The performance of subgrid-scale models in large-eddy simulation in counter rotating flows

Langmuir turbulence is driven by Langmuir circulation (LC), which forms at the surfaces of rivers, lakes, bays, and oceans due to the interaction between wind-induced shear and surface gravity waves. In uniform shallow waters, LC can reach the bottom and influence the boundary layer near the bed. A large-eddy simulation (LES) study of LC in shallow water was conducted using the finite volume method and several subgrid-scale (SGS) models, each with different approaches to modeling near-wall eddy viscosity. The ratio of wave to wind forcing used in the LES was based on field data of full-depth LC observed in a 15-meter-deep coastal ocean column, as documented in prior studies. The analysis showed that the choice of SGS model significantly affects LC structure in the lower part of the water column. The simulations were assessed based on (1) velocity statistics of Langmuir turbulence and (2) the crosswind length scale and overall structure of LC cells. Among the models tested, the S-Omega SGS model—using eddy viscosity based on strain rate and vorticity magnitudes, with a Van Driest wall damping function—most closely matched pseudo-spectral LES results in terms of lateral cell size and structure. In contrast, the dynamic Smagorinsky model and the wall-adapting local eddy-viscosity model showed less agreement, exhibiting less coherent bottom convergence and weaker upward transport of slower downwind-flowing water. The S-Omega model also aligned well with real-world observations, accurately capturing the decay of Langmuir turbulence during surface heating events in coastal ocean settings.