Explore the vast range of titles available.

In this study, we evaluated the performance of machine learning (ML) models (XGBoost) in predicting low‐cloud fraction (LCF), compared to two generations of the community atmospheric model (CAM5 and CAM6) and ERA5 reanalysis data, each having a different cloud scheme. ML models show a substantial enhancement in predicting LCF regarding root mean squared errors and correlation coefficients. The good performance is consistent across the full spectrums of atmospheric stability and large‐scale vertical velocity. Employing an explainable ML approach, we revealed the importance of including the amount of available moisture in ML models for representing spatiotemporal variations in LCF in the midlatitudes. Also, ML models demonstrated marked improvement in capturing the LCF variations during the stratocumulus‐to‐cumulus transition (SCT). This study suggests ML models' great potential to address the longstanding issues of “too few” low clouds and “too rapid” SCT in global climate models. Plain Language Summary Low clouds impose a strong radiative cooling effect on Earth's climate. Predicting low‐cloud fraction (LCF) is, however, challenging in global climate models (GCMs), partly due to some deficiencies in cloud parameterization schemes. Machine learning (ML) models might fill this gap as it is recognized as an efficient, economical, and accurate method to make predictions. In this study, we find that ML models (XGBoost) exhibit superior proficiency in predicting LCF regarding root mean squared errors and correlation coefficients compared to two generations of the community atmospheric model (CAM5 and CAM6) and ERA5 reanalysis data, each having a different cloud scheme. This improvement helps address one important issue of “too few” low clouds in GCMs. Furthermore, ML models demonstrate marked improvement in representing LCF variations when stratocumulus clouds transition to cumulus clouds, as opposed to too rapid decreases in LCF simulated by two CAMs and ERA5. Such findings testify to the unique role of ML models in refining the parameterization of LCF within GCMs. Key Points Machine learning (ML) models substantially improve “too few” low‐cloud problems in the subtropics compared to traditional cloud schemes They also show marked improvement in representing low‐cloud fraction (LCF) variations during the stratocumulus‐to‐cumulus transition Including the effect of moisture source in ML models is crucial to representing spatiotemporal variations in LCF in the midlatitudes

The response of marine low‐cloud mesoscale morphologies to climate change and emission reductions remains poorly understood. Here, we link long‐term trends in six cloud morphologies to variations in large‐scale meteorology and aerosols. The trends show strong spatial heterogeneity, with closed and disorganized mesoscale cellular convection decreasing in the Northeast Pacific and Southeast Atlantic. We develop a deep learning model (UMorNet) to predict instantaneous cloud morphologies from meteorology and cloud droplet number concentration (Nd), a proxy for aerosols. UMorNet achieves an average test accuracy of 0.55 and captures spatial patterns of climatology and long‐term trends. Out‐of‐sample test with a marine heatwave event further demonstrates the model's performance. Sensitivity experiments identify Nd, marine cold‐air outbreak index, sea surface temperature, and inversion strength as key drivers. Different responses of clustered Cu and suppressed Cu to Nd was identified. These findings highlight the potential role of aerosols in shaping cloud morphological changes.

Tropical low‐cloud feedback is the largest source of uncertainty in climate sensitivity, yet multi‐century records of surface shortwave radiation are scarce. We calibrate Porites coral δ13C against satellite photosynthetically available radiation (PAR) and reconstruct monthly PAR for the northern South China Sea during the Medieval Climate Anomaly (1129–1264 CE) and the Little Ice Age (1631–1771 CE). After correcting for the Suess effect and propagating errors via Monte Carlo resampling techniques, annual PAR during the Little‐Ice‐Age is ∼22% lower and seasonality slightly weaker. The dimming aligns with regional proxies for cooler, wetter conditions and is best explained by brighter low clouds, likely boosted by volcanic aerosol–cloud interactions. CMIP6/PMIP4 past1000 simulations, however, yield <0.2% change over the same interval, indicating that current models understate volcanic microphysics and tropical low‐cloud sensitivity. The coral PAR record thus provides a quantitative pre‐industrial target for evaluating tropical cloud processes and reducing uncertainty in equilibrium climate sensitivity.

Sulfate aerosol greatly contributes to wintertime haze pollution in emission‐intensive regions like the North China Plain (NCP) in China. Fast sulfate increase and accumulation are usually recorded during winter haze; however, the multiphase oxidation of sulfur dioxide (SO2) and the physical processes affecting near‐surface sulfate are not fully understood. By combining in situ observations and numerical simulations, we found that high sulfur oxidation ratios (>0.6) under heavily polluted conditions are associated with low clouds and fog over NCP, induced by the moist southerly airflow. Thick low clouds and high SO2 levels in NCP provide a reaction environment for sulfate production. The sulfate production rate in cloud water can reach 0.5–1.3 μg m−3 h−1. The results demonstrate that the vertical mixing of sulfate generated within the cloud water to the surface plays a significant role in rapid sulfate production, highlighting the importance of understanding cloud‐water processes in haze pollution. Plain Language Summary Sulfate has been recognized as an important chemical component of atmospheric aerosols, especially during winter haze events. Rapid increases in sulfate concentrations are frequently observed during heavy pollution in the North China Plain (NCP) of China. However, the processes involved in the multiphase oxidation of sulfur dioxide (SO2) and the physical processes influencing sulfate variations near the surface remain unclear. In particular, the contribution of traditional in‐cloud sulfate production to the surface sulfate has been considered to be negligible in the NCP. Here, we revisited the role of in‐cloud sulfate production in the NCP by using ground‐based observations, radiosonde measurements, and model simulation. Our results indicate that the rapid conversion of SO2 to sulfates during heavy pollution is associated with the presence of low clouds and fog. We find that high sulfate production rates in cloud water lead to the rapid accumulation of sulfate, which is vertically mixed to the surface, resulting in a rapid increase in surface sulfate concentrations. This work sheds a new perspective on understanding the role of sulfate production in cloud water and its impact on air pollution. Key Points High sulfur oxidation ratios under heavy pollution conditions in the North China Plain are associated with low clouds and fog formation Sulfate production rates within cloud water can reach 0.5–1.3 μg m−3 h−1, with NO2 and O3 oxidation pathways dominating Vertical mixing of sulfate produced in cloud water to the surface causes a rapid increase in near‐surface sulfate concentration

Low, marine clouds cool the Earth system, reflecting sunlight back to space. Low cloud response to environmental change is a key uncertainty in future climate projections. It is especially uncertain how much warming amplification will occur due to tropical cumulus feedback. A potentially important feedback modulator is the ability for cumulus to cluster through mesoscale circulations. Janssens et al. (2025, ) demonstrate that moisture convergence in ascending circulation branches organizes clouds into fewer and brighter structures while moisture divergence dries descending branches, reducing cloud and increasing longwave cooling. These offsetting effects result in a small net radiative effect due to organization. Janssens et al. (2025, )'s results imply that the influence of organization on cumulus feedback is insignificant. The proposed offsetting of radiative effects across mesoscale organization patterns, or “symmetry,” is worthy of continued research. Observational support and further investigations into whether cumulus organization has other climate impacts is encouraged.

The double‐Intertropical Convergence Zone (ITCZ) bias is a common model bias, which has puzzled the climate model community for several decades. Here, by comparing a high‐ and low‐resolution state‐of‐the‐art model CESM1, it is found that the double‐ITCZ bias is largely reduced in the high‐resolution CESM1. The key reason is the realistic colder sea surface temperature (SST) over the southeast Pacific (SEP) in the high‐resolution model. This realistic SEP SST is mainly due to a spuriously deeper mixed layer with a more realistic wind, as the sensitivity of mixed layer depth to wind is overestimated in both versions of CESM1. The better representation of terrain, such as Andes Mountains, elevates warm advection from inland to the coastal region, which maintains the inversion structure favorable for low cloud. The resultant increased coastal cloud in the high‐resolution CESM1 causes the colder coastal SST, thus improving the wind and deepening the mixed layer. Plain Language Summary The double‐Intertropical Convergence Zone (ITCZ) bias is a common issue in climate models that has puzzled scientists for many years, featuring as two incorrect ITCZs on both sides of the equator instead of one to the north of the equator viewed from the annual mean precipitation. In our study, we compared two versions of a climate model called CESM1, one with high resolution and one with low resolution. We find that the double‐ITCZ bias in the high‐resolution CESM1 is significantly reduced. The main reason for this improvement is that the high‐resolution model simulates the colder SST in the southeast Pacific more accurately. The realistic SST is caused by an unrealistically deeper mixed layer, as the sensitivity of mixed layer depth to the wind is much higher than the observation in the two versions of CESM1 model. The improved wind is associated with changes in the SST and cloud cover along the coast of South America, originated from the atmospheric model by improving the ability to depict the terrain, such as Andes Mountains, which elevates the warm advection and maintains the temperature inversion favorable for low cloud formation in the high‐resolution models. Key Points The reduced double‐ITCZ bias in high‐resolution CESM1 (CESM1‐H) than low‐resolution one is linked to the colder southeastern Pacific (SEP) The colder SEP is due to a spuriously deeper mixed layer caused by the overestimated sensitivity to a more realistic wind in the CESM1‐H The more realistic SEP wind is caused by the colder coast with more low clouds, tied to the better‐depicted terrain over South America

In this study, the impact of different configurations of the Goddard radiation scheme on convection-permitting simulations (CPSs) of the West African monsoon (WAM) is investigated using the NASA-Unified WRF (NU-WRF). These CPSs had 3 km grid spacing to explicitly simulate the evolution of mesoscale convective systems (MCSs) and their interaction with radiative processes across the WAM domain and were able to reproduce realistic precipitation and energy budget fields when compared with satellite data, although low clouds were overestimated. Sensitivity experiments reveal that (1) lowering the radiation update frequency (i.e., longer radiation update time) increases precipitation and cloudiness over the WAM region by enhancing the monsoon circulation, (2) deactivation of precipitation radiative forcing suppresses cloudiness over the WAM region, and (3) aggregating radiation columns reduces low clouds over ocean and tropical West Africa. The changes in radiation configuration immediately modulate the radiative heating and low clouds over ocean. On the 2nd day of the simulations, patterns of latitudinal air temperature profiles were already similar to the patterns of monthly composites for all radiation sensitivity experiments. Low cloud maintenance within the WAM system is tightly connected with radiation processes; thus, proper coupling between microphysics and radiation processes must be established for each modeling framework.

The relative dispersion of the cloud droplet spectrum (ε) is a key parameter that characterizes the spectral shape. Uncertainties in its parameterization can introduce significant biases in the simulation of cloud and fog microphysical and optical properties, making its accurate diagnosis in models critically important. In this study, we conduct sensitivity experiments using the large‐eddy simulation model to evaluate the performance of a newly developed nonlinear ε parameterization based on volume‐mean diameter of cloud droplets (Dv). Four typical surface types associated with cloud and fog processes are tested. Compared to conventional linear parameterizations based on droplet number concentration or Dv, the new parameterization significantly improves the simulation of droplet spectral shape parameters and effective radius, with average improvements of 90.59% and 78.49%, respectively. Moreover, the new parameterization captures the complex relationship between ε and liquid water content observed across different surface types and exhibits distinct impacts on cloud‐to‐rain autoconversion rate from those of default parameterization with a net mean change of +72.21%, thereby affecting the formation and intensity of precipitation. The new parameterization also leads to reduced cloud and fog optical thickness, which weakens cloud radiative cooling. These findings highlight the importance of improved ε parameterizations for the studied regions in China and underscore the potential for enhancing model performance in representing microphysics, radiation, and precipitation processes in low clouds and fog. Assessing the generality of the conclusions across different geographic regions and climatic regimes would be a valuable focus for future work. Plain Language Summary Clouds and fog are made up of many tiny water droplets. The way these droplet sizes are spread out, known as droplet spectrum dispersion, plays an important role in how clouds reflect radiation and produce rain. However, many weather and climate models use simplified ways to represent this dispersion, which can cause errors in simulating cloud properties and precipitation. In this study, we introduce a new method for describing droplet dispersion based on the volume‐mean diameter and test it in a high‐resolution weather model. Compared to traditional methods, the new approach more accurately captures cloud droplet characteristics and improves the simulation of cloud microphysical and optical properties. These improvements can help make weather and climate predictions more reliable, especially for low clouds and fog. Key Points Novel parameterization improves the simulations of cloud droplet spectral shape and effective diameter It exerts distinct impacts on cloud‐to‐rain autoconversion process over different surface types It mitigates cloud radiative cooling strength in model simulations

The response of marine low cloud systems to changes in aerosol concentration represents one of the largest uncertainties in climate simulations. Major contributions to this uncertainty are derived from poor understanding of aerosol under natural conditions and the perturbation by anthropogenic emissions. The eastern North Atlantic (ENA) is a region of persistent but diverse marine boundary layer (MBL) clouds, whose albedo and precipitation are highly susceptible to perturbations in aerosol properties. In this study, we examine MBL aerosol properties, trace gas mixing ratios, and meteorological parameters measured at the Atmospheric Radiation Measurement Climate Research Facility's ENA site on Graciosa Island, Azores, Portugal, during a 3-year period from 2015 to 2017. Measurements impacted by local pollution on Graciosa Island and during occasional intense biomass burning and dust events are excluded from this study. Submicron aerosol size distribution typically consists of three modes: Aitken (At, diameter Dp<∼100 nm), accumulation (Ac, Dp within ∼100 to ∼300 nm), and larger accumulation (LA, Dp>∼300 nm) modes, with average number concentrations (denoted as NAt, NAc, and NLA below) of 330, 114, and 14 cm−3, respectively. NAt, NAc, and NLA show contrasting seasonal variations, suggesting different sources and removal processes. NLA is dominated by sea spray aerosol (SSA) and is higher in winter and lower in summer. This is due to the seasonal variations of SSA production, in-cloud coalescence scavenging, and dilution by entrained free troposphere (FT) air. In comparison, SSA typically contributes a relatively minor fraction to NAt (10 %) and NAc (21 %) on an annual basis. In addition to SSA, sources of Ac-mode particles include entrainment of FT aerosols and condensation growth of Aitken-mode particles inside the MBL, while in-cloud coalescence scavenging is the major sink of NAc. The observed seasonal variation of NAc, being higher in summer and lower in winter, generally agrees with the steady-state concentration estimated from major sources and sinks. NAt is mainly controlled by entrainment of FT aerosol, coagulation loss, and growth of Aitken-mode particles into the Ac-mode size range. Our calculation suggests that besides the direct contribution from entrained FT Ac-mode particles, growth of entrained FT Aitken-mode particles in the MBL also represent a substantial source of cloud condensation nuclei (CCN), with the highest contribution potentially reaching 60 % during summer. The growth of Aitken-mode particles to CCN size is an expected result of the condensation of sulfuric acid, a product from dimethyl sulfide oxidation, suggesting that ocean ecosystems may have a substantial influence on MBL CCN populations in the ENA.

CGILS—the CFMIP‐GASS Intercomparison of Large Eddy Models (LESs) and single column models (SCMs)—investigates the mechanisms of cloud feedback in SCMs and LESs under idealized climate change perturbation. This paper describes the CGILS results from 15 SCMs and 8 LES models. Three cloud regimes over the subtropical oceans are studied: shallow cumulus, cumulus under stratocumulus, and well‐mixed coastal stratus/stratocumulus. In the stratocumulus and coastal stratus regimes, SCMs without activated shallow convection generally simulated negative cloud feedbacks, while models with active shallow convection generally simulated positive cloud feedbacks. In the shallow cumulus alone regime, this relationship is less clear, likely due to the changes in cloud depth, lateral mixing, and precipitation or a combination of them. The majority of LES models simulated negative cloud feedback in the well‐mixed coastal stratus/stratocumulus regime, and positive feedback in the shallow cumulus and stratocumulus regime. A general framework is provided to interpret SCM results: in a warmer climate, the moistening rate of the cloudy layer associated with the surface‐based turbulence parameterization is enhanced; together with weaker large‐scale subsidence, it causes negative cloud feedback. In contrast, in the warmer climate, the drying rate associated with the shallow convection scheme is enhanced. This causes positive cloud feedback. These mechanisms are summarized as the “NESTS” negative cloud feedback and the “SCOPE” positive cloud feedback (Negative feedback from Surface Turbulence under weaker Subsidence—Shallow Convection PositivE feedback) with the net cloud feedback depending on how the two opposing effects counteract each other. The LES results are consistent with these interpretations. Key Points Reasons of negative and positive cloud feedbacks in SCMs are explained A framework is provided to interpret cloud feedbacks in models SCM results are compared with LES simulations