Assessing impact of mesoscale eddy processes in coarse resolution ocean models

Mesoscale eddies, not only in the Southern Ocean but globally, play a vital role in mixing and transporting climatically important tracers such as heat and carbon dioxide. Nearly all the ocean models used for climate studies run at a non-eddy-resolving horizontal resolution of 1° or coarser, and therefore the effects of eddies (advection and mixing) are parameterised in these models. In some models, this parameterisation is employed as a spatially uniform value while other models employ varying advection and mixing parameterisations. However, observations show that it is spatially non-uniform. Recently developed, suppressed mixing length theory takes into account eddy propagation relative to the background mean flow which suppresses the mixing rates - Ferrari and Nikurashin, (2010). Observations have also shown the evidence of suppression of mixing (Naveira Garabato et al., 2011; Roach et al., 2018). The research on suppressed eddy mixing has revealed that mixing rates vary spatially and impact tracer transport. An improved parameterisation of eddy mixing which considers the spatial variability of eddy mixing could improve our understanding of the response of the global ocean to tracer uptake and climate change. Here we test the competing parameterisations. We do this assessment in Python Ocean Model (pyOM2.2), with a closed energy cycle where the energy available for the mixing in the ocean is only controlled by the external energy input from the atmosphere, tidal system, and internal exchanges. Currently none of the other ocean models used for climate projections use such an energy consistency framework. First, the parameterisation of suppressed eddy mixing is implemented in an idealised model configuration of the Antarctic Circumpolar Current in a zonally symmetric periodic channel. This study of an idealised channel model suggests that suppressed eddy mixing parameterisation performs better than the traditionally used parameterisations (spatially uniform and unsuppressed), in the context of tracer uptake. Suppressed mixing parameterisation relies on the energy consistency framework of pyOM2.2 and captures the wind sensitivity information and suppression effect. Therefore, suppression coupled with improved eddy advection parameterisation reduces the tracer uptake difference with high-resolution simulation to less than 2%. Secondly, we found from the findings of the first scientific chapter that the inclusion of improved eddy advection term (spatially uniform) improves the slopes of isotherms and leads to accurate estimates of the tracer transport. This motivated us to use a realistic model configuration of the global ocean with 2° resolution and test different values of eddy advection and eddy mixing (which are spatially uniform) to determine which eddy effect, advection, or mixing, plays the main role in controlling the tracer distribution on different time scales. We found on decadal time scales eddy mixing and eddy advection both play equally important roles in controlling tracer distribution. However, on the time scale of a hundred-year eddy advection dominates eddy mixing to control the global tracer uptake. Finally, suppressed eddy mixing is parameterised in the global ocean. The suppressed eddy mixing parameterisation shows a rich spatially varying structure of eddy mixing in the western boundary currents due to the strong presence of jets and eddies. Suppression plays an important role not only in the Southern Ocean but also in the global ocean. Overall, this work provides evidence that in a 2°x2° pyOM2.2 model that mixing parameterisation schemes give limited scope to improve the tracer uptake estimates on decadal and century time scales and that the key uncertainty from Chapter 2 and Chapter 3 is in eddy advection. Implementing improved parameterisations of advection and mixing in an energetically consistent ocean model will lead to improved representation of mesoscale eddy effects in the coarse resolution ocean models and help to develop a better understanding of the response of the Southern and global oceans to climate change and to accurate projections of carbon feedbacks in the future.