Although much of this paper reviews published work, additional calculations are reported based on ocean and wind observations. These data sets and their processing are described in detail in the Appendix. The principal thermal data set used to construct the average annual cycle of thermocline depth and the resulting geostrophic currents (Section 4.2) is a compilation of historical XBT profiles by Donoso et al. (1994), and referred to here as the "AOML XBT data" (Appendix A.1). In some regions, directly measured velocities are available from the ships servicing the TAO moorings and from a few of the moorings themselves (Appendix C.1). The ship velocities are referred to here as the "Johnson ADCP data" (Johnson et al., 2002), and the moored velocities as the "TAO data". Surface drifters (Appendix C.4) also produce useful direct velocity measurements, however in many cases the very-near-surface currents sampled by the drifters are quite different than the much thicker currents due to thermocline topography, and do not represent the major transport patterns, especially under strong winds (e.g., the difference between Figs. 2 and 4, south of the equator). We also note that very few drifters have been deployed near the coast of central Mexico, so these data do not provide a good picture of the crucial region where the California Current meets the tropics. Satellite scatterometer winds are used to estimate Ekman pumping and to force a Rossby wave model (Section 4.2.1). These winds were sampled by the European Research Satellite (ERS) over the period 1991–2001, used to construct an average annual cycle, and also by the Quikscat (Seawinds) instrument for the period August 1999 through July 2002 (Appendix B). These are referred to here as the "ERS" or the "QuikSCAT" winds, respectively. Supplementary data sets (satellite SST, and sea surface height (SSH) from satellite altimetry) are used to augment the principal data sources for particular purposes (Appendix C).
In several plots, the geostrophic circulation is shown as contours of dynamic height, which is a convenient quantity because its contours are streamlines of the geostrophic flow, and its cross-stream gradient measures the flow speed; these relations are seen as the current vectors parallel to the dynamic height contours in Fig. 2. Dynamic height measures the vertically integrated density anomaly, expressing the fact that a less dense (i.e., warmer or fresher) water column of a given total mass stands taller than a denser one. Because deep flows are observed to be small in most regions, it is reasonable to assume that there are no pressure gradients at some deep reference level, otherwise these gradients would drive currents. Since pressure measures the mass of water above that level, equal pressure implies that the mass of water columns above the reference level must be the same, and thus the density of each column determines its height. Dynamic height is scaled to accurately represent sea surface height relative to a reference level. In the tropics, the density contrast between adjacent columns arises primarily because of thermocline variations: a water column with deep thermocline has a thick upper warm layer and is overall warmer than a column with shallow thermocline; thus it has higher dynamic height.
Return to previous section or go to next section
