Numerical modeling study of the circulation of the Greenland Sea

Abstract

This study is a simulation of the circulation of the Greenland Sea aimed at modeling some of the issues related to the Great Salinity Anomaly (GSA) and deep water formation using a primitive equation ocean general circulation model (Semtner, 1974). The features of the model include: (1) a high resolution, (2) real topography, (3) open boundaries at the south and north, and (4) temporally variable wind and thermohaline forcing. The model is used to study: (1) the spreading of a fresh water anomaly, (2) the mechanisms of cross frontal mixing that lead to deep water formation, (3) the general circulation of the deep and upper layers of the ocean and their dependence on wind and thermohaline forcing, and (4) the possible implications of meso-scale and large-scale variability on climate change. One of the major results of this work is the simulation of continental shelf waves propagating along the shelf slope of Greenland between 77$\sp\circ$N and 72$\sp\circ$N. Waves with a subinertial period of 17.2 hrs, a wavelength of 363 km, a phase speed of 586 cm/s and a group velocity of 409 cm/s, are found. Possible mechanism for generation of shelf waves is presented. It is suggested that some energy related with wave activity may support cross-frontal mixing in the East Greenland Current (EGC), where formation of the two main sources of North Atlantic Deep Water (e.g. Norwegian Sea Deep Water and Denmark Strait Overflow Water) have been reported. The results from the GSA simulation suggest that during the early stage of the GSA (e.g. during its propagation with the EGC to the south, in the late 1960s) when no observations are available, the fresh water signal is not being mixed into the interior circulation of the Greenland Sea gyre. The second experiment, representing recirculation of the GSA from the North Atlantic back into the Greenland Sea, in the late 1970s, shows freshening in the Greenland Sea gyre of comparable magnitudes ($-$0.05 to $-$0.1 psu) to the observed ones. These results agree with the earlier indirect measurements (Rhein, 1991; Schlosser et al., 1991) indicating dramatic reduction of deep water renewal in the Greenland Sea in the late 1970s and early 1980s. From the general circulation experiments it has been found that the ocean response to seasonal forcing is mainly barotropic. This implies a strong topographic control in the distribution of currents and hydrographic variables. Most of the areas of topographic steering which are simulated in the region have been reported in the literature. The so-called Molloy Deep eddy shows its direct dependence on the large scale dynamics affecting the northward flow of the West Spitsbergen Current (WSC), controlling this way a net mass transport into the Arctic Ocean. Simulations with different wind forcing suggest dependence of the Greenland Sea gyre circulation on the variations with time of the local wind forcing. Results indicate that monthly mean wind stress forcing probably underestimate wind forcing in the model. Analysis of surface, intermediate and deep ocean velocity fields compare reasonably well with observations.