ABSTRACT:Results are presented from an intercomparison of single-column and cloud-resolving model simulations of a cold-air outbreak mixed-phase stratocumulus cloud observed during the Atmospheric Radiation Measurement (ARM) programme's Mixed-Phase Arctic Cloud Experiment. The observed cloud occurred in a well-mixed boundary layer with a cloud-top temperature of −15 • C. The average liquid water path of around 160 g m −2 was about two-thirds of the adiabatic value and far greater than the average mass of ice which when integrated from the surface to cloud top was around 15 g m −2 .Simulations of 17 single-column models (SCMs) and 9 cloud-resolving models (CRMs) are compared. While the simulated ice water path is generally consistent with observed values, the median SCM and CRM liquid water path is a factor-of-three smaller than observed. Results from a sensitivity study in which models removed ice microphysics suggest that in many models the interaction between liquid and ice-phase microphysics is responsible for the large model underestimate of liquid water path.Despite this underestimate, the simulated liquid and ice water paths of several models are consistent with observed values. Furthermore, models with more sophisticated microphysics simulate liquid and ice water paths that are in better agreement with the observed values, although considerable scatter exists. Although no single factor guarantees a good simulation, these results emphasize the need for improvement in the model representation of mixed-phase microphysics.

[1] A size-segregated multicomponent aerosol algorithm, the Canadian Aerosol Module (CAM), was developed for use with climate and air quality models. It includes major aerosol processes in the atmosphere: generation, hygroscopic growth, coagulation, nucleation, condensation, dry deposition/sedimentation, below-cloud scavenging, aerosol activation, a cloud module with explicit microphysical processes to treat aerosol-cloud interactions and chemical transformation of sulphur species in clear air and in clouds. The numerical solution was optimized to efficiently solve the complicated size-segregated multicomponent aerosol system and make it feasible to be included in global and regional models. An internal mixture is assumed for all types of aerosols except for soil dust and black carbon which are assumed to be externally mixed close to sources. To test the algorithm, emissions to the atmosphere of anthropogenic and natural aerosols are simulated for two aerosol types: sea salt and sulphate. A comparison was made of two numerical solutions of the aerosol algorithm: process splitting and ordinary differential equation (ODE) solver. It was found that the process-splitting method used for this model is within 15% of the more accurate ODE solution for the total sulphate mass concentration and <1% accurate for sea-salt concentration. Furthermore, it is computationally more than 100 times faster. The sensitivity of the simulated size distributions to the number of size bins was also investigated. The diffusional behavior of each individual process was quantitatively characterized by the difference in the mode radius and standard deviation of a lognormal curve fit of distributions between the approximate solution and the 96-bin reference solution. Both the number and mass size distributions were adequately predicted by a sectional model of 12 bins in many situations in the atmosphere where the sink for condensable matter on existing aerosol surface area is high enough that nucleation of new particles is negligible. Total mass concentration was adequately simulated using lower size resolution of 8 bins. However, to properly resolve nucleation mode size distributions and minimize the numerical diffusion, a sectional model of 18 size bins or greater is needed. The number of size bins is more important in resolving the nucleation mode peaks than in reducing the diffusional behavior of aerosol processes. Application of CAM in a study of the global cycling of sea-salt mass accompanies this paper [Gong et al., 2002].

Abstract. The Canadian Earth System Model version 5 (CanESM5) is a global model developed to simulate historical climate change and variability, to make centennial-scale projections of future climate, and to produce initialized seasonal and decadal predictions. This paper describes the model components and their coupling, as well as various aspects of model development, including tuning, optimization, and a reproducibility strategy. We also document the stability of the model using a long control simulation, quantify the model's ability to reproduce large-scale features of the historical climate, and evaluate the response of the model to external forcing. CanESM5 is comprised of three-dimensional atmosphere (T63 spectral resolution equivalent roughly to 2.8∘) and ocean (nominally 1∘) general circulation models, a sea-ice model, a land surface scheme, and explicit land and ocean carbon cycle models. The model features relatively coarse resolution and high throughput, which facilitates the production of large ensembles. CanESM5 has a notably higher equilibrium climate sensitivity (5.6 K) than its predecessor, CanESM2 (3.7 K), which we briefly discuss, along with simulated changes over the historical period. CanESM5 simulations contribute to the Coupled Model Intercomparison Project phase 6 (CMIP6) and will be employed for climate science and service applications in Canada.

No abstract