5.5.2.4 Projection of sea salt emissions in 2100

The production of sea salt aerosol is also a strong function of wind speed. The semi-empirical formulation of Monahan et al. (1986) was used to produce global monthly sea salt fluxes for eight size intervals (dry diameter of 0.06 to 16 µm) using procedures discussed in Gong et al. (1997a,b). In order to project sea salt emissions for the workshop, the ratio of daily average wind speed in the ten years prior to 2100 and 2000 was used to scale the 2000 daily average sea salt flux to the one in 2100. Because this method may overestimate emissions if the product of the daily average ratio with the daily average winds in 2000 produces high wind speeds with a high frequency, the calculation was checked using the ratio of the monthly average wind speeds. This produced a total sea salt flux that was 13% smaller than the projections given in the workshop specifications.

Predicted sea salt emissions were 3,340 Tg in 2000 and increased to 5,880 Tg in 2100. These increases point to a potentially important negative climate feedback. For example, the present day direct radiative impact of sea salt is estimated to be between -0.75 and -2.5 Wm^–2 using the clear-sky estimates from Haywood et al. (1999) or -0.34 Wm^–2 using the whole-sky estimates from Jacobson (2001). Assuming the ratio of whole-sky to clear-sky forcing from Jacobson (2000b), we project that these changes in sea salt emissions might lead to a radiative feedback in 2100 of up to -0.8 Wm^–2. If we assume that the near-doubling of the sea salt mass flux would result in a proportional increase in the number flux, a significant increase in reflected radiation may also result from the indirect effect. The LLNL/Umich model was used to evaluate the possible impact of these emissions. It was found that the changes in natural emissions in 2100 result in a radiative feedback of -1.16 Wm^–2 in the 2100 A2 scenario. These projected climate changes rely on the projected wind speed changes from the NCAR CSM model. Because projections from other models may not be as large, we also calculated the projected sea salt emissions from three other climate models using the ratio of monthly average winds for the time period from 2090 to 2100 to that for the time period 1990 to 2000 to scale the 2000 sea salt fluxes. Compared to the projections from the NCAR CSM, these models resulted in annual average fluxes that were 37% higher (GFDL model), 13% smaller (Max Planck model), and 9% smaller (Hadley Centre UK Met Office model). The monthly average temperature change associated with these wind speed and sea salt projections was 2.8°K (GFDL model; Knutson et al. 1999), 2.8°K (Max Planck model; Roeckner et al. 1999) and 2.15°K (Hadley Centre UK Met Office model; Gordon et al., 2000) compared to the temperature projection of 1.8°K from the NCAR CSM model (Dai et al., 2001).

In order to project future aerosol concentrations, we formed the average burdens from the models that gave reasonable agreement with observations for the 2000 scenario. As noted above, the differences in the burdens calculated by the different models point to substantial uncertainties in the prediction of current burdens and these translate into similar uncertainties in projecting future burdens. Except for sulphate and black carbon, however, the future projected global average burdens scaled approximately linearly with emissions. Thus, we may assume that the projected uncertainty in future burdens is mainly determined by the uncertainty in the emissions themselves together with the uncertainty in the burdens associated with different model treatments. For SO₄^2-, future anthropogenic concentrations were not linear in the emissions. For example, some models projected increases in burden relative to emissions while others projected decreases. The range of projected changes in anthropogenic burden relative to emissions was -14% to +25% depending on the scenario. Use of the average of the models for the projection of anthropogenic SO₄^2- and total SO₄^2- may therefore bias the results somewhat, but the uncertainties in the projected SO₄^2- concentrations are smaller than those introduced from the range of estimates for the 2000 scenario itself. Table 5.14 gives our projected average burdens for each draft SRES scenario. Results for the burdens associated with the final SRES scenarios are reproduced in Appendix II and are shown in Figure 5.13 (see Chapter 9 and Nakic´enovic´ et al. (2000) for scenario definitions).

Table 5.14: Projected future aerosol burden for draft SRES scenarios. The range predicted from the models that participated in the workshop are given for 2000.
	2000 (SC1)	A2 2030 (SC2)	A2 2100 (SC3)	B1 2100 (SC4)	A2 2100 with natural aerosols (SC5)	B1 2100 with natural aerosols (SC8)
Sulphate Natural (TgS)	0.26 0.15 - 0.36	0.26	0.26	0.26	0.28	0.28
Sulphate Anthr. (TgS)	0.52 0.35 - 0.75	0.90	0.55	0.28	0.54	0.26
Nitrate Natural (TgN)	0.02		0.02
Nitrate Anthr. (TgN)	0.07		0.38
Ammonium Nat. (TgN)	0.09		0.09
Ammonium Anthr. (TgN)	0.33		0.72
BC (Tg)	0.26 0.22 - 0.32	0.33	0.61	0.25	0.61	0.25
OC Natural (Tg)	0.15 0.08 - 0.28	0.15	0.15	0.15	0.22	0.22
OC Anthr. (Tg)	1.52 1.05 - 2.21	1.95	2.30	0.90	2.30	0.90
Dust (D<2 µm) (Tg)	12.98 6.24 - 17.73	13.52	13.52	13.52	13.54	13.54
Dust (D>2 µm) (Tg)	19.58 7.39 - 33.15	19.58	19.58	19.58	20.91	20.91
Sea salt (D<2 µm) (Tg-Na)	2.74 1.29 - 7.81	2.74	2.74	2.74	4.77	4.77
Sea salt (D>2 µm) (Tg-Na)	3.86 1.22 - 6.51	3.86	3.86	3.86	6.68	6.68

Continues on next page

IPCC Homepage