Figure 10. Same as Fig. 9 except for the second EOF. Note the pronounced increase in JFM rainfall in the region north of Hawaii from 1990 to 1997.
The wave mode involves low frequency Rossby waves generated by anomalous wind stress curl or which are radiated from the eastern boundary after forcing in the equatorial or coastal waveguides. This mode contributes anomalous velocities which distort the gyre, resulting in differential meridional advection of mean gyre properties over a space scale set by the wavelength of the Rossby wave.
The freshening trend observed at ALOHA is temperature compensated. Thus, if it is due to the subduction mode, the increased rainfall in the subduction zone must be compensated by greater heat losses from the upper ocean. This is consistent with the idea that the storm tracks either shifted or the storms intensified. Is the delay between the time history of EOF2 consistent with this scenario?
Huang and Russell (1994) computed the transit time from the ventilation regions for various isopycnals to ALOHA, finding that the delay is from about 1 year for densities between 24 and 24.6 σθ, to about 3-4 years for density levels between 25 and 26 σθ. (Bingham's [1998] estimates are somewhat longer.) The penetration time for the salinity signals at ALOHA if they are foced in the storm track of the midlatitude North Pacific is consistent with this calculation. These estimates assume that the geostrophic gyre circulation and the subduction zones do not vary, however it is unlikely that the trajectories for anomalous water parcels are stationary because the wave mode may also be present. It is difficult to take the interaction of the two modes into account without recourse to a more sophisticated model, such as that of Zhang and Liu (1998).
Is the freshening trend also consistent with gyre wobble? For the density surfaces affected by the freshening trend (24.2-26.6 σθ), first baroclinic wave velocity perturbations would be of the same sign and in phase vertically, though of decreasing magnitude with depth. Given the mean salinity distributions on this range of density surfaces in the broad region around Hawaii, the wave perturbation velocity would have to be from the south or southeast. Is this consistent with the signals seen in other variables at ALOHA, given their mean spatial distribution?
A multivariate EOF analysis was performed on the HOT salinity, potential vorticity and dissolved oxygen observations as a function of potential density.
