Modeling the North Atlantic Circulation:
From Eddy Permitting to Eddy Resolving*
Frank O. Bryan1, Richard D. Smith2,
Mathew E. Maltrud2, Matthew W. Hecht1
-
Oceanography Section, Climate and Global Dynamics Division,
National Center for Atmospheric Research, Boulder, CO. 80307
-
Theoretical Division, Group T-3, Los Alamos National Laboratory,
University of California, Los Alamos, New Mexico 87545
Introduction
During the WOCE era, the computational resources available to the ocean
modeling community have increased from a level that allowed eddy permitting
basin scale simulations with horizontal resolutions of 35-40 km and
20-30 vertical levels to the present capability of eddy resolving
simulations at 8-10 km horizontal resolution and 40-50 vertical levels.
Model Configuration
Three experiments have been carried out with the POP ( Dukowicz
and Smith, 1994) primitive equation ocean model that differ only in
their horizontal resolution and dissipation parameters. A Mercator grid
is used with equatorial resolution of 0.1°, 0.2°
or 0.4°. The domain includes the Atlantic Ocean from
20°S to 73°N latitude, the Gulf of Mexico,
Caribbean Sea and western Mediterranean Sea. The 0.1° case
has grid spacing smaller than the first baroclinic radius of deformation
at all latitudes.
The horizontal hyperviscosity (
)
and diffusivity (
) are scaled with
the grid spacing (on each grid and between grids) according to:
[,
]=
[
0,0]
(dx/dx0)3
where dx is the local grid spacing, 0
= -27x1017 cm4 s-1, 0=
-9x1017 cm4s-1, and dx0 = 11
km. The hyperviscosity is shown below as a function of grid spacing, along
with the values used in several other recent North Atlantic simulations.
All cases are forced with daily wind stress derived from 1985-1995 ECMWF
surface analyses, Barnier et al (1995) climatological
monthly heat flux, and restoring of surface salinity to the World Ocean
Atlas 1994 (Levitus et al, 1994a) monthly
climatology. Sponge layers are located at the north and south boundaries,
and in the eastern Mediterranean. Topography is derived from the 5" ETOPO5
data set. A 5 year spin-up initialized with temperature and salinity from
the World Ocean Atlas 1994 (Levitus et al, 1994a, 1994b) is followed by a 10 year experiment at each
resolution.
Large Scale Mean Circulation
The mean sea surface height illustrates the strong sensitivity of the simulated
mean general circulation to resolution. The differences between the cases
are particularly striking in the Gulf Stream region and subtropical-subpolar
gyre boundary. In the lower resolution cases the Gulf Stream separates
from the coast well north of Cape Hatteras and is dominated by a tight
anticyclonic circulation at the coast, a characteristic of many previous
simulations (Dengg et al, 1996). In contrast, the
highest resolution experiment has a realistic separation latitude and structure.
Likewise, the lower resolution cases do not properly simulate the path
of the North Atlantic Current to the east of the Grand Banks, while the
highest resolution case shows a realistic simulation of the "Northwest
Corner", secondary separation, and subsequent northeastward flow into the
Nordic Seas. (Click on thumbnails below for full size image; Note that the Gulf of Mexico is pasted into North Africa in the images below to conserve space).
These differences are also evident in the barotropic transport streamfunction.
The northern recirculation gyre of the Gulf Stream emerges and the southern
recirculation gyre strongly intensifies at the highest resolution.
The circulation in the vertical-meridional plane is also strongly modified
with increasing resolution. An intense zonally integrated vertical recirculation
develops at the latitude of the Gulf Stream, and the Deep Western Boundary
Current (DWBC) intensifies, narrows and deepens with increasing resolution.
In the highest resolution case, the breadth, speed and depth of the
DWBC core to the east of the Bahamas at 26.5°N latitude
are in good agreement with the observations of Lee et
al (1996).
Heat Transport
The meridional heat transport reaches a maximum between 25°N
and 30°N latitude in each case, but increases in magnitude
by 50% from the lowest to highest resolution case. There is only a small
increase in the eddy heat transport with increasing resolution (not shown),
the majority of the increase in the total transport is accounted for by
changes in the time mean flow.
Large Scale Variability
Changes in the amplitude and distribution of variability accompany the
changes in the mean flow with increasing resolution. The rms sea surface
height distribution for each case along with the corresponding field derived
from the Topex/Poseiden altimeter are shown below. The 0.1°
case shows good agreement with the observations over most of the domain,
including the Gulf Stream region, North Atlantic Current to the east of
the Grand Banks, and the Azores Current. In the lower resolution cases
the distribution of variability reflects the biases in the mean path of
the North Atlantic Current, and the variability in the vicinity of the
Azores Current is missing.
|
TOPEX/POSEIDON Observations
|
0.1°
|
|
|
|
0.2°
|
0.4°
|
|
|
An increase in the eddy kinentic energy (EKE) is seen in the vicintity
of the Gulf Stream as expected from the differences in the mean flow characteristics,
but also in the eastern basin., well away from the direct influence of
the boundary currents. The distribution of EKE along 48°N
latitude in the highest resolution case is in good agreement with the observations
of Colin de Verdiere et al (1989).
Gulf Stream Dynamics
In the Gulf Stream region the character of the variability, in addition
to its amplitude, changes dramatically with resolution. The position of
the Gulf Stream, as defined by the 12°C isotherm at 400m,
over a four year period is shown below for each case along with the mean
position and meander envelope. The lower resolution cases are characterized
by large amplitude standing meanders near the coast, and mean paths that
are displaced well north of that observed. The highest resolution case
shows a more realistic, tighter, meander envelope west of 65°W
longitude, with an increase in meander amplitude near the New England Sea
Mounts as well as a more realistic mean path position
For the 0.1° resolution case, we have computed the properties
of the Gulf Stream in stream-averaged coordinates, following the method
of Johns et al (1995). The stream-averaged downstream
velocity at 73°W, 68°W, and 55°W
longitude show good agreement in maximum speed, horizontal shear, subsurface
distribution, and width of the jet with observations from the SYNOP program
(Halkin and Rossby, 1985; Johns
et al, 1995; Bower and Hogg, 1996)..
The net, stream-averaged, transport of the Gulf Stream, obtained by integrating
the downstream velocity between the 0 m/sec isotachs is shown below, along
with observations compiled by Johns et al (1995).
The observed increase in downstream transport is modeled well.
Conclusions
-
The 0.1° resolution experiment shows remarkable improvements
over the lower resolution experiments in many aspects of the simulation
of both the time mean circulation and variability of the North Atlantic.
-
These results suggest that we are closely approaching a "realistic" solution
at 0.1° resolution, though we cannot demonstrate true convergence
without yet higher resolution simulations.
-
These short duration experiments demonstrate that we are able to obtain
realistic simulations of the circulation given the observed water mass
distribution. They cannot tell us if the models are capable of properly
simulating the formation and long term evolution of water masses.
-
The degree of qualitative agreement between the simulated and observed
circulation at the highest resolution have allowed us to begin forming
direct quantitative comparisons with in situ and remotely sensed
observations at both local and basin scales. The next step will be to compare
higher order statistics of the variability, such as eddy fluxes and their
divergences, with available observations.
-
The results from this and other simulations of comparable resolution will
provide an important resource that can be used in combination with diverse
observations in exploring the dynamics of the general circulation, in designing
future observational systems, and developing sub-grid scale parameterization
for lower resolution models.
References
Barnier, B., L. Siefridt and P. Marchesiello, 1995:
Thermal forcing for a global ocean circulation model using a three-year
climatology of ECMWF analyses. J. Marine Systems., 6, 363-380.
Bower, A.S. and N.G. Hogg, 1996: Structure of
the Gulf Stream and its recirculation at 55W. J. Phys. Ocean., 26,
1002-1022.
Colin de Verdière, A., H. Mercier, and
M. Ahran, 1989: Mesoscale variability transition from the western to the
eastern Atlantic along 48N. J. Phys. Ocean., 19, 1149-1170.
Dengg, J., A. Beckmann, and R. Gerdes, 1996: The
Gulf Stream separation problem. In: The Warmwatersphere of the North
Atlantic Ocean. pp 253-290. W. Krauss (Ed.) Gebrüder Borntraeger.
Berlin.
Dukowicz, J.K. and R.D. Smith, 1994: Implicit free-surface
method for the Bryan-Cox-Semtner ocean mode. J. Geophys. Res., 99,
7991-8014.
Halkin, D. and T. Rossby, 1985: The structure and
transport of the Gulf Stream at 73W. J. Phys. Ocean., 15,
1439-1452.
Johns, W.E., T.J. Shay, J.M. Bane and D.R. Watts,
1995: Gulf Stream structure, transport and recirculation near 68W. J.
Geophys. Res., 100, 817-838.
Lee, T.N., W.E. Johns, R.J. Zantopp and E.R. Fillenbaum,
1996: Moored observations of western boundary current variability and thermohaline
circulation at 26.5N in the subtropical North Atlantic. J. Phys. Ocean.,
26, 962-983.
Levitus, S., R. Burgett and T.P. Boyer, 1994: World
Ocean Atlas 1994. Volume 3: Salinity. NOAA Atlas NESDIS 3. U.S. Deptartment
of Commerce. Washington, D.C.
Levitus, S. and T.P. Boyer, 1994: World
Ocean Atlas 1994. Volume 4: Temperature. NOAA Atlas NESDIS 3. U.S.
Deptartment of Commerce. Washington, D.C.
Acknowledgements
This work was supported by the Department of Energy CHAMMP Program, NASA
contract #1987-100 to NCAR, and the National Science Foundation through
its support of NCAR.
* This material is based on a poster presented
at the International World Ocean Circulation Experiment Conference on Ocean
Circulation and Climate, Halifax, Nova Scotia, Canada, 24-29 May 1998.
If you have comments or questions, e-mail one of us at
