Development of a next-generation general ocean circulation model for the Great Lakes

Abstract. The Laurentian Great Lakes share several physical characteristics with the coastal ocean, including atmosphere–water interactions, rotational dynamics, and ice cover processes. However, their weak density stratification, relatively small surface area, and distinct seasonal mixing cycles pose unique challenges for numerical modeling. Modeling approaches and parameterizations developed for global applications, however, may yet provide valuable pathways for addressing persistent biases in lake models. To examine these possibilities, we develop a 3D hydrodynamic model for Lake Michigan-Huron (LMH) using the Modular Ocean Model version 6.0 coupled with the Sea Ice Simulator version 2.0 (MOM6-SIS2). Originally designed for global ocean and earth system modeling, MOM6 offers flexible vertical coordinate systems (VCSs) to maintain density gradients and improved handling of complex bathymetry, both potential advantages for application in inland water bodies like the Great Lakes. This is the first study to investigate MOM6-SIS2's ability to simulate key features of hydrography and circulation in freshwater systems under different VCSs. This study tested z∗ (depth-based) and hybrid (depth and isopycnal) VCSs. Simulations were performed for the years 2017 and 2018 and evaluated against in situ and remote sensing observations, as well as outputs from a contemporary Finite Volume Community Ocean Model (FVCOM) of LMH (LMH-FVCOM), used in an operational forecast system. MOM6-SIS2-LMH skillfully simulated daily averaged lake surface temperature (LST), vertical thermal structure, and ice concentration, with biases in LST and ice concentration generally below 0.5 °C and 2 %, respectively. It also produced comparable results to LMH-FVCOM in terms of LST, vertical thermal structure, and ice concentration. Both VCSs (z∗ and hybrid) successfully captured large-scale circulation patterns and seasonal overturning. The hybrid VCS, reduced excessive thermocline diffusion in deep waters, observed in both FVCOM and MOM6-SIS2 with z∗ VCS and allowed the model to maintain ecologically important deep cold water in the summer months. These improvements highlight the potential of MOM6-SIS2 to successfully simulate lake dynamics and offer the potential to more accurately resolve the delicate balance of thermal structure and mixing in stratified lake environments. However, the limited nearshore resolution resulting from MOM6's structured grid degraded the simulation of flows through the Straits of Mackinac, as well as nearshore temperature and water level variability.