Terence O'Kane, Russell Fiedler, Mark A Collier, Vassili Kitsios

In climate model configurations, standard approaches to the representation of unresolved, or subgrid scales, via deterministic closure schemes are being challenged by stochastic approaches inspired by statistical dynamical theory. Despite gaining popularity, studies of various stochastic subgrid scale parameterizations applied to atmospheric climate and weather prediction systems have revealed a diversity of model responses, including degeneracy in the response to different forcings and compensating model errors, with little reduction in artificial damping of the small scales required for numerical stability. Due to the greater range of spatio-temporal scales involved, how to best sample subgrid fluctuations in a computationally inexpensive manner, with the aim of reduced model error and improvements to the simulated climatological state of the ocean, remains an open question. While previous studies have considered perturbations to the surface forcing or subsurface temperature tendencies, we implement an energetically consistent, simple, stochastic subgrid eddy parameterization of the momentum fluxes in regions of the three-dimensional ocean typically associated with high eddy variability. We consider the changes in the modelled energetics of low-resolution simulations in response to stochastically forced velocity tendencies whose perturbation statistics and amplitudes are calculated from an eddy resolving ocean configuration. Kinetic energy spectra from a triple-decomposition reveal a systematic redistribution from the seasonal (climatological minus mean) potential energy to preferentially generate small scale transient kinetic energy while the total energy spectra remains largely unchanged. We show that stochastic parameterization generally improves model biases, noticeably so for the simulated energetics of the Southern Oceans.