but additionally, provides a statistically reliable quantification of the transport. The
mean relative mismatch between the reconstructed surface velocities (Fig. 2a) and
assimilated drifter velocities is 0.67.
Figure 2b provides a comparison of the optimized velocity field at 1000 m with
independent (not assimilated) velocity data derived from the Argo floats. The
average speed of the Argo floats is 4.6 cm/s, which is close to the mean optimized
velocity amplitude of 3.7 cm/s. This 25% difference can be considered to be in
reasonable agreement with the observations since a significant fraction of the
ARGO drifters are involved in transient eddy motions resulting in higher Lagrangian mean velocities compared to the Eulerian mean speed estimates.
Figure 2c, d shows the summer TS distributions in the BS. In comparison with
the optimal interpolation results (e.g., https://www.nodc.noaa.gov/OC5/
PACIFIC2009/), the reconstructed fields are less smoothed, especially in the
Northern Bering Sea (NBS) and along the Aleutian Arc. These are signatures of the
strong topographically controlled flows through the Bering and Anadyr Straits and
through the island system of the Aleutian Arc. The most important advantage of
these TS distributions is that they are dynamically balanced with corresponding
velocity fields (Fig. 2a, b), which allows for identifying finer structures of regional
circulation while simultaneously providing estimates of climatological mass, heat
and salt transports.
Except for the summer season, Pacific water masses entering the Bering Sea
within the upper subsurface layer are characterized by higher temperature and
salinity than the water mass residing in the Bering Sea at the same depth (Fig. 2c,
d). The major cause of such temperature/salinity signal is due to the relatively high
heat and salt content of the Pacific Water compared to the Bering Sea Water, as well
as intense vertical and lateral mixing in the Aleutian Passes.
In the summer season, the warmest and least salty subsurface layer waters are
formed in the nearshore regions of Bristol Bay, Norton Sound, and the Gulf of
Karaginski. The presence of warm and relatively fresh salty waters in these areas is
a consequence of strong stratification observed after the melting of the sea-ice cover
and relatively weak winds, which are not able to mix the upper warm layer with the
waters underneath. A somewhat colder and saltier upper subsurface layer water is
present in the summer in several regions of the Bering Sea (e.g. in the central and
eastern parts of the Aleutian Arc, and in the Chirikov Basin, Fig. 2c, d). Due to
intense tidal and non-periodical currents in these regions, surface waters are mixed
with the underlying cold waters formed in the winter and fall.
The optimized estimates of the mean volume transport through the major
Aleutian Passes are found to be 2.5–7 times larger than those in the dynamical
method [28]. This discrepancy is likely due to significant underestimation of the
barotropic velocity component in the Aleutian Straits by the dynamical method.
The latter suggestion is well supported by recent velocity observations in several
Aleutian Passes, which reveal significant northward flow at a 100–200 m depth
[29], indicating the importance of the regional barotropic flow in the overall Bering
Sea volume balance.
The Bering Sea Regional Data Assimilation System …
505
mean relative mismatch between the reconstructed surface velocities (Fig. 2a) and
assimilated drifter velocities is 0.67.
Figure 2b provides a comparison of the optimized velocity field at 1000 m with
independent (not assimilated) velocity data derived from the Argo floats. The
average speed of the Argo floats is 4.6 cm/s, which is close to the mean optimized
velocity amplitude of 3.7 cm/s. This 25% difference can be considered to be in
reasonable agreement with the observations since a significant fraction of the
ARGO drifters are involved in transient eddy motions resulting in higher Lagrangian mean velocities compared to the Eulerian mean speed estimates.
Figure 2c, d shows the summer TS distributions in the BS. In comparison with
the optimal interpolation results (e.g., https://www.nodc.noaa.gov/OC5/
PACIFIC2009/), the reconstructed fields are less smoothed, especially in the
Northern Bering Sea (NBS) and along the Aleutian Arc. These are signatures of the
strong topographically controlled flows through the Bering and Anadyr Straits and
through the island system of the Aleutian Arc. The most important advantage of
these TS distributions is that they are dynamically balanced with corresponding
velocity fields (Fig. 2a, b), which allows for identifying finer structures of regional
circulation while simultaneously providing estimates of climatological mass, heat
and salt transports.
Except for the summer season, Pacific water masses entering the Bering Sea
within the upper subsurface layer are characterized by higher temperature and
salinity than the water mass residing in the Bering Sea at the same depth (Fig. 2c,
d). The major cause of such temperature/salinity signal is due to the relatively high
heat and salt content of the Pacific Water compared to the Bering Sea Water, as well
as intense vertical and lateral mixing in the Aleutian Passes.
In the summer season, the warmest and least salty subsurface layer waters are
formed in the nearshore regions of Bristol Bay, Norton Sound, and the Gulf of
Karaginski. The presence of warm and relatively fresh salty waters in these areas is
a consequence of strong stratification observed after the melting of the sea-ice cover
and relatively weak winds, which are not able to mix the upper warm layer with the
waters underneath. A somewhat colder and saltier upper subsurface layer water is
present in the summer in several regions of the Bering Sea (e.g. in the central and
eastern parts of the Aleutian Arc, and in the Chirikov Basin, Fig. 2c, d). Due to
intense tidal and non-periodical currents in these regions, surface waters are mixed
with the underlying cold waters formed in the winter and fall.
The optimized estimates of the mean volume transport through the major
Aleutian Passes are found to be 2.5–7 times larger than those in the dynamical
method [28]. This discrepancy is likely due to significant underestimation of the
barotropic velocity component in the Aleutian Straits by the dynamical method.
The latter suggestion is well supported by recent velocity observations in several
Aleutian Passes, which reveal significant northward flow at a 100–200 m depth
[29], indicating the importance of the regional barotropic flow in the overall Bering
Sea volume balance.
The Bering Sea Regional Data Assimilation System …
505