of high-resolution models since the first successful
studies for idealized basins in the mid-1970s.
A particular focus has been on the questions of
the generation mechanisms of the observed eddy
spectrum and its role in the time-mean transports
of heat and other properties.
High-resolution global and basin-scale models
have had growing success in reproducing the eddy
variability that accounts for most of the signal in
SSH variability maps from T/P. Early WOCE studies with spherical-coordinate models qualitatively
reproduced the localized regions of strong variability in the North Atlantic at 1/3°, in the Southern
Ocean at 1/4°, and in the World Ocean at 1/2°
(Semtner and Chervin, 1988; Bryan and Holland,
1989; FRAM Group, 1991). Grid refinements
have been made by switching to Mercator and/or
finer grids to obtain an average of 1/6° or 1/8°
globally and below 1/10° in the North Atlantic.
Results show that the magnitudes of variability
converge toward the observed values as grid size is
decreased. Figure 2.2.8 is an example from an
evaluation of the circulation in the Pacific Ocean
seen in two versions (1/4° and 1/8° mean grids) of a
global model (Saunders et al., 1999). As in similar
comparisons made by McClean et al. (1997), all
model cases roughly reproduce the distribution of
SSH variability as derived from the T/P altimeter.
Increased resolution yields an enhanced realism in
regions of strong currents in mid-latitudes, but the
models tend to underestimate the T/P variability in
the subtropical oceans, away from the strong currents. The reader is referred to various maps of
SECTION 2 OBSERVATIONS AND MODELS
70
Fig. 2.2.6 Transport of the DWBC between 1000 m
and 3100 m depth in the equatorial Atlantic, from several
year-long time series of moored current meters (Fischer
and Schott, 1997) and a DYNAMO model simulation
with daily wind fields from ECMWF. Moored results
shown are the combined annual and semiannual
harmonics of transport variability for two different
observational periods. Model curves show the mean
seasonal cycle as obtained by averaging over 5 years of
integration and a harmonical analysis similar to the
observations.The annual mean southward transports are
13.0 Sv for the model, 12.4<6.8 Sv and 13.5<5.3 Sv for
the two measurement periods.
Fig. 2.2.7 Difference in 5-year mean northward heat transport between winter (Jan–Feb–Mar) and summer
(Jul–Aug–Sep) for three models based on different vertical coordinate schemes and, associated with that, different
representations of topography and sub-grid-scale mixing (Böning et al., 2000).
studies for idealized basins in the mid-1970s.
A particular focus has been on the questions of
the generation mechanisms of the observed eddy
spectrum and its role in the time-mean transports
of heat and other properties.
High-resolution global and basin-scale models
have had growing success in reproducing the eddy
variability that accounts for most of the signal in
SSH variability maps from T/P. Early WOCE studies with spherical-coordinate models qualitatively
reproduced the localized regions of strong variability in the North Atlantic at 1/3°, in the Southern
Ocean at 1/4°, and in the World Ocean at 1/2°
(Semtner and Chervin, 1988; Bryan and Holland,
1989; FRAM Group, 1991). Grid refinements
have been made by switching to Mercator and/or
finer grids to obtain an average of 1/6° or 1/8°
globally and below 1/10° in the North Atlantic.
Results show that the magnitudes of variability
converge toward the observed values as grid size is
decreased. Figure 2.2.8 is an example from an
evaluation of the circulation in the Pacific Ocean
seen in two versions (1/4° and 1/8° mean grids) of a
global model (Saunders et al., 1999). As in similar
comparisons made by McClean et al. (1997), all
model cases roughly reproduce the distribution of
SSH variability as derived from the T/P altimeter.
Increased resolution yields an enhanced realism in
regions of strong currents in mid-latitudes, but the
models tend to underestimate the T/P variability in
the subtropical oceans, away from the strong currents. The reader is referred to various maps of
SECTION 2 OBSERVATIONS AND MODELS
70
Fig. 2.2.6 Transport of the DWBC between 1000 m
and 3100 m depth in the equatorial Atlantic, from several
year-long time series of moored current meters (Fischer
and Schott, 1997) and a DYNAMO model simulation
with daily wind fields from ECMWF. Moored results
shown are the combined annual and semiannual
harmonics of transport variability for two different
observational periods. Model curves show the mean
seasonal cycle as obtained by averaging over 5 years of
integration and a harmonical analysis similar to the
observations.The annual mean southward transports are
13.0 Sv for the model, 12.4<6.8 Sv and 13.5<5.3 Sv for
the two measurement periods.
Fig. 2.2.7 Difference in 5-year mean northward heat transport between winter (Jan–Feb–Mar) and summer
(Jul–Aug–Sep) for three models based on different vertical coordinate schemes and, associated with that, different
representations of topography and sub-grid-scale mixing (Böning et al., 2000).