of the coordinate system in which the velocity
components were computed.
The large eddy energy regions were: along ocean
western boundaries and their seaward extensions;
the equatorial zones of the Tropical Instability
Waves (Baturin and Niiler, 1997); and winddriven jets (Ralph et al., 1997). The subtropical
frontal zones in the western Pacific and Atlantic
displayed relative maxima, as did the Canary Current system in the North Atlantic. The Circumpolar Current system displayed regions of high
energy, both due to the eastward extensions of
Malvinas/Brazil Confluence and the Aghulas Current and the effects of the topographic ridges
southeast of New Zealand and the central South
Indian Ocean. The lowest eddy energy regions
were in the northeastern North Pacific, the region
north of Iceland and the central tropical South
Atlantic. The South Atlantic minimum is based on
a small number of observations. Figure 4.1.6 presents a challenge to eddy-resolving ocean models,
as these should be able to duplicate the observed
regions of both high and low eddy energy.
The eddy energy distribution obtained by
drifters can differ from those developed by ship
drift or from satellite altimeters in several important ways. Direct observations of velocity with
drifters have the capability to resolve very low levels of currents with a variance of about 2 cm
2 s
92
,
or the sum of the squares of the instrumental
errors due to ARGOS fixes (1 cm s
91 , averaged
over a day) and unknown slip through the water
due to the shear of currents between the drogue
and the float (1 cm s
91
). This estimate is a lower
bound, because the drifter eddy energy uncertainty
is also a function of the number of independent
samples within the resolving grid and the biases
due to over-sampling convergent regions. Unknown
scales of the mean circulation, on scales smaller
than a bin, can alias the mean shear within a bin
into eddy energy. When comparing the ship-driftobserved eddy energy to drifter observations in the
northeast Pacific, the ship-drift observations show
a very large area of nearly constant eddy energy of
about 400 cm
2 s
92 (Wyrtki et al., 1976), while the
drifter data show that region to contain an eddy
minimum of less than 25 cm
2 s
92
. Thus, ship-drift
random error of about <20 cm s
91 appeared to
have produced large regions of constant eddy
energy that have been interpreted as ocean current
variability. The eddy energy computed from
altimeters can miss the 100–300 cm
2 s
92 variability
due to wind-driven currents. The uncertainty of
eddy energy computed from satellite altimeter data
is restricted to the geostrophic part of the currents.
This uncertainty should be a function of latitude,
the filtering length along the satellite track, the
number of independent estimates of sea-level
slope at the track cross-over, and the angle of that
cross-over (Strub et al., 1997).
The quantitative comparisons of drifter and
satellite altimeter-derived eddy distributions of the
world are ongoing areas of fruitful research. Comparisons of individual satellite altimeter-derived
currents with drifter-observed currents from the
Gulf of Mexico demonstrate an 85% coherence
between the two observations, provided an appropriate spatial scale of smoothing (110 km) is
adopted for the satellite data processing (Ohlmann
et al., 2001). This spatial scale of smoothing is
different in different latitude bands, increasing
from 55 km at 50°N to 120 km at 5°N (Barth
et al., 2001). The combined satellite sea-level
observations and the direct observations of circulation can now produce truly global, accurate
maps of ocean eddy energy.
4.1.4 The wind-driven Ekman currents
Ekman (1905) presented the first modern fluid
mechanical theory of wind-driven ocean currents
on the rotating globe. His calculations showed
that the large spatial scale, steady winds drive a
surface current at an angle to the right of the wind
in the northern hemisphere and to the left in the
southern hemisphere. The Coriolis force, coupled
to ocean turbulence, restricted the currents to a
shallow surface layer, now called the ‘Ekman
layer’. In the northern hemisphere this wind-driven
current rotates progressively to the right with
increasing depth through this turbulent layer. The
vertical average of the horizontal velocity was precisely at right angles to the wind, independent of
the depth of the Ekman layer or the nature of the
turbulence therein. Observations with current
meters have verified certain aspects of Ekman’s
theory, such as the rotation of currents with depth
and the vertical average velocity to the right of the
surface winds (Davis et al., 1981; Price et al.,
1987; Chereskin, 1995). Drifter data offer further
tests of Ekman’s theory, especially regarding the
strength of the wind-driven currents, the depth of
4.1 The World Ocean Surface Circulation
201
Niiler
components were computed.
The large eddy energy regions were: along ocean
western boundaries and their seaward extensions;
the equatorial zones of the Tropical Instability
Waves (Baturin and Niiler, 1997); and winddriven jets (Ralph et al., 1997). The subtropical
frontal zones in the western Pacific and Atlantic
displayed relative maxima, as did the Canary Current system in the North Atlantic. The Circumpolar Current system displayed regions of high
energy, both due to the eastward extensions of
Malvinas/Brazil Confluence and the Aghulas Current and the effects of the topographic ridges
southeast of New Zealand and the central South
Indian Ocean. The lowest eddy energy regions
were in the northeastern North Pacific, the region
north of Iceland and the central tropical South
Atlantic. The South Atlantic minimum is based on
a small number of observations. Figure 4.1.6 presents a challenge to eddy-resolving ocean models,
as these should be able to duplicate the observed
regions of both high and low eddy energy.
The eddy energy distribution obtained by
drifters can differ from those developed by ship
drift or from satellite altimeters in several important ways. Direct observations of velocity with
drifters have the capability to resolve very low levels of currents with a variance of about 2 cm
2 s
92
,
or the sum of the squares of the instrumental
errors due to ARGOS fixes (1 cm s
91 , averaged
over a day) and unknown slip through the water
due to the shear of currents between the drogue
and the float (1 cm s
91
). This estimate is a lower
bound, because the drifter eddy energy uncertainty
is also a function of the number of independent
samples within the resolving grid and the biases
due to over-sampling convergent regions. Unknown
scales of the mean circulation, on scales smaller
than a bin, can alias the mean shear within a bin
into eddy energy. When comparing the ship-driftobserved eddy energy to drifter observations in the
northeast Pacific, the ship-drift observations show
a very large area of nearly constant eddy energy of
about 400 cm
2 s
92 (Wyrtki et al., 1976), while the
drifter data show that region to contain an eddy
minimum of less than 25 cm
2 s
92
. Thus, ship-drift
random error of about <20 cm s
91 appeared to
have produced large regions of constant eddy
energy that have been interpreted as ocean current
variability. The eddy energy computed from
altimeters can miss the 100–300 cm
2 s
92 variability
due to wind-driven currents. The uncertainty of
eddy energy computed from satellite altimeter data
is restricted to the geostrophic part of the currents.
This uncertainty should be a function of latitude,
the filtering length along the satellite track, the
number of independent estimates of sea-level
slope at the track cross-over, and the angle of that
cross-over (Strub et al., 1997).
The quantitative comparisons of drifter and
satellite altimeter-derived eddy distributions of the
world are ongoing areas of fruitful research. Comparisons of individual satellite altimeter-derived
currents with drifter-observed currents from the
Gulf of Mexico demonstrate an 85% coherence
between the two observations, provided an appropriate spatial scale of smoothing (110 km) is
adopted for the satellite data processing (Ohlmann
et al., 2001). This spatial scale of smoothing is
different in different latitude bands, increasing
from 55 km at 50°N to 120 km at 5°N (Barth
et al., 2001). The combined satellite sea-level
observations and the direct observations of circulation can now produce truly global, accurate
maps of ocean eddy energy.
4.1.4 The wind-driven Ekman currents
Ekman (1905) presented the first modern fluid
mechanical theory of wind-driven ocean currents
on the rotating globe. His calculations showed
that the large spatial scale, steady winds drive a
surface current at an angle to the right of the wind
in the northern hemisphere and to the left in the
southern hemisphere. The Coriolis force, coupled
to ocean turbulence, restricted the currents to a
shallow surface layer, now called the ‘Ekman
layer’. In the northern hemisphere this wind-driven
current rotates progressively to the right with
increasing depth through this turbulent layer. The
vertical average of the horizontal velocity was precisely at right angles to the wind, independent of
the depth of the Ekman layer or the nature of the
turbulence therein. Observations with current
meters have verified certain aspects of Ekman’s
theory, such as the rotation of currents with depth
and the vertical average velocity to the right of the
surface winds (Davis et al., 1981; Price et al.,
1987; Chereskin, 1995). Drifter data offer further
tests of Ekman’s theory, especially regarding the
strength of the wind-driven currents, the depth of
4.1 The World Ocean Surface Circulation
201
Niiler