Abstract Template



Regional high resolution analyses and forecasts

at the mesoscale

M. Kamachi1, P. DeMey2, F. Davidson3, Y.-H. Kim4, J.-M. Lellouche5, C. Rowley6 D. Storkey7, K. Thompson8, and J.-H. Yoon9

 

1MRI, Tsukuba, Japan

2 LEGOS, Toulouse, France

3 DFO, St. John's, Canada

4 KORDI, Seoul, Korea

5 MERCATOR, Toulouse, France

6 NRL, Stennis Space Center, USA

7 UKMO, Exeter, UK

8 Dalhousie U., Halifax, Canada

9 Kyushu U., Fukuoka, Japan

Abstract

GODAE is associated with real time high resolution products of global ocean model assimilation and prediction, usually space-time gridded data. Several GODAE partners currently operate regional systems that are usually basin scale systems with high resolution and are nested to global systems. The regional system has higher resolution with less computational burden compared with global system. The intent of the paper is not to provide comprehensive reviews but rather a scope of the operational/research status of basin scale nowcasting and forecasting of ocean currents and mesoscale eddies. Two examples of coordination of North Atlantic and North Pacific systems, which have been developed in the GODAE period, are reported with different systems of GODAE partners. The paper also briefly shows linkage between regional and coastal/shelf-sea systems.

1. Introduction

The initial objective of GODAE was the demonstration of nowcasting and forecasting to cover global (focus on the global capability) and also regional systems (e.g. basin scale at high resolution - zoom on a specific region such as the Kuroshio and Gulf Stream) that are the core of GODAE systems. Both global and regional assimilation/prediction systems give the product as the initial and boundary conditions for the coastal/shelf sea assimilation/prediction systems. Coastal/shelf sea systems are not part of GODAE core, but are recently addressed as the Pilot Project. Scientific developments needed are downscaling techniques (see the review paper of De Mey et al. in this Proceedings) and biogeochemical processes (such as IMBER project).

Several GODAE partners/centers, in France, UK, Canada, USA, Korea and Japan, have developed and currently operate regional assimilation-prediction systems that are usually basin scale systems with high resolution and are nested to global systems. The systems have specialities for site dependent experiences that closely related to site-specific phenomena. The regional system has higher resolution with less computational burden compared with global system. It gives products of mesoscale ocean state such as ocean currents and mesoscale eddies. It also gives open boundary conditions to coastal/shelf sea system.

The intent of the paper shows a scope of the operational/research status of basin scale nowcasting and forecasting of ocean currents and mesoscale eddies. Each basin has site-specific phenomena in each systems such as Gulf Stream with mesoscale eddy in the North Atlantic, Kuroshio large meander in the North pacific, Loop current in the Gulf of Mexico, regional and coastal (relocatable) ocean prediction models, European shelf-sea system, East Asian Seas systems, real-time Intra-Americas Sea system, eddy activity in the Arabian Sea, Korean shelf-sea state with Tsushima current variability, Oyashio intrusion in the confluence zone of subtropical and subpolare gyres boundary etc. The systems and relationship with coordinated systems in the North Atlantic and North Pacific, as examples of GODAE results, are shown. Some examples of coastal/shelf sea systems are also briefly reported.

Structure of this paper is as follows: Examples of regional systems are explained briefly in section 2. Section 3 gives examples of highlighted examples, and section 4 gives connections to coastal/shelf sea systems. Section 5 dedicates summary.

2. Overview of the system

GODAE centers/partners which systems are related to the North Atlantic in France, UK, Canada, USA, and other EU countries have several coordinations. Geographical relationship of the systems of Canadian group, UK Met Office, French MERCATOR OCEAN and related EU countries is shown in Fig. 1. Similarly GODAE centers/partners related to the North Pacific in Japan, Korea, China, Australia and USA have also some coordinations. An example of the relationship among the systems in Japan and Korea is shown in Fig. 2. In this section, we introduce overview of the systems that are shown in the Figures 1 and 2.

Fig.1 Geographical relations of global, regional, and coastal/shelf sea systems related to the North Atlantic

Fig.2 Geographical relations of global, regional, and coastal/shelf sea systems related to the North Pacific

2.1 North Atlantic

(MERCATOR OCEAN)

Since the beginning of GODAE and also in the framework of the European projects MERSEA and now Kopernikus/MyOcean, Mercator-Ocean has been designing a hierarchy of operational oceanography analysis and forecasting systems. These systems are based on numerical models of the ocean and data assimilation systems which interpolate in an optimal way all available observations of the ocean. The real time operation of these systems began in 2001, in order to produce each week realistic 3-dimensional oceanic conditions (temperature, salinity, currents…) two weeks back in time and a two weeks forecast, driven at the surface by atmospheric conditions.

Since April 2008, the Mercator-Ocean forecasting system demonstrates that the use of the ocean and sea ice model NEMO and of the data assimilation system SAM2 (Système d’Assimilation Mercator version 2) based on the SEEK filter, can produce high quality real time analyses and forecast of the ocean at the “eddy resolving” resolution. We are ready to consider the challenge of high resolution raised by strong user demands and the transition to a strong operational activity oriented towards social needs. Developing this new generation of ocean service is the key objective of Mercator-Ocean operational oceanography. After almost 8 years of continuous operations, the Mercator-Ocean products have been involved in a wide range of application sectors. Four categories are considered: (1) Institutional Operational applications; (2) Research; (3) Private sector Operational Recreational and Commercial applications and (4) Environment Policy Makers. Under category 1, Mercator-Ocean has been involved in various experiments concerning: Oil Spill drifts experiments (Météo-France, Met.No, CEDRE), Navy operations (SHOM), Ocean inputs for Seasonal Forecasting system (Météo-France) and Education (schools, user training sessions, individual requests, ...). Under category 2, Mercator-Ocean has been requested to provide boundary conditions to coastal models, ocean inputs for biogeochemical models and seasonal forecasting systems, and involved in various Research Sea campaign (IRD, Ifremer, CNRS, IFM Kiel, ...). Mercator-Ocean has been serving under category 3 commercial activities (offshore and fisheries) and many recreational marine activities (sailing races, rowing races, ...), and a growing range of activities enter today category 4 amongst with: assessment on observation network (satellite and in situ) for decision makers, monitoring and expertise on extreme ocean climate events or new indicators for ocean pollution risk.

The new high resolution regional Mercator-Ocean operational system (hereafter called PSY2v3) currently comprises an ocean model and the SAM2 data assimilation scheme based on the SEEK filter. The PSY2v3 ocean model component is built from the OGCM NEMO 1.09 (Madec, 2008). It consists of an eddy resolving regional ocean model coupled to the sea ice model LIM2 (Fichefet and Gaspar, 1998). The geographical domain of this model configuration is the North Atlantic between 20°S and 80°N and the Mediterranean Sea (see Figure 1). The horizontal grid is an extraction of the global quasi isotropic tripolar ORCA grid (Madec and Imbard, 1996) with 1/12°x1/12°cosinus(latitude) resolution. The vertical resolution is based on 50 levels. The vertical coordinate is z-level with partial steps (Barnier et al., 2006).

The PSY2v3 assimilation system is based on the SAM2 tool which is a multivariate assimilation algorithm derived from the Singular Extended Evolutive Kalman (SEEK) filter analysis method (Pham et al., 1998) with the First Guess at Appropriate Time (FGAT) approximation. The error statistics are represented in a subspace of small dimension where the background error covariances are modelled by an ensemble of 3D anomalies. The formulation of the assimilation algorithm relies on a low-rank error covariance matrix, which makes the calculations tractable even with state vectors of very large dimension. The extrapolation of the data from observed to non-observed variables is performed along the directions represented by these error modes which connect all dynamical variables and grid points of the numerical domain. Weekly dependant multivariate 3D anomalies (HBAR, TEM, SAL, U, V)[1], computed from an interannual simulation, have been used to estimate the background error. The analysis provides a 3D oceanic correction (TEM, SAL, U, V), which is applied to the model at each analysis cycle. Unlike the original SEEK filter, SAM2 does not evolve the error statistics according to the model dynamics. However, some form of evolutivity of the background error is taken into account by adapting the error variance at each analysis cycle. Numerically, the analysis step of the conventional Kalman Filter is reformulated to take advantage of the low-rank approximation, leading to a more efficient inversion in the reduced space than in the observation space (Testut et al., 2003). To minimise the computational requirements, the analysis kernel in SAM2 is massively parallelized and integrated in the operational platform hosting both the SAM2 kernel families via the PALM software (Piacentini et al., 2003). Lastly, the system assimilates in a fully multivariate way, conjointly the SLA from satellites Jason, Envisat and GFO, SST and the T/S vertical profiles from Coriolis including ARGO.

(UK Met Office)

Operational ocean modelling at the UK Met Office is based on a hierarchy of nested models with a global 1/4 degree configuration and basin-scale 1/12 degree configurations in the North Atlantic, Mediterranean and Indian Ocean. The global system is based on the "ORCA025" configuration which is shared with Mercator Ocean in France. Shelf seas and coastal models are nested within the North Atlantic configuration to cover the north-west European shelf and UK coastal waters. Boundary conditions are also provided from the deep-ocean models for high-resolution local models run by met.no in Norway and ESEOO in Spain. The global and basin-scale models are based on the NEMO ocean modelling system (Madec, 2008) with the LIM2 sea-ice model (Fichefet et al., 1997). They run on a daily cycle assimilating temperature and salinity profile data (including ARGO floats), satellite altimeter and GHRSST SST data, and SSMI sea-ice concentration. The assimilation scheme is currently an optimal interpolation scheme and will transition to a variational system over the next 2-3 years. Surface forcing is 6-hourly from the Met Office NWP model. Figure 3 shows the realistic sea suraface temperature field. All models run daily in the operational suite and produce output for customers and downstream services as well as automated verification statistics.

Fig. 3 Sea-surface temperature analysis (deg C) for 00Z 7/10/2008 from the UKMO FOAM 1/12 degree North Atlantic model.

(CANADA)

The approach taken in Canada to develop an operational ocean forecasting capability for the North Atlantic is to use results from a global model to drive a regional ocean forecast system in which, in turn, can be used to drive higher resolution models of the adjacent shelf seas. The present region of interest, and the associated model domains, are shown in Figure 1. One-way nesting is presently used for the Northwest Atlantic domain although the system has been tested in two-way nesting mode using smaller regions of interest. (The AGRIFF package was used.) The C-NOOFS pre-operational forecast system started on 1 January, 2007. It is based on a ¼ degree regional model that has been one-way coupled to a ¼ degree global model run by Mercator Ocean (i.e., the PSY3V2 system). Weekly output from the global model is used to provide initial and boundary conditions for the regional model. The atmospheric forcing fields are provided with a horizontal resolution of 33km by Environment Canada (based on their general environmental multiscale (GEM) atmospheric forecast model). Presently the C-NOOFS system updates a public website (c-noofs.gc.ca) daily with images of predicted sea level and also maps of velocity, temperature and salinity at various depths. Clients for ftp data use include the Canadian Coast Guard and Canadian Forces.

The present version of the C-NOOFS operational forecast system does not assimilate ocean data (although the open boundary conditions used to drive the system are the result of an assimilative model developed, and run, by the Mercator Ocean as described above.) There are plans to use the SEEK filter to assimilate ocean data into C-NOOFS in the near future. In parallel with the development of C-NOOFS, new operational schemes have been developed for assimilating altimeter and Argo data into an eddy resollving model of the North Atlantic. For example Liu and Thompson (2008) have demonstrated useful forecast skill with lead times that range from 10 days in the Gulf Stream region, and over 20 days for the rest of the North Atlantic, using a physically-motivated scheme that builds on the water property scheme.

2.2 North Pacific

(JMA-MRI)

The Meteorological Research Institute (MRI) multivariate ocean variational estimation (MOVE/) Systems in japan Meteorological Agency (JMA) has been developed for research system to study mesoscale phenomena and climate variability, and for the operational ocean data assimilation system in Japan Meteorological Agency (JMA). The MOVE/ systems have three varieties, the global (MOVE/-G), and North Pacific (MOVE/-NP), and western North Pacific (MOVE/-WNP) systems (Usui et al., 2005). The aims of MOVE/-G are initialization of MRI coupled GCM for seasonal-interannual forecasting and analysis/reanalysis project, which is related to CLIVAR/GSOP project. The period of the analysis/reanalysis product is 1948 to 2007. The aims of MOVE/-NP and -WNP are initialization of ocean forecasting in the North pacific (esp. around Japan) and analysis/reanalysis, which period is 1993-2007 and which is related to GODAE project.

MOVE/ systems are composed of OGCMs and a variational analysis scheme which synthesizes the observed information (i.e., temperature, salinity and sea surface height) together with the OGCMs. The numerical code for the OGCMs used in the system is the MRI community ocean model (). has been developed in JMA/MRI and is independent of any other popular OGCM code. It is a multilevel model code that solves the primitive equations under the hydrostatic and Boussinesq approximations. The vertical coordinate is a terrain following-depth (σ−z) hybrid, i.e., the levels near the surface follow the surface topography. It enables us to adopt a fine vertical resolution near the surface because it prevents the uppermost layer from vanishing during integration when the free surface variation is explicitly solved. For momentum advection, uses the generalized enstrophy-preserving scheme along with the Takano–Oonishi scheme, which contains the concept of diagonally upward/downward mass momentum fluxes along a sloping bottom (Ishizaki and Motoi, 1999). Detailed model structure with numerical schemes are reported by Ishikawa et al. (2005) and Tsujino et al. (2006). For ocean weather, MOVE/ employs two models, namely the North Pacific and western North Pacific models (models NP and WNP). Model WNP is nested into model NP (one-way nesting). The model domain of the model G extends from 75°S to 75°N globally. The grid spacing in the zonal direction is 1° and that in the meridional direction is 0.3° within 5°S–5°N, and 1° poleward of 15°S and 15°N. There are 50 levels in vertical. The domain of the model WNP extends from 15°N to 65°N, and 117 °E to 160°W, with a grid spacing of 1/10° × 1/10° around Japan. This model is nested into model NP, whose model region is from 15°S to 65°N, and 100°E to 75°W with a grid spacing of 1/2°x1/2°. NP and WNP have the same vertical grid spacing (54 levels).

The analysis fields for models G, NP, and WNP are calculated separately. The analysis scheme adopted in the MOVE/ system is a multivariate 3DVAR analysis scheme with vertical coupled T–S EOF modal decomposition of a background error covariance matrix. The scheme is based on Fujii and Kamachi, 2003a, 2003c and Fujii et al., 2005. The amplitudes of the coupled EOF modes are employed as control variables and the analyzed temperature and salinity fields are represented by the linear combination of the EOF modes in the scheme. A preconditioned optimizing utility for large-dimension analysis (POpULar; Fujii and Kamachi, 2003b and Fujii, 2005) is developed and applied for minimizing the nonlinear cost function as the descent method. This scheme can minimize a cost function including a constraint of the background without inversion of the background error covariance matrix, even if the function is nonlinear. It is useful for handling the correlation among background errors. The regions of the models G, NP, and WNP are divided into 40, 12 and 13 subregions, respectively. EOF modes are calculated in each subregion for each model from world ocean database 2001 (WOD2001), as well as the representativeness error covariance matrix, according to Fujii and Kamachi (2003b). We retained 12 dominant modes in each subregion. In fact, more than 85% of the total variance can be explained by the dominant 12 modes although this estimate will differ from one in a different subregion. The Gaussian function is adopted as the horizontal correlation model applied in the background covariance matrix B. The e-folding scales along latitude and longitude lines are also different in different subregions and are decided from Kuragano and Kamachi (2000). The model temperature and salinity fields are corrected by the analysis result through the incremental analysis updates (IAU) technique. The assimilation period is 1/3 month.

Observing systems assimilated are temperature, salinity and along-track SSH observations. The temperature and salinity observations were collected from WOD2001 and the global temperature–salinity profile program (GTSPP) database that contains ship hydrography, XBT, drifting buoy, Argo float and tropical array systems such as TAO/Triton. We also adopted the along track SSH anomaly data of TOPEX/Poseidon (T/P), Jason-1, ERS-1/2, ENVISAT, which are extracted from the SSALTO/DUACS delayed time multimission altimeter products, after adding it to the mean SDH calculated from a preliminary analysis using temperature and salinity observations alone. Real time global merged SST analysis MGDSST (i.e., Japan GHRSST) combined ship and satellite data by JMA is also adopted.

(KORDI)

The East/Japan Sea has recently received considerable attention since amount of data has been accumulated through regular and intense observation programs, and satellite measurments. Partly, the typical oceanic features in the circulation and hydrography such as western boundary currents and mesoscale variabilities of the Ulleung Warm Eddy and Dok Cold Eddy has been remarked. KORDI (Korean Ocean Research and Development Institute) and Korean unversities groups have developed the ocean data assimilation system based on the 3-dimensional data assimilation technique and implemented it to the regional ocean model in the East/Japan Sea as a prototype of the ocean forecast system in the East/Japan Sea.

Recently, the roles of the ocean has been considered to take a key role on the typhoon prediction, especially, in terms of typhoon intensity prediction. Most of typhoons which went through the Korea Peninsula passed across the Northwestern Pacific. To understand the ocean response to the typhoon in the Northwestern Pacific, we developed a data assimilative NorthWestern Pacific Regional Ocean Model (NWPROM), which will be coupled with the GFDL hurricane model. In present stage, we are producing the ocean reanalysis to hindcast the ocean state and typhoon intensity from 2000 to 2007.

Korean researchers have developed data assimilative regional ocean models in the East/Japan Sea and the Northwestern Pacific. The models are based on the MOM3. The domain of the East/Japan Sea Regional Ocean Model(ESROM) is 127.5 to 142.5°E and 33.0 to 52.0°N wich the varing horizontal resolution from 0.06° near the western boundary to 0.1° in the east of 130°E. To resolve the bottom geometry more accurately, the partial bottom cell scheme was used and high resolution bathymetry of 1/60° (Choi et al., 2002) was was adopted for the model topography. The number of vertical levels is 42. The ESROM has been driven by monthly mean windstress from the ECMWF reanalysis, monthly mean heat flux calculated by bulk air-sea flux formulation using ECMWF reanalysis meteorological variables, and open boundary conditions. The hydrographic condition at the open boundary was restored to the climatological hydrography of the World Ocean Atlas and the barotropic velocity through the inlet was given by the volume transport monitored by the submarine cable (Kim et al., 2004). For the open boundary condition, a radiation condition is applied to the tracers and barotrpic currents.

The domain of the Northwestern Pacific Regional Ocean Model (NWPROM) is 115°E to 150°E and 15°N to 52°N with 1/12° horizontal resolution. To resolve the bottom geometry more accurately, the partial bottom cell scheme was used and high resolution bathymetry of 1/60° (Choi et al., 2002) and ETOPO2 bathymetry were adopted for the model topograph. The vertical resolution is varying from 10 m at the surface to about 438 m at the bottom with 40 vertical levels. The NWPROM has been forced by monthly mean surface and open boundary conditions. Surface windstress, heat flux has been interpolated from the NCEP meteorological datasets and for the salt flux, the surface salinity in the model is relaxed to that of the WOA. A radiation condition with a nudging term is applied to the ocean reanalysis of the tracers and barotropic current from the HYCOM Consortium data products along the open boundary.

The 3-dimensional variational assimilation routine (3DVAR) has been fully coupled with the regional ocean model, and temperature profiles, and satellite-borne sea surface temperature and surface height anaomaly have been assimilated using diffusion equation for correlation modeling. Among several different methods to assimilate the sea surface height (SSH) into ocean models, we adopted and modified a method of water property conservation scheme to assimilate the SSH anomaly.

3. Highlighted examples

3.1 Examples in the North Atlantic

(MERCATOR OCEAN)

The PSY2v3 system has been initialized by Levitus climatology in Atlantic and Medatlas in Mediterranean the October 11th 2006. It has been run using a 7-day assimilation cycle and has been comprehensively validated over the year 2007. The realism of the description of the ocean physics, water masses, and volume transports has been assessed. Although some biases develop in regions where complex interactions take place between the different limitations of the system (mostly the tropics), the results show a “qualitative jump” of the physical and statistical skills of the system. They reinforce the scientific feasibility of the future upgrade of the system into a global high resolution configuration at 1/12°.

Many statistics have been calculated to check the consistency of the system. Some of them constitutes analysis and forecast scores at seven days. For instance, Figure 4 represents temporal series of RMS value of the innovation (observation minus model) allowing to check the performance of the system continuously with a set of independent, i.e. not yet assimilated observations. This RMS moves very little during the simulation and is of the order of the observations errors, as well for salinity as for the temperature, with a maximum at the thermocline depth. Fluctuations of the scores are limited to the seasonal cycle.

A comparison of the PSY2v3 currents map with the Chlorophyll-A concentration (independent data, i.e. not assimilated) in the Gulf Stream region (Figure 5), shows a good agreement between the model and the observations. Chlorophyll fronts are coherent with the model streamlines. The subsurface drifters confirm that the model current are good in term of direction and intensity.

|[pic] |[pic] |

Fig. 4 RMS of vertical temperature profiles innovation (left) and salinity profiles innovation (right) until 2000 meters from November 2006 to February 2008.

|[pic] |[pic] |

Fig. 5. Chlorophyll-A concentration from MODIS (left) and PSY2v3 currents at 15 meters (right) on July 11th 2007. The coloured dots represent the position and velocity measurements of the drifters (the dots are shrinking as the measurement gets further from the model date).

(Canada)

For the North Atlantic University and Department of Fisheries and Oceans Researchers are developing data assimilation schemes for the North Atlantic as part of the GOAPP project. Emphasis is on maintaining prediction skill over longer periods of time (15-45 days). Figure 16 shows an example of assimilation results (ODA), 15 and 45 days prediction (at the same analysis time) with a snapshot of the observed state. Predicted state (15 day and 45 day in the figure) shows how the observed/analysis field (Obs/ODA) will be kept well or partly deteriorated in 15 and 45 days prediction period.

[pic]

Fig. 6 Snapshots of sea level in cm for part of the North Atlantic model domain of Liu and Thompson (2008). The upper left panel shows the along track altimeter anomalies from within 5 days of the analysis time (7 august, 2004). The upper right is the analysis after the assimilation of all available Argo profile and altimeter data using the assimilation method of Liu and Thompson (2008). The bottom panels 15 day and 45 days forecasts of sea level for the verifying time.

3.2 Examples in the Western North Pacific

(JMA-MRI)

The assimilation experiment (analysis/reanalysis) was conducted from January 1948 to December 2007 for global and North Pacific systems, and from January 1985 to September 2007 for western North Pacific system. 138 cases of prediction experiments for the Kuroshio path variability south of Japan were also conducted from February 1993 to July 2004. Predictions start at the first day of every month and are integrated for 90 days. The wind-stress and heat fluxes used in the prediction experiments are NCEP R-2, the same as in the assimilation experiment.

One example of products is about velocity field. Figure 7 shows comparison of the velocity field by the assimilation result and independent ADCP observation. Velocity field of the assimilation shows the correct Kuroshio path and mesoscale eddies. The correlation coefficients of the zonal (meridional) velocity between the two datasets is 0.84 (0.47). Using the analysis/reanalysis dataset prediction experiments are conducted. Results of the prediction are shown in Figs. 8 and 9. Figure 8 shows the comparison of predicted and assimilated Kuroshio path. It shows the successful prediction. 138 cases of the prediction are compared with persistency and climatological variability of sea surface field (Fig. 9). It shows 40-60 days predictability (Usui et al. 2006, 2008a,b).

Fig. 7. Comparison of near surface velocity fields. (a): plan view, (b): zonal velocity, (c): meridional velocity. Black: assimilation, Red: independent observation by ADCP.

Fig. 8. Prediction of the 2004 Kuroshio large meander. Color bar shows the speed, and arrow shows the horizontal velocity field. (a): initial condition (1 July, 2008), (b): assimilation (25 July), (c): assimilation (25 August), (d) 25days prediction (25 July), (e): 55days prediction (25 August).

Fig. 9. Predictability diagram. RMS error is calculated from the sea surface height in the south of Japan (see map at the upper left corner for the evaluation area). Blue line: prediction; red line: persistency; broken line: mean sea surface height variability.

(KORDI)

The reanalysis through the ESROM with data assimilation routine has been performed in the East/japan Sea over the period of 1993 to 2005 in KORDI. The reanalysis by the ESROM was verified through comparing with a independent observation dataset by the PIES (Pressure-equipped Inverted Echo Sounder) in the Ulleung Basin located in the western side of the East/Japan Sea. It is evident that this system reproduces the mesoscale variability as well as the general circulation in the East/Japan Sea for the mean correlation between the 100m/100-dbar temperature fields of the reanalyzed products and from the PIES measurments is relatively high with mean value of 0.79. The Figure 10 shows that the ESROM has well represented the deveopment and movement of the Ulleng Warm Eddy and the formation and westward propagation of the Doc Cold Eddy. The ESROM also reproduces the seasonal variability of the North Korean Cold Current (NKCC), which is well-known subsurface western boundary current flowing southward off the eastern coast of Korea in summer. The climatological monthly mean volume transport of the NKCC is maximum of about 0.8 Sv in August to September, which could not be resolved by previous East/Japan Sea numerical models (Kim and Min, 2008). Even though the NWPROM is now on development stage and is not verified to compare with any independent observation yet, it seems to well represent the general circulation of the Northwestern Pacific. In particular, it is noticiable that the NWPROM well resolves the sub-surface Ryukyu Current south of the Okinawa Island which has been reported recently (Fig. 11).

[pic]

Fig. 10. 100m Temperature fields (a) measured by PIES, and (b) produced by the ESROM from December, 1999 to June, 2000 (from Kim et al., 2008).

Fig. 11. Example of flow field obtained with the NPROM. (a) Surface current and (b) vertical distribution of the meridional velocity in December, 2002.

4. Connections to coastal/shelf sea systems

4.1 North Atlantic

(MERCATOR OCEAN)

The operational PSY2v3 system in MERCATOR provides the initial conditions and/or the boundary conditions in real time to several embedded sub-regional or coastal systems as the Mohid Azores system of the University of the Azores (UAC), the IMI-NE-Atlantic system of the Marine Ireland Institute (IMI), the Mohid Iberian coast system of Technical University of Lisbon (IST/Maretec) and the MANGA Bay of Biscay system of IFREMER. The data exchanged through the boundaries is carried out in 1-way mode. But some systems such as the IST/Maretec one can be nested in two levels: the “father” embedding sub-regional Iberian coast system and a “child” coastal system interact in 2-way mode, with a feedback from the coastal system to the larger one. Figure1 shows the domain of these different systems and the downscaling scheme.

Moreover, the PSY2v3 system is coherent, in term of computational grid, with the Mercator-Ocean global system. The use of the global system to force the regional one is under development and will be soon in an operational mode.

(UK Met Office)

The shelf and coastal modelling is currently based on the POLCOMS system (Holt and James, 2001) but will transition to use NEMO in the near future. Surface forcing is hourly for winds and pressures and 3-hourly for heat fluxes. There is currently no assimilation in these systems, though the use of NEMO will facilitate this in the future. Nesting of the basin-scale models within the global model is done using the Flow Relaxation Scheme (FRS) by McDonald (1997) for temperature, salinity, currents and ice variables, whereby the fields are fixed to the outer model values at the edge of the domain and there is a damping zone just inside the domain to minimise spurious reflections. The bathymetry of the inner model is matched to the outer model around the boundaries. The integrated volume flux across the open boundaries is constrained to be zero to prevent drift in the mean surface height in the limited domain. The shelf model takes temperature, salinity, depth-mean currents and surface height from the North Atlantic model as well as tidal harmonic components. The temperature and salinity boundary conditions are applied using FRS and the currents and SSH using the Flather condition (Flather, 1994). Figure 12 shows the SST field. It represents more realistic frontal structure than the regional system (see also Figure 2).

Fig. 12. Sea-surface temperature nowcast (deg C) for 00Z 7/10/2008 from the UKMO Shelf Seas 12km Atlantic Margin model.

(Canada)

In Canada, the C-NOOFS system is the downscaling exercise with MERCATOR-OCEAN PSY3V2 and PSY3V1 as boundary conditions. It is validated through the use of the Atlantic Monitoring Program where CTD’s transects are performed roughly every 20km across the shelf break. This is seen below in Figure 13. While no assimilation is performed yet in the C-NOOFS ocean forecast system, the water masses represent the effects of data assimilation in the PSY3V1/PSY3V2 system that drives initial conditions and 1 way boundary conditions. Presently the PSY3V2 system is used. The system runs a daily 6 day forecast at ¼ of a degree with envrionemtal forcing from the Canadain Meteorological Centre 33 km GLOBAL GEM system. A 1/12th degree 1 way nested version of C-NOOFS is in preparation and will be in pre-operational model by December 2008. Consideration is being given to include all of Artic Archipelago and add open 1way nested boundary with MERCATOR PSY3V2 global ocean model. Use of AGRIF downscalling to enhance resolution to 1/36th is also in preparation. Data assimilation for the C-NOOFS system is envisaged both for ice and the ocean. Downscalling to near shore environments is planned using unstructured grid models, but no formal projects to implement such systems operationally is currently in the works.

[pic]

Fig. 13. Observed temperature transect along the Flemish Cap Line from Atlantic Zonal Monitoring Program (left panel) and from C-NOOFS Ocean Forecast Sytem for 27 th of April 2008. The initial conditions for the model came from PSY3V1.

4.1 North Pacific

(RIAM and Japan Weather Association)

Development of a regional ocean prediction system for the Tokyo Bay and its adjacent seas has been progressed. This project is a joint study between the Japan Weather Association, Kyushu University and Tokyo Maritime University with participation of some other Agencies as data and service providers including JMA, Japan Coast Guard, Japan Oceanographic Data Center and some others, aiming the development of the prototype of operational ocean prediction system for regional oceans. For this purpose we chose the Tokyo Bay and its adjacent seas as a test case.

This system utilizes a high resolution 3D RIAM Ocean Model (RIAMOM; see the detailed description of RIAMOM to Zhu et al., in this Proceedings) with free surface and 66 vertical levels in the deepest place for the short-term prediction of ocean conditions and circulation in the Tokyo bay and it adjacent seas with a nested grid system and provide approximately 2 days ocean prediction using the JMA meteorological forecasts (MSM) and output of operational ocean models as surface and lateral boundary conditions, respectively. The model domain consists of three nested sub regions; TB, TSB and R2 with horizontal grid intervals of 280 m, 840 m and 3000 m, respectively as shown in Fig. 2. The outer boundary of the R2 is forced by sea level height of 16 tidal harmonics and model output of operational ocean models by the JMA MOVE/(1/10 degree) or the FRSGC Coastal Ocean Predictability Experiment (JCOPE) (1/12 degree). The river discharges are also imposed at various river mouths in the Tokyo Bay.

The system tried to hindcast the circulations of the Tokyo Bay and it adjacent seas during September 23 to 25 in 2006 and October 11 to 12 in 2007. Figure 14 shows comparisons of vertical sections of northward velocity on Sept. 25, 2006 along the Line C in Fig. 2 in the case using MOVE- as boundary conditions. The vertical section of RIAMOM-T 1/36 (lower left) with 3km resolution reproduces a finer structure than the interpolated field of MOVE- (upper left), showing a good agreement with the observation (lower right). Figure 15 shows current vector at 2 m depth in the Tokyo Bay. At the sea surface sharp tidal fronts, as well as meander and eddies are formed. The temperature distribution demonstrates the penetration of colder waters into the Tokyo bay with the formation of some mushroom-like structures and spots of cold and warm water. The comparison between model and observation in the Tokyo Bay was not done due to the lack of observation during the period except sea level variations which are simulated generally nicely although model sea levels are underestimated with a standard err of 14.5 cm (Fig. 16).

Two serious problems have been illuminated during the course of hindcast. One is the poor resolution of MOVE- and JCOPE along the coast of the Boso Peninsula along which the along shore wind generate coastal waves and coastal jets, which propagate into the Tokyo Bay, giving great impacts on the circulation there. The poor resolutions of these coastal processes in the MOVE- and JCOPE unable the system to incorporate realistic coastal process into the model domain through boundary conditions. The other is the lack of incident internal tides which propagate into the model domain through the boundary of the R2 from the Pacific Ocean with a great energy, giving great impacts on vertical motions and temperature variations, since the MOVE- and JCOPE do not incorporate tidal motions. Beside those problems above, the further finer horizontal resolution beyond 100 m will require the replacement of the model basic equation from the Reynolds Averaged Navier-Stokes Equation used in this study to the equation system with a realistic reproduction of homogenous turbulence, such as the large eddy simulation (LES).

[pic]

Fig. 14. Comparisons of vertical sections of northward velocity along the Line C in Fig. 2 in the case using MOVE- as boundary conditions.

[pic]

Fig. 15. Current vector at 2 m depth in the Tokyo Bay. Temperature distribution demonstrates penetration of colder waters into the Tokyo bay with formation of some mushroom-like structures and spots of cold and warm water.

[pic]

Fig. 16. Sea level anomaly (cm) at Shibaura in the Tokyo Bay. Red color indicates model and blue one observation.

5. Summary and future issues

We showed some examples of the regional systems in the North Atlantic and North Pacific with coupled coastal/shelf sea systems. General future issues are clarified from each projects under GODAE. We summarise the issues or future developments needed though the detailed discussion is omitted:

(1): Sensitivity study of a coastal/shelf sea system.

(2): developing model insertion method such as IAU to avoid model shock.

(3): Improving error variance/covariance matrices.

(4): developing advanced assimilation method such as 4DVAR.

(5): OSSE/OSE type of observation sensitivity experiments.

(6): coupling to the thermodynamic-dynamic sea ice model/assimilation.

(7): regional air-sea coupled model/assimilation system for improving prediction of local air-sea interaction such as typhoon.

6. References

Barnier, B., G. Madec, T. Penduff, J.M. Molines, A.M. Treguier, J. Le Sommer, A. Beckmann, A. Biastoch, C. Boening, J. Dengg, C. Derval, E. Durand, S. Gulev, E. Remy, C. Talandier, S. Theetten, M. Maltrud, J. McClean, B. De Cuevas (2006): Impact of partial steps and momentum advection schemes in a global ocean circulation model at eddy permitting resolution, Ocean Dynamics, 543-567, DOI: 10.1007/s10236-006-0082-1.

Choi, B.H., K.O. Kim and H.M. Eum, 2002. Digital bathymetric and topographic data for neighboring seas of Korea. J. Korean Soc. Coastal Ocean Eng., 14, 41–50 (in Korean with English abstract).

Fichefet T. and Gaspar P. (1998): A model study of upper ocean-sea ice interaction, J. Phys. Oceanogr., 18, 181-195.

Fichefet, T., and M.A. Morales Maqueda (1997): "Sensitivity of a global sea ice model to the treatment of ice thermodynamics and dynamics" J. Geophys. Res., 102, 12,60912,646

Flather, R.A. (1994), A storm surge prediction model for the northern Bay of Bengal with application to the cyclone disaster in April 1991, J. Phys. Ocean. 24, pp 172-190

Fujii Y. (2005): Preconditioned optimizing utility for large-dimensional analyses (POpULar). J. Oceanogr. 61, 167-181.

Fujii Y, Kamachi M. (2003a): Three-dimensional analysis of temperature and salinity in the equatorial Pacific using a variational method with vertical coupled temperature-salinity EOF modes. J. Geophys Res., 108(C9), 3297, doi:10.1029/2002JC001745.

Fujii Y, Kamachi M. (2003b): A reconstruction of observed profiles in the sea east of Japan using vertical coupled temperature-salinity EOF modes. J. Oceanogr., 59, 173-186.

Fujii Y, Kamachi M. (2003c): A nonlinear preconditioned quasi-Newton method without inversion of a first-guess covariance matrix in variational analysis. Tellus, 55A, 450-454.

Fujii Y, Ishizaki S, Kamachi M. (2005): Application of nonlinear constraints in a three-dimensional variational ocean analysis. J. Oceanogr., 61, 655-662.

Holt, J.T. and James, I.D. (2001), An s-coordinate density evolving model of the northwest European continental shelf: 1. Model description and density structure, J. Geophys. Res. 106(C7) pp 14015-14034

Ishikawa I, Tsujino H, Hirabara M, Nakano H, Yasuda T, Ishizaki H. (2005): Meteorological Research Institute Community Ocean Model () Manual. Technical Reports of the Meteorological Research Institute, No.47 (in Japanese).

Ishizaki H, Motoi T. (1999): Reevaluation of the Takano-Oonishi schemes for momentum advection on bottom relief in ocean models. J. Atmos Oceanic Technol., 16, 1994-2010.

Kim, Y. H. and H. S. Min, 2008. Seasonal and Interannual Variability of the North Korean Cold Current in the East Sea Reanalysis Data. Ocean and Polar Research, 30(1), 21-31. (Korean with English abstract).

Kuragano T, Kamachi M. (2000): Global statistical space-time scales of oceanic variability estimated from the TOPEX/POSEIDON altimetry data. J. Geophys. Res., 105, 955-974.

Liu, Y. And K. R. Thompson. (2008) Predicting Mesoscale Variability of the North Atlantic Using a Physicall-Motivated Scheme for Assimilating Altimeter and Argo Observations. Under review. Monthly Weather Review.

McDonald, A. (1997), Lateral boundary conditions for operational regional forecast models; a review , HIRLAM Tech Report 32, HIRLAM-5 Project, c/o Per Unden, SMHI, S-601 76 Norrkoping, Sweden

Madec G. 2008: "NEMO ocean engine". Note du Pole de modarisation, Institut Pierre-Simon Laplace (IPSL), France, No 27 ISSN No 1288-1619

Madec, G., and M. Imbard (1996): A global ocean mesh to overcome the North Pole singularity, Clim. Dyn., 12, 381-388.

Pham, D., Verron, J., and Roubaud, M. (1998): A Singular Evolutive Extended Kalman filter for data assimilation in oceanography, J. Mar. Syst., 16(3-4), 323-340.

Piacentini A., Buis S., Declat D. and the PALM group (2003): A computational Framework for assembling high performance computing applications,Concurrency and Computat.: Pract. Exper.,00,1-7.

Testut C.E., Brasseur P., Brankart J.M. and Verron J. (2003): Assimilation of sea surface temperature and altimetric observations during 1992-1993 into an eddy-permitting primitive equation model of the North Atlantic Ocean, J. Mar. Syst., 40-41 (april 2003), pp 291-316.

Tsujino, H., Hasumi, H., Suginohara, N. (2000): Deep pacific circulation controlled by vertical diffusivity at the lower thermocline depth. J. Phys. Oceanogr. 30, 2853-2865, 2000.

Tsujino H, Usui N, Nakano H (2006): Dynamics of Kuroshio path variations in a high-resolution GCM. J. Geophys Res., 111 , C11001, doi:10.1029/ 2005JC003118

Usui N., S. Ishizaki, Y. Fujii, H. Tsujino, T. Yasuda and M. Kamachi (2006): Meteorological Research Institute Multivariate Ocean Variational Estimation (MOVE) System: Some Early Results. Adv. Space Res., vol. 37, 806-822.

Usui, N., H. Tsujino, Y. Fujii, M. Kamachi (2006): Short-range prediction experiments of the Kuroshio path variabilities south of Japan. Ocean Dynamics, vol.56, 607-623, DOI 10.1007/s10236-006-0084-z

Usui, N., H. Tsujino, Y. Fujii, M. Kamachi (2008a): Generation of a trigger meander for the 2004 Kuroshio large meander. J. Geophys. Res. 113, C01012, doi:10.1029/2007JC004266.

Usui, N., H. Tsujino, H. Nakano, and Y. Fujii (2008b): Formation process of the Kuroshio large meander in 2004. J. Geophys. Res. (in press)

-----------------------

[1] HBAR: Barotropic Height, TEMP: Temperature, SAL: Salinity U: Zonal Velocity component, V: Meridional Velocity component.

-----------------------

Ryukyu Current

Kuroshio

(b)

Deep counter current

(a)

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download