www.ecco-group.org/abstracts/woce_abstract_1.txt

ECCO 1 degree global WOCE Synthesis:  The Method

E. Remy, A. Koehl, Y. Lu, P. Heimbach, B. Cornuelle, 
D. Stammer and C. Wunsch for the ECCO Consortium.

A global WOCE data synthesis  describing the three-dimensional 
oceanic state in the 10-year period 1992 - 2001 now exists.  
The synthesis results from the combination of  many of the WOCE 
data sets with a general circulation model.  This combination 
solves the complete global time-dependent ocean state estimation 
problem, up to remaining approximations in the model and underlying 
statistics.  The MIT ocean model used represents the global ocean 
between +/- 80 degrees with 1 degree horizontal resolution, 23 
vertical layers,  and includes parameterizations for the surface 
boundary layer (KPP) and eddy tracer transfers (GM).

At the cost of increased computational load, the estimation 
(assimilation) method is chosen specifically so that the resulting 
state estimate is dynamically consistent with the model equations. 
No   artificial sources or sinks are introduced by the model/data 
combination, and the resulting time-evolution satisfies the equations 
of motion. The combination method used is based upon the method of 
Lagrange multipliers (adjoint method), in which the adjustable parameters 
(controls) include the initial temperature and salinity conditions, 
and the time-dependent surface fluxes of momentum, heat and freshwater.
An important test of the result is passed: adjustments in these control 
variables are consistent with prior estimates of the expected errors in
the oceanographic data and the meteorological analyses
1	ECCO 1 degree global WOCE Synthesis: The Method
2
3	E. Remy, A. Koehl, Y. Lu, P. Heimbach, B. Cornuelle,
4	D. Stammer and C. Wunsch for the ECCO Consortium.
5
6	A global WOCE data synthesis describing the three-dimensional
7	oceanic state in the 10-year period 1992 - 2001 now exists.
8	The synthesis results from the combination of many of the WOCE
9	data sets with a general circulation model. This combination
10	solves the complete global time-dependent ocean state estimation
11	problem, up to remaining approximations in the model and underlying
12	statistics. The MIT ocean model used represents the global ocean
13	between +/- 80 degrees with 1 degree horizontal resolution, 23
14	vertical layers, and includes parameterizations for the surface
15	boundary layer (KPP) and eddy tracer transfers (GM).
16
17	At the cost of increased computational load, the estimation
18	(assimilation) method is chosen specifically so that the resulting
19	state estimate is dynamically consistent with the model equations.
20	No artificial sources or sinks are introduced by the model/data
21	combination, and the resulting time-evolution satisfies the equations
22	of motion. The combination method used is based upon the method of
23	Lagrange multipliers (adjoint method), in which the adjustable parameters
24	(controls) include the initial temperature and salinity conditions,
25	and the time-dependent surface fluxes of momentum, heat and freshwater.
26	An important test of the result is passed: adjustments in these control
27	variables are consistent with prior estimates of the expected errors in
28	the oceanographic data and the meteorological analyses