This paper introduces the Flexible Global Ocean‐Atmosphere‐Land System Model: Grid‐Point Version 3 (FGOALS‐g3) and evaluates its basic performance based on some of its participation in the sixth phase of the Coupled Model Intercomparison Project (CMIP6) experiments. Our results show that many significant improvements have been achieved by FGOALS‐g3 in terms of climatological mean states, variabilities, and long‐term trends. For example, FGOALS‐g3 has a small (−0.015°C/100 yr) climate drift in 700‐yr preindustrial control (piControl) runs and smaller biases in climatological mean variables, such as the land/sea surface temperatures (SSTs) and seasonal soil moisture cycle, compared with its previous version FGOALS‐g2 during the historical period. The characteristics of climate variabilities, for example, Madden‐Julian oscillation (MJO) eastward/westward propagation ratios, spatial patterns of interannual variability of tropical SST anomalies, and relationship between the East Asian Summer Monsoon and El Niño–Southern Oscillation (ENSO), are well captured by FGOALS‐g3. In particular, the cooling trend of globally averaged surface temperature during 1940–1970, which is a challenge for most CMIP3 and CMIP5 models, is well reproduced by FGOALS‐g3 in historical runs. In addition to the external forcing factors recommended by CMIP6, anthropogenic groundwater forcing from 1965 to 2014 was incorporated into the FGOALS‐g3 historical runs.

We present a new framework for global ocean–sea-ice model simulations based on phase 2 of the Ocean Model Intercomparison Project (OMIP-2), making use of the surface dataset based on the Japanese 55-year atmospheric reanalysis for driving ocean–sea-ice models (JRA55-do). We motivate the use of OMIP-2 over the framework for the first phase of OMIP (OMIP-1), previously referred to as the Coordinated Ocean–ice Reference Experiments (COREs), via the evaluation of OMIP-1 and OMIP-2 simulations from 11 state-of-the-science global ocean–sea-ice models. In the present evaluation, multi-model ensemble means and spreads are calculated separately for the OMIP-1 and OMIP-2 simulations and overall performance is assessed considering metrics commonly used by ocean modelers. Both OMIP-1 and OMIP-2 multi-model ensemble ranges capture observations in more than 80 % of the time and region for most metrics, with the multi-model ensemble spread greatly exceeding the difference between the means of the two datasets. Many features, including some climatologically relevant ocean circulation indices, are very similar between OMIP-1 and OMIP-2 simulations, and yet we could also identify key qualitative improvements in transitioning from OMIP-1 to OMIP-2. For example, the sea surface temperatures of the OMIP-2 simulations reproduce the observed global warming during the 1980s and 1990s, as well as the warming slowdown in the 2000s and the more recent accelerated warming, which were absent in OMIP-1, noting that the last feature is part of the design of OMIP-2 because OMIP-1 forcing stopped in 2009. A negative bias in the sea-ice concentration in summer of both hemispheres in OMIP-1 is significantly reduced in OMIP-2. The overall reproducibility of both seasonal and interannual variations in sea surface temperature and sea surface height (dynamic sea level) is improved in OMIP-2. These improvements represent a new capability of the OMIP-2 framework for evaluating process-level responses using simulation results. Regarding the sensitivity of individual models to the change in forcing, the models show well-ordered responses for the metrics that are directly forced, while they show less organized responses for those that require complex model adjustments. Many of the remaining common model biases may be attributed either to errors in representing important processes in ocean–sea-ice models, some of which are expected to be reduced by using finer horizontal and/or vertical resolutions, or to shared biases and limitations in the atmospheric forcing. In particular, further efforts are warranted to resolve remaining issues in OMIP-2 such as the warm bias in the upper layer, the mismatch between the observed and simulated variability of heat content and thermosteric sea level before 1990s, and the erroneous representation of deep and bottom water formations and circulations. We suggest that such problems can be resolved through collaboration between those developing models (including parameterizations) and forcing datasets. Overall, the present assessment justifies our recommendation that future model development and analysis studies use the OMIP-2 framework.

This paper presents global comparisons of fundamental global climate variables from a suite of four pairs of matched low- and high-resolution ocean and sea ice simulations that are obtained following the OMIP-2 protocol (Griffies et al., 2016) and integrated for one cycle (1958–2018) of the JRA55-do atmospheric state and runoff dataset (Tsujino et al., 2018). Our goal is to assess the robustness of climate-relevant improvements in ocean simulations (mean and variability) associated with moving from coarse (∼ 1^∘) to eddy-resolving (∼ 0.1^∘) horizontal resolutions. The models are diverse in their numerics and parameterizations, but each low-resolution and high-resolution pair of models is matched so as to isolate, to the extent possible, the effects of horizontal resolution. A variety of observational datasets are used to assess the fidelity of simulated temperature and salinity, sea surface height, kinetic energy, heat and volume transports, and sea ice distribution. This paper provides a crucial benchmark for future studies comparing and improving different schemes in any of the models used in this study or similar ones. The biases in the low-resolution simulations are familiar, and their gross features – position, strength, and variability of western boundary currents, equatorial currents, and the Antarctic Circumpolar Current – are significantly improved in the high-resolution models. However, despite the fact that the high-resolution models “resolve” most of these features, the improvements in temperature and salinity are inconsistent among the different model families, and some regions show increased bias over their low-resolution counterparts. Greatly enhanced horizontal resolution does not deliver unambiguous bias improvement in all regions for all models.

The second version of Chinese Academy of Sciences Earth System Model (CAS‐ESM 2) is described with emphasis on the development process, strength and weakness, and climate sensitivities in simulations of the Coupled Model Intercomparison Project (CMIP6) DECK experiments. CAS‐ESM 2 was built as a numerical model to simulate both the physical climate system as well as atmospheric chemistry and carbon cycle. It is a newcomer in the international modeling community to provide sufficiently independent solutions of climate simulations from those of other models. Performances of the model in simulating the basic states of the radiation budget of the atmosphere and ocean, precipitation, circulations, variabilities, and the 20^th Century warming are presented. Model biases and their possible causes are discussed. Strength includes horizontal heat transport in the atmosphere and oceans, vertical profile of the Atlantic Meridional Overturning Circulation; weakness includes the double Inter‐Tropical Convergence Zone (ITCZ) and stronger amplitude of the El Niño‐Southern Oscillation (ENSO) that are also common in many other models. The simulated the 20^th Century warming shares a similar discrepancy with observations as in several other models—less warming in the 1920s and stronger cooling in the 1960s than in observation—at the time when there was a steep increase of anthropogenic aerosols. As a result, the 20^th Century warming is about 60% of the observed warming despite that the model simulated a similar slope of warming trend after 1980 to observation. The model has an equilibrium climate sensitivity of 3.4 K with a positive cloud feedback from the shortwave radiation.

The ocean mixed layer is the interface between the ocean interior and the atmosphere or sea ice and plays a key role in climate variability. It is thus critical that numerical models used in climate studies are capable of a good representation of the mixed layer, especially its depth. Here we evaluate the mixed-layer depth (MLD) in six pairs of non-eddying (1^∘ grid spacing) and eddy-rich (up to $\mathrm{1}/\mathrm{16}$ ^∘) models from the Ocean Model Intercomparison Project (OMIP), forced by a common atmospheric state. For model evaluation, we use an updated MLD dataset computed from observations using the OMIP protocol (a constant density threshold). In winter, low-resolution models exhibit large biases in the deep-water formation regions. These biases are reduced in eddy-rich models but not uniformly across models and regions. The improvement is most noticeable in the mode-water formation regions of the Northern Hemisphere. Results in the Southern Ocean are more contrasted, with biases of either sign remaining at high resolution. In eddy-rich models, mesoscale eddies control the spatial variability in MLD in winter. Contrary to a hypothesis that the deepening of the mixed layer in anticyclones would make the MLD larger globally, eddy-rich models tend to have a shallower mixed layer at most latitudes than coarser models do. In addition, our study highlights the sensitivity of the MLD computation to the choice of a reference level and the spatio-temporal sampling, which motivates new recommendations for MLD computation in future model intercomparison projects.