Combined Ocean-Geodetic Analysis of Global and Regional Ocean

Combined Ocean-Geodetic Analysis of Global and Regional Ocean Mass-, and Freshwater Transport Divergences D. Stammer, A. Köhl, V. Romanova, F. Sigismund, W. Wang (Institut für Meereskunde, Universität Hamburg) C. Böning, J. Dengg, K. Lorbacher (IFM-GEOMAR) M. Visbeck, J. Karstensen, U. Neuman (IFM-GEOMAR)

Overarching Project Goals Use satellite-based data of gravity and altimetry, in situ ocean data, river-discharge and ocean circulation models to determine ocean circulation, its mass transports and mass changes in the ocean. Central focus: redistribution of mass and freshwater on the seasonal to decadal time scale, its connection to atmospheric and terrestrial water transports and storages, and the oceans role in the global freshwater cycle.

The GECCO State Estimate Ocean Synthesis, performed over the period 1952 through 2001 on a 1º global grid with 23 layers in the vertical, using the ECCO/MIT adjoint. The models adjoint (obtained using TAF) is used to bring the model into consistency with most of the available ocean observations over the full period by adjusting control parameters. Control parameter are the models initial temperature and salinity field as well as the time varying surface forcing, leading to a dynamically self-consistent solution (next step is to include mixing).

Schematic of GECCO Optimization 1952 - 2001

SSH Drift Changes in SSH on inter-decadal time scale have been “observed”. Unclear is what is causing them. In fact, the data base is hardly sufficient to infer global and regional inter-decadal sea level trends with any statistical significance. We can use again the results from the 50 -year long ocean state estimation performed in Germany as part of the ECCO project (GECCO) to investigate changes in SSH for the period 1952 through 2001 and infer dynamical causes. See Köhl and Stammer, J. Clim, in press, 2007 for details.

ECCO-50 y

Comparison of maximum MOC at 25 N with Bryden et al. (2005) (Köhl and Stammer, 2007, submitted)

Max. K-7 MOC o. N 25 Bryden et al. (2005) ECMWF

K-7 Heat transport 25 o. N

The MOVE Array plus RAPID (Neuman, Visbeck, Carstens) Absolute transport fluctuations below 1180 dbar (across 16｡N) / 1110 dbar (across 26. 5｡N), combining geostrophic transports and continental slope contributions, as measured by two mooring arrays. The light gray and bold coloured curves represent 2 and 30 day low-passed filtered time series.

green: MOVE 1 red: OMCT black: GECCO Correlation: 2002 2003 2004 Correlation: OMCT and MOVE: 0. 04, -0. 01, 0. 12 OMCT and MOVE: -0. 01 GECCO and MOVE: 0. 31 0. 27, 0. 31, 0. 50

r=0. 68 r=0. 69 r=0. 76 r=0. 74 r=0. 54

Seasonal pb Variations Amplitude GRACE GAD GECCO OMCT Phase

Vinogradova, Ponte and Stammer, GRL, 2007 STD daily SSH STD bottom pressure SSH and pb correspondence ratio

1/6° Model Resolution (Serra, Köhl and Stammer, 2007)

Zonally averaged annual mean net heat fluxes (W/m 2) Seasonality of the globally averaged heat fluxes (W/m 2) GECCO NCEP SOC GECCO lies mostly between NCEP and SOC; largest adjustments toward SOC occur in high latitudes. Biggest differences between SOC and other fields during winter months. (Romanova et al. , 2007)

Zonally integrated freshwater fluxes (mm/day) GECCO NCEP HOAPS GECCO corrects away from HOAPS and NCEP in subtropics and high latitudes; HOAPS has much stronger ITCZ. (Romanova et al. , 2007)

Amazonas River discharge and corrected freshwater fluxes Trenberth GECCO MIT

Effect of Greenland, Arctic Ice Melting on SSH (Stammer, 2007)

(Stammer, 2007)

Long term trends of the Thermohaline Circulation (Lorbacher and Dengg, 2007) Left Figure: Annual mean differences of the streamfunction of the meridional overturning circulation (MOC) are around 5 Sv at most while the general structure – maximum transport between 30°-40°N at 1000 m depth – does not change significantly. Right Figure: Time series of the MOC at 26. 5°N show interannual variability with amplitudes around 2 Sv and while the MOC in KAB 109 remains almost stable over the five decades the MOC in KAB 110 exhibits a 50% (~5 Sv) reduction.

… Sea Surface Height and Ocean Bottom Pressure Difference between the 5 th and 1 st decade (1996 -2004 minus 1960 -1968) of ocean bottom pressure (in cm, equivalent water thickness). (Left and right figure: KAB 109 and KAB 110, respectively). SSH analyses show the same spatial structure. Reflection of similar interannual and decadal behaviour of the MOC variability (and since we removed a global mean of the parameters the changes are mainly due to changes of large-scale meridional mass distribution changes)

Some Statistics Figures: Linear regression between SSH and MOC (in cm/Sv) of monthly mean anomalies (climatological seasonal cycle removed) for the two experiments. Pictures for ocean bottom pressure show similar structures. Strong linear correlation between SSH and MOC mainly due to the long-term trend. (Trends of the last decade (1993 to 2004) and an outlook are presented at the poster)

Global Assimilation Future Cube Sphere Topology In collaboration with MITgcm primitive eq. ocean model Sea Ice model Bulk flux formulae llc (long. -lat. -polar cap) grid Variable resolution: horiz. (1/3 -1°) and vert. (min: 10 m; max: 550 m) 1948 -2007 50 levels, 2 x 90 + 2 x 90 x 360 + 2 x 360 x 90 horizontal grid points 1800 s-2400 s time step Mercator grid telescopic equatorial Arctic cap

Preliminary forward run 1948 -2007 SST and Seaice height Problems ●Spindown of the gyres ●Decreasing transport through all major passages ●Warming of the upper ocean Llc-Model Performance Tornado (2. 6 Ghz opteron cluster) tile size - # cores - elapsed time for 1 yr simulation (45 x 45) - 72 - 2: 20 h => one iteration over 60 years: 30 -40 days Timestep/vert. grid/seaice: factor 8 more expensive than GECCO-50 yr although still a on deg. Model. => 1/3 deg. Cube Sphere set-up instead?