We document and evaluate the aerosol schemes as implemented in the physical and Earth system models, the Global Coupled 3.1 configuration of the Hadley Centre Global Environment Model version 3 (HadGEM3-GC3.1) and the United Kingdom Earth System Model (UKESM1), which are contributing to the sixth Coupled Model Intercomparison Project (CMIP6). The simulation of aerosols in the present-day period of the historical ensemble of these models is evaluated against a range of observations. Updates to the aerosol microphysics scheme are documented as well as differences in the aerosol representation between the physical and Earth system configurations. The additional Earth system interactions included in UKESM1 lead to differences in the emissions of natural aerosol sources such as dimethyl sulfide, mineral dust and organic aerosol and subsequent evolution of these species in the model. UKESM1 also includes a stratospheric–tropospheric chemistry scheme which is fully coupled to the aerosol scheme, while GC3.1 employs a simplified aerosol chemistry mechanism driven by prescribed monthly climatologies of the relevant oxidants. Overall, the simulated speciated aerosol mass concentrations compare reasonably well with observations. Both models capture the negative trend in sulfate aerosol concentrations over Europe and the eastern United States of America (US) although the models tend to underestimate sulfate concentrations in both regions. Interactive emissions of biogenic volatile organic compounds in UKESM1 lead to an improved agreement of organic aerosol over the US. Simulated dust burdens are similar in both models despite a 2-fold difference in dust emissions. Aerosol optical depth is biased low in dust source and outflow regions but performs well in other regions compared to a number of satellite and ground-based retrievals of aerosol optical depth. Simulated aerosol number concentrations are generally within a factor of 2 of the observations, with both models tending to overestimate number concentrations over remote ocean regions, apart from at high latitudes, and underestimate over Northern Hemisphere continents. Finally, a new primary marine organic aerosol source is implemented in UKESM1 for the first time. The impact of this new aerosol source is evaluated. Over the pristine Southern Ocean, it is found to improve the seasonal cycle of organic aerosol mass and cloud droplet number concentrations relative to GC3.1 although underestimations in cloud droplet number concentrations remain. This paper provides a useful characterisation of the aerosol climatology in both models and will facilitate understanding in the numerous aerosol–climate interaction studies that will be conducted as part of CMIP6 and beyond.

Producing a global and comprehensive description of atmospheric aerosols requires integration of ground-based, airborne, satellite and model datasets. Due to its complexity, aerosol monitoring requires the use of several data records with complementary information content. This paper describes the lessons learned while developing and qualifying algorithms to generate aerosol Climate Data Records (CDR) within the European Space Agency (ESA) Aerosol_cci project. An iterative algorithm development and evaluation cycle involving core users is applied. It begins with the application-specific refinement of user requirements, leading to algorithm development, dataset processing and independent validation followed by user evaluation. This cycle is demonstrated for a CDR of total Aerosol Optical Depth (AOD) from two subsequent dual-view radiometers. Specific aspects of its applicability to other aerosol algorithms are illustrated with four complementary aerosol datasets. An important element in the development of aerosol CDRs is the inclusion of several algorithms evaluating the same data to benefit from various solutions to the ill-determined retrieval problem. The iterative approach has produced a 17-year AOD CDR, a 10-year stratospheric extinction profile CDR and a 35-year Absorbing Aerosol Index record. Further evolution cycles have been initiated for complementary datasets to provide insight into aerosol properties (i.e., dust aerosol, aerosol absorption).

Aerosol–cloud interactions (ACIs) are considered to be the most uncertain driver of present-day radiative forcing due to human activities. The nonlinearity of cloud-state changes to aerosol perturbations make it challenging to attribute causality in observed relationships of aerosol radiative forcing. Using correlations to infer causality can be challenging when meteorological variability also drives both aerosol and cloud changes independently. Natural and anthropogenic aerosol perturbations from well-defined sources provide “opportunistic experiments” (also known as natural experiments) to investigate ACI in cases where causality may be more confidently inferred. These perturbations cover a wide range of locations and spatiotemporal scales, including point sources such as volcanic eruptions or industrial sources, plumes from biomass burning or forest fires, and tracks from individual ships or shipping corridors. We review the different experimental conditions and conduct a synthesis of the available satellite datasets and field campaigns to place these opportunistic experiments on a common footing, facilitating new insights and a clearer understanding of key uncertainties in aerosol radiative forcing. Cloud albedo perturbations are strongly sensitive to background meteorological conditions. Strong liquid water path increases due to aerosol perturbations are largely ruled out by averaging across experiments. Opportunistic experiments have significantly improved process-level understanding of ACI, but it remains unclear how reliably the relationships found can be scaled to the global level, thus demonstrating a need for deeper investigation in order to improve assessments of aerosol radiative forcing and climate change.

The question of how to derive and present uncertainty information in climate data records (CDRs) has received sustained attention within the European Space Agency Climate Change Initiative (CCI), a programme to generate CDRs addressing a range of essential climate variables (ECVs) from satellite data. Here, we review the nature, mathematics, practicalities, and communication of uncertainty information in CDRs from Earth observations. This review paper argues that CDRs derived from satellite-based Earth observation (EO) should include rigorous uncertainty information to support the application of the data in contexts such as policy, climate modelling, and numerical weather prediction reanalysis. Uncertainty, error, and quality are distinct concepts, and the case is made that CDR products should follow international metrological norms for presenting quantified uncertainty. As a baseline for good practice, total standard uncertainty should be quantified per datum in a CDR, meaning that uncertainty estimates should clearly discriminate more and less certain data. In this case, flags for data quality should not duplicate uncertainty information, but instead describe complementary information (such as the confidence in the uncertainty estimate provided or indicators of conditions violating the retrieval assumptions). The paper discusses the many sources of error in CDRs, noting that different errors may be correlated across a wide range of timescales and space scales. Error effects that contribute negligibly to the total uncertainty in a single-satellite measurement can be the dominant sources of uncertainty in a CDR on the large space scales and long timescales that are highly relevant for some climate applications. For this reason, identifying and characterizing the relevant sources of uncertainty for CDRs is particularly challenging. The characterization of uncertainty caused by a given error effect involves assessing the magnitude of the effect, the shape of the error distribution, and the propagation of the uncertainty to the geophysical variable in the CDR accounting for its error correlation properties. Uncertainty estimates can and should be validated as part of CDR validation when possible. These principles are quite general, but the approach to providing uncertainty information appropriate to different ECVs is varied, as confirmed by a brief review across different ECVs in the CCI. User requirements for uncertainty information can conflict with each other, and a variety of solutions and compromises are possible. The concept of an ensemble CDR as a simple means of communicating rigorous uncertainty information to users is discussed. Our review concludes by providing eight concrete recommendations for good practice in providing and communicating uncertainty in EO-based climate data records.

This paper discusses a best-practice representation of uncertainty in satellite remote sensing data. An estimate of uncertainty is necessary to make appropriate use of the information conveyed by a measurement. Traditional error propagation quantifies the uncertainty in a measurement due to well-understood perturbations in a measurement and in auxiliary data – known, quantified "unknowns". The under-constrained nature of most satellite remote sensing observations requires the use of various approximations and assumptions that produce non-linear systematic errors that are not readily assessed – known, unquantifiable "unknowns". Additional errors result from the inability to resolve all scales of variation in the measured quantity – unknown "unknowns". The latter two categories of error are dominant in under-constrained remote sensing retrievals, and the difficulty of their quantification limits the utility of existing uncertainty estimates, degrading confidence in such data. This paper proposes the use of ensemble techniques to present multiple self-consistent realisations of a data set as a means of depicting unquantified uncertainties. These are generated using various systems (different algorithms or forward models) believed to be appropriate to the conditions observed. Benefiting from the experience of the climate modelling community, an ensemble provides a user with a more complete representation of the uncertainty as understood by the data producer and greater freedom to consider different realisations of the data.

Recent years have seen the increasing inclusion of per-retrieval prognostic (predictive) uncertainty estimates within satellite aerosol optical depth (AOD) data sets, providing users with quantitative tools to assist in the optimal use of these data. Prognostic estimates contrast with diagnostic (i.e. relative to some external truth) ones, which are typically obtained using sensitivity and/or validation analyses. Up to now, however, the quality of these uncertainty estimates has not been routinely assessed. This study presents a review of existing prognostic and diagnostic approaches for quantifying uncertainty in satellite AOD retrievals, and it presents a general framework to evaluate them based on the expected statistical properties of ensembles of estimated uncertainties and actual retrieval errors. It is hoped that this framework will be adopted as a complement to existing AOD validation exercises; it is not restricted to AOD and can in principle be applied to other quantities for which a reference validation data set is available. This framework is then applied to assess the uncertainties provided by several satellite data sets (seven over land, five over water), which draw on methods from the empirical to sensitivity analyses to formal error propagation, at 12 Aerosol Robotic Network (AERONET) sites. The AERONET sites are divided into those for which it is expected that the techniques will perform well and those for which some complexity about the site may provide a more severe test. Overall, all techniques show some skill in that larger estimated uncertainties are generally associated with larger observed errors, although they are sometimes poorly calibrated (i.e. too small or too large in magnitude). No technique uniformly performs best. For powerful formal uncertainty propagation approaches such as optimal estimation, the results illustrate some of the difficulties in appropriate population of the covariance matrices required by the technique. When the data sets are confronted by a situation strongly counter to the retrieval forward model (e.g. potentially mixed land–water surfaces or aerosol optical properties outside the family of assumptions), some algorithms fail to provide a retrieval, while others do but with a quantitatively unreliable uncertainty estimate. The discussion suggests paths forward for the refinement of these techniques.