GODAE is sponsored by
List of abstracts (WS1)
(Alphabetical order)
GODAE-OOPC OSSE-OSE meeting, 5-7 November, IOC Paris, France
OSE's in the ECMWF operational ocean analysis: Impact on the ocean mean state, seasonal forecasts and climate variability |
Magdalena Balmaseda1
D.L.T. Anderson1 and A. Vidard 1
(1) European Centre for Medium Range Weather Forecast (ECMWF), UK
The impact of Argo and Altimeter data on the new ECMWF ocean analysis system during the period 2001-2006 has been assessed by conducting a set of observing system experiments. The experiments evaluate the information content of Argo temperature and salinity data, and its synergy with the altimeter data. The impact of the different observing system is gauged in terms of its influence on the ocean state and its impact on skill of the seasonal forecasts.
The observing system experiments are also used to assess the robustness of recent climate signals, such as the 2003 global cooling and sea level rise.
A discussion of the limitation of the OSE technique will be provided, as well as a description of alternatives already used in meteorology.
Recent progress under a NOPP-CODAE project on boundary conditions, data assimilation, and predictability of coastal ocean flows is reviewed. This research focuses on numerical model simulations of ocean circulation in the Oregon coastal transition zone (CTZ), the region extending several hundred kim offshore from the Oregon coast into the northern portion of the California Current System and the eastern interior North Pacific. The CTZ model is nested in the NRL NCOM-CCS model, and impact of these open boundary conditions on the CTZ model is examined. Atmospheric forcing is obtained from the NRL COAMPS product. Validation of the simulated coastal ocean circulation is provided by comparison with in-situ data sets from regional observational programs during 2000-2003 and with satellite and land-based remote-sensing observations.
NOAA's OSE/OSSE Requirement |
Eric Bayler1
(1) NESDIS/NOAA, USA
NOAA's leadership views Observing System Evaluations (OSEs) and Observing System Simulation Experiments (OSSEs) as powerful and relatively inexpensive ways of assessing the impact of potential new observations, for determining the impact of losing current observing systems, and for refining and redirecting current observing practices. NOAA critically requires ocean OSE/OSSE development and implementation for satellite and in-situ data integration, as well as optimization of the U.S. Integrated Ocean Observing System (US IOOS), the U.S. contribution to the Global Ocean Observing System (GOOS). NOAA leadership's budgetary guidance is that NOAA should carry out an OSSE as part of the preparation for certain observation system deployments, and as a way of redefining the overall observing strategy to fill critical mission needs. Further, with the advent of new forecast techniques, including high-resolution models, OSSEs would provide an excellent way of fine-tuning the observing system to the forecast needs. With OSE/OSSEs a specific focus for NOAA ocean data assimilation and being an area where NOAA ocean modeling has minimal exposure/development, it is crucial for NOAA to establish participation at the ground floor of national and international efforts.
Optimal design of observing systems: review of methods based on assimilative models |
Pierre Brasseur1
(1) CNRS, France
Data assimilative systems provide useful tools to conduct OSSEs for the optimal design of observing systems. In the ocean, a crucial question is to identify the best complementarity between in situ and satellite observations. The information content of multi-component observing systems can be efficiently evaluated using diagnostic quantities such as: degrees of freedom for system and degrees of freedom for noise. In this presentation, the basic concepts behind OSSE, impact studies and diagnostics of information content will be overview and illustrated using oceanographic and atmospheric examples.
Assessment of Observation Impact Using a Variational Assimilation Adjoint System |
Jim Cummings1
Eric Chassignet2 and Harley Hurlburt 3
(1) NRL, Monterey, CA, USA
(2) Florida State University, Tallahassee, FL, USA
(3) NRL, Stennis Space Center, MS, USA
An adjoint-based procedure for assessing the impact of any or all observations used in a data assimilation and forecast system is described. The procedure has been developed at the Naval Research Laboratory (NRL) for use in Navy numerical weather prediction (NWP). The method can be directly applied in ocean prediction systems as well. Observation impact is calculated using adjoint sensitivity gradients and actual innovations (model-data differences) to estimate the impact of each observation on the short term forecast error or cost function. Impact assessment is sensitive to the selected cost function, which is typically defined as a measure of forecast error at different forecast periods. The method is computationally inexpensive and allows impacts to be partitioned for any set of observations: by instrument type, observed variable, geographic region, vertical level, or individual reporting platform. It is not necessary to add or remove observations from the assimilation to estimate their impact on the forecast. The procedure can be routinely applied in an operational analysis/forecast system as a diagnostic tool to monitor the assimilation and quality control of the observations as well as perform predictability/data impact studies. This paper describes plans for implementing an adjoint observation impact system in the HYbrid Coordinate Ocean Model (HYCOM) prediction system that is being transitioned into operations at the Naval Oceanographic Office. Observations are assimilated in HYCOM using the Navy Coupled Ocean Data Assimilation (NCODA) system. NCODA is being ported to a 3-dimensional variational analysis (3DVar). Adjoints for the HYCOM forecast model and the NCODA 3DVar assimilation are readily available.
Review of OSSE/OSE performed at Mercator-Ocean |
Eric Dombrowsky1
Fabrice Hernandez1, Mounir Benkiran1, Eric Greiner1, Sylvie Giraud 2 and Benoit Tranchant1
(1) Mercator-Ocean, France
(2) CLS, France
Mercator Océan is the French operational oceanography center. It operates systems based on assimilation of earth observation data into OGCM configurations. Observations assimilated are altimeter sea level anomalies, in situ T and S profiles, and SST products. The OGCM is based on NEMO code, deployed at the global scale at ¼° and at the regional scale (North Atlantic and Mediterranean) at a 5 to 7 km horizontal resolution. The assimilation schemes are all based on multivariate statistics allowing the fully multivariate assimilation of these data in a single step, with the analysis step being the optimal interpolation for the regional and the SEEK filter for the global. Mercator Océan has conducted several impact studies both with real data (OSEs) and simulated (OSSE) data with its prediction systems, either for the evaluation of the impact of existing observing systems (altimeter data, tropical T/S profiles, etc.) or to evaluate the potential of other observing systems (SMOS, surface velocities,etc.).
We will give and overview talk presenting the major results obtained, and the main conclusions.
OSSE/OSE activities with Multivariate Ocean Variational Estimation (MOVE) System.
|
Yosuke Fujii1
H. Tsujino1, N. Usui1, H. Nakano1 and M. Kamachi1
(1) Meteorological Research Institute / Japan Meteorological Agency
Singular vector analysis is one of promising tools to identify a key region important for predicting a certain target phenomenon. It is, therefore, useful for designing an efficient observing system. Singular vector analysis is applied to the formation of the Kuroshio large meander south of Japan. The analysis reveals that an anticyclonic eddy approaching the Kuroshio southeast of Kyushu (southwestern part of Japan) from east activates the baroclinic instability, and generates an anticyclonic eddy in the deep layer (deeper than 1500m). This eddy interacts with a small meander of the Kuroshio in the upper layer, and promotes the growth of it to the large meander. This result implies the importance of observing around southeast of Kyushu and the deep layer south of Japan.
Feasibility of a 100-year long ocean reanalysis |
Ben Giese1
Jim Carton2
(1) Texas A & M University, USA
(2) Department of Atmospheric and Oceanic Science, University of Maryland, USA
The historical ocean profile and SST data sets have been substantially expanded and cleaned through the data archaeology efforts of GODAR, contained in WOD05 and ICOADS. Although limited in regions outside of the Atlantic, there are hydrographic records that date back to the early 1900s. A new effort at the NOAA Earth System Research Laboratory/CIRES to produce a 100-year atmospheric reanalysis is now underway. The NCEP reanalysis extends back through the first half of the 20th Century based on sea level pressure and SST. Several additional long-term data sets are also available. For example, tide gauge sea level records are available from the Permanent Service for Mean Sea Level for some 1800 stations. Although only a much smaller number are available during the first half of the 20th Century, some are located appropriately to reflect large scale ocean conditions, and are fairly continuous. Advances have also been made in the field of data assimilation that match recent increases in computational resources. This all raises the possibility that there could be ocean reanlayses that span the 20th Century. The feasibility and the efforts needed to proceed with a 100-year long ocean reanalysis will be discussed.
Contributions of the ARGO array and complementarities with the altimeter observing system |
Stephanie Guinehut1
Marie-Hélène Rio1, Gilles Larnicol 1, Pierre-Yves Le Traon2
(1) CLS, France
(2) Ifremer, France
Contributions of the ARGO array and complementarities with the altimeter observing system. The presentation follows twofold objectives. The first one concern the analysis of the contribution of different profiling float arrays to the description of the 3-D large-scale thermohaline fields and the second one is focus on the complementarities between in-situ and altimeter observing systems. Firstly, results from an OSSE study (Guinehut et al., 2002) will be presented. It uses outputs and profiling float simulations derived from a primitive equation model of the North Atlantic Ocean to simulate different ARGO array (1°x1°, 3°x3°, 5°x5°) and their ability to retrieve the large scale and low frequency ocean signal variability. Comparison between Eulerian and Lagrangian arrays have been done. Results from a similar study led on the Mediterranean Sea will be also presented (Guinehut, 2006). Secondly, the complementary of the two observing systems will be discussed. The methodology of combination will be described and the respective contribution of those observing system for the retrieval of the 3D-thermohaline fields will be assessed (Guinehut et al., 2004; Larnicol et al, 2006). In particular, the problem of aliasing of the mesoscale signal is analysed. Issues concerning the methodology and future plans will be discussed. discussed.
Impact studies on the altimeter observing system: review of the work done by the SSALTO/DUACS center |
Gilles Larnicol1
Marie-Hélène Rio1, Gérald Dibarboure 1, Ananda Pascual 2 and Pierre-Yves Le Traon3
(1) CLS, France
(2) IMEDEA (UIB-CSIC), Spain
(3) Ifremer, France
The aim of the presentation is to give a review of the studies done in the past to characterise the impact of the different satellite configurations to monitor the ocean variability. Several kinds of studies will be distinguished. OSSE studies that evaluate the potential impact of a future mission or the design of the satellite configuration, OSE studies that evaluate the performance of an observing system with real data. In particular, the impact of four or two satellite configuration on the representation of the mesoscale activity will be presented (see Pascual et al, 2004, 2006), both for the global ocean and Mediterranean Sea. The role and the impact of the Mean Dynamic Topography (Rio et al, 2004, 2007) will be also described as well as the improvement of the altimeter processing (impact of the geophysical correction, of the high frequency signal aliasing, Volkov et al, 2006). Issues concerning the methodology and future plans will be discussed.
Observation bias correction in altimeter ocean data assimilation in FOAM comparing two MDTs |
Daniel Lea1
(1) Met Office, UK
Altimeter observations provide a useful constraint to the ocean circulation. However, while the sea surface level anomalies are accurately measured, the mean dynamic topography (MDT) is less well constrained. We compare the original FOAM MDT which is a combination of the model mean sea surface height (SSH) and the Singh and Kelly mean SSH in the Gulf Stream region with a new MDT developed by R. Bingham which is derived from GRACE gravity observations and drifter data. This comparison is performed using 5 year hindcasts of the FOAM (Unified Model) OI ocean data assimilation system which assimilates temperature and salinity data as well as SSH data. Various diagnostics including observation minus background statistics are used to assess the results. These show that the Bingham MDT produces superior results compared to the original FOAM MDT.
In addition to assimilating SSH with the two MDTs described above, we run experiments with an online observation bias correction scheme which aims to correct the errors in the MDTs. The results from both MDTs converge to an optimal MDT which produces more accurate results than either of the non-bias corrected experiments.
Argo and other profile data assimilation on z and density levels in FOAM |
Daniel Lea1
Matthew Martin1
(1) Met Office, UK
Argo data is an important aspect of the ocean observing system. Two five-year hindcast experiments, using the UK Met Office FOAM (Unified Model) OI ocean data assimilation system, one with Argo data included and one without, have been carried out. Both systems assimilate non-Argo in situ temperature and salinity, and sea surface temperature. The assimilation experiment with Argo data is shown to be substantially more realistic than the experiment excluding it.
The network of Argo observations with temperature and salinity profile data provides an opportunity to consider a novel data assimilation scheme, based on density levels, in order to maximise the information extracted from the data. In this method the model background field is converted to spiciness on a set of specified density levels and the depths of those density levels are calculated. The profile observations are also converted to spiciness and density level depth for the same set of specified density levels. We perform two analyses, first of spiciness on density levels with the data being spread over large spatial scales, and second of density level depth with smaller spatial scales. The order of the analyses is not important. The resulting increment fields are converted back to temperature and salinity on model z-levels. The results of this are compared to the standard FOAM assimilation on z-levels over a one year period assimilating the same observations. Future work includes using these systems to assess the impact of the assimilation scheme on OSSE/OSE results. For example, we can test different observation densities in order to test the ability of density level assimilation to effectively spread the observation information further than z-level assimilation.
What do we gain having an additional SeaWinds-like scatterometer? |
Tong Lee1
(1) NASA Jet Propulsion Laboratory, California Institute of Technology, USA
The tandem mission of SeaWinds scatterometers on QuikSCAT and on ADEOS II satellites from April to October of 2003 provides vector wind stress measurements to evaluate the impact having two SeaWinds-like scatterometers versus one. One such instruments samples about 60% of the world ocean daily. The two together, with local equatorial crossing times of 6am/pm and 10:30am/pm, sample about 90% of the world ocean daily. The effects of such enhanced sampling of the wind field on the structure of the upper ocean is evaluated by using the wind stress measurements from one and both to force an ocean model in separate experiments. In particular, the impact on sea surface temperature, mixed-layer depth, and vertical mixing are described. The implications to data assimilation, air-sea interaction, and biogeochemistry are discussed.
The impact of assimilating sea surface height data from one, two and three altimeters on the surface currents in the 1/9° North Atlantic FOAM system |
Matthew Martin1
(1) Met Office, UK
The FOAM deep ocean forecasting system assimilates in situ temperature and salinity data, in situ and satellite sea surface temperature (SST) data, and satellite altimeter sea surface height (SSH) data. The altimeter data is the along-track sea level anomaly data provided by CLS and is obtained by processing data from the Jason, Envisat and GeoSat Follow-On (GFO) satellites. A Mean Dynamic Topography (MDT) is added to the data so that the analysis scheme can produce a field of increments to the surface height, and the Cooper and Haines (1996) scheme is used to update the model's temperature and salinity fields.
A number of integrations of the 1/9° resolution North Atlantic FOAM system have been run for a four month period from 1st December 2005 to 31st March 2006, the last three months of which have been used for assessment. The runs include one in which no altimeter data was assimilated, a run in which only Jason altimeter data were assimilated, a run in which both Jason and Envisat data were assimilated, and one in which data from all three satellites were assimilated. All of these runs include assimilation of the other temperature and salinity data sources. The output of these integrations has been assessed by comparing the SSH of the model to the data, and also by comparing the surface currents to independent data from surface drifters. These results indicate that a single satellite altimeter significantly improves the results compared with no altimeter data, and that errors are further reduced when data from both the second and third altimeters are assimilated.
OSSE/OSE activities with Multivariate Ocean Variational Estimation (MOVE) System.
|
Satoshi Matsumoto1
Y. Fujii1, T. Soga1, T. Yasuda1, S. Ishizaki1, N. Usui1 and M. Kamachi1
(1) Meteorological Research Institute / Japan Meteorological Agency
Multivariate Ocean Variational Estimation (MOVE) System is the ocean data assimilation system developed in MRI. In MOVE System, not only temperature but also salinity is analyzed using temperature and salinity profiles and satellite altimetry, and directly corrected in the model. In order to assess the importance of salinity observation by Argo floats, impact of the salinity correction in MOVE system is examined. Salinity correction affects the stability of the density stratification, and therefore has an impacts on the temperature field in the near-surface and subsurface layers. We also examine the effect of assimilating TAO/TRITON on analysis fields and ENSO forecasts. TAO/TRITON plays an important role in keeping the sharp thermocline at the equator in the central Pacific, but the impact on the forecast is not clear.
Status of the 1999 XBT Recommendations Made to the Ocean Obs99 Meeting |
Robert L. Molinari1
Gustavo Goni 1, Molly Baringer1, Claudia Schmid1, Silvia Garzoli1
(1) RSMAS/University of Miami, USA
(2) NOAA/AOML, USA
Smith et al. (1999) generated nine recommendations designed both to improve the global XBT observing network and to integrate this network with Argo and satellite altimetry. Recommendations included discontinuing the low-density network; increasing the frequently sampled and high-density sampling; ensuring that Argo and altimetry can resolve climatically important signals such as accomplished by the low?density network and improving data management. Herein, we review the present day status of these recommendations and modifications that might have been made in the nine-year interim. We also propose additional modifications based on analyses of existing data. We show that some low-density lines can be extended back to the late 1940's using mechanical bathythermograph data and provide valuable information on decadal scale upper ocean variability. The ability of the Argo array to resolve oceanic features such as the southern recirculation gyre of the Gulf Stream is illustrated. An evaluation of the ocean observing system for upper ocean heat storage that utilizes XBT, Argo and altimetry observations is also shown. Additional analyses of results from high-density lines provide support for the original 1999 recommendation to maintain this sampling approach.
Objective Array Design: Application to the tropical Indian Ocean |
Peter Oke1
Pavel Sakov 1
(1) CSIRO, Marine and Atmospheric Research, Australia
We describe a simple, versatile, computationally efficient ensemble-based method for objectively designing an observation array. The method seeks to compute the observation array that minimises the analysis error variance, according to Kalman Filter theory. While most elements of the method have been described elsewhere, we attempt to present a simple yet comprehensive recipe for array design based on an ensemble of anomalies that represents the background error covariance. The versatility of the method is demonstrated through a series of applications to the tropical Indian Ocean (TIO). The first application uses model-generated fields of high-pass filtered mixed layer depth to design an array to monitor intraseasonal variability. The second uses gridded observations of sea-level anomaly to design an array to monitor intraseasonal to interannual variability. For both applications, the objectively designed arrays are compared to an array that will soon be implemented under the auspices of the CLIVAR-GOOS Indian Ocean Panel (CG-IOP). We conclude that the CG-IOP array produces results that compare well to the objectively designed arrays for intraseasonal variability, and that observations to the east and north-east of the TIO and south of India are most important for resolving intraseasonal variability. We also find that observations near 9oS, where seasonal Rossby waves dominate, are important for observing seasonal to interannual variability. The described method for objective array design can be applied to a wide range of geophysical applications where time series of gridded modelled or observed fields are available.
Impact of Argo, SST and altimeter data on an eddy-resolving ocean reanalysis |
Peter Oke1
Andreas Schiller 1
(1) CSIRO, Marine and Atmospheric Research, Australia
We perform a series of Observing System Experiments (OSEs), where components of the Global Ocean Observing System (GOOS) are systematically withheld from a data assimilating ocean reanalysis. We assess the relative importance of Argo temperature (T) and salinity (S) profiles, sea-surface temperature (SST) and altimetric sea-level anomalies (SLA) for constraining upper-ocean T and S properties and mesoscale variability of SLA in an eddy resolving ocean reanalysis in the Australian region. Each OSE is assessed by comparing modelled fields with assimilated and withheld observations. We show that each observation type brings complementary information to the GOOS, and demonstrate that while there is some redundancy for representing broad-scale circulation, mesoscale circulation requires all observation types to be assimilated.
Design of surface drifting buoy deployments using eddy resolving reanalyses and representing a vortex merger in an ocean prediction system |
Gary B. Brassington1
Nicolas Summons 1
(1) Centre for Weather and Climate Research, BMRC Melbourne, Australia
(to be presented by Peter Oke)
Ensemble Lagrangian drifter trajectories are derived from an eddy resolving ocean reanalysis to optimise the deployment strategy of surface drifting buoys for the East Australian Current. The deployment was constrained to be performed from two Volunteer Observing Ships that occupy the PX30 line and traverses the EAC reducing the design to selecting the longitude. Critical design criteria include: (a) low probability of convergence into the coast, (b) high probability of observing the EAC and (c) high probability of a long deployment period. The BLUElink ocean reanalysis provided three dimensional daily average velocities at 1/10 degree resolution. The Lagrangian model was developed and calibrated against historical observations during the reanalysis period. The analysis guided the deployment location of 8 drifting buoys between Feb-March 2007 with a 100% success rate of initial deployment. Experiment design, results and future plans are reviewed.
One pair from the experiment were deployed on an anticyclonic eddy which was forming and interacting with the core EAC. The pair of drifters remained on the eddy for a total of 8 orbits of the eddy showing long-lived stable surface flow. During this period the newly formed eddy came under the influence of an offshore deep anticyclonic eddy and separated from the EAC and merged. The merger occurred with minimal shear and Argo profiles suggest the eddy superimposed onto the offshore eddy over the top 150m. We conclude that this was a stratified merger of a deep old anticyclonic eddy and a more buoyant newly formed eddy with examples to be found in recent Quasi-geostrophic simulations. Deep, large anticyclones have been observed to persist in the Tasman Sea for 1-2 years while new buoyant anticyclones are generated from the EAC each summer season. This type of vortex merger may occur frequently and have a significant influence on the circulation in the region. We show how this merger was represented in OceanMAPS, the Bureau's operational ocean prediction system and discuss the implications and observing system requirements.
(Last Updated: 20-02-2009)