The Response of the Indo–Pacific Throughflow to Interannual Variations in the Pacific Wind Stress. Part II: Realistic Geometry and ECMWF Wind Stress Anomalies for 1985–89

Abstract

The effect of interannual variability in wind stress on the transport between the Pacific and Indian Oceans through the Indonesian archipelago is investigated using a multilevel, numerical, general circulation model (GCM). The experiments are conducted in a spinup, anomaly mode: the GCM is initially at rest with a horizontally uniform stratification typical of tropical oceans. The wind stress anomalies are derived from the European Centre for Medium-Range Weather Forecasts 1000-mb winds for 1985?89. The modeled variability is sensitive to the specification of bottom topography and archipelago geometry. Two model configurations are considered: (i) flat-bottomed with the option of sills within a simplified, idealized archipelago and (ii) realistic bottom topography and a complex archipelago. In the first model, there is northward depth-integrated transport anomaly lasting 2½ years with phase 6 months ahead of the collapse of the equatorial easterlies prior to the 1986?87 El Niño. The peak transport is about 5 Sv (Sv ≡ 106 m3 s?1), which is reduced to 3 Sv if shallow sills are present in the archipelago. There follows a southward transport anomaly, peaking at about 6 Sv during the ensuing U Nina. The baroclinic transport is consistent with cold and warm equatorial Rossby waves partially scattered into the archipelago during El Niño and La Nina, respectively. The reduction in depth-integrated throughflow generated by the reduction in southern midlatitude westerlies over 1985?86 results in a distinct signal in the baroclinic transport, and similarly in their increase again over 1988. Contrasting models with open and wholly blocked archipelagos yields heat content differences, measured as the temperature averaged over the upper 300 m, of typically O(0.05°C) in the west equatorial Pacific and O(0.3°C) in the Pacific western boundary layers and Indian Ocean. This suggests that throughflow variations could have implications for the timing of ENSO. In contrast, the second model has a very weak throughflow with peak-to-peak amplitude of only 3 Sv. Its archipelago is a relatively poor transmitter of equatorial waves, and its f/H contours are such that the depth-integrated throughflow is driven in part by the southerly extent of the South Pacific midlatitude wind stress anomalies. However, interestingly, the model still exhibits a decrease in throughflow in excess of 1 Sv prior to the onset of the 1986?87 El Niño due to variations in the southern midlatitude wind stress over the Pacific, which is also accompanied by a signature in the baroclinic transport. Also, contrasting homogeneous and stratified versions of the models demonstrates the importance of JEBAR over direct wind stress forcing of the depth-integrated throughflow. A closed form analytical expression for the transport in terms of an integral over a closed path formed by uniquely defined f/H contours, an island rule, is derived and used to help interpret the results.