Towards high-fidelity simulations of coastal submesoscale baroclinic instabilities with MPAS-O (vE3SM3.0.0) Part II: Realistic experiments

Abstract. River plumes can strongly influence coastal dynamics but are poorly represented in global ocean models due to stringent resolution requirements. Following the idealized simulations presented in a companion paper (Hinson et al., 2026), this study presents submesoscale-permitting realistic simulations using unprecedented regional refinement with MPAS-O. We assess MPAS-O's ability to represent river plume dynamics by comparing with a validated ROMS configuration focused on the Mississippi–Atchafalaya (M-A) plume in the northern Gulf of Mexico (GoM). The variable-resolution mesh spans 1.4 km in the GoM to 100 km in the Pacific and Indian Oceans. We incorporate several improvements to the representation of river forcing in MPAS-O: the pseudo point source treatment of runoff, river temperature prescribed from air temperature, and passive tracers to track freshwater transport. MPAS-O generates vigorous submesoscales in the open GoM, as quantified by probability density functions of surface relative vorticity and divergence. While the model qualitatively captures the M-A plume's seasonal evolution, it does not realistically reproduce the summer submesoscale eddy field on the Texas–Louisiana shelf. A persistent brackish lens associated with reduced vertical mixing forms and is present throughout the M-A plume. We hypothesize that this arises from interacting parameterization and structural design choices within MPAS-O that affect plume stratification and instability growth under realistic forcing, rather than from deficiencies in numerics. Our results demonstrate MPAS-O's potential for simulating submesoscale processes in the open ocean while highlighting challenges in using regional refinement to model coastal dynamics.