Quoted from: Ren, Shihe, Xi Liang, Qizhen Sun, Hao Yu, L. Bruno Tremblay, Xiaoping Mai, Fu Zhao et al. "A fully coupled Arctic sea ice-ocean-atmosphere model (ArcIOAM v1. 0) based on C-Coupler2: model description and preliminary results." Geoscientific Model Development Discussions (2020): 1-28.  https://doi.org/10.5194/gmd-14-1101-2021 

The oceanic and sea ice component model

        The ocean and sea ice component of ArcIOAM is an Arctic configuration of the MITgcm (Nguyen et al., 2011; Liang and Losch, 2018; Liang et al., 2019, 2020). The model has an average horizontal resolution of 18 km and covers the whole Arctic Ocean with open boundaries close to 55◦ N in both the Atlantic and Pacific sectors (Losch et al., 2010). The ocean model includes 420×384 horizontal grid points and 50 vertical model layers based on Arakawa C grid and Z coordinates and a time step of 1200 s. The ocean model uses curvilinear coordinates, and the model grid is locally orthogonal. Vertical resolution of the ocean model layers increases from 10 m near the surface to 456 m near the bottom. The K-profile parameterization (KPP) (Large et al., 1994) is used as the vertical mixing scheme.

        The sea ice model shares the same horizontal grid with the ocean model and divides each model grid into two parts: ice and open ocean. In the open-ocean area, ocean-atmosphere heat and momentum fluxes are calculated following the standard bulk formula (Doney et al., 1998). In the ice-covered area, the ice surface and bottom heat and momentum fluxes are calculated according to viscous-plastic dynamics and zero-layer thermodynamics (Hibler, 1980; Semtner, 1976). The so-called zero-layer thermodynamic model assumes one layer of ice underneath one layer of snow and assumes ice does not store heat and therefore tends to exaggerate the seasonal variability in ice thickness. Snow modifies ice surface albedo and conductivity. If enough snow accumulates on top of the ice, its weight submerges the ice and the snow is flooded. In order to parameterize a sub-grid-scale distribution for sea ice thickness, the mean sea ice thickness in each grid can be split into as many as seven thickness categories in the MITgcm sea ice model. In our coupled model for simplicity, we use two thickness categories: open water and sea ice.

The atmospheric component model

        The atmospheric component of ArcIOAM is based on the Polar WRF (Bromwich et al., 2013; Hines and Bromwich, 2008) model, which is an optimized version of the WRF model (Skamarock et al., 2008) for use in polar regions. Previous researchers have made several specific modifications for polar environments, which primarily encompass the land surface model and sea ice to adapt to the particular conditions in Arctic regions. The Noah land surface model is embedded inside the Polar WRF. The changes made in the Noah land surface model (LSM; Chen and Dudhia, 2001) include using the latent heat of sublimation for calculating latent heat flux over ice surface, increasing the snow albedo and the emissivity value for snow, adjusting snow density, modifying thermal diffusivity and snow heat capacity for the subsurface layer, changing the calculation of skin temperature, and assuming ice saturation in calculating the surface saturation mixing ratio over ice. Other modifications of the Polar WRF include a fix to allow specified sea ice quantities and the land mask associated with sea ice to update during a simulation. These modifications improve model performance over the pan-Arctic for short-term forecasts.

        The Arctic configuration of the Polar WRF model has been tested and evaluated by a set of simulations over several key surface categories, including large permanent ice sheets with the Greenland/North Atlantic grid and Arctic land (Hines et al., 2011; Hines and Bromwich, 2008) and the production of the Arctic System Reanalysis (ASR) (Bromwich et al., 2010). In this study, the Polar WRF model covers the Arctic regions with a horizontal resolution of 27 km. The model has 306 × 306 horizontal grid points and 60 vertical layers and a time step of 120 s. The Polar WRF model employed physics options that included the Mellor–Yamada–Janjic boundary layer scheme in conjunction with the Janjic Eta Monin–Obukhov surface layer scheme (Janjic, 2002), the WRF single-moment 6-class microphysics scheme for microphysics, the Grell–Devenyi scheme for clouds (Grell and Dévényi, 2002), and the new version of the rapid radiative transfer model for both shortwave and longwave radiation.

The coupler

        We have implemented the C-Coupler2 to couple the MITgcm and the Polar WRF model. The first version (C-Coupler1) includes features such as a flexible coupling configuration and 3-D coupling capability (Liu et al., 2014). Two coupled models have been built using the C-Coupler1. The first is a coupled climate system model version FGOALS-gc at the Institute of Atmospheric Physics, Chinese Academy of Sciences. The FGOALS-gc can achieve exactly the same (bitwise identical) simulation results as the same model components with a different coupler, the CPL6 (Liu et al., 2014). The second is a regional coupled model FIO-AOW (Zhao et al., 2017) which includes an atmosphere model WRF, an ocean model POM (Princeton Ocean Model) and a wave model MASNUM (Yang et al., 2005).

        The second version of the C-Coupler family, the CCoupler2 (Liu et al., 2018), is equipped with many advanced functions, including (1) a common, flexible, user-friendly coupling configuration interface; (2) the capability of coupling within one executable or the same subset of Message Passing Interface (MPI) processes; (3) flexible and automatic coupling procedure generation for any subset of component models; (4) dynamic 3-D coupling that enables convenient coupling of the field on 3-D grids with time-evolving vertical coordinate values; (5) non-blocking data transfer; (6) model nesting; (7) increment coupling; and (8) adaptive restart capability (Liu et al., 2018).