Modelling of Ocean Turbulence Effects of resolution on predictive power of ocean models

The oceans play a large role in global climate models by distributing physical quantities like heat around the Earth. It is therefore important that ocean models used in large-scale climate simulations produce accurate and reliable results. Leading global ocean models like the Community Earth System Model feature a horizontal resolution of 1 ◦ and use turbulent viscosity parameterization of small-scale turbulence grounded theoretically in Prandtl’s mixing length model. In the present work, an idealized ocean has been numerically modeled using a coarse-resolution model and a high-resolution model with horizontal resolutions of 1 ◦ and 0 . 1 ◦ , respectively. The coarse-resolution model uses turbulent viscosity parameterization of small-scale turbulence. The results of these simulations have been used to compare the general mean flow characteristics of high- and low-resolution models and to investigate whether the assumptions behind Prandtl’s mixing length model are representative for an ocean based on the high-resolution model. The study finds that significant mean-flow structure found in the high-resolution model is absent in the coarse-resolution model and that the width of the western boundary current is underestimated in the coarse-resolution model. This study further finds that Prandtl’s mixing length model fails to predict the magnitude and structure of small-scale turbulence based on the turbulent viscosity parameterization via the mean flow.