Aucun titre de diapositive

1
ESTIMATION OF A GLOBAL MEAN DYNAMIC TOPOGRAPHY
USING ALTIMETRY, GRAVIMETRY AND IN-SITU DATA -
VALIDATION AND APPLICATIONS
Slope ratios
RMS differences ratios
Comparison to existing solutions
V A L I D A T I O N
We now investigate which mean topography (the
Levitus climatology, the OCCAM MSSH or our SMDT)
provides the more realistic absolute topography
when added to SLAs. Regressions analysis are
performed between the various mean topograpies
added to the SLAs and absolute in-situ data.
Regression parameters (slope A and rms
differences E) are computed using a minimal least
square method .
zonal velocities
meridian velocities
For in-situ data spanning the years 1993-1999,
respective results are compared looking at the
slopes ratios (AOCCAM/ASMDT) and the rms
differences ratios (ESMDT/EOCCAM) on a 5 degree
resolution grid. The use of the SMDT allows to
increase the regression slopes (slopes quotient lt
1) and to reduce the rms differences (rms
differences quotient lt1) in most regions. This
means that the obtained absolute circulation is
more realistic when using the SMDT.
Dynamic Heights
Independent in-situ data are available for the
year 2000. Regressions are presented here for the
Gulfstream area. Again, regression slopes are
enhanced and rms differences are significantly
reduced when using the SMDT instead of the OCCAM
MSSH or the Levitus climatology.
Accessing absolute circulation maps
See Movies
A P P L I C A T I O N
As part of the CLS project DUACS (Developing Use
of Altimetry for Climate Studies) maps of Sea
Level Anomalies are produced for the global ocean
every week. We used the Synthetic Mean Dynamic
Topography to compute absolute circulation maps
for the Gulfstream, the Kuroshio and the Aghulas
current areas. Oceanic fronts, which were
difficult to identify on SLA maps are clearly
visible on the absolute circulation maps.
Besides, the production of eddies associated to
these main currents and their temporal evolution
are easier to follow. The analyse of these
absolute circulation maps will allow to better
understand the variability of the main currents
fronts and associated transports and will make
easier the eddies detection and follow-up.
Comparison between MERCATOR model runs with NEW
Synthetic Mean Dynamic Topography and
operational MSSH
See Movies
REFERENCE RUN The MERCATOR operational
MSSH is used. This MSSH is derived from the mean
sea surface height from a forced model run from
Jan 1992 to Dec 1995. The MSSH is then modified
to reduce model mean SST and satellite observed
SST mismatch. The operational MSSH is then
adjusted to account for the use of SLA data (in
the assimilation) referenced to a 7 year mean
state (1992-1998) rather than a 3 year mean state
(1992-1995).
New
Gulf of Mexico March 13 2002 Snap Shot at 100m
depth on 13 March 2002 The new MSSH restrains
the loop current from entering far into the Gulf
while still allowing eddies to spin off from the
loop current. Separation of the Gulf Stream at
Cape Hatteras is better represented using the new
MSSH. Note that water at 100m is cooler in the
new run.
A P P L I C A T I O N
NEW
Reference
MSSH
MSSH
Gulf Stream and Labrador Current Snap Shot at
100m depth on 17 April 2002 The main
improvements due to the new MSSH lie in regions
of strong mean currents. In particular the model
now resolves counter currents north of the Gulf
Stream. Note the southward flow on the western
flank of the Grand Banks and inshore from the
Gulf Stream north of Cape Hatteras. Furthermore
the Labrador Current is now clearly shown with
flow around the Flemish Cap
(m)
NEW MSSH RUN The model starts from an
operational run restart file on August 29th 2001
(as above). The model is then run onwards to May
8th 2003 with the NEW MSSH derived by the
synthetic method instead of the operational MSSH.
Assimilating MSSH into an ocean model impacts
water mass and momentum characteristics over
the entire ocean water column1 in comparison to
assimilation of SLA data which affects mostly
near surface circulation. On the left we see
model-climatology temperature differences
averaged for the month of April 2002. The
climatology stems from Reynaud et al.2 and covers
50 years. At 1000 m the new MSSH improves the
models correspondence with climatology while
still allowing eddy scale activity (note cooler
spots in the Gulf Stream). The new model run over
emphasized the extent of cool water off Portugal.

(NEW Reference) MSSH The main contribution of
the Synthetic Mean Dynamic Height approach lies
in regions of strong mean currents and fronts
(Gulf Stream Labrador Current). There are also
strong differences in the North Atlantic Drift
region.
(m)
TEMPERATURE oC

Conclusion A Synthetic Method is described and
applied to compute a Mean Dynamic Topgraphy for
the global ocean. Where in situ data is
available, this approach is useful for providing
a realistic MDT for assimilating into an ocean
model. Preliminary results are shown here. A
detailed evaluation of model-observation
inter-comparisons which will look at the impact
on model forecasts and lead to a refinement of
MMSH estimation techniques is underway.
Bibliography (1) Killworth, P.D., C. Dieterich,
C. Le Provost, A.Oschlies and J.Willebrand
(2001). Assimilation of altimetric data and
mean sea surface height into an eddy-permitting
model of the North Atlantic. Progress in
Oceanography 48313-335. (2) Reynaud, T., P.
Legrand, et al. (1998). "A new analysis of
hydrographic data in the Atlantic and its
application to an inverse modeling study."
International WOCE Newsletter 32 29-31.
Acknowledgements Lucas Nouel is thanked for
running the MERCATOR ocean model. Nathalie
Verbruge, Eric Greiner and Pierre-Yves le Traon
nurtured useful discussions. This work was
carried out as part of the French MERCATOR
science working team activities (Groupe Mission
Mercator).