Why much of the Atlantic circulation enters the Caribbean Sea and very little of the Pacific circulation enters the Sea of Japan

Abstract

A peculiar aspect of western marginal seas is that some of them accommodate large transports whereas others, apparently similar seas, accommodate relatively small transports. For instance, about 30–40 Sv of Atlantic upper water passes through the Caribbean and yet only 2 Sv or so passes through the Sea of Japan. The passages connecting the Caribbean to the Atlantic are somewhat larger than their Sea of Japan counterparts, but the dimensions of the gaps are too similar to each other to account for the large difference in the transports. Four different baroclinic models are used to investigate the problem. First, we used a wind-driven, layer-and-a-half isopycnal model with an idealized basin and an adjacent marginal sea separated from the main basin by a thin and long island. In this scenario, thermocline water enters the marginal sea via a gap to the south of the island and exits the sea via a gap to the north of the island. The mean latitudinal position of the marginal sea is gradually increased and the numerical transport across the sea is measured. We then compared the transport determined this way to that computed using three analytical baroclinic models: Godfrey's island rule (Godfrey, J. S. (1989). A Sverdrup model of the depth-integrated flow for the world ocean allowing for island circulations. Geophysical and Astrophysical Fluid Dynamics, 45, 89–112.), the ‘ β-controlled’ transport formula (Nof, D. (1993). The penetration of Kuroshio water into the Sea of Japan. Journal of Physical Oceanography, 23, 797–807), and the general gap transport formula (Nof, D., (1995). Choked flows and wind-driven interbasin exchange. Journal of Marine Research, 53, 23–48). The results were also compared to Minato and Kimura's (Minato, S., & Kimura, R. (1980). Volume transport of the western boundary current penetrating into a marginal sea. Journal of the Oceanographic Society of Japan, 36, 185–195) barotropic calculations. All four baroclinic models show that, when the marginal sea is situated in low and mid latitude, the penetration of thermocline water into the marginal sea decreases rapidly with increased latitude. By contrast, in equatorial marginal seas, the penetration of thermocline water increases with increasing latitude. Minato and Kimura's (1980) barotropic model shows a similar behavior except that the maximum transport occurs at a much higher latitude. Excellent agreement between the numerics and Godfrey's island rule formula is found in equatorial seas. As the latitude of the marginal sea increases to low and mid latitudes, the deviations from the numerics increase dramatically though the agreement is still very reasonable. Both the β-controlled formula and the gap formula are relatively accurate in mid latitudes when the sea is close to the separation latitude. The importance of frictional forces within the western boundary current results in the β-controlled formula failing for marginal seas situated far away from the separation latitude. Similarly, the general gap formula fails in equatorial latitudes where there is no western boundary current south of the island. On the basis of all four baroclinic models, it is suggested that the reason for the small transport within the Sea of Japan (relative to the Caribbean) is its relatively high latitude and not the size of the gaps connecting it to the Pacific Ocean (or frictional effects).