The Stability of NADMF under Mixed Boundary Conditions with an Improved Diagnosed Freshwater Flux

Abstract

Ocean general circulation models can be run using Haney boundary conditions (BCs) for both temperature and salinity, or using mixed boundary conditions, which consists of a Haney BC for temperature and a flux BC for salinity. A switch from Haney BCs to mixed BCs often causes the model North Atlantic Deep Water Formation (NADWF) to either collapse or intensify. Recently, Tziperman et al. found that the collapse was due to an unrealistic freshwater flux field diagnosed from a spinup using a too short relaxation timescale for salinity. They replaced the unrealistic freshwater flux with a more realistic freshwater flux diagnosed from a spinup using a longer relaxation timescale for salinity and found that NADWF stabilized. In this study, the author shows that mixed BCs are not suitable for studying the stability of the present ocean climate, regardless whether a realistic freshwater flux is realistic or not. Further, the instability associated with mixed BCs is due more to the use of a Haney BC for temperature than to an unrealistic freshwater flux. This is shown in a series of numerical experiments using a global Bryan?Cox ocean general circulation model. In these experiments, although a more realistic freshwater flux is used, NADWF is still very sensitive to a perturbation in high-latitude freshwater flux and to an enhancement of the implied hydrological cycle. This is because a Haney BC for temperature, when used with a flux BC for salinity, promotes a positive feedback between surface salinity and overturning. When the Haney BC for temperature is replaced by a Schopf BC, the overturning circulation associated with NADWF is quite stable.