The Response of the Indo-Pacific Throughflow to Interannual Variations in the Pacific Wind Stress. Part I: Idealized Geometry and Variations

Abstract

The effect of interannual variations in the Pacific wind stress on the barotropic and baroclinic components of the flow between the Pacific and Indian Oceans through the Indonesian Archipelago, the Indo-Pacific Throughflow, is investigated using a numerical ocean general circulation model (GCM) with a simplified geometry. In agreement with the modified Island Rule, variations in the depth-integrated throughflow are generated by zonal wind stress variations over the Pacific at the latitudes of the tips of the Australian continent, assuming the Pacific basin is flat bottomed. Wind stress variations at other latitudes generate variations in the depth-integrated transport only if they produce a depth-integrated pressure drop along the oceanic eastern boundary through the archipelago. From the Island Rule, alongshore wind stress variations on the west coasts of Australia and South America produce direct variations, but the observed signal is weak at interannual periods. Baroclinic variations in the throughflow are generated by baroclinic waves entering the archipelago, or by interaction between the barotropic component and the sills within the archipelago. Shallow sills within the archipelago are found to only partially block the throughflow. The dynamical constraint, that quasi-steady flow is parallel to f/H contours, is relaxed by weak friction; the reduction in throughflow is only 30% for sills blocking 70% of the water column at 9°S. In the absence of sills, the throughflow response to a southern midlatitude wind stress anomaly is purely barotropic at interannual periods. Sills within the archipelago induce a baroclinic adjustment resulting in a surface trapping of the transport, which is in phase with the wind stress variations. Also, the depth-integrated pressure gradient through the archipelago produced by the topographic upwelling and downwelling is always directed to reduce the magnitude of the throughflow. For equatorial wind stress variations, the associated equatorial baroclinic Rossby waves are partially scattered into the archipelago. Sills within the archipelago block the transmitted equatorial Rossby waves, which would enhance the throughflow except that the baroclinic response to the topographic upwelling and downwelling negates the effect for part of the forcing cycle. Nonlinearity in the eastern equatorial response to an oscillating equatorial wind stress anomaly results in a mean throughflow from the Pacific to Indian Ocean. The combined effect of equatorial and southern midlatitude wind stresses, as typified by climatological mean values, yields a throughflow that is reduced by the inclusion of sills within the archipelago. Finally, in a comparison with the response in a GCM with a wholly blocked archipelago, the heat content anomaly (measured as the temperature averaged over the upper 300 m) in the equatorial Pacific is similar. However, the heat content anomaly in the archipelago and Indian Ocean is typically five to ten times larger than the equatorial anomaly difference between GCMs with an open/partially open and blocked archipelago. This is attributable to the difference in widths between equatorial and coastal baroclinic waveguides. The result suggests that the effect of variations in the throughflow on the Southern Oscillation is most likely felt in the archipelago and Indian Ocean.