Geologic studies indicate that prior to ~40 Ma the Drake Passage was closed and the Central American Isthmus was opened. The effect of these changes has been examined in an ocean general circulation model. Several sensitivity experiments were conducted, all with atmospheric forcing and other boundary conditions from the present climate, but with different combinations of closed and open gateways. In the first experiment, the only change involved closure of the Drake Passage. In agreement with earlier studies the barrier modified the geostrophic balance that now maintains the circumpolar flow in the southern ocean, with the net effect being decreased transport of the Antarctic Current and an approximate fourfold increase in outflow of Antarctic deep-bottom waters. The very large increase in Antarctic outflow suppresses North Atlantic Deep Water (NADW) formation. In addition to corroboration of results from earlier studies, our simulations provide several new insights into the role of a closed Drake Passage. A more geologically realistic closed Drake/open central American isthmus experiment produces essentially the same pattern of deepwater circulation from the first experiment, except that Antarctic outflow is about 20% less than the first experiment.

The resultant unipolar deepwater circulation pattern for the second experiment is consistent with paleoceanographic observations from the early Cenozoic. A third experiment involved an open Drake and open central American isthmus. In this experiment, Antarctic outflow is diminished to slightly above present levels but NADW production is still low due to free exchange of low-salinity surface water between the North Pacific and North Atlantic. The low level of thermohaline overturn should have reduced oceanic productivity in the Oligocene (~30 Ma), a result in agreement with geologic observations. Finally, simulations with an energy balance model demonstrate that the changes in surface heat flux south of 60¿S due to breaching of the Drake barrier do not result in temperature changes large enough to have triggered Antarctic glaciation. This last result suggests that some other factor (CO2?) may be required for Antarctic ice sheet expansion in the Oligocene (~30--34 Ma). Our results lend further support to the concept that even in the absence of changing boundary conditions due to ice sheet growth, variations in the geometry of the ocean basins can significantly influence ocean circulation patterns and the sediment record. The results also suggest that the primary polarities of the Cenozoic deepwater circulation may have been controlled by opening and closing of these two gateways. ¿ American Geophysical Union 1993