In response to continued greenhouse gas (GHG) increases, the Atlantic Meridional Overturning Circulation (AMOC) is expected to weaken through the 21st century. However, AMOC impacts associated with efforts to improve air quality are less well understood. Here, eight models from the Regional Aerosol Model Intercomparison Project are examined to quantify mid-21st century AMOC changes resulting from global and regional anthropogenic aerosol and precursor gas (AA) emissions reductions (industrial and biomass burning), by comparing strong air pollution control shared socioeconomic pathway (SSP1-2.6) to a baseline with weak air pollution control (SSP3-7.0). Global AA reductions and subsequent warming yield multi-model mean AMOC weakening of 6% ( (Formula presented) (Formula presented) Sv; 1 Sv = 106 m3 s−1) by the last 12 years of the simulation (2039–2050). This is ⅓ of the magnitude of the corresponding weakening associated with the high GHG emissions scenario SSP3-7.0. Of the regional perturbations, combined North American and European AA reductions drive the largest AMOC weakening, followed by combined African and Middle Eastern reductions and then East Asian reductions, with South Asian reductions yielding non-significant weakening. Across these experiments, AMOC weakening is significantly correlated with the North Atlantic Ocean aerosol effective radiative forcing ( (Formula presented) (Formula presented) ) and aerosol optical depth response ( (Formula presented) (Formula presented) ). AMOC weakening under AA reductions is associated with a thermally driven reduction in buoyancy in the subpolar North Atlantic, which is largely driven by surface shortwave radiation increases, consistent with the forcing from AA reductions. Africa + Middle East AA reductions also involve excitation of a negative North Atlantic Oscillation pattern, which contributes to AMOC weakening. Our results show that efforts to improve air quality, particularly around the Atlantic basin but also far away in East Asia, will contribute to future AMOC weakening.

The Regional Aerosol Model Intercomparison Project (RAMIP) is designed to quantify the forcing and climate impacts of mid-21st century anthropogenic aerosol and precursor gas (AA) emissions reductions (both industrial and biomass burning), by comparing a weak (SSP3-7.0) versus strong (SSP1-2.6) level of air quality control aerosol emissions pathway. AA emissions reductions experiments include global (GLO), East Asia (EAS), South Asia, Africa and the Middle East (AFR), and North America and Europe (NAE). Here, we use RAMIP time-slice simulations with fixed sea surface temperatures and sea-ice distributions from nine models to quantify the aerosol effective radiative forcing (ERF), including aerosol radiation (ERF_ari) and aerosol cloud interactions (ERF_aci). The multi-model global mean net ERF_ari+aci is 0.77 ± 0.25 W m⁻² for GLO, and three of the four regional perturbations yield a significant positive net ERF_ari+aci (up to 0.15 ± 0.07 W m⁻² for EAS). In all cases, net ERF_ari+aci is dominated by aerosol-cloud interactions, which are largely due to reduced cloud scattering. Of the four regions, NAE yields the largest forcing efficiency whereas AFR yields the weakest. Although the areas outside our four target regions contribute 25% to the GLO aerosol optical depth reduction, they disproportionately contribute 44% to the GLO net ERF_ari+aci. The multimodel regional mean net ERF_ari+aci for three regional perturbations is much larger (up to  1.64 ± 1.36 W m⁻² for EAS) than the corresponding global mean value. However, these regional values are even larger (up to 2.69 ± 1.72 W m⁻² for EAS) under global aerosol reductions, implying remote emission reductions represent a sizable contribution (up to 1.05 ± 0.56 W m⁻² for EAS). These large regional ERFs will in turn drive relatively large regional climate impacts, which continue to be underappreciated in most policy discussions.

Decreases in anthropogenic aerosols will reduce fine particulate matter (PM2.5); however, meteorological feedbacks alter dust emissions, modifying air quality gains. We use eight Earth System Models from the Regional Aerosol Model Intercomparison Project (RAMIP) simulations to assess African climate and air quality responses to anthropogenic aerosol emission perturbations, including meteorological feedbacks on dust emissions. By 2050, African and global emissions reductions drive the largest continent-average PM2.5 decrease (0.92 ± 0.17 μg m−3; 5% and 1.35 ± 0.50 μg m−3; 7%, respectively) relative to SSP3-7.0, though regional dust increases partially offset these reductions. Anthropogenic emissions reductions in the U.S. and Europe also lower African PM2.5 by 0.29 ± 0.32 μg m−3 (2%) due to teleconnections of Northern Hemisphere warming influencing the Intertropical Convergence Zone. Inter-model variability in dust and total PM2.5 reflects differences in meteorological responses and dust emission parameterizations. Meteorological responses explain 90% of dust emissions variability across regions. Aerosol-driven climate feedbacks on dust account for up to 70% of total PM2.5 changes in the Sahara and Namib, offsetting up to 20% of anthropogenic PM2.5 reductions across Africa. Under 2050 global and Africa-wide anthropogenic aerosol reductions, 96,000 (95% CI: 54,000–137,000) and 84,000 (95% CI: 43,000–125,000) PM2.5-related deaths are avoided in Africa, respectively. Dust PM2.5 contributes an uncertain 3.4% of the avoided deaths under global reductions and has no net effect under Africa-wide reductions. Aerosol-driven climate feedbacks may partially offset direct air quality gains, though their continental-scale contribution remains small and uncertain.

Aerosol particles from both natural and anthropogenic sources play a critical role in the Earth's climate by interacting with solar radiation and clouds. Anthropogenic aerosol and precursor emissions have historically exerted a global cooling effect, which has partially offset the warming from concurrent greenhouse gas emissions. Recent reductions and shifts in aerosol and precursor emission patterns may reduce this offset and introduce spatially and temporarily varying climate impacts. Investigating aerosol-climate effects is typically done with computationally expensive Earth System Models, which include complex representations of physical, chemical, biological, and geological processes and their coupled interactions for the entire global climate system. In this study, we develop a machine-learning climate emulator using Gaussian processes, called AeroGP, that can be used to quickly assess, for example, the impact of different policy decisions on future climate mitigation strategies. The emulator is trained on a unique data set from the Norwegian Earth System Model (NorESM), analyzed as an ensemble here for the first time. AeroGP accounts for the joint spatial covariance of the output variables and captures the complex, heterogeneous impacts of aerosols on surface temperature using coregionalization. We believe this is the first time this method has been used to account for the spatial correlation of such climate data. We show that AeroGP retains the spatial complexity of NorESM at a fraction of the computational cost and demonstrate its usefulness to assess the sensitivity of temperature to idealized future aerosol emission scenarios.

Global surface warming has accelerated since around 2010, relative to the preceding half century. This has coincided with East Asian efforts to reduce air pollution through restricted atmospheric aerosol and precursor emissions. A direct link between the two has, however, not yet been established. Here we show, using a large set of simulations from eight Earth System Models, how a time-evolving 75% reduction in East Asian sulfate emissions partially unmasks greenhouse gas-driven warming and influences the spatial pattern of surface temperature change. We find a rapidly evolving global, annual mean warming of 0.07 ± 0.05 °C, sufficient to be a main driver of the uptick in global warming rate since 2010. We also find North-Pacific warming and a top-of-atmosphere radiative imbalance that are qualitatively consistent with recent observations. East Asian aerosol cleanup is thus likely a key contributor to recent global warming acceleration and to Pacific warming trends.

Ice formation remains one of the most poorly represented microphysical processes in climate models. While primary ice production (PIP) parameterizations are known to have a large influence on the modeled cloud properties, the representation of secondary ice production (SIP) is incomplete and its corresponding impact is therefore largely unquantified. Furthermore, ice aggregation is another important process for the total cloud ice budget, which also remains largely unconstrained. In this study, we examine the impact of PIP, SIP, and ice aggregation on Arctic clouds, using the Norwegian Earth System Model, version 2 (NorESM2). Simulations with both prognostic and diagnostic PIP show that heterogeneous freezing alone cannot reproduce the observed cloud ice content. The implementation of missing SIP mechanisms (collisional breakup, drop shattering, and sublimation breakup) in NorESM2 improves the modeled ice properties, while improvements in liquid content occur only in simulations with prognostic PIP. However, results are sensitive to the description of collisional breakup. This mechanism, which dominates SIP in the examined conditions, is very sensitive to the treatment of the sublimation correction factor, a poorly constrained parameter that is included in the utilized parameterization. Finally, variations in ice aggregation treatment can also significantly impact cloud properties, mainly through their impact on collisional breakup efficiency. Overall, enhancement in ice production through the addition of SIP mechanisms and the reduction in ice aggregation (in line with radar observations of shallow Arctic clouds) result in enhanced cloud cover and decreased TOA radiation biases, compared to satellite measurements, especially during the cold months.

The Arctic climate system is host to many processes which interact vertically over the tightly coupled atmosphere, sea ice and ocean. The coupled Atmosphere-Ocean Single-Column Model (AOSCM) allows to decouple local small-scale and large-scale processes to investigate the model performance in an idealized setting. Here, an observed Arctic warm air intrusion event is used to show how to identify model deficiencies using the AOSCM. The AOSCM allows us to effectively produce a large number of perturbation simulations, around 1,000, to map sensitivities of the model results due to changes in physical and model properties as well as to the large-scale tendencies. The analysis of the summary diagnostics, that is, aggregated results from sensitivity experiments evaluated against modeled physical properties, such as surface energy budget and mean sea ice thickness, reveals sensitivities to the chosen parameters. Further, we discuss how the conclusions can be used to understand the behavior of the global host model. The simulations confirm that the horizontal advection of heat and moisture plays an important role for maintaining a low-level cloud cover, as in earlier studies. The combined cloud layers increase the energy input to the surface, which in turn enhances the ongoing melt. The clouds present an additional sensitivity in terms of how they are represented but also their interaction with the large-scale advection and the model time step. The methodology can be used for a variety of other regions, where the coupling to the ocean is important.

The effective radiative forcing, which includes the instantaneous forcing plus adjustments from the atmosphere and surface, has emerged as the key metric of evaluating human and natural influence on the climate. We evaluate effective radiative forcing and adjustments in 17 contemporary climate models that are participating in the Coupled Model Intercomparison Project (CMIP6) and have contributed to the Radiative Forcing Model Intercomparison Project (RFMIP). Present-day (2014) global-mean anthropogenic forcing relative to pre-industrial (1850) levels from climate models stands at 2.00 (+/- 0.23) W m(-2), comprised of 1.81 (+/- 0.09) Wm(-2) from CO2, 1.08 (+/- 0.21) Wm(-2) from other well-mixed greenhouse gases, -1.01 (+/- 0.23) W m(-2) from aerosols and -0.09 (+/- 0.13) W m(-2) from land use change. Quoted uncertainties are 1 standard deviation across model best estimates, and 90 % confidence in the reported forcings, due to internal variability, is typically within 0.1 W m(-2). The majority of the remaining 0.21 W m(-2) is likely to be from ozone. In most cases, the largest contributors to the spread in effective radiative forcing (ERF) is from the instantaneous radiative forcing (IRF) and from cloud responses, particularly aerosol-cloud interactions to aerosol forcing. As determined in previous studies, cancellation of tropospheric and surface adjustments means that the stratospherically adjusted radiative forcing is approximately equal to ERF for greenhouse gas forcing but not for aerosols, and consequentially, not for the anthropogenic total. The spread of aerosol forcing ranges from -0.63 to -1.37 W m(-2), exhibiting a less negative mean and narrower range compared to 10 CMIP5 models. The spread in 4 x CO2 forcing has also narrowed in CMIP6 compared to 13 CMIP5 models. Aerosol forcing is uncorrelated with climate sensitivity. Therefore, there is no evidence to suggest that the increasing spread in climate sensitivity in CMIP6 models, particularly related to high-sensitivity models, is a consequence of a stronger negative present-day aerosol forcing and little evidence that modelling groups are systematically tuning climate sensitivity or aerosol forcing to recreate observed historical warming.

Aerosol radiative forcing can influence climate both locally and far outside the emission region. Here we investigate black carbon (BC) aerosols emitted in four major emission areas and evaluate the importance of emission location and magnitude as well as the concept of the absolute regional temperature-change potentials (ARTP). We perform simulations with a climate model (NorESM) with a fully coupled ocean and with fixed sea surface temperatures. BC emissions for year 2000 are increased by factors of 10 and 20 in South Asia, North America, and Europe, respectively, and by 5 and 10 in East Asia (due to higher emissions there). The perturbed simulations and a reference simulation are run for 100 years with three ensemble members each. We find strikingly similar regional surface temperature responses and geographical patterns per unit BC emission in Europe and North America but somewhat lower temperature sensitivities for East Asian emissions. BC emitted in South Asia shows a different geographical pattern in surface temperatures, by changing the Indian monsoon and cooling the surface. We find that the ARTP approach rather accurately reproduces the fully coupled temperature response of NorESM. Choosing the highest emission rate results in lower surface temperature change per emission unit compared to the lowest rate, but the difference is generally not statistically significant except for the Arctic. An advantage of high-perturbation simulations is the clearer emergence of regional signals. Our results show that the linearity of normalized temperature effects of BC is fairly well preserved despite the relatively large perturbations but that regional temperature coefficients calculated from high perturbations may be a conservative estimate. Regardless of emission region, BC causes a northward shift of the ITCZ, and this shift is apparent both with a fully coupled ocean and with fixed sea surface temperatures. For these regional BC emission perturbations, we find that the effective radiative forcing is not a good measure of the climate response. A limitation of this study is the uncertainties in BC-cloud interactions and the amount of BC absorption, both of which are model dependent.

Clean air policies can have significant impacts on climate in remote regions. Previous modeling studies have shown that the temperature response to European sulfate aerosol reductions is largest in the Arctic. Here we investigate the atmospheric and ocean roles in driving this enhanced Arctic warming using a set of fully coupled and slab‐ocean simulations (specified ocean heat convergence fluxes) with the Norwegian Earth system model (NorESM), under scenarios with high and low European aerosol emissions relative to year 2000. We show that atmospheric processes drive most of the Arctic response. The ocean pathway plays a secondary role inducing small temperature changes mostly in the opposite direction of the atmospheric response. Important modulators of the temperature response patterns are changes in sea ice extent and subsequent turbulent heat flux exchange, suggesting that a proper representation of Arctic sea ice and turbulent changes is key to predicting the Arctic response to midlatitude aerosol forcing.