Sea-level measurements along the western coast of the Americas have shown that there is a strong signal at ENSO frequencies (approximately 2π/2 yr−1 to 2π/5 yr−1) that propagates poleward at about 40 to 90 cm s−1. This ENSO sea level signal must be associated with ENSO coastal currents, but because adequate interannual current time series are unavailable, the structure and strength of these currents are not known. Coastal ENSO currents must be fundamentally affected by bottom topography and bottom friction, but previous theory has not taken these effects into account. A near-boundary numerical model with realistic bottom friction, stratification, and shelf and slope topography was therefore constructed to study ENSO coastal flow.
(i) At ENSO frequencies, previous results for models with no bottom topography and no bottom friction suggest that sea level should not propagate poleward. With realistic bottom friction and bottom topography coastal sea level propagates poleward at speeds similar to those observed. A simple mechanism based on the effect of bottom friction on offshore propagating geostrophic alongshore flow explains why coastal poleward alongshore propagation occurs. Calculations also show that the depths of the 20°C and 15°C isotherms also propagate poleward at approximately the observed speeds.
(ii) Sea levels change little across the shelf and slope at lower latitudes and slowly decrease in amplitude alongshore due to bottom friction. The small sea level change across the shelf and slope implies that the long sea level records available are useful for analysing the nearby deep ocean variability.
(iii) Lower-order deep-sea vertical modes incident at the equator are rapidly scattered mainly by bottom friction into other (higher) vertical modes. Scattering has two main effects. Those vertical modes equatorward of their critical latitudes propagate offshore as Rossby waves interfere with each other and produce a complicated deep-sea velocity field, especially at low latitudes where most of the vertical modes are propagating offshore. Those vertical modes poleward of their critical latitudes only exist as coastally trapped motion and give rise to a trapped ENSO jet over the continental slope. This jet has an amplitude peak of order 20 cm s−1 and is trapped within about 500 m of the bottom. The jet core is approximately 180° out of phase with near-surface currents over the continental shelf and slope. Therefore, during (say) the El Niño part of the ENSO cycle when the sea level is high and the near-surface flow over the continental shelf and slope tends to be poleward, the flow in the jet core tends to be equatorward. Present observations are inadequate to prove or disprove the existence of this ENSO continental slope jet. Due to bottom friction, the alongshore velocity decreases shoreward of the shelf break and is negligible at the coast.
(iv) Critical latitudes for vertical modes change when the coastline angle changes and so motion near the boundary is affected by coastline angle. When the coastline is less meridional, coastal sea level and the 20°C and 15°C isotherm depths propagate poleward more rapidly (although still at approximately the observed speeds). The ENSO jet has its maximum amplitude nearer the equator.
(v) In the biologically important top 100 m or so of the ocean alongshore particle displacements seaward of the shelf can be ∼1000 km. Interannual near-surface alongshore currents over the continental shelf and slope lead coastal sea level by several months.