al. (2004) provided an updated, alternative estimate by comparing observations of visibility, as a proxy for dust events, from over 2,000 surface stations with model results, and suggested that only 5 to 7% of mineral dust comes from anthropogenic agricultural sources. Yoshioka et al. (2005) suggested that a model simulation best reproduces the North African TOMS aerosol index observations when the cultivation source in the Sahel region contributes 0 to 15% to the total dust emissions in North Africa. A 35-year dust record established from Barbados surface dust and satellite observations from TOMS and the European geostationary meteorological satellite (Meteosat) show the importance of climate control and Sahel drought for interannual and decadal dust variability, with no overall trend yet documented (Chiapello et al., 2005). As further detailed in Section 7.3, climate change and CO2 variations on various time scales can change vegetation cover in semi-arid regions. Such processes dominate over land use changes as defined above, which would give rise to anthropogenic dust emissions (Mahowald and Luo, 2003; Moulin and Chiapello, 2004; Tegen et al., 2004). A best guess of 0 to 20% anthropogenic dust burden from these works is used here, but it is acknowledged that a very large uncertainty remains because the methods used cannot exclude either a reduction of 24% in present-day dust nor a large anthropogenic contribution of up to 50% (Mahowald and Luo, 2003; Mahowald et al., 2004; Tegen et al., 2005). The RF efficiency of anthropogenic dust has not been well differentiated from that of natural dust and it is assumed that they are equal. The RF due to dust emission changes induced by circulation changes between 1750 and the present are difficult to quantify and not included here (see also Section 7.5).

In situ measurements of the optical properties of local Saharan dust (e.g., Haywood et al., 2003c; Tanré et al., 2003), transported Saharan mineral dust (e.g., Kaufman et al., 2001; Moulin et al., 2001; Coen et al., 2004) and Asian mineral dust (Huebert et al., 2003; Clarke et al., 2004; Shi et al., 2005; Mikami et al., 2006) reveal that dust is considerably less absorbing in the solar spectrum than suggested by previous dust models such as that of WMO (1986). These new, spectral, simultaneous remote and in situ observations suggest that the single scattering albedo (ωo) of pure dust at a wavelength of 0.67 µm is predominantly in the range 0.90 to 0.99, with a central global estimate of 0.96. This is in accordance with the bottom-up modelling of ωo based on the haematite content in desert dust sources (Claquin et al., 1999; Shi et al., 2005). Analyses of ωo from long-term AERONET sites influenced by Saharan dust suggest an average ωo of 0.95 at 0.67 µm (Dubovik et al., 2002), while unpolluted Asian dust during the Aeolian Dust Experiment on Climate (ADEC) had an average ωo of 0.93 at 0.67 µm (Mikami et al., 2006 and references therein). These high ωo values suggest that a positive RF by dust in the solar region of the spectrum is unlikely. However, absorption by particles from source regions with variable mineralogical distributions is generally not represented by global models.

Measurements of the DRE of mineral dust over ocean regions, where natural and anthropogenic contributions are indistinguishably mixed, suggest that the local DRE may be extremely strong: Haywood et al. (2003b) made aircraft-based measurements of the local instantaneous shortwave DRE of as strong as –130 W m–2 off the coast of West Africa. Hsu et al. (2000) used Earth Radiation Budget Experiment (ERBE) and TOMS data to determine a peak monthly mean shortwave DRE of around –45 W m–2 for July 1985. Interferometer measurements from aircraft and the surface have now measured the spectral signature of mineral dust for a number of cases (e.g., Highwood et al., 2003) indicating an absorption peak in the centre of the 8 to 13 µm atmospheric window. Hsu et al. (2000) determined a longwave DRE over land areas of North Africa of up to +25 W m–2 for July 1985; similar results were presented by Haywood et al. (2005) who determined a peak longwave DRE of up to +50 W m–2 at the top of the atmosphere for July 2003.

Recent model simulations report the total anthropogenic and natural dust DRE, its components and the net effect as follows (shortwave / longwave = net TOA, in W m–2): H. Liao et al. (2004): –0.21 / +0.31 = +0.1; Reddy et al. (2005a): –0.28 / +0.14 = –0.14; Jacobson (2001a): –0.20 / +0.07 = –0.13; reference case and [range] of sensitivity experiments in Myhre and Stordal (2001a, except case 6 and 7): –0.53 [–1.4 to +0.2] / +0.13 [+0.0 to +0.8] = –0.4 [–1.4 to +1.0]; and from AeroCom database models, GISS: –0.75 / (+0.19) = (–0.56); UIO-CTM*: –0.56 / (+0.19) = (–0.37); LSCE*: –0.6 / +0.3 = –0.3; UMI*: –0.54 / (+0.19) = (–0.35). (See Table 2.4, Note (a) for model descriptions.) The (*) star marked models use a single scattering albedo (approximately 0.96 at 0.67 µm) that is more representative of recent measurements and show more negative shortwave effects. A mean longwave DRE of 0.19 W m–2 is assumed for GISS, UMI and UIO-CTM. The scatter of dust DRE estimates reflects the fact that dust burden and τaer vary by ±40 and ±44%, respectively, computed as standard deviation from 16 AeroCom A model simulations (Textor et al., 2006; Kinne et al., 2006). Dust emissions from different studies range between 1,000 and 2,150 Tg yr–1 (Zender, 2004). Finally, a major effect of dust may be in reducing the burden of anthropogenic species at sub-micron sizes and reducing their residence time (Bauer and Koch, 2005; see Section

The range of the reported dust net DRE (–0.56 to +0.1 W m–2), the revised anthropogenic contribution to dust DRE of 0 to 20% and the revised absorption properties of dust support a small negative value for the anthropogenic direct RF for dust of –0.1 W m–2. The 90% confidence level is estimated to be ±0.2 W m–2, reflecting the uncertainty in total dust emissions and burdens and the range of possible anthropogenic dust fractions. At the limits of this uncertainty range, anthropogenic dust RF is as negative as –0.3 W m–2 and as positive as +0.1 W m–2. This range includes all dust DREs reported above, assuming a maximum 20% anthropogenic dust fraction, except the most positive DRE from Myhre and Stordal (2001a).