Characteristics of atmospherically forced mid-latitude,
subinertial barotropic motions in the oceans are studied
with measurements of bottom pressure, barotropic
currents, surface winds and air pressure from the central
North Pacific. The bottom pressure is found to be
dominated by atmospherically forced variability within the
periods of about 1.5 to 300 days. The bottom pressure
exhibits characteristics of evanescent, non-free-wave
motions at periods shorter than the shortest allowable free
Rossby wave (from linear theory this cutoff period is about
2.6 days for the study area). At longer periods, the bottom
pressure variability is consistent with the propagation of
topographic Rossby waves. The main features of the
bottom pressure energy densities and the patterns of
coherence of bottom pressure with local and non-local
atmospheric variables are predicted reasonably well by
simple analytical models, both above and below the Rossby
wave cutoff period. The spatial patterns of coherence
between bottom pressure and the wind stress curl are
shown to be the result of the generation of Rossby waves
from specific locations, as opposed to a uniform generation
of a spectrum of Rossby waves throughout the North
Pacific. There is clear evidence for the existence of "hot
spots" at which the atmosphere is strongly forcing oceanic
Rossby waves that then propagate to all five bottom
pressure measurement sites.

In contrast to the bottom pressure, the barotropic current
energy densities and their coherences with the atmospheric
variables are predicted poorly by the simple models.
Filtering the barotropic currents by spatial averaging results
in a current field more clearly related to atmospheric
forcing that exhibits the signatures of large scale, westward
propagating topographic Rossby waves. Contrary to results
of previous observational studies and model predictions,
there is no clear indication that at longer periods, where
relative vorticity is weaker, a Sverdrup balance emerges
between the currents and local wind stress curl with large
area averages.

As opposed to simple analytical models, the sophisticated
POP numerical model predicts energy densities of the
observed bottom pressure exceptionally well and shows
high significant coherences between model and
observational bottom pressure with near zero phases. On
the other hand, energy densities of the barotropic currents
are not predicted as well by the POP model, although the
model's ability to reproduce the barotropic current field
improves significantly when the currents are filtered using
large spatial averages.

The models' inability to reproduce the barotropic currents
and the finer features of the observed bottom pressure well
are thought to be due to the models' simplified forcing, lack
or inadequate resolution of topography, and possibly
inaccuracies in the models' dissipation relative to that of the
real ocean.

More information is available at http://
www.soest.hawaii.edu/oceanography/gradstudents_html/
reka/