ECCO (Estimating the Circulation and Climate of the Ocean)

Initially funded by NOPP (National Oceanographic Partnership Program) to demonstrate practicality and utility of state estimation for global-scale physical oceanography.

ECCO produced a demonstration ocean state estimate for period 1992-present on a 1-deg grid, excluding sea ice and Arctic Ocean.

With support from GODAE and now CLIVAR, the coarse-resolution ECCO state estimate has become quasi-operational.

Despite shortcomings, this solution has proved scientifically useful for a large number of oceanographic and interdisciplinary studies on ocean circulation, ocean biogeochemical cycles, air-sea fluxes, sub-grid-scale parameterizations, earth rotation studies, etc.

Over 100 publications have originated from ECCO:http://www.ecco-group.org/publications.htm

Decadal sea-level patterns and their top-to-bottom partition

• Vertical partition in density trends due to:

– trends in temperature T– trends in salinity S– trends in T, S

TS S

T

Wunsch et al. (2007)

Need for high resolution

Eddy-parameterizations are not based upon fundamental principles and fail to adequately account for known anisotropies. Eddy flux errors accumulate and change large scale ocean circulation in important ways. Horizontal grid-spacing of ~2 km is required to resolve restratification processes in deep convection.

In climate, scalar property transports (heat, fresh water, carbon, oxygen,etc.) are of central interest. Narrow western and eastern boundary currents make major contributions but are not parameterizable. Until they are resolved, there will be doubts that ocean models carry property transports realistically.

Water mass properties are important to climate and climate change. In theabyssal ocean, inability to resolve major topographic features (e.g., fracture zones, sills, overflows) leads to systematic errors in movement of deep water masses with consequences for accuracy of computation of carbon uptake, etc.

Ocean current speed at 15 m depth from 1/16° ECCO2 integration

Hill et al. (2007)

Need for sea ice

Sea ice affects radiation balance due to its high albedo, surface heat and massfluxes due to its insulating properties, ocean convection by preconditioningsites of deep water formation, freshwater fluxes during ablation, ice marginprocesses related to small-scale ocean phenomena, human operations, etc.

Sea-ice processes impact high-latitude oceanic uptake and storage of anthropogenic CO2 and other greenhouse gases both directly, by affecting exchange of these gases across the air-sea interface, and indirectly, byinfluencing the biological, chemical, and physical processes that transportgreenhouse gases from the surface ocean into the interior.

Inclusion of a dynamic/thermodynamic sea ice model permits fuller utilization of high-latitude satellite data collected by NASA and other agencies.

Variability of sea ice simulations assessed with RGPS kinematics

Kwok et al. (2008)

ECCO2: Ocean State Estimation in the Presence of Eddies and Ice

A first ECCO2 solution was obtained for 1992-2002 period using a Green’s function approach to adjust a small number (~80) of model parameters.

Early science applications include improved error estimates and eddy parameterizations for coarse-resolution ocean simulations and estimations, impact of mesoscale eddies on large-scale ocean circulation and its variability, ocean biogeochemistry, and studies of the polar oceans.

A follow-on ECCO2 solution is currently being obtained for the 2004-present period using the adjoint method to adjust ~109 model parameters.

Green’s functions applied to global problem

Menemenlis et al. (2008)

1st baroclinic mode Long Rossby waves

Eddying ocean circulation studies

Eddy zonal propagation from altimetry

Eddy zonal propagation from ECCO2

ECCO2 simulation of the eddy zonal propagation speed (purple) and the altimetry observation (black)

km/d

ay

km/day

km/day

Fu (2006); Volkov and Fu (2008)

Arctic Ocean studies

Modeling transport, fate and lifetime of riverine DOC in the Arctic Ocean. Manizza et al., in press.

Response of the Arctic freshwater budget to extreme NAO forcing. Condron et al., 2009

Sea ice model and adjoint development.Campin et al., 2008; Nguyen et al., in press.Losch et al., submitted; Heimbach et al., submitted.

Variability of sea ice simulations assessed with RGPS kinematics. Kwok et al., 2008.

Arctic Ocean studies

Nguyen et al., in preparation

Arctic cost function reduction

Canada Basin Hydrography

2007-2008 summer sea ice minima

Atlantic water pathways

Ice Shelf Ocean Interaction

ICESAT/GLAS observation estimates (E Rignot)

ECCO2 Model estimates

Melt rate dh/dt 2004 (m/yr)

Schodlok et al., in preparation

Antarctic Ice Shelves in ECCO2

Towards global, eddying, adjoint method ocean/sea ice state estimation

Adjoint method ocean and sea ice optimization.Heimbach et al., submitted;Fenty, PhD thesis, 2009;Fenty et al., in preparation.

92-93 optimized/baseline heat flux difference

Adjoint sensitivity of ice flow through Lancaster Sound

Adjoint method optimization in the presence of eddies.Gebbie, PhD thesis, 2005;Gebbie et al., 2006; 2007;Mazloff, PhD thesis, 2008;Mazloff et al., submitted;Abernathey et al., submitted.

SOSE surface speed and tracer concentration

CS510 adjoint optimization during ARGO period (2004-present)

Cost function reduction

Iteration 3 vs baseline SST change on Aug 27, 2004

Cost function reduction vs ARGO salt

Cost function reduction vs ARGO temp

Plans for last 2 years of ECCO2

1. Initiate CS510 adjoint optimization during ARGO period (2004-present).

2. Test, calibrate and adopt new parameterizations: downslope package, Winton thermodynamics, salt plume, restratification, shelf ice, etc.

3. Ultra high-resolution global ocean plus ice integration.

4. Continue evaluation.

5. Support science applications: parameterizations, error estimates, mesoscale eddy processes, regional studies, air-sea fluxes, mode water formation, biogeochemical, geodetic, acoustic, electromagnetic, etc.

ECCO2 publications: http://ecco2.org/manuscripts