Anthropogenic aerosol forcing under the Shared Socioeconomic Pathways

6 Emissions of anthropogenic aerosols are expected to change drastically over the coming decades, 7 with potentially significant climate implications. Using the most recent generation of harmonized 8 emission scenarios, the Shared Socioeconomic Pathways (SSPs) as input to a global chemistry 9 transport and radiative transfer model, we provide estimates of the projected future global and 10 regional burdens and radiative forcing of anthropogenic aerosols under three contrasting pathways 11 for air pollution levels: SSP1-1.9, SSP2-4.5 and SSP3-7.0. We find that the broader range of future 12 air pollution emission trajectories spanned by the SSPs compared to previous scenarios translates 13 into total aerosol forcing estimates in 2100 relative to 1750 ranging from -0.04 W m-2 in SSP1-1.9 14 to -0.51 W m-2 in SSP3-7.0. Compared to our 1750-2015 estimate of -0.55 W m-2, this shows that 15 depending on the success of air pollution policies and socioeconomic development over the 16 coming decades, aerosol radiative forcing may weaken by nearly 95% or remain close to the pre17 industrial to present-day level. In all three scenarios there is a positive forcing in 2100 relative to 18 2015, from 0.51 W m-2 in SSP1-1.9 to 0.04 W m-2 in SSP3-7.0. Results also demonstrate significant 19 differences across regions and scenarios, especially in South Asia and Africa. While rapid 20 weakening of the negative aerosol forcing following effective air quality policies will unmask 21 more of the greenhouse gas-induced global warming, slow progress on mitigating air pollution 22 will significantly enhance the atmospheric aerosol levels and risk to human health in these regions. 23 In either case, the resulting impacts on regional and global climate can be significant. 24

properties of the clouds, the approach from Quaas et al. (2006) is used. This method has also been 136 applied in earlier studies (Bian et al., 2017).

139
In the following, we first document the future global emissions and abundances of aerosols, 140 according to our three chosen SSP scenarios. We then show the resulting regional aerosol burden 141 levels, and finally global and regional radiative forcing. documented in previous studies as well (Bauer et al., 2007;Bellouin et al., 2011). In SSP1-1.9 and 174 SSP2-4.5, there is negligible net change in NH3 emissions over the century, while NOx emissions 175 decline, resulting in a lower burden also of nitrate. Figure 1i-j shows the simulated anthropogenic 176 global-mean aerosol optical depth (AOD) and absorption aerosol optical depth (AAOD) 177 (calculated as the difference between each year and the 1750 value, with meteorology and hence 178 contribution from natural aerosols constant). The anthropogenic AOD falls from 0.026 in 2015 to 179 0.0005 in 2100 in SSP1-1.9 and 0.006 in SSP2-4.5. These changes correspond to a reduction of 180 the total AOD of 20% (15%) in 2100 in SSP1-1.9 (SSP2-4.5) from the present-day level of 0.13.

181
In SSP3-7.0, the anthropogenic AOD increases by 12% to 2050 before returning approximately to 182 its present-day value. Similar magnitude decreases in anthropogenic AAOD are found. The decline 183 in anthropogenic AOD is stronger than implied by the burden changes. We note that the sulfate 184 and nitrate burdens include also smaller contributions from natural (ocean and vegetation) sources 185 that remain constant to 2100. In SSP1-1.9 we find a small, negative AAOD value in 2100. This 186 results from emissions on BC and OC being lower in 2100 than in 1750. The stronger decline in 187 anthropogenic AAOD relative to AOD in SSP1-1.9 is reflected in the total (anthropogenic and 188 natural aerosols) Single Scattering Albedo (SSA) (Fig. 1k) which increases to above pre-1970s 189 levels by mid-century and is notably higher than in SSP3-7.0 by the end of the century. As the 190 mechanisms that link aerosol emissions to climate impacts are markedly different for scattering 191 and absorbing aerosols (Ocko et al., 2014;Samset et al., 2016;Smith et al., 2018), this reduction 192 highlights a need for regional studies of aerosol impacts that go beyond the total top-of-atmosphere 193 effective radiative forcing.

194
The global-mean time series hide significant spatiotemporal differences in aerosol trends. Methods) for the historical period. We calculate a net aerosol-induced RF in 2015, relative to 1750, 224 of -0.55 W m -2 , whereof -0.14 W m -2 is due to aerosol-radiation interactions, as also shown in 225 Lund et al. (2018), and -0.42 W m -2 due to aerosol-cloud interactions. Due to the rapid emission 226 reductions projected over the next couple of decades, the RF is less than half in magnitude to its 227 present-day value in SSP1-1.9 already by 2030, continuing to weaken at a slower rate after. In

255
The spatiotemporal differences in trend documented above translates into effects on global and 256 regional RF. In Fig. 4 we therefore show the change in RFtotal over four time periods, 1750-2015, 257 1750-1990, 1990-2015and 2015-2030. Figure S3 show the corresponding results   (Li et al., 2017;Zheng et al., 2018). In contrast, emissions of India are 285 projected to increase, at least initially. The potential implications of this feature are discussed in a 286 separate paper (Samset et al., 2019). Weak RF is found over the African continent in the SSP2-4.5 287 and SSP3-7.0 scenarios. However, as shown in Figure 2, aerosols will continue to affect local 288 climate and air quality in this region. begin to decline rapidly also in other high emitting regions, whereas the mainly residential, more 296 challenging, BC sources remain largely unchecked, the aerosol forcing may follow a different path 297 than estimated here. As an illustrative example, we calculate the contribution to RFari in 2020 and 298 2050 (relative to 1750) from individual components under SSP1-1.9 and SSP3-7.0 (Table S1).

299
Taking the sum of the sulfate forcing from SSP1-1.9 and the remaining components from SSP3-300 7.0, the total RFari is -0.018 W m -2 in 2020, i.e., significantly weaker than when all emissions 301 follow SSP1-1.9, and 0.15 W m -2 in 2050. Continuing along the recent emission development of 302 declining SO2 emissions and increasing BC could imply a different development in the total 303 aerosol effect relative to pre-industrial than shown by the three scenarios here, at least towards the 304 mid-century. We emphasize that these numbers are meant to be illustrative and note that significant 305 uncertainties surround the climate impact of both BC and the co-emitted organic aerosols. As noted 306 above, our estimates do not account for the rapid adjustments which might reduce the global 307 surface temperature response to BC perturbations. Additionally, the role of absorption by so-called 308 brown carbon remains an important uncertainty (Samset et al., 2018b). Previous work has also 309 pointed to the possibility of substantial increases in radiative forcing by nitrate over the 21st 310 century (Bauer et al., 2007;Bellouin et al., 2011;Shindell et al., 2013), and a notable increase in 311 nitrate burden is also estimated in the present study when emissions follow SSP3-7.0. This 312 translates into a nitrate RF that is almost a factor 2 stronger in 2100 than in 2020 and constitutes a 313 correspondingly larger fraction of the RFtotal in this scenario (Table S1) 314 We present projected future aerosol RF based on single-model simulations. Aerosols, however, to aerosol perturbations. Under SSP1-1.9, aerosol emissions are projected to decline even more 341 rapidly than in RCP2.6 over the coming couple of decades (Fig. 1). If in fact associated with a 342 rapid warming, this development could further hinder the realization of the already ambitious 343 temperature goals of the Paris agreement and this feature hence needs to be better quantified.

344
Previous work also demonstrate effects of falling aerosol emissions also other climate variables 345 such as mean and extreme precipitation (Navarro et al., 2017;Pendergrass et al., 2015) and 346 atmospheric dynamics (Rotstayn et al., 2014). The numerous and significant impacts of aerosols 347 underline the need to encompass the full range of projected emissions, regionally and globally, in 348 future assessment, in particular in light of the crucial role of aerosols in shaping regional climate, 349 regional assessments are needed to capture the impact of different trends.

350
It is well-established that future changes in aerosols will critically affect local air quality. Silva et  Anthropogenic AOD and AAOD, and total (anthropogenic and natural) SSA at 550nm. 718  1750-1990, 1990-2015, 1750-2015, and 2015-2030 for 734 each of the three SSP scenarios considered here. 735