The y-axis in Fig. 4 may be regarded as a measure of variability in the Pacific, and it is interesting to consider the time spiral on Fig. 4, the streamfunction in Fig. 3 would produce. The result for two choices of y_N is shown in Fig. 5, in which the fractional composition is contoured as a function of time and throughflow magnitude. For y_N = 2.25｡?, the latitude of the northern tip of Halmahera, the time spiral would increasingly move into the region where the throughflow was wholly fed by the SEC, see Fig.5a. If the throughflow magnitude remains small, then the composition is very sensitive to conditions in the western Pacific and easily flips between extremes of wholly SP- and NP-fed flows. This seems somewhat unrealistic, and inconsistent with observations. If y_N = 4.25｡?, corresponding to nonlinear effects shifting the northern latitude of the SEC retroflection by 2｡?f latitude, then the throughflow is predominantly fed by the Mindanao Current for a wide range of throughflow magnitudes, see Fig. 5b, and so the time spiral in Fig. 4 would tend to lie below the diagonal.

Fig. 5　The fractional composition resulting from applying (1) to the FSU streamfunction shown in Fig. 3 is contoured as a function of time and throughflow. In (a) y_N is the northern tip of Halmahera at 2.25°N corresponding to the assumption that the dynamics are linear and relative vorticity is destroyed at the latitude of creation in the western boundary layer. In (b) y_N is 2° of latitude further north under the assumption that nonlinear overshooting of the western boundary currents and distortion of the streamlines by the Mindanao and Halmahera Eddies effectively shifts the latitude of the northern extent of the SEC retroflection by this amount.

BACK　　　CONTENTS　　　NEXT