Diapycnal mixing across density surfaces provides the main way to warm the cold deep waters which are formed by convection and sink at high latitudes, thus allowing them to rise through the thermocline and complete the lower limb of the "conveyor belt". Diapycnal mixing is an essential part of the ocean circulation, in particular the THC, and can affect surface conditions and climate change on decadal to centennial and longer time-scales.
The processes leading to mixing in the main thermocline involve random internal wave breaking (Gregg, 1989; Polzin et al., 1995), and to some degree also double-diffusive mixing (St. Laurent and Schmitt, 1999). Diffusivities in most of the main thermocline are typically of the order 10-5 m2s-1 or less (Gregg, 1989; Ledwell et al., 1993), whereas diffusivities of an order of magnitude larger are required to close the THC (Munk, 1966). Recent observations in the deep ocean have, however, found that turbulence is greatly enhanced 1,000 to 2,000 m above the bottom in regions of rough bottom topography (Polzin et al., 1997), with mixing rates reaching values of 1 to 3 x 10-4 m2s-1 (Ledwell et al., 2000). The most likely cause is internal waves generated by tidal flows over kilometre-scale bathymetric features. In addition to tidal-driven mixing, the calculations of Munk and Wunsch (1998) indicate that wind-driven mixing in the Antarctic Circumpolar Current contributes substantially. Preliminary estimates of mixing rates based on the internal wave parametrization suggest that the wind-driven mixing is more uniformly distributed throughout the water column (Polzin and Firing, 1997). Whereas the classical assumption of uniform mixing and upwelling led to an expectation of poleward interior flow in the deep ocean, the new data suggest downwelling in the interior, with upwelling confined to the many ocean bottom fracture zone valleys perpendicular to the ridge axis (Ledwell et al., 2000).
Evidence for a significant role for double diffusion in ocean mixing has emerged. The "salt finger" and "diffusive convection" processes transfer heat and salt at different rates resulting in a density transport against its mean gradient. New data suggest that salt fingers may have a somewhat subtle role over widespread areas of the sub-tropical gyres (St. Laurent and Schmitt, 1999). The implications of widespread double diffusion for the general circulation are, however, controversial. Gargett and Holloway (1992) found in a model study with a simplistic representation of double-diffusive mixing that there were dramatic changes in the circulation and water masses. Recent studies using a more conservative parametrization of double-diffusion suggest modest though still significant changes in the circulation (Zhang et al., 1998).
Diapycnal mixing in the ocean is usually associated with energy conversions, and parametrizations are often based on energy arguments. In their most simple form, these suggest formulating mixing coefficients in terms of stability frequency and/or Richardson number. Double-diffusive mixing has been included in a few models only to very a limited extent. The higher mixing rates found in the abyssal ocean over rough topography have so far not been included in the vertical mixing scheme of coupled climate models, but have been used with some effect on the deep circulation in ocean models (e.g., Hasumi and Suginohara, 1999).
The sensitivity of climate model results to the mixing parametrization is not fully clear. In a state of thermohaline equilibrium, the intensity of the meridional overturning circulation should be strongly dependent on the average internal mixing rate, and in ocean-only models the meridional overturning circulation often varies with a certain power of the average mixing rate (Zhang et al., 1999a). Some models indicate, however, that significant transports may be involved in wind-driven flows along isopycnals which outcrop in different hemispheres. The buoyancy changes are confined to the surface mixed layers, and little or no interior mixing is required. The scaling laws and dynamics of such flows have yet to be clarified, but they do raise the possibility of a meridional heat transport independent of the interior mixing rate.
In summary, the uncertainties associated with interior ocean mixing parametrizations are likely to be small for climate projections over a few decades but could be considerable over longer time scales.
Other reports in this collection