Contact: Nils Brüggemann (firstname.lastname@example.org)
The effect of small- to sub-mesoscale turbulence on the ocean-atmosphere gas exchange
Variability on scales smaller than the local Rossby radius of 10 - 50 km -- i.e. the submesoscale -- are of major importance in the surface mixed layer of the ocean for realistic large-scale simulations of air-sea gas exchange. Therefore, they have to be parameterized in coarse ocean models to simulate realistic air-sea gas exchanges (e.g. Oschlies (2002)).
In this study, we investiagte two different parameterizations for mixed layer eddies. One is based on a scaling of the potential energy release, PER, (following Fox-Kemper et al., 2008). The other one is based on linear stabiltiy analysis, ALS, (Stone, 1966) and a scaling that uses the time and length scale of the fastest growing mode (Killworth, 1997). Highly resolved simulations of two different mixed layer type scenarios are used to evaluate the parameterizations for a wide range of dynamical regimes that may occur. The first scenario consists of an equilibrated statistically steady flow and the second on a re-stratifying density front.
Fig.1: Idealized simulation to evaluate MLEs in an equilibrated scenario (a) - (d) and the scenario of re-stratifying density front (e) - (h). (a), (e) and (f) denote temperature snapshots. (b) denotes surface vorticity (upper pannel) and divergence of the horizontal velocity (lower pannel) at the surface. (c) and (g) denote eddy stream function (upper pannel) and diapycnal diffusivity (lower pannel). (d) and (h) denote time series of global averages of the mean kinetic energy (MKE), eddy kinetic energy (EKE), available potential energy (APE) and total energy (TE).
It turns out that PER is better suited to capture the Ri-dependency of the magnitude of the eddy fluxes. On the other hand, the vertical structure of the meridional eddy fluxes predicted by ALS is more accurate than that of PER, while the vertical structure of the vertical eddy fluxes is well predicted by both parameterizations. Therefore, we suggest to use the magnitude of PER and the vertical structure functions of ALS for an improved parameterization of mixed layer eddy fluxes.PI's: C. Eden, H. Burchard
Fig.2: Evaluating the parameterizations in the equilibrated scenario (a) - (d) and the scenario of the re-stratifying density front (e) - (h). Red dots denote values for the diagnosed MLEs, black dots denote the PER parameterization, green dots the LSA parameterization and blue dots a numerically obtained higher order solution to the LSA problem. (a), (b), (e) and (f) show global mean values of the meridional (<v'b'>) and the vertical (<w'b'>) eddy fluxes. (c), (d), (g) and (h) show the vertical profiles of the meridional and vertical eddy fluxes for different Richardson numbers normalized by the maximum of the profile.