Partially-Averaged Navier-Stokes Equations Model for Prediction of
Turbulent Ocean Flows
Abstract
The accurate simulation of oceanic turbulence is crucial to
understanding global ocean circulation, impacting accurate prediction of
global warming effects and national security problems. Nonetheless, high
fidelity simulations are difficult due to the ocean’s complex physics
and broad range of spatial and temporal scales. These range from the
smallest dissipative scales with typical lengths of millimeters, to the
largest energetic mesoscale eddies with characteristic wavelengths
exceeding tens of kilometers. Whereas the accurate parameterization of
all flow scales with the Reynolds-averaged Navier-Stokes equations
(RANS) is nearly impossible, resolving all scales of turbulence through
a direct numerical simulation (DNS) is beyond the capabilities of the
most powerful supercomputers in the foreseeable future. Hence, efficient
parameterizations are needed. In this work, we extend the bridging
partially-averaged Navier-Stokes equations (PANS) model to ocean flows.
This parameterization operates between RANS and DNS, and aims only to
resolve the scales not amenable to modeling in order to increase the
efficiency of ocean computations. We also propose a PANS scale-aware
closure to model the unresolved scales. The initial validation tests of
the new PANS model include two representative ocean channel flows. The
results confirm the potential of the new parameterization to predict
oceanographically relevant turbulence efficiently.