To do so, we shall consider the model shown in Fig. 3ab, where the top view (shown in the central panel of Fig. 3b) corresponds to a typical observational view of the top of the pool (shown in the upper panel of Fig. 1), and the side view (shown in the lower panel of Fig. 3b) corresponds to a typical observational view of the side of the pool (shown in the upper panel of Fig. 2). For simplicity, we focus on the most intense El Ninos of the century (e.g, those of 1982 and 1997) because these events correspond to an almost complete relaxation of the winds. The model consists of a warm water (σt＜ 22.0) tongue which is usually held against the western boundary of the Pacific by the westward wind drag but is now free to move eastward. The warm water tongue overlies an ocean containing an“intermediate water”(22 ＜ σt ＜ 25) which is allowed to move1 and an infinitely deep inactive cold lower layer (25 ＜ σt). The nonh-south extent of the pool is a few Rossby radii (which, for a typical density anomaly of one part per thousand and a depth of 80m, is roughly 212km).
Fig. 3a. A schematic three-dimensional view of the two-and-a-half-layer model under study. The light advancing pool (red) penetrates into a two-layer ocean consisting of an intermediate upper layer (clear) and an infinitely deep lower layer (blue). Far downstream (x → ∞) both the intermediate water and the deep water are at rest. To compensate for the volume displaced by the eastward moving pool, the intermediate water is forced to dive under the pool and move westward. The deep water is at rest everywhere. “A” is the “nose” of the intruding pool.
In our scenario, the eastward motion results from excess pressure on the westem side which is created when the winds subside. It is ultimately balanced by a form-drag exerted on the pool by the intermediate water which, in order to compensate for the volume displaced by the eastward moving pool, is forced to dive under the pool and flow westward. In accordance with the observations that during strong El Ninos the Equatorial Undercurrent is drastically reduced (see e.g., Firing et al. 1983, Halpern 1987) or even completely disappears (the 1982 El Nino) we shall neglect the motions in the deep layer (σt ＞ 25).
1 Note that this“intermediate water”has no relationship to the much deeper water frequently referred to as“intermediate water”. Furthermore, note that the intermediate water under the pool is sometimes referred to (by other authors) as the“barrier layer.” We chose to use the term “intermediate water”“rather than”barrier layer”because, in our work, the extension of the layer ahead of the pool and to the sides of the pool is exposed to the atmosphere and, consequently, does not serve as a“barrier.
