Sea surface topography over the past decade observed from satellite altimeters
C.J. Koblinsky, S. Howden, C. Boone and B. Beckley
NASA/GSFC/Oceans and Ice Branch, Greenbelt, Maryland, USA
A decade long time series of sea surface topography has been constructed from three satellite altimeter missions: Geosat, ERS, and TOPEX/POSEIDON. Monthly estimates of sea surface topography have been constructed relative to the 1993 mean topography or the EGM96 geopotential reference.
The altimeter data are processed with a consistent set of models and algorithms based on the TOPEX/POSEIDON experience to provide a consistent reference frame. The data have been validated against long term observations at tide gauge stations. Preliminary analysis shows evidence of a decadal variability in the mid-latitude oceans. This variation, in terms of topography and geostrophic flow, will be described using a variety of comparisons with in situ observations and simulations from general circulation models.
dual-satellite altimetry crossovers: achievements and prospects
Jaroslav Klokocnik, Carl Wagner, Jan Kostelecky and Wolfgang Bosch
Astronom. Inst., Academy of Sciences of the
Czech Republic, Ondrejov, CZECHOSLAVAKIA
Research Institute Geodesy, Topography & Cartography, Munchen
GERMANY,
Dual-satellite crossover differences of altimetric sea heights (DSC), have been found useful to improve orbit determination of the lower satellite of an altimetry pair, for accuracy assessments of the Earth’s gravity field models, and for gravitational field modelling. Recently, a coordinate system offset of GEOSAT with respect to that of TOPEX/POSEIDON (T/P) has been discovered from DSC residuals of this pair of satellites, and removed for subsequent use in oceanography. DSC’s provide a link between two and more altimetry missions, eventually over decades. For decadal sea level analysis and ocean circulation modelling, multi-mission studies with their use will become more and more important.
We review current achievements with dual-satellite crossover altimetry and discuss the prospects. First, we define the DSCs and their useful "spectral" adjuncts, the Latitude Lumped Coefficients (LLC), with regard to their principal systematic source from geopotential-orbit error. Then we outline the method and results of the projection of calibrated variance-covariance matrix of recent gravity models (JGM-2,-3) to estimate the standard error of DSCs or LLCs for diverse orbits, including orbits of T/P, GEOSAT, and ERS-1(2).
In the second part, we focus on some systematic errors remaining in the DSC residuals (after all available corrections are applied as well as significant 1 cycle per revolution orbit error removed), namely for orders m = 0 and 1. For example, we give two approaches to evaluation of the coordinate system shift between GEOSAT & T/P, and ERS-1, -2 & T/P, using the observed DSC/LLC residuals, JGM-2 based, and present numerical results of (relative values of) the so-called ‘forbidden’ harmonic coefficients C¥11,S¥11 and C¥10. We present sea-height surfaces before and after elimination of this effect. We intend to repeat this analysis with the new DSC residuals models, to investigate the long-term oceanographic signal within the DSC residuals.
The DSC residuals, especially of pass-pairs close in time, are uniquely sensitive to the geopotential through its effects on the satellite’s orbit. We prove it by a sensitivity analysis. We then assess the role of DSC and SSC in the gravity field recovery using a design matrix analytically derived from "geostationary" definitions for these quantities based on the work of Rosborough and Klokocnik. Long term averages of SSC and DSC residuals from GEOSAT, ERS-1, and T/P (corrected for media, sea-level, 1~ cycle per revolution orbit, geocenter and time-tag errors) show surprisingly that the lower order harmonics are the prime candidates for improvement. Also, the results from crossover altimetry are shown to be complementary to those from traditional analyses of resonant phenomena in satellite orbits.
ON THE TROPICAL ATLANTIC DYNAMICS USING TOPEX/POSEIDON
Laboratoire d'Oceanographie Dynamique et de Climatologie (LODYC), UMR 121 CNRS/ORSTOM/UPMC, FRANCE
The TOPEX/POSEIDON data (1992-1996) are used together with WOCE in-situ measurements and general circulation model results to investigate the Tropical Atlantic ocean dynamics. The seasonal cycle is first described to emphasise the interannual event which occurred in 1995 in the Gulf of Guinea. The North Equatorial Countercurrent area is also particularly described in terms of heat and mass transports.
Mean Sea Surface and Gravity Anomaly Maps From ERS-1 and Geosat Altimetry
P. Mazzega, M. Berge-Nguyen, A. Cazenave and P. Schaeffer
UMR5566/CNRS/CNES, Toulouse Cedex, France
Global maps of the geoid heights and gravity anomalies (GA) are recovered from an analysis of altimetry data from the ERS-1 and GEOSAT Geodetic Missions. After various geophysical corrections, each altimetric sub-cycle from both satellites is first adjusted by a least squares procedure to the TOPEX/POSEIDON mean profiles. Then each profile is fitted to a MSS model previously computed from ERS-1, using a second order polynomial of the along track coordinate. A model of the stationary Sea Surface Topography (SST) is removed from the data, as well as the DMA (1996) earth’s gravity field model complete up to degree and order 360 which constitutes our reference surface for the analysis.
The medium and short wavelengths of the geoid are mapped (1/16ƒ x 1/16ƒ meshes) by a co-location method applied to the ERS-1 and GEOSAT pre-processed data. The implementation of our algorithm is sub-optimal as only the data included in a small cap around each grid point are simultaneously analysed. The map of the residual gravity anomalies (GA) is built with the same procedure using geoid versus gravity anomaly covariance functions. The signal covariance functions are empirically deduced from a small set of altimeter data surrounding the mapping point: the residual geoid power spectrum is computed from a 2D-FFT and then the isotropic covariance functions obtained via an Hankel transform of the spectrum and the usual propagation rules. The DMA geoid and corresponding GA are finally added to our geoid and GA solutions respectively. A MSS model can also be obtained from the geoid solution by restoring the SST model.
Our MSS solution is tested by comparing its heights and gradients with independent TOPEX/POSEIDON data and with other equivalent models. The GA solution is compared to models deduced from satellite altimetry by other authors using different analysis methods. These competing GA maps are then compared to extended sets of ship track’s gravity data in various parts of the world ocean.
Mauricio M. Mata and Carlos A.E. Garcia
Departamento de Fisica, FURG, Rio Grande,
BRAZIL
School of Earth Sciences, Flinders University, Adelaide,
AUSTRALIA
TOPEX altimetry data, obtained during 1993 and 1994, was used to determine the ocean surface geostrophic circulation in the southwestern Atlantic region between latitudes 28ƒS and 42ƒS and longitudes 35ƒW and 60ƒW. Due to lack of knowledge of marine geoid undulations, ocean surface circulation extracted from altimetry is limited to spatial scale of greater than 2000 km. However, the fluctuating portion of surface geostrophic currents can be estimated quite accurately on smaller scales under the assumption that the marine geoid does not vary its position, on the time scales in which we are interested. To estimate the surface geostrophic circulation on smaller scales, we used the composite sea surface dynamic topography (CSSDT). In this approach, the fluctuating portion of surface geostrophic currents, extracted from altimetry, is added to the climatological dynamic heights, from Levitus climatology, which is assumed as the mean portion of the surface geostrophic currents. The results were mapped and compared with the dynamic topography obtained from the Brazilian navy R/V Antares CTD casts in the region during April of 1994. In spite of the limitations of the CSSDT method (assuming a reference level, joining data from different periods and spatial scales and relatively poor Levitus data over the South Atlantic Ocean), there is a good agreement between both sets. The major differences were found near the continental slope, where the geostrophic velocities from hydrographic data are much stronger. Nevertheless, the overall results showed that the combination of altimetry and climatology could be very useful in obtaining a picture of the mesoscale surface geostrophic currents.
Assimilation of large scale satellite altimeter data in a PE model of the Southern Ocean
Jens Schr–ter, J–rg-Olaf Wolff, Nathan Bindoff and Richard Coleman
Alfred-Wegener-Institut f¸r Polar- und
Meeresforschung, Bremerhaven, Germany
Antarctic CRC, Hobart, Australia
The long wavelength component of TOPEX/POSEIDON altimeter data, expressed in spherical harmonics, is assimilated into an ocean general circulation model. The model is a version of the Hamburg Ocean Primitive Equation model (HOPE) which is focused on the Southern Ocean south of Australia.
The aim of our study is to find out how well the barotropic circulation of the model can be constrained by data assimilation and which transports of the Antarctic Circumpolar Current the assimilation yields.
Three different techniques are applied. The first one is nudging of the model’s sea surface height (SSH). The second method consists of a least squares fit to the model’s barotropic dynamics and the observational data.
This technique involves the adjoint of the dynamic operator. The third method is the simplified version of this technique, where the change in the model’s SSH is restricted to predefined structures.
All methods are then compared and evaluated according to
John Wahr, Mery Molenaar and Frank Bryan
Dept of Physics and CIRES, University of
Colorado, Boulder, CO, USA
NCAR, Boulder, CO, USA
Changes in the distribution of mass within the ocean can cause time-dependent fluctuations in the earth's gravitational field, which are large enough to be detectable using proposed dedicated-gravity satellites. The GRACE satellite mission (proposed to NASA under the ESSP Program by the University of Texas, JPL, and GFZ-Potsdam) is especially well suited to determining time-dependent gravity, due both to its relatively long mission length (nominally 5 years), and its high level of accuracy in delivering gravity estimates at spatial scales from a few hundred to tens of thousands of km. Using results from such a satellite, it would be possible to infer time-dependent changes in mass, integrated vertically through the atmosphere and oceans, a quantity proportional to sea floor pressure. The ability of satellite gravity data to provide sea floor pressure depends not only on the satellite's error spectrum, but also on how effectively the oceanic signal can be separated from the contributions of other processes, such as changes in the distribution of water/snow/ice on continents, and the movement of mass within the solid earth. We have used model output from a number of sources and for a number of geophysical processes, along with estimates of the GRACE error spectrum, to assess the accuracy with which GRACE could determine sea floor pressure. We conclude that the satellite could deliver monthly values of sea floor pressure, averaged over scales of about 500 km, to an accuracy approaching a tenth of a mbar over the entire globe (the results tend to be worse than this within a few hundred km of the coast, due to contamination from the gravitational effects of changes in water storage on continents). The accuracy improves further at larger spatial scales and for longer averaging times. Results degrade rapidly at spatial scales below about 300 km.
Donna L. Witter and Arnold L. Gordon
Lamont-Doherty Earth Observatory of Columbia University, Palisades, USA
As part of a larger study of the relation between South Atlantic circulation, interbasin exchange, and global thermohaline fluxes, variations of South Atlantic circulation at seasonal and year-to-year time scales, and their relation to wind forcing, are investigated using more than three years of TOPEX/POSEIDON observations. Co-variability between South Atlantic ocean circulation and wind forcing is identified from contemporaneous altimeter sea level observations and National Center for Environmental Prediction/National Center for Atmospheric Research wind reanalysis fields. The TOPEX/POSEIDON observations show a tongue of high variability extending northwest from the Agulhas retroflection into the central South Atlantic. This arc of high variability may mark a wind-forced, low-frequency shift in the location of the boundary between the South Atlantic subtropical gyre and the tropical waters to the north. The correspondence between the location of this signal and the path of water mass transport from the Indian to the Atlantic basin suggests that this signal may be related to time-dependent variations of interbasin exchange via the "warm water route". Along the edges of the South Atlantic subtropical gyre, low-frequency variations contribute significantly to the total sea level variability of the boundary current regimes of the Brazil, Malvinas, Benguela, Agulhas, and Agulhas Return Currents. The relation between the surface circulation and interbasin exchange is investigated in these boundary current regions, and in the gyre interior, from a comparison of the altimeter data and in situ observations collected during the WOCE and BEST (Benguela Sources and Transport) programs.
Very Low Frequency Sea Level Change Observed by TOPEX/POSEIDON
Center for Space Research, The University of
Texas at Austin, Austin, USA
Department of Marine Sciences, University of South Florida, St.
Petersburg, USA
Over the past four years, the TOPEX/POSEIDON (T/P) satellite altimeter mission has provided 10-day maps of sea level with a point-to-point precision of 2-3 cm. While this is a great achievement in itself, the long term stability of these measurements is even more remarkable, and it allows us to begin studying the very low frequency (VLF) components of sea level change. In particular, the long-term veracity of the measurements of global mean sea level now allow us to place bounds on the rate of sea level rise over the past four years. Studies of VLF sea level change require careful monitoring of the instrument behaviour. In this regard, we have monitored the altimeter performance using comparisons to sea level measured simultaneously at island tide gauges. It has been demonstrated that the tide gauges can monitor the long-term stability of the altimeter to an accuracy of 1-2 mm/year. We will present the latest T/P measurements of changes in mean sea level in the context of the long term instrument behaviour as measured by the tide gauges. These results will also include a discussion of the effects of the recently discovered error in the T/P oscillator correction, which largely accounted for the extreme rates of sea level rise previously observed in the T/P data, as well as the effects of other error sources. We will also discuss the geographical characteristics of the VLF sea level change observed by T/P. We will present maps of VLF changes in sea level, although these maps are currently dominated by interannual changes in sea level, such as the ENSO. We will also present the results of a principal component analysis for extracting the largest modes of VLF sea level change.
The Gravity Field and Steady-State Ocean Circulation Mission (GOCE)
J.A.Johannessen, K. Hasselmann and J.-F. Minster
Earth Sciences Division, ESA-ESTEC, 2200 AG
Noordwijk, Netherlands
Max-Planck-Instit¸t f¸r Meteorologie, Hamburg, Germany
INSU, Paris, France
The European Space Agency (ESA) is presently considering a Gravity Field and Steady-State Ocean Circulation (GOCE) Mission with tentative launch date around 2004 as one of four candidate Earth Explorer Missions. The aim of the GOCE mission is to provide global and regional models of the Earth’s gravity field and of the geoid with high spatial resolution and accuracy. For oceanographic application this would contribute to the correction of a major deficiency in respect to the presently insufficient accuracy of the marine geoid. The GOCE mission would therefore directly respond to some of the requirements put forth by many international scientific programmes such as WOCE, CLIVAR, GCOS and GOOS.
In this presentation, the scientific objectives and observational requirements will be discussed. In particular, the impact of an accurate, high spatial resolution geoid on global and regional ocean circulation retrievals will be addressed in respect to determination of improved estimates of: a) the mean sea surface height (MSSH); and b) ocean circulation properties such as the transport of heat. It will be shown that improvements in the determination of the MSSH field translate directly into comparable improvements in the quantitative estimates of the transport and storage of heat and water masses. It is therefore argued that the GOCE mission would be highly beneficial for ocean circulation studies. This is further illustrated in a brief comparison of the various scientific objectives and measurement requirements associated with different gravity field missions presently under consideration.
Detection of Seasonal to Interannual Oceanic Subsurface Variability via SST and SSH measurements
Robin Tokmakian and Albert J. Semtner
Department of Oceanography, Naval Postgraduate School, Monterey, USA
Long period changes (seasonal and greater) to the ocean thermohaline circulation occur when significant changes in the surface forcing of the ocean (momentum, heat, or fresh water) also occur. This paper discusses how a high resolution global ocean model with high frequency forcing can be a useful tool to determine where such changes in SST and SSH correspond to long period changes below the surface. The Semtner/Chervin Parallel Ocean Climate Model (POCM_4B) (Stammer, et al., 1996, Tokmakian, 1996), with a time series of nine years available for analysis, is initially used for the comparisons. Later, a longer model run with new forcing from reanalysed ECMWF meteorological data will extend the study to a period of about 20 years. The model's relationship between the surface and subsurface processes is explored through the comparisons of time series of its SST and SSH (along with the forcing functions) with subsurface quantities of temperature and salinity variation. The model surface fields are also compared with TOPEX/POSEIDON altimeter data and data from the Advanced Very High Resolution Radiometer (AVHRR) satellites to show how representative the model is of the real world. Specific quantities such as the variation in the model's overturning in each ocean basin are examined along with long used indices such as the North Atlantic Oscillation (NAO) and the variation in transport through passages such as Drake Passage.
The effect of the long period thermohaline changes on the shape of the mean sea surface (relative to the geoid) is also examined. These initial analyses will give an indication of specific surface processes that might be indicative of changes in the subsurface fields both in the real world and in the model.
Propagation and Decay of Forced and Free Baroclinic Rossby Waves in Off-Equatorial Oceans
Bo Qiu, Weifeng Miao and Peter Muller
Department of Oceanography, University of Hawaii at Manoa, Honolulu, USA
Baroclinic Rossby wave motions in the off-equatorial oceans are investigated with emphasis on how eddy dissipation can influence the propagation of the height anomalies when both the forced wave response in the interior ocean and the free-wave response originating along the ocean's eastern boundary are present. By explicitly estimating the decay scale for the long baroclinic Rossby waves, we show that the forced wave patterns at all off-equatorial regions appear to propagate westward at 2 cr, where cr is the phase speed of the long baroclinic wave. The presence of the boundary-generated, free Rossby waves in the low-latitudes, however, can reinforce the 1 cr phase propagation in the combined height anomaly fields. Towards higher latitudes, this reinforcement weakens as the boundary-generated, free waves become highly dissipative; as a result, the forced wave motion becomes more dominant, which works to increase the apparent phase speed up to 2 cr. In the subpolar regions where the annual baroclinic Rossby waves become evanescent, an apparent phase speed higher than 2 cr is observed when an annual, standing-wave response and a propagating baroclinic wave response with an interannual frequency coexist. Stronger annual and interannual wind fluctuations over the southern hemisphere subpolar regions than over the northern hemisphere subpolar regions suggest that this coupling, and the phase speed higher than 2 cr, are more likely to be detected in the southern hemisphere subpolar oceans.
John A. Church,
Richard Coleman, Neil J. White
and Stephen R. Rintoul
Antarctic CRC and CSIRO Division of Marine
Research, Hobart, Australia
Antarctic CRC and Department of Geography and Environmental
Studies, University of Tasmania, Hobart, Australia
The 130 Sv transport carried by the Antarctic Circumpolar Current (ACC) is the primary means by which mass, heat, salt and freshwater are exchanged between the ocean basins. For this reason, the ACC is a critical link in the global pattern of ocean currents that strongly influences the earth’s climate on time-scales of years to centuries. Due to the isolation, vast size and typically extreme weather of the Southern Ocean, the region remains poorly sampled. Historical data generally have insufficient sampling in space and/or time to resolve the fronts which carry most of the transport of the ACC. Satellite altimetry provides global coverage of sea surface height, but additional information is needed to infer transports from measurements of sea surface height variability.
A unique program of repeat CTD and XBT sections initiated in 1991 and moored measurements being carried out south of Australia are providing the only measurements across the ACC with sufficient temporal resolution to monitor seasonal and interannual changes in the transport of the ACC and the location and movement of fronts. Six CTD sections have been completed between 1991 and 1996, covering all seasons. Six high density XBT lines have been occupied along a similar track each year since 1992 during the austral summer (November to March).
The repeat CTD sections have been used to derive a relationship between upper ocean temperatures and dynamic height, despite the fact that the T-S curve varies dramatically across the Southern Ocean. The relationship was used to estimate dynamic height from the repeat XBT sections. These in situ data sets were used to define a synthetic geoid which was combined with TOPEX/POSEIDON altimeter data mapped onto the repeat XBT and CTD tracks. There is good agreement between the variability in the CTD/XBT transects and the TOPEX/POSEIDON satellite altimeter data. The combination of repeat XBTs and satellite altimetry provides a powerful tool for monitoring changes in the strength and location of the major ACC fronts. Results of analysis on the variability of the location and structure of the fronts and the transport of the ACC over the period from 1991 through to 1996 will be presented.
L. Gourdeau, J. Verron, R. Murtugudde, A.J. Busalacchi and J. Picaut
ORSTOM, BP A5, Noumea, New Caledonia
NASA Goddard Space Flight Center, Greenbelt, USA
First results of numerical simulation experiments realised with the primitive-equation, reduced-gravity, sigma-coordinate, Gent & Cane model with assimilation of TOPEX/POSEIDON data are shown over the entire Tropical Pacific during the period encompassing the 1992-1996 El NiÒo/La NiÒa events.
The ocean model is at high resolution on the horizontal and has nine levels on the vertical. It is forced by monthly FSU winds and heat fluxes from an advective atmospheric model. The assimilation procedure is a quite new method derived from the extended Kalman filter and recently proposed by our group under the name of SEEK filter for Singular Evolutive Extended Kalman filter. Assimilated data are TOPEX/POSEIDON sea-level anomalies. Eventually the 0-500 m temperature and current data from the TAO mooring array are used, first for the purpose of validation against the satellite predictions, second in conjunction with the satellite data and in assimilating mode.
Extensive validation has been conducted in simulated twin experiments with synthetic or real data. In this regard, the capability of vertically transferring information from only the altimeter surface data through the assimilation process was seen as very promising.
At this stage, the work being presented in this talk will now deal with a better scrutiny of ENSO mechanisms in relation with the zonal displacement of the eastern edge of the Pacific warm pool and reflection of equatorial waves on ocean boundaries.
Quality of the new tide models over continental shelves
Ocean Modelling Group, Laboratoire des Ecoulements GÈophysiques et Industriels, CNRS-UniversitÈ Joseph Fourier, Grenoble, France
The new tide models recently made available to the scientific community, most of them issued from the analysis of satellite altimeter data (TOPEX/POSEIDON and ERS1/2), have shown major improvements by reference to the previous global ocean tide models published in the literature. Over the deep oceans, these new models have reached a high level of accuracy. A lot of them are indeed very similar (within one centimetre RMS difference, see Shum et al, 1997). But large differences between these models have been observed over continental shelves, of the order of tens of centimetres.
The evaluation of the quality of any global ocean tide model over shallow water areas is a difficult task, because of the complexity of the tidal characteristics over shelves and coastal basins. There, the typical wavelengths are shorter. Regional amplifications, often due to local resonance, result in sharp gradients which are difficult to catch in the models. And non-linearities are taking place, which lead to more complex tidal spectra than over the deep oceans.
The aim of this talk is to propose a rationale for evaluating the quality of the available global ocean tide models over continental shelves and shallow water areas. The method relies on the use of the enormous amount of coastal data available in the International Hydrographic Bureau data bank. A selection of stations uniformly distributed along the coastlines of the world oceans will be proposed, based on a careful evaluation of the intrinsic quality of the data issued from these stations, their spatial coherence, and their level of agreement with the ocean tide models themselves.
Benchmarks will be given on the quality of some of the ocean tide models which are now used for satellite altimetric data corrections, and for other geophysical applications.
Mean and Time-Varying Long-Wavelength Topography from TOPEX/POSEIDON Altimetry
B.D. Tapley, D.P. Chambers, J.C. Ries and C.K. Shum
Center for Space Research, University of Texas at Austin, Austin, USA
Over four years of TOPEX/POSEIDON data have been used to study long-wavelength dynamic ocean topography. Mean topography models with respect to the JGM-3 and TEG-3 geoids are presented, and the accuracy of the topography and geoid models is discussed. The time-varying topography is also studied, focusing on changes coherent over wavelengths of 1500 km or longer. Variations in the global averaged sea level are presented, as well as local variations at intraseasonal, annual, and interannual time-scales. Variability associated with annual and interannual heating and annual Rossby waves in the tropical Pacific and Indian Oceans is discussed in particular.
Impact of an Improved Gravity Field on Surface Heat Flux Estimates
Earth and Space Research, Seattle, USA
It is well known that an accurate static geoid will directly benefit ocean circulation studies through the determination of a mean geostrophic surface current in conjunction with precise satellite altimeter measurements. At present the geoid uncertainty limits this calculation to wavelengths longer than about 2000 km. Spaceborne gravity missions are expected to reduce this uncertainty so that resolutions of a few hundred km will be obtained. One important need for the mean surface velocity is the estimate of the surface layer advective heat flux, which is a significant component of the upper ocean heat budget in many ocean regions. Because geostrophic current is proportional to the dynamic height gradient, the uncertainty in the surface heat advection can be estimated from the velocity uncertainty, which is in turn related to the geoid slope uncertainty. The local gravity anomaly is proportional to the geoid slope (1 mGal is equivalent to approximately 1 mm per km).
It is desirable to resolve the net surface heat flux and upper ocean heat budget with an uncertainty of less than 10 Wm-2 for climate studies. The mid latitude North Pacific SST undergoes significant interannual and decadal climate variability in which surface advection may play an important role. Based on mean climatological SST gradients and assuming a mixed layer depth of 50 m, it is estimated that gravity errors less than 0.05 mGal are required to achieve less than 10 Wm-2 heat flux error in the surface layer advection term. In addition, the discussion will include recent results of surface circulation and heat flux studies in the mid latitude and tropical Pacific using a hybrid Ekman/geostrophic model, TOPEX/POSEIDON data, and a reference dynamic height from Levitus climatology and various GCM simulations.