This study provides comprehensive insight into the notable differences in clouds and precipitation simulated by the Energy Exascale Earth System Model Atmosphere Model version 0 and version 1 ...(EAMv1). Several sensitivity experiments are conducted to isolate the impact of changes in model physics, resolution, and parameter choices on these differences. The overall improvement in EAMv1 clouds and precipitation is primarily attributed to the introduction of a simplified third‐order turbulence parameterization Cloud Layers Unified By Binormals (along with the companion changes) for a unified treatment of boundary layer turbulence, shallow convection, and cloud macrophysics, though it also leads to a reduction in subtropical coastal stratocumulus clouds. This lack of stratocumulus clouds is considerably improved by increasing vertical resolution from 30 to 72 layers, but the gain is unfortunately subsequently offset by other retuning to reach the top‐of‐atmosphere energy balance. Increasing vertical resolution also results in a considerable underestimation of high clouds over the tropical warm pool, primarily due to the selection for numerical stability of a higher air parcel launch level in the deep convection scheme. Increasing horizontal resolution from 1° to 0.25° without retuning leads to considerable degradation in cloud and precipitation fields, with much weaker tropical and subtropical short‐ and longwave cloud radiative forcing and much stronger precipitation in the intertropical convergence zone, indicating poor scale awareness of the cloud parameterizations. To avoid this degradation, significantly different parameter settings for the low‐resolution (1°) and high‐resolution (0.25°) were required to achieve optimal performance in EAMv1.
Plain Language Summary
The Energy Exascale Earth System Model (E3SM) is a new and ongoing U.S. Department of Energy (DOE) climate modeling effort to develop a high‐resolution Earth system model specifically targeting next‐generation DOE supercomputers to meet the science needs of the nation and the mission needs of DOE. The increase of model resolution along with improvements in representing cloud and convective processes in the E3SM atmosphere model version 1 has led to quite significant model behavior changes from its earlier version, particularly in simulated clouds and precipitation. To understand what causes the model behavior changes, this study conducts sensitivity experiments to isolate the impact of changes in model physics, resolution, and parameter choices on these changes. Results from these sensitivity tests and discussions on the underlying physical processes provide substantial insight into the model errors and guidance for future E3SM development.
Key Points
CLUBB along with the companion changes in EAMv1 primarily account for the overall improvements in clouds and precipitation simulation
Underestimate of coastal Sc in EAMv1 is due to CLUBB and model tuning; increased vertical resolution partially offsets this degradation
The poor scale awareness of EAMv1 requires retuning as resolution increases, which has a large impact on model cloud behavior
The global source–receptor relationships of sulfate concentrations, and direct and indirect radiative forcing (DRF and IRF) from 16 regions/sectors for years 2010–2014 are examined in this study ...through utilizing a sulfur source-tagging capability implemented in the Community Earth System Model (CESM) with winds nudged to reanalysis data. Sulfate concentrations are mostly contributed by local emissions in regions with high emissions, while over regions with relatively low SO2 emissions, the near-surface sulfate concentrations are primarily attributed to non-local sources from long-range transport. Regional source efficiencies of sulfate concentrations are higher over regions with dry atmospheric conditions and less export, suggesting that lifetime of aerosols, together with regional export, is important in determining regional air quality. The simulated global total sulfate DRF is −0.42 W m−2, with −0.31 W m−2 contributed by anthropogenic sulfate and −0.11 W m−2 contributed by natural sulfate, relative to a state with no sulfur emissions. In the Southern Hemisphere tropics, dimethyl sulfide (DMS) contributes 17–84 % to the total DRF. East Asia has the largest contribution of 20–30 % over the Northern Hemisphere mid- and high latitudes. A 20 % perturbation of sulfate and its precursor emissions gives a sulfate incremental IRF of −0.44 W m−2. DMS has the largest contribution, explaining −0.23 W m−2 of the global sulfate incremental IRF. Incremental IRF over regions in the Southern Hemisphere with low background aerosols is more sensitive to emission perturbation than that over the polluted Northern Hemisphere.
Secondary organic aerosols (SOA) are large contributors to fine‐particle loadings and radiative forcing but are often represented crudely in global models. We have implemented three new detailed SOA ...treatments within the Community Atmosphere Model version 5 (CAM5) that allow us to compare the semivolatile versus nonvolatile SOA treatments (based on some of the latest experimental findings) and to investigate the effects of gas‐phase fragmentation reactions. The new treatments also track SOA from biomass burning and biofuel, fossil fuel, and biogenic sources. For semivolatile SOA treatments, fragmentation reactions decrease the simulated annual global SOA burden from 7.5 Tg to 1.8 Tg. For the nonvolatile SOA treatment with fragmentation, the burden is 3.1 Tg. Larger differences between nonvolatile and semivolatile SOA (up to a factor of 5) exist in areas of continental outflow over the oceans. According to comparisons with observations from global surface Aerosol Mass Spectrometer measurements and the U.S. Interagency Monitoring of Protected Visual Environments (IMPROVE) network measurements, the FragNVSOA treatment, which treats SOA as nonvolatile and includes gas‐phase fragmentation reactions, agrees best at rural locations. Urban SOA is underpredicted, but this may be due to the coarse model resolution. All three revised treatments show much better agreement with aircraft measurements of organic aerosols (OA) over the North American Arctic and sub‐Arctic in spring and summer, compared to the standard CAM5 formulation. This is mainly due to the oxidation of SOA precursor gases from biomass burning, not included in standard CAM5, and long‐range transport of biomass burning OA at high altitudes. The revised model configurations that include fragmentation (both semivolatile and nonvolatile SOA) show much better agreement with MODerate resolution Imaging Spectrometers (MODIS) aerosol optical depth data over regions dominated by biomass burning during the summer compared to standard CAM5, and predict biomass burning and biofuel as the largest global source of OA, followed by biogenic and fossil fuel sources. The large contribution of biomass burning OA in the revised treatments is supported by these measurements, but the emissions and aging of SOA precursors and POA are uncertain, and need further investigation. The nonvolatile and semivolatile configurations with fragmentation predict the direct radiative forcing of SOA as −0.5 W m−2 and −0.26 W m−2 respectively, at top of the atmosphere, which are higher than previously estimated by most models, but in reasonable agreement with a recent constrained modeling study. This study highlights the importance of improving process‐level representation of SOA in global models.
Key Points
Fragmentation is an important sink of SOA
Biomass burning is the largest SOA source
SOA DRF is higher than most models
The Community Atmosphere Model Version 5 is run at horizontal grid spacing of 2, 1, 0.5, and 0.25°, with the meteorology nudged toward the Year Of Tropical Convection analysis, and cloud simulators ...and the collocated A‐Train satellite observations are used to explore the resolution dependence of aerosol‐cloud interactions. The higher‐resolution model produces results that agree better with observations, showing an increase of susceptibility of cloud droplet size, indicating a stronger first aerosol indirect forcing (AIF), and a decrease of susceptibility of precipitation probability, suggesting a weaker second AIF. The resolution sensitivities of AIF are attributed to those of droplet nucleation and precipitation parameterizations. The annual average AIF in the Northern Hemisphere midlatitudes (where most anthropogenic emissions occur) in the 0.25° model is reduced by about 1 W m−2 (−30%) compared to the 2° model, leading to a 0.26 W m−2 reduction (−15%) in the global annual average AIF.
Key Points
Using simulators and constrained meteorology facilitates fair model evaluation
Spop and AIF decrease, but SRe increases with increasing resolution
Droplet nucleation and precipitation processes exhibit resolution sensitivity
Previous studies have noticed that the Coupled Model Intercomparison Project Phase 6 (CMIP6) models with a stronger cooling from aerosol‐cloud interactions (ACI) also have an enhanced warming from ...positive cloud feedback, and these two opposing effects are counter‐balanced in simulations of the historical period. However, reasons for this anti‐correlation are less explored. In this study, we perturb the cloud ice microphysical processes to obtain cloud liquid of varying amounts in two Earth System Models (ESMs). We find that the model simulations with a larger liquid water path (LWP) tend to have a stronger cooling from ACI and a stronger positive cloud feedback. More liquid clouds in the mean‐state present more opportunities for anthropogenic aerosol perturbations and also weaken the negative cloud feedback at middle to high latitudes. This work, from a cloud state perspective, emphasizes the influence of the mean‐state LWP on effective radiative forcing due to ACI (ERFACI).
Plain Language Summary
Since the preindustrial era, emissions of greenhouse gases (GHGs) and aerosols have both increased substantially. Planetary warming from the elevated GHGs causes changes in cloud distribution and properties, thereby imposing feedbacks on the climate system. At the same time, aerosols from air pollution exert a cooling effect by modifying cloud properties and lifetime. Cloud feedback and aerosol‐cloud interactions (ACI) are two critical factors for understanding the past and projecting the future climate change. In this study, we examine the relationships of ACI and cloud feedback with the mean‐state cloud liquid water amount predicted from simulations with two different Earth system models. We find that both the effective radiative forcing due to aerosol‐cloud interactions (ERFACI) and the cloud feedback are modulated by mean‐state liquid water path (LWP). The warming induced by cloud feedback and cooling by ACI counteracts each other. Our study suggests that the mean‐state LWP is an important factor influencing both ERFACI and the cloud feedback, and is a useful index for the future climate projection.
Key Points
Effective radiative forcing due to aerosol‐cloud interactions strengthens with the increase of mean‐state LWP in two Earth System Models
Cloud radiative feedback increases monotonically with the increase of mean‐state LWP
Mean‐state LWP is a good predictor for both ERFACI and cloud radiative feedback
Wintertime aerosol pollution in the North China Plain has increased over the past several decades as anthropogenic emissions in China have increased, and has dramatically escalated since the ...beginning of the 21
century, but the causes and their quantitative attributions remain unclear. Here we use an aerosol source tagging capability implemented in a global aerosol-climate model to assess long-term trends of PM
(particulate matter less than 2.5 μm in diameter) in the North China Plain. Our analysis suggests that the impact of China's increasing domestic emissions on PM
concentrations over the last two decades of 20
century was partially offset (13%) by decreasing foreign emission over this period. As foreign emissions stabilized after 2000, their counteracting effect almost disappeared, uncovering the impact of China's increasing domestic emissions that had been partially offset in previous years by reductions in foreign emissions. A slowdown in the impact from foreign emission reductions together with weakening winds explain 25% of the increased PM
trend over 2000-2014 as compared to 1980-2000. Further reductions in foreign emissions are not expected to relieve China's pollution in the future. Reducing local emissions is the most certain way to improve future air quality in the North China Plain.
The direct radiative forcing of black carbon aerosol (BC) on the Earth system remains unsettled, largely due to the uncertainty with physical properties of BC throughout their lifecycle. Here we show ...that ambient chamber measurements of BC properties provide a novel constraint on the crude BC aging representation in climate models. Observational evidence for significant absorption enhancement of BC can be reproduced when the aging processes in the four‐mode version of the Modal Aerosol Module (MAM4) aerosol scheme in the Community Atmosphere Model version 5 are calibrated by the recent in situ chamber measurements. An observation‐based scaling method is developed in the aging timescale calculation to alleviate the influence of biases in the simulated model chemical composition. Model sensitivity simulations suggest that the different monolayer settings in the BC aging parameterization of MAM4 can cause as large as 26% and 24% differences in BC burden and radiative forcing, respectively. We also find that an increase in coating materials (e.g., sulfate and secondary organic aerosols) reduces BC lifetime by increasing the hygroscopicity of the mixture but enhances its absorption, resulting in a net increase in BC direct radiative forcing. Our results suggest that accurate simulations of BC aging processes as well as other aerosol species are equally important in reducing the uncertainty of BC forcing estimation.
Plain Language Summary
Following carbon dioxide, black carbon particle (BC) is the second climate warming agent in the Earth system. However, BC radiative effect in most global climate models (GCM) varies by a factor of 2. The uncertainty mainly arises from the transformation of BC in the atmosphere during its lifecycle and the corresponding treatment in models. Recently, a series of BC chamber experiments were conducted over two polluted regions around the world, Beijing, China, and Houston, USA, to characterize BC aging processes. In this study, we capitalize those direct observations to calibrate a key parameter associated with BC lifetime and physical property in a GCM. Our work also illustrates the importance of the coating materials of aged BC on its radiative and climate impacts. Hence, our study calls on the deployment of the ambient smoke chamber over more places and development of a reliable aerosol composition data set with a global coverage to better constrain the BC forcing assessment.
Key Points
Atmospheric black carbon aging processes in CAM5 are constrained by chamber experiments
The calibrated model shows that coating results in net enhancement in BC forcing in spite of a reduction in BC lifetime
Modeled BC direct radiative forcing varies up to 26% due to the uncertainties in BC aging parameterization and coating aerosol mass
We compare various radiative feedbacks over the Arctic (60–90°N) estimated from short‐term climate variations occurring in reanalysis, satellite, and global climate model data sets using the combined ...Kernel‐Gregory approach. The lapse rate and surface albedo feedbacks are positive, and their magnitudes are comparable. Relative to the tropics (30°S–30°N), the lapse rate feedback is the largest contributor to Arctic amplification among all feedbacks, followed by surface albedo feedback and Planck feedback deviation from its global mean. Both shortwave and longwave water vapor feedbacks are positive, leading to a significant positive net water vapor feedback over the Arctic. The net cloud feedback has large uncertainties including its sign, which strongly depends on the data used for all‐sky and clear‐sky radiative fluxes at the top of the atmosphere, the time periods considered, and the methods used to estimate the cloud feedback.
Plain Language Summary
The Arctic has warmed dramatically in recent decades, with temperature increasing at a rate of about twice as fast as the global mean value. This phenomenon, commonly known as Arctic amplification, has been found in the observed and modeled climate changes. Several feedback mechanisms have been shown to contribute to Arctic amplification, but their relative importance is still very uncertain. Here we use a variety of reanalysis and satellite data sets to quantify the Arctic local feedbacks based on short‐term climate variations, evaluate the feedbacks simulated in a global climate model, diagnose the impact of data set choices on the feedback estimates, and identify the sources of main uncertainties. The most disagreement is found in the estimate of cloud feedback.
Key Points
Lapse rate feedback is the largest contributor to Arctic amplification relative to the tropics
Lapse rate feedback is consistently comparable to surface albedo feedback over the Arctic
Estimated net cloud feedback over the Arctic is largely uncertain including its sign
The vertical distribution of biomass burning aerosol (BBA) is important in regulating their impacts on weather and climate. The plume‐rise process affects the injection height of BBA and interacts ...with the air parcel lifting and cloud processes. However, these processes are not represented in most global climate models. In this study, we replaced the fixed vertical profiles of monthly BBA emissions in the Department of Energy's Energy Exascale Earth System Model version 1 (E3SMv1) with an interactive fire plume‐rise model. The vertical distribution of BBA emissions was calculated as a function of ambient thermodynamic conditions from the host E3SMv1, with distributions of fire sizes and sensible heat fluxes derived from the observations. The maximum fire radiative power (FRP) technique was used to determine the fire size. Scaling‐FRP technique is used to calculate the wildfire heat release. Daily BBA emission, superimposed with a fire diurnal cycle retrieved from the satellite observation, was included in model simulations. The model shows improved agreement with satellite retrievals and in situ measurement during the National Oceanic and Atmospheric Administration Wildfire Experiment for Cloud chemistry, Aerosol absorption, and Nitrogen campaign. The model‐observation comparison demonstrates the importance of the plume‐rise model and fire diurnal cycle assumption in determining the BBA fields. We also find that E3SMv1 with new features produces a larger carbonaceous aerosol burden, leading to 0.13 W m−2 warming at the top of atmosphere compared to the default E3SMv1. This highlights the importance of accurately representing the BBA injection height and indicates a no‐linear nature in the BBA‐induced radiative effect.
Plain Language Summary
In this study, we replaced fixed vertical profiles of biomass burning aerosols (BBA) used in the default doe's Energy Exascale Earth System Model version 1 (E3SMv1) with an interactive fire plume‐rise model and fire diurnal cycle assumption. The BBA vertical profiles predicted by the model with the new features are in better agreement with the satellite observations and in situ aircraft observations compared to the default version. In addition, the inclusion of the plume‐rise model and fire diurnal assumption also lead to a warming effect as strong as 0.13 W m−2.
Key Points
We incorporate an interactive fire plume‐rise model and fire diurnal cycle in the Energy Exascale Earth System Model version 1 (E3SMv1)
The new features in E3SMv1 generate biomass burning aerosol vertical profiles that agree better with the observations than the default ones
The E3SMv1 with plume‐rise model and fire diurnal cycle causes a 0.13 W m−2 warming effect compared to the default one