Another complication comes from the way in which different chemical species are mixed in aerosols (e.g., Li and Okada, 1999). Radiative properties can change depending on whether different chemicals are in the same particles (internal mixtures) or different particles (external mixtures). Also, combining species may produce different aerosol size distributions than would be the case if the species were assumed to act independently. One example is the interaction of sulphate with sea salt or dust discussed in Section 5.2.2.6.

Fortunately, studies of the effects of mixing different refractive indices have yielded a fairly straightforward message: the type of mixing is usually significant only for absorbing material (Tang, 1996; Abdou et al., 1997; Fassi-Fihri et al., 1997). For non-absorbing aerosols, an average refractive index appropriate to the chemical composition at a given place and time is adequate. On the other hand, black carbon can absorb up to twice as much light when present as inclusions in scattering particles such as ammonium sulphate compared with separate particles (Ackerman and Toon, 1981; Horvath, 1993; Fuller et al., 1999). Models of present day aerosols often implicitly include this effect by using empirically determined light absorption coefficients but future efforts will need to explicitly consider how black carbon is mixed with other aerosols. Uncertainties in the way absorbing aerosols are mixed may introduce a range of a factor of two in the magnitude of forcing by black carbon (Haywood and Shine, 1995; Jacobson, 2000).

To assess uncertainties associated with the basic aerosol parameters, a compilation is given in Table 5.1, stratified by a crude geographic/aerosol type differentiation. The values for the size distribution parameters given in the table were derived from the references to the table. The mass scattering efficiency and upscatter fraction shown in the table are derived from Mie calculations for spherical particles using these size distributions and a constant index of refraction for the accumulation mode. The scattering efficiency dependence on relative humidity (RH) and the single-scattering albedo were derived from the literature review of measurements.

Table 5.1: Variation in dry aerosol optical properties at 550 nm by region/type.
Aerosol type	Dg (mode) (µ m)	Geometric standard dev. sg	asp (m 2 g -1)	Asymmetry parameter g		f (RH) (at RH = 80%)	fb(RH) (at RH = 80%)	Single scattering albedo wo (dry)	Single- cattering albedo wo (80%)
Pacific marine
w/single mode	0.19 ±0.03	1.5 ±0.15	3.7 ±1.1	0.616 ±0.11	0.23 ±0.05	2.2 ±0.3		0.99 ±0.01	1.0 ±0.01
w/accum & coarse	(0.46)	(2.1)	1.8 ±0.5	0.661 ±0.01	0.21 ±0.003	2.2 ±0.3		0.99 ±0.01	1.0 ±0.01
Atlantic marine	0.15 ±0.05	1.9 ±0.6	3.8 ±1.0	0.664 ±0.25	0.21 ±0.11	2.2 ±0.3		0.97 ±0.03	1.0 ±0.03
Fine soil dust	0.10 ±0.26	2.8 ±1.2	1.8 ±0.1	0.682 ±0.16	0.20 ±0.07	1.3 ±0.2		0.82 ±0.06	0.83 ±0.06
Polluted continental	0.10 ±0.08	1.9 ±0.3	3.5 ±1.2	0.638 ±0.28	0.22 ±0.12	2.0 ±0.3	0.81 ±0.08	0.92 ±0.05	0.95 ±0.05
Background
Continental
w/single mode	0.08 ±0.03	1.75 ±0.34	2.2 ±0.9	0.537 ±0.26	0.27 ±0.11	2.3 ±0.4		0.97 ±0.03	1.0 ±0.03
w/accum. & coarse	(1.02)	(2.2)	1.0 ±0.08	0.664 ±0.07	0.21 ±0.03	2.3 ±0.4		0.97 ±0.03	1.0 ±0.03
Free troposphere	0.072 ±0.03	2.2 ±0.7	3.4 ±0.7	0.649 ±0.22	0.22 ±0.10	-		-
Biomass plumes	0.13 ±0.02	1.75 ±0.25	3.6 ±1.0	0.628 ±0.12	0.23 ±0.05	1.1 ±0.1		0.87 ±0.06	0.87 ±0.06
Biomass regional haze	0.16 ±0.04	1.65 ±0.25	3.6 ±1.1	0.631 ±0.16	0.22 ±0.07	1.2 ±0.2	0.81 ±0.07	0.89 ±0.05	0.90 ±0.05
Literature references: Anderson et al. (1996); Bodhaine (1995); Carrico et al. (1998, 2000); Charlson et al. (1984); Clarke et al. (1999); Collins et al. (2000); Covert et al. (1996); Eccleston et al. (1974); Eck et al. (1999); Einfeld et al. (1991); Fitzgerald (1991); Fitzgerald and Hoppel (1982), Frick and Hoppel (1993); Gasso et al. (1999); Hegg et al. (1993, 1996a,b); Hobbs et al. (1997); Jaenicke (1993); Jennings et al. (1991); Kaufman (1987); Kotchenruther and Hobbs (1998); Kotchenruther et al. (1999); Leaitch and Isaac (1991); Le Canut et al. (1992, 1996); Lippman (1980); McInnes et al. (1997; 1998); Meszaros (1981); Nyeki et al. (1998); O’Dowd and Smith (1993); Quinn and Coffman (1998); Quinn et al. (1990; 1993, 1996); Radke et al. (1991); Raes et al. (1997); Reid and Hobbs (1998); Reid et al. (1998b); Remer et al. (1997); Saxena et al. (1995); Seinfeld and Pandis (1998); Sokolik and Golitsyn (1993); Takeda et al. (1987); Tang (1996); Tangren (1982); Torres et al. (1998); Waggoner et al. (1983); Whitby (1978); Zhang et al. (1993).

The uncertainties given in the table for the central values of number modal diameter and geometric standard deviation (D_g and s_g) are based on the ranges of values surveyed in the literature, as are those for f(RH) and w_o. Those for the derivative quantities a_sp and b, however, are based on Mie calculations using the upper and lower uncertainty limits for the central values of the size parameters, i.e., the propagation of errors is based on the functional relationships of Mie theory. The two calculations with coarse modes require some explanation.

While the accumulation mode is generally thought to dominate light scattering, recent studies – as discussed below – have suggested that sea salt and dust can play a large role under certain conditions. To include this possibility, a sea salt mode has been added to the Pacific marine accumulation mode. The salt mode extends well into the accumulation size range and is consistent with O’Dowd and Smith (1993). It is optically very important at wind speeds above 7 to 10 ms^-1. For the case shown in the table, the sea salt mode accounts for about 50% of the local light scattering and could contribute over a third of the column optical depth, depending on assumptions about the scale height of the salt. Similarly, a soil dust coarse mode based on work by Whitby (1978) was added to the continental background accumulation mode. When present, this mode often dominates light scattering but, except for regions dominated by frequent dust outbreaks, is usually present over so small a vertical depth that its contribution to the column optical depth is generally slight. The importance of these coarse modes points to the importance of using size-resolved salt and dust fluxes such as those given in this report.

IPCC Homepage