www.ecco-group.org/abstracts/woce_abstract_11.txt

Testing the Mid-depth Circulation in ECCO Ocean Estimates

Russ Davis (rdavis@ucsd.edu), Detlef Stammer (dstammer@ucsd.edu),
Scripps Institution of Oceanography

Tong Lee (tlee@pacific.jpl.nasa.gov)
Jet Propulsion Laboratory

Improved model physics, better surface forcing fields and higher numerical resolution 
have made ocean circulation models more accurate. In addition, data assimilation has 
improved the agreement of model simulations with observations and provides a rational 
procedure for model testing using the very data that are assimilated as well as withheld 
information.  Most data used for assimilation are satellite altimetry and temperature 
profiles from XBTs, P-ALACE and Argo that are most descriptive of the upper ocean.  
In this study we use absolute mid-depth velocities from WOCE floats as independent 
data to test the results from ECCO ocean circulation estimates at depths removed from
the majority of assimilated data.

The test will be performed at several stages of the assimilation: Comparisons of float 
data with adjoint and Kalman-filter models having nominal resolution of 1o to 2o generally 
show mean flows of generally similar shape but with the measured field being significantly 
stronger.  Some high-resolution simulations show zonal flows that cannot be identified in 
the float data. Results from runs with and without insitu data will be used to identify 
their impact on the estimated mean flow field.   Although the float measurement density 
is marginal for resolving variabili ty of narrow currents, comparisons with annual and 
specific interannual variability is possible.

1	Testing the Mid-depth Circulation in ECCO Ocean Estimates
2
3	Russ Davis (rdavis@ucsd.edu), Detlef Stammer (dstammer@ucsd.edu),
4	Scripps Institution of Oceanography
5
6	Tong Lee (tlee@pacific.jpl.nasa.gov)
7	Jet Propulsion Laboratory
8
9	Improved model physics, better surface forcing fields and higher numerical resolution
10	have made ocean circulation models more accurate. In addition, data assimilation has
11	improved the agreement of model simulations with observations and provides a rational
12	procedure for model testing using the very data that are assimilated as well as withheld
13	information. Most data used for assimilation are satellite altimetry and temperature
14	profiles from XBTs, P-ALACE and Argo that are most descriptive of the upper ocean.
15	In this study we use absolute mid-depth velocities from WOCE floats as independent
16	data to test the results from ECCO ocean circulation estimates at depths removed from
17	the majority of assimilated data.
18
19	The test will be performed at several stages of the assimilation: Comparisons of float
20	data with adjoint and Kalman-filter models having nominal resolution of 1o to 2o generally
21	show mean flows of generally similar shape but with the measured field being significantly
22	stronger. Some high-resolution simulations show zonal flows that cannot be identified in
23	the float data. Results from runs with and without insitu data will be used to identify
24	their impact on the estimated mean flow field. Although the float measurement density
25	is marginal for resolving variabili ty of narrow currents, comparisons with annual and
26	specific interannual variability is possible.
27