1. Introduction
Since the strong influence of the Southern Oscillation (El Niño) on the extratropical climate began to unfold (Bjerknes 1966), much attention focused on the investigation of the tropical Pacific ocean circulation (Johnson et al. 2002; Kessler et al. 2003; McPhaden et al. 1998; McCreary et al. 2002) using both observations and numerical models (e.g., Bonjean and Lagerloef 2002; Durand and Delcroix 2000; Lagerloef et al. 1999; Vialard et al. 2001). The global climate impacts associated with the strong 1997–98 El Niño event further underscored the need for a better understanding of tropical Pacific dynamics and for accurate descriptions of the space–time structure of the tropical Pacific Ocean circulation. Many studies have been based primarily on surface datasets such as sea surface temperature (SST), wind stress, and sea surface height (SSH) fields. With the advent of Argo (Gould et al. 2004) and the presence of the Tropical Ocean Global Atmosphere-Tropical Atmosphere Ocean (TOGA-TAO) buoy network (McPhaden et al. 1998), the availability of subsurface data is now also becoming more complete. However, a detailed description of the tropical Pacific dynamics still requires the support of model simulations driven by surface fluxes of heat, momentum, and freshwater.
Complete spatial and temporal coverage of surface fluxes of momentum, heat, and freshwater is available from atmospheric analysis centers. However, their forcing fields are prone to random uncertainties and significant model biases (Yang et al. 1999; Putman et al. 2000). Wind stress errors can be partially remedied by using satellite wind stress fields (Grima et al. 1999). However, Auad et al. (2001) concluded from a study of a tropical Pacific model response to National Centers for Environmental Prediction (NCEP) and Comprehensive Ocean–Atmosphere Data Set/Florida State University (COADS/FSU) forcing fields that NCEP fluxes overall yielded a better basinwide ocean hindcast, except for serious problems related to weak interannual variability in the NCEP forcing fields. A number of GCM representations of the tropical Pacific show bias in the cold tongue SST (Sun et al. 2003). Improvement of model simulations, such as reduction of model bias errors, is expected to arise by improving the model physics, or by adjusting forcing and boundary conditions to match the model to ocean observations through the use of data assimilation.
Many assimilation systems in the tropical Pacific were based on techniques such as optimal interpolation or the so-called three-dimensional variational methods (e.g., Behringer et al. 1998; Giese and Carton 1999). These methods provide the analysis of the system state using the observations at a given time without enforcing smoothness and dynamical consistency (i.e., conservation properties) in the time dimension. An example of a more advanced assimilation scheme is implemented by Bennett et al. (2000), who used the representer method to assimilate TAO data into an intermediate coupled ocean–atmosphere model. Bonekamp et al. (2001) used the adjoint method to adjust the model wind stress over the tropical Pacific while constraining the model to temperature data profiles in the framework of an “identical twin” experiment. Recently, Weaver et al. (2003) successfully tested the incremental four-dimensional variational data assimilation (4DVAR) method (Courtier et al. 1994) using a primitive equation model to assimilate temperature profiles by adjusting initial conditions.
The present study is part of the Consortium for Estimating the Circulation and Climate of the Ocean (ECCO) project (Stammer et al. 2002a). It is intended to develop a nested eddy-permitting regional adjoint assimilation procedure for the tropical Pacific Ocean. In cooperation with the Consortium for the Ocean’s Role in Climate (CORC) project, we intend to constrain the model by all available data of the tropical Pacific and thereby to study the dynamics of the tropical Pacific circulation. As a prerequisite to such a data assimilation effort, we present a detailed model–data comparison without assimilation to test the model skill and to demonstrate that the model is consistent with the observations, within the expected uncertainties of both model and data. For that purpose we explore the sensitivity of the model to the horizontal resolution and to the forcing fields and study its dynamics as a function of model resolution. A first assimilation study based on the model is provided by Hoteit et al. (2005).
Forcing fields used during our study were drawn from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCAR) reanalysis project (Kalnay et al. 1996) as well as from the ECCO results (Köhl et al. 2007). The NCEP forcing is available on a 1° × 1° global grid and contains twice-daily wind stress vectors and daily net heat flux, net shortwave radiation, and water flux at the sea surface. The ECCO forcing is the NCEP forcing adjusted by a global state estimation procedure using the Massachusetts Institute of Technology general circulation model (MITgcm) with a 1° × 1° global grid and the adjoint method assimilating a multivariate global ocean dataset, as described in detail by Stammer et al. (2002b, 2004. In the current experiments, we used the ECCO solution from iteration 69, which is nearly identical to that reported in Köhl et al. (2007). The ECCO winds are given twice per day, and the heat and freshwater fluxes are daily, the same as NCEP. The assimilation system significantly improved the agreement of the 1° × 1° grid global model to SSH and SST observations, but the skill of the ECCO forcing for nested higher-resolution model runs remains unclear and will be tested here. This question is of interest to the ocean community at large since it is part of the general question of how beneficial it is to optimize low-resolution models and then use the adjusted control parameters (forcing and boundary conditions in our case) for higher-resolution process models.
The paper is organized as follows. Section 2 describes the model configuration adopted for our experiments. Section 3 describes the evaluation of the model with respect to the different datasets and presents sensitivity studies of the model to the forcing fields and the horizontal resolution. Section 4 discusses the momentum balance of the equatorial circulation as simulated by the model. Concluding remarks are given in section 5.
2. The model
a. Model description
The numerical model (MITgcm) is described by Marshall et al. (1997). It is based on the primitive (Navier–Stokes) equations on a sphere under the Boussinesq approximation operated in a hydrostatic mode with an implicit free surface. No-slip conditions are imposed at the lateral boundaries while the friction condition is quadratic at the bottom with a drag coefficient equal to 0.002 m−1. The subgrid-scale physics is approximated by a tracer diffusive operator of second order in the vertical and by the K-profile parameterization (KPP) in the surface mixed layer (Large et al. 1994). The horizontal diffusive and viscous operators are either of second or fourth order, depending on the experiment (see Table 1). The mixing parameters are adjusted as the cube of the horizontal grid spacing to account for meridional varying horizontal grid spacing. The model domain covers the tropical Pacific basin from 26°S to 26°N and from 104°E to 68°W. The vertical resolution is 10 m from the surface to 300 m in depth, with spacing gradually increasing below 300 m for a total of 39 layers. The maximum bottom depth is 6000 m and the bathymetry is extracted from the global topography prepared by Smith and Sandwell (1997).
The study covers a 9-yr period from 16 January 1992 to 31 December 2000, which includes a strong El Niño event. The model is initialized with January potential temperature and salinity fields provided by the ECCO global optimization procedure on a 1° × 1° grid (Köhl et al. 2007), with the velocity fields adjusted over a 2-week period. NCEP and ECCO forcing fields were interpolated to match the horizontal resolution of the model grid and were applied with the same temporal resolution. In both cases, a relaxation term is included in the surface layer tracer equations, relaxing surface temperature and salinity fields toward a Levitus monthly climatology (Levitus and Boyer 1994; Levitus et al. 1994) with a 30-day time scale.
b. Open-boundary conditions
To account for large-scale circulation changes, the model is nested into the ECCO global 1° optimization (Köhl et al. 2007), which provides an estimate of the complete ocean state for the period 1992 through 2002. Open boundaries (OBs) for the regional model are prescribed at 26°S and 26°N, as well as at four straits in the Indonesian throughflow (ITF), using monthly ECCO fields. The Lombok Strait and Sumba Strait are southern open boundaries, while the Ombai Strait and the Timor Passage are western open boundaries. The OBs are specified as described by Zhang and Marotzke (1999) using monthly mean results of the ECCO global estimate.
The normal velocity fields across the open boundaries have been further adjusted (after interpolation to the higher-resolution domains) (i) to impose the same transport at 26°N and 26°S as in the global ECCO model, and (ii) to bring the time-mean volume transports across all open boundaries into balance. The adjustments at each boundary are added as a barotropic velocity uniformly distributed over all grid points. At the northern boundary, the net, time-mean, meridional transport over the entire period is negligible and fluctuations in monthly averages are smaller than 1 × 106 m3 s−1. This is a result of a closed Bering Strait in the global ECCO model and the use of a virtual salt flux for evaporation minus precipitation (E − P). The northward transport entering the model at 26°S leaves the model again through the ITF. The 11.4-Sv mean transport for 1992–2000 is partitioned among the different straits as follows: 3 × 106 m3 s−1 in the Timor Passage, 5.8 × 106 m3 s−1 in the Ombai Strait, 0.3 × 106 m3 s−1 in the Sumba Strait, and 2.3 × 106 m3 s−1 in the Lombok Strait. The transport through the Sumba Strait was neglected. The adjustment to the barotropic velocities at the boundaries is not tuned to keep the temperature transports across the boundaries identical to the ECCO values. At the northern boundary, where the net volume transport is zero, the time-mean meridional heat flux over the entire period is 0.56 PW, compared to 0.6 PW before the barotropic adjustment, and 0.46 PW in the ECCO mean. The volume transport does not vanish at the other boundaries so the heat fluxes cannot be calculated, but the net heat flux out of the domain combining the southern boundary and the ITF is about 0.6 PW.
3. Open-loop sensitivity studies
a. Experiments
To study the sensitivity of the model’s state to details in the forcing fields and to the model horizontal resolution, several 9-yr integrations were performed (Table 1). These experiments help to determine the best first-guess state for future assimilation experiments. At the same time they are intended to quantify the minimum model resolution required for acceptable physical accuracy. One important question is why some global OGCMs have a bias in SST in the cold tongue region of the eastern tropical Pacific (Sun et al. 2003). A specific goal, therefore, is to identify to what extent forcing changes can control model biases in the tropical Pacific (i.e., identify whether model biases could result from incorrect forcing) and to see if ECCO forcing can reasonably replicate the cold tongue.
For a test of model sensitivity to horizontal resolution, the model was run with three different spatial grids: 1° × 1°, 1/3° × 1/3°, and 1/6° × 1/6° (vertical resolution was held constant at 39 levels). For this comparison, only the time step and the subgrid-scale horizontal mixing of both heat and momentum were changed; all other parameters were held fixed. In particular, the model was driven by the ECCO forcing during all model resolution sensitivity experiments. The run of the regional model at 1° × 1° resolution with ECCO forcing was intended to test the open boundary conditions (the solution should and does match the global 1° × 1° solution except for the effects of the sponge layer at the boundary). It can also evaluate the effect of enhanced vertical resolution on the skill of the model (39 levels in the regional model versus 23 levels in the global model). Since the global ECCO optimization was constrained by SSH and SST, the 1DEG model run can be expected to compare best to these observations in the analysis.
Forcing fields were varied for 1/3° model resolution, ranging from original NCEP forcing to ECCO forcing. In addition various combinations of both forcing fields were used, such as NCEP winds combined with ECCO heat and freshwater fluxes. Also used were NCEP forcing fields with the time mean of the ECCO adjustments added. The experiments ECCOF and NCEPF (see Table 1) will be used to assess the quality of the adjustments produced by the ECCO optimization procedure with regard to the NCEP forcing fields. The experiment NCEPW differs from ECCOF only in its wind stress forcing and is intended to assess the quality of the wind forcing. ECCOM was performed to test the skills of the mean ECCO adjustments. Finally, the comparison between the experiments NCEPF and NCEPW, which are forced with different heat and freshwater fluxes but identical wind stress, will be used to assess the effects of the thermohaline forcing.
The ECCO adjustments increase the (1992–2000 time) mean easterly wind stress in the tropical Pacific while making much smaller changes in the meridional wind component (Fig. 1). This agrees with Josey et al. (2002), who reported that NCEP wind stress in the tropics was weak. There are also strong north/south gradients in the ECCO zonal wind adjustments between 5° and 10°N, increasing the wind stress curl in the region of the North Equatorial Counter Current (NECC). The ECCO adjustments increase the northward wind stress north of the equator and east of 240°E. There is also an increase in the southward wind stress just south of the equator in the eastern Pacific, increasing the divergence at the equator. The standard deviations of the NCEP and ECCO wind stress are displayed in Figs. 1c,d. For the ECCO winds, the wind stress variability is significantly higher in the tropical Pacific than provided by NCEP.
Stammer et al. (2004) investigated the ECCO heat and wind stress adjustments in comparison to the NCEP forcing fields. As shown in their study, the ECCO heat flux into the ocean is significantly decreased near the eastern edge of the domain. The freshwater flux shows increased evaporation (by about 10%) south of the equator and in the warm pool, while precipitation is increased in the central Pacific north of the equator.
b. Model evaluation
In the following we will evaluate each model run against ocean observations to determine the model skill in simulating the ocean as a function of resolution and forcing. Observations from the tropical Pacific Ocean are available from the TOGA-TAO moorings (McPhaden et al. 1998) and from several additional datasets. Those include surface drifter velocity data (Lagerloef et al. 1999), and satellite observations of SSH from the Ocean Topography Experiment (TOPEX)/Poseidon (T/P; Fu et al. 1994) and of SST from Reynolds and Smith (1994). The metric of goodness used throughout is the weighted RMS misfit between the model and the observed mean and time-varying ocean state. In that sense it is the metric that is usually being minimized during ECCO optimization runs. However, we also include parameters such as the basinwide zonal slope in sea level in the comparison with T/P results. Comparisons will be presented separately for TOPEX/Poseidon SSH data, for Reynolds and Smith (1994) SST fields, for TAO subsurface temperatures and velocities, and for surface drifter velocities and ADCP subsurface velocities.
1) TOPEX/Poseidon
A comparison with sea level data is provided here in terms of the cross-basin zonal slope of the sea surface along the equator observed by the time-mean T/P SSH observations minus the most recent Gravity Recovery and Climate Experiment (GRACE) geoid, and in terms of the RMS misfit between observed and simulated SSH variability. Although details in the simulated and observed time-mean SSH structure on the equator differ substantially at the western side of the section, the cross-basin slope in SSH appears well reproduced in most model runs (Fig. 2). However, we note that the basinwide slope is best reproduced in runs using the ECCO forcing and is weakest in the NCEPF and NCEPW runs. The difference in modeled slope, stemming from model differences in eastern basin SSH, is in part due to the increased mean easterly wind stress in the ECCO forcing, as can be seen from the differences between ECCOF and NCEPW runs: the mean ECCO easterlies within 5° of the equator are 50% stronger than NCEP winds and produce a 30% larger SSH height difference (Fig. 2) along the equator between 160° and 240°E. The difference between ECCOM and NCEPW is due to both the addition of the ECCO mean to the NCEP winds and to the difference between the ECCO and NCEP heat and freshwater flux variability. That the adjusted heat and freshwater fluxes play some role in setting the SSH slope along the equator is documented by the difference between the NCEPF and NCEPW runs. Finally, the difference between ECCOF and ECCOM shows that the effect of the ECCO variability in all forcing components also plays a role in shaping the mean equatorial SSH.
In contrast to the time-mean cross-basin SSH slope, which appears less sensitive to the horizontal model resolution, the time-varying SSH signal is significantly affected by the model’s horizontal resolution as well as by details of the forcing fields, especially the wind stress (Fig. 3). Time and spatial means were removed from the SSH before the calculation, but the seasonal cycle was not removed. Again the 1° model run (1DEG) shows the best RMS agreement with observed variability (about 4.5-cm RMS misfit, or about 20% of the signal) throughout the period, followed by experiments NCEPF and NCEPW. In contrast, model runs with 1/3° and 1/6° resolution forced by ECCO fields lead to the largest discrepancies relative to T/P monthly mean SSH observations. This indicates that the skill of the 1° model does not carry over to higher resolution because the ECCO forcing fields estimated with lower resolution partly compensate the higher viscosity and sluggishness of the 1° model. The ECCO fields therefore make the higher-resolution runs too energetic, especially in the western part of the basin (see below). The misfit for the ECCOM run is only slightly larger than for the NCEPF run, which suggests that the variability of the ECCO winds is responsible for much of the loss in skill in SSH horizontal structure. The small differences between NCEPW and NCEPF show the relatively small effect of the heat and freshwater fluxes on the horizontal SSH structure.
For reference, the observed RMS SSH calculated for 1-yr sliding windows (not shown) varies interannually, starting at about 15 cm in 1993 and increasing to about 16–17 cm at the end of 2000, reaching a maximum of about 18 cm during the 1997–98 El Niño event. Using those numbers, the RMS misfits in Fig. 3 could be converted into a fractional RMS by dividing by the reference RMS, and would range from a minimum of about 0.15 for the 1DEG run around 1995–96 to about 0.5 for the 1SIXTH run in 1993 and late 1997.
The RMS misfits (in space) between the time-mean SSH from the model runs and T/P data over the whole domain show a similar tendency, in that the 1DEG run is best, but the other runs do not differ significantly (4.4 cm for the 1DEG run, 6.5 cm from the NCEPW run, 6.7 cm from the 1SIXTH run, 6.9 cm for the NCEPF run, 7.05 cm from the ECCOF run, and 7.3 cm for the ECCOM run) because part of the misfit is due to errors in the time-mean SSH estimate (i.e., due to geoid errors).
Finally, we compare the observed point-by-point standard deviation for the monthly mean SSH with similar fields obtained from experiments 1DEG, NCEPF, and ECCOF, respectively. SSH anomalies were calculated relative to the 1993–2000 mean SSH for observations and model simulations alike. Because the results from the runs NCEPF, ECCOF, and 1° × 1° roughly (leaving out 1SIXTH) define the envelope of all results in Fig. 3, we will restrict all further comparisons to those runs.
There are four main “hot spots” for SSH variability (Fig. 4): on the equator in the cold tongue/Niño-3 region, north of the equator in the eastern Pacific where wind jets through orographic gaps such as the Tehuantepec trigger mesoscale eddies (Chelton et al. 2001; Kessler 2002), near the east coast of Mindanao where the NECC begins and the North Equatorial Current (NEC) splits, and off the east coast of New Guinea in the region of horizontal shear between the South Equatorial Current (SEC) and the South Equatorial Counter Current (SECC; Qiu and Chen 2004). The run 1DEG reproduces these hot spots reasonably well, albeit with somewhat weak variability. The variability of the SSH obtained from the NCEP forcing is generally weaker than observed, especially in the eastern Pacific. The dominance of the wind stress as a source of variability of SSH is demonstrated by the similarity of the NCEPF and NCEPW runs, which differ only in the buoyancy forcing. The importance of the time-dependent winds can be assessed by comparing ECCOF and ECCOM. The largest differences reside in the eastern Pacific, where the Tehuantepec eddies are nearly absent from the ECCOM run and the cold tongue variability is similar to NCEPF. This suggests that enhanced variability of the ECCO zonal and meridional wind stress in the eastern Pacific is responsible for the increased cold tongue variability and for triggering the eddies in the eastern Pacific. In that region all model runs show a dynamic height field that resembles that shown by Kessler (2002), with the ECCOF run having the best agreement in the strength and location of the Costa Rica Dome and the anticyclone to the northwest (not shown). Excess SSH variability in the western Pacific, present in the ECCOF and the 1SIXTH run (the SSH variability of the 1SIXTH run is comparable but about 15% stronger than what is found in ECCOF), is reduced in the ECCOM, suggesting that the ECCO wind stress variability in the western Pacific is too strong or that variability is propagating westward.
To summarize the results of the SSH comparison, the 1DEG run does best, showing that the optimized forcing retains good skill in the regional model even after a change to higher vertical resolution. The ECCO forcing fields have mixed results when applied to higher-resolution models, generally giving better agreement in the eastern Pacific and worse agreement in the west. Wind stress has the largest effect on SSH variability, but heat and freshwater fluxes have a significant effect on the time-mean SSH structure, suggesting that errors in heat fluxes and freshwater fluxes could also be a significant source of model biases in the tropical Pacific. Of course, incorrect physics might also contribute to model biases.
2) Reynolds SST data
Typically SST fields are used to evaluate the performance of models in specific regions of the tropical Pacific (e.g., Niño-1–Niño-4). We will use the SST fields as a measure of model skill over the entire model domain so as to compare with the earlier SSH comparisons. For later reference, the observed RMS of the monthly mean SST anomalies varies between 0.8° and 2°C across the model domain. The RMS differences between the model and Reynolds and Smith (1994) monthly mean SST fields with the temporal means removed (Fig. 5) reveal that the SST fields from different runs appear more similar as compared to the previous SSH analysis. This is partly due to the restoring of SST toward the Levitus monthly mean SST climatology. Over the cold tongue this can lead to an extra heat forcing on the order of 20% of the atmospheric heat flux signal there. In spite of the restoring, the model runs still differ significantly in SST. Later runs without the restoring term verify that it does not qualitatively change the differences between runs.
Overall, the RMS model data difference is less than 50% of the observed RMS SST variability. As before, the best results are obtained from the 1DEG run, directly followed by the higher-resolution ECCO runs. This indicates that the optimized ECCO heat fluxes retain skill in the higher-resolution runs. In contrast, the NCEPF and ECCOM runs produce the worst agreement with the data, followed by the NCEPW run. The ECCOM run is significantly worse than the ECCOF run at several points in the record (1998 and 1999, in particular), underlining the importance of the time variability in the ECCO forcing for controlling the SST. The SST and SSH (Fig. 4) model data misfits from all runs show a strong temporal structure, highest during the first two years and the 1997/98 ENSO event. This misfit is significantly higher in NCEP results and less pronounced in ECCO results. The initial misfit is most likely due to spinup adjustment in all model runs.
Important features in the observed time-mean SST field are the cold tongue in the east and the warm pool in the west (Fig. 6). The runs with ECCO forcing lead to the best agreement with the cold tongue in the Reynolds time-mean SST field both in amplitude and spatial structure. These runs (ECCOF, ECCOM, 1DEG, and 1SIXTH) all have roughly similar structures in the eastern Pacific. In contrast, the cold tongue in the NCEP runs is over 2° too warm and too narrow in latitude. The NCEPW run (not shown) is of intermediate quality, showing a cold tongue that is significantly cooler than NCEPF but still warmer than ECCOF or the observations. This means that both wind stress and ECCO heat flux forcing anomalies affect the cold tongue in this model. ECCOM forcing (ECCO mean corrections added to NCEP) shows a slightly better (colder) cold tongue than ECCOF. We speculate that since the enhanced wind stress variability in ECCO forces increased eddy energy, these eddies may tend to mix the cold tongue away (Swenson and Hansen 1999). In summary it appears that not just the ECCO wind stress but also the ECCO heat flux are essential ingredients for this success and that the time-varying forcing is required to reproduce the time-mean SST field. There are two obvious hypotheses for explaining the improvement of the representation of the cold tongue using ECCO forcing. Either (i) the NCEP forcing is correct and the model has errors that cause a bias in the SST (and other fields) or (ii) the model is correct and the ECCO adjustments to the NCEP forcing result in better agreement with the observations because they are more correct. As mentioned above, Stammer et al. (2004) supported (ii) by arguing for the plausibility of the ECCO forcing adjustments. Further data are required to distinguish between the hypotheses. For example, if the forcing fields were made less uncertain using new satellite products, then any model data misfit resulting from these fields would have to be explained as model error or as initial and boundary condition effects, which would need to be ruled out by observations.
3) TAO subsurface temperature data
The Tropical Atmosphere and Ocean buoy array consists of about 70 Autonomous Temperature Line Acquisition System (ATLAS) wind and thermistor chain moorings and current meter moorings in the tropical Pacific between 8°S and 8°N. Data from the TAO array allow a detailed comparison of temperature and velocity with the model results along the equator at the surface and subsurface. A time–longitude plot of the SST anomalies along the equator from selected runs compared to the TAO data shows that the 1DEG model agrees best with the TAO SST data, as expected from previous comparisons (rhs of Fig. 6). However, the 1DEG run simulates El Niño and La Niña events that are slightly stronger in SST than observed in the east but somewhat weaker than observed in the west. In the eddy-permitting model, the ECCO winds produce ENSO events with approximately the right strength in the east but with significantly less strength in the west than produced by NCEP winds, which gives SST values closer to the observations in the west. The 1SIXTH run (not shown) has similar (but about 15% stronger) variability than the ECCOF run.
A more quantitative measure of the similarity between model and data comes from correlating the model temperature with the TAO temperature time series at various longitudes, latitudes, and depths (Fig. 7). Statistical tests (based on the probability of getting a correlation as large as the observed value by random chance, when the true correlation is zero) revealed that significant correlations at a 95% confidence level exist between the observed and the simulated temperature signal at almost all TAO locations (insignificant correlations are marked in the figure). Model performance is good at 5- and 100-m depth on the equator, but poor at 250 m, perhaps because of the smaller signal at depth (RMS of temperature at 100 m is above 2°C compared to about 0.5°C or less at 250 m). The curves are complicated, but the summary of the comparisons off the equator is that the 1DEG is usually among the best, except for 8°N, 5 m in the western Pacific where NCEPF and NCEPW are clearly the best. This suggests that, even for the 1DEG model, ECCO winds are not good in the west, as we have seen before. Because ECCOM does poorly in the west compared to NCEPF, the mean ECCO winds are adversely affecting the comparison in the west. Plots of cross-equatorial temperature and salinity sections at 170°W (not shown) show no clear distinctions between the different model runs.
4) Current observations
The strong currents of the tropical Pacific Ocean are well studied and provide a benchmark for model simulations. The strongest currents in the region are the NECC, Equatorial Undercurrent (EUC), NEC, and SEC, which are mainly wind driven (Halpern et al. 1995; Blanke and Raynaud 1997; Pedlosky 1996). The EUC is associated with the thermocline and halocline structure on the equator (Pedlosky 1996) and is part of a 3D wind-driven circulation pattern that rises toward the ocean surface as it travels from west to east along the equator.
The standard World Ocean Circulation Experiment (WOCE)-TOGA drifters with drogues at 15-m depth (Yu et al. 1995) provide a good in situ mean surface velocity dataset for model data comparisons (Fig. 8). Time-mean drifter velocities were computed as averages over 5-day intervals in 1/2° latitude by 2° longitude bins from the entire time of the drifter record. Overall the velocity structure observed by the drifters is qualitatively simulated by all the model runs. Once again, the ECCOF and ECCOM currents are too strong in the western tropics, where the NCEPF and NCEPW runs show better agreement with the observed velocity structures. In contrast, ECCOF and ECCOM are best in the eastern tropics, including the banded structure south of the equator. Meridional velocities (not shown) show similar tendencies. The agreement between the model and drifter results suggests that a possible bias between them due to the different averaging periods and the significantly reduced observation count during the model period (Lagerloef et al. 1999) is not dominant.
Mean subsurface zonal model currents can be compared against shipboard ADCP measurements (Johnson et al. 2002) available along the equator (Fig. 9). Consistent with the the sea level slopes (Fig. 2), the NCEPF and NCEPW runs show less west–east thermocline and EUC tilt than are observed but each produce an EUC that is too strong west of the date line. In contrast, the ECCOF and ECCOM runs show the best agreement with the observed EUC structures, both in speed and location. This is also consistent with TAO moored ADCP observations (not shown). For example, on average, the velocities of the ECCOF EUC reach about 1.0 m s−1 at 110-m depth at 220°E, in agreement with the Johnson et al. (2002) analysis and the mean from the TAO data, whereas the maximum EUC simulated by the NCEP winds is around 0.85 m s−1. The 1DEG run has a very sluggish EUC, showing the effects of the poor horizontal resolution. The 1DEG run has higher vertical resolution than the global ECCO solution, which has even weaker EUC maxima, showing the importance of vertical resolution. The 1SIXTH run has a slightly more compact EUC, but the differences are not large, and the agreement with the data is not better than for ECCOF. The maximum current speed in the EUC is about 15% higher for ECCOF, which has to be compared to a 30% larger SSH height difference, indicating that the dynamics are more complicated than a simple frictional balance.
TAO profiles of RMS zonal and meridional velocity variability compared with respective model results show that overall the 1DEG runs are too weak in velocity variability (Fig. 10). With a 1/3° resolution, the ECCO wind stress (case ECCOF) has good agreement at 110°W for both zonal and meridional velocity, while NCEP winds produce relatively low variability. Some of this variability comes from tropical instability waves (TIWs), which seem to be more realistic in runs with ECCO forcing when comparing frequency spectra (not shown). As can be expected from the previous discussion, NCEP winds lead to more realistic variability in the western model domain; there the ECCOF and 1SIXTH runs show too much shallow variability in zonal and meridional direction, due to the effects of the eddies growing on the stronger currents driven by the ECCO forcing.
Based on the correlation between observed and simulated zonal velocities, most model runs are of comparable quality in simulating the zonal variability at 110°W and 147°E (Fig. 11). An exception is the 1DEG model in the core of the EUC at 110°W, which may be related to the weakness of the EUC in that run. Interestingly, the 1DEG run has the best skill between 50- and 200-m depth at 165°E. Note also the small differences between the NCEPF and NCEPW curves in most regions of significant correlation. The two runs differ only by the ECCO heat and freshwater fluxes, which do not apparently contribute strongly to the velocity variability. ECCOF does somewhat worse than NCEPW for zonal velocity at 165°E, but is better at 147°E, although no model has correlation >0.6 there.
4. Equatorial momentum balances
Zonal (left panels) and meridional (right panels) acceleration terms (m s−1) for the 1°, 1/3°, and 1/6° numerical model forced by ECCO winds were evaluated along the section 189°E at a depth of 140 m (i.e., at the depth of the EUC at that longitude; Fig. 12). This location is chosen to be away from complex topography and near the center of the basin. The zonal momentum budget shows the importance of the zonal pressure gradient and the nonlinear terms. The meridional budget shows a nearly geostrophic balance, as expected (Pedlosky 1996), with the matching slopes of the dp/dy and fu terms suggesting that the βu = ∂2u/∂y2 relationship holds, although it is not as good in the 1° model.
When sufficiently resolved, the zonal-momentum balance on the equator is dominated by a balance between the pressure gradient and the zonal advection of the zonal gradient of zonal velocity (Uux), reflecting the increasing zonal velocity downstream in the EUC at that longitude. Between the equator and 3° off the equator, the meridional advection of the meridional gradient of zonal velocity (Vuy) takes over in the balance, showing the acceleration of the water converging into the undercurrent. The vertical advection of the vertical gradient of zonal velocity (Wuz) is relatively small, but shows acceleration of the water moving upward. The Wuz term shows asymmetry of the interaction between the EUC and the shallow overturning cell near the equator, and the EUC is stronger north of the equator than south at 140-m depth. There is a residual difference between the Coriolis term (green line) and the summed pressure gradient and advective accelerations (red line). This is due to both viscosity and Reynolds stress. The 1/3° resolution model has a smaller residual than the 1/6° model outside the EUC, due to the lower level of eddy mixing seen in the 1/3° simulation.
The zonal momentum balance on the equator and especially the role of nonlinear terms in the balance as a function of depth have been examined previously by Qiao and Weisberg (1997) from moorings on the equator at 140°W. Comparisons of 1/3° and 1/6° model runs with the Qiao and Weisberg (1997) results (not shown) reveal qualitative agreement, although the momentum flux divergences seen in the model are about one-half the size of means seen by Qiao and Weisberg (1997). We also see a negative lobe in w(∂u/∂z) below the EUC, which is absent in their data. Kessler et al. (2003) examined the dynamics of vertically integrated equatorial currents using the Gent and Cane (1989) model and found zonal momentum flux divergence with a qualitatively similar shape to that found here, although the figures show the flux form of the balance, which must be compared to the sum of the individual term estimates shown here (the MITgcm uses the flux form in the numerics, but the individual terms have been estimated separately here for discussion).
The 1° run shows significantly poorer balances, which is expected. The 1/3° run seems to have just sufficient resolution to resolve the mean zonal momentum terms, and turbulent stresses (which are the residuals in the curves) are of similar sizes in each case. Results from the 1/6° run show further changes in the dynamical balances; however, the changes relative to the 1/3° run are much smaller than what can be seen between the 1/3° run and the 1° run. We conclude from the figure that a model resolution of 1/3° should be barely sufficient to perform data assimilation in the tropical Pacific.
5. Discussion
In this paper we evaluated a regional general circulation model for the tropical Pacific Ocean by varying horizontal resolution and surface forcing and comparing against several datasets during 1992–2000. The model was run with 1°, 1/3°, and 1/6° grid resolutions and with combinations of forcing fields obtained from NCEP and ECCO products. The latter are the NCEP forcings adjusted by a global 1° model optimization procedure (Stammer et al. 2004). The goal of the study is to evaluate the model and to determine the model configuration that can be used in a quantitative data assimilation approach at reasonable computational cost. The study also closely examined the benefit of using optimized global coarse-resolution products (ECCO) to force our nested higher-resolution tropical Pacific model (boundary conditions and surface forcing). Our main conclusions can be summarized as follows.
As seen in many other studies, the current system of the tropical Pacific Ocean is strongly sensitive to the wind stress forcing. The ECCO forcing appears to produce better model outputs than the NCEP forcing in the eastern Pacific, but has too strong a wind stress curl around the NECC in the western Pacific, forcing an overstrong NECC. We note that the heat flux is important for improving the mean SST of the simulations. ECCO heat fluxes produce a more accurate cold tongue, and the stronger ECCO wind stress in the eastern tropical Pacific produces better velocity variability on the equator. Adjusted surface heat and freshwater fluxes appear important for removing apparent SST biases in the tropical Pacific, but this does not distinguish between model errors and errors in the forcing dataset.
The ECCO forcing fields have been determined so that a 1° × 1° global model leads to the best agreement with ocean observations. As a result, the forcing seems to be too strong in some regions for a high-resolution model with lower friction and better physical representation, because the assimilation strengthened the forcing to counter deficiencies in the physics of the 1° model. This was particularly evident in the northeastern Pacific where the strong ECCO wind stress curl may have compensated for an overly damped 1° model. Nevertheless, agreement was enhanced in many ways by using the forcing optimized for a 1° × 1° model as a starting point for a further optimization with the higher-resolution model. Sensitivity to boundary conditions was not explored; they were always taken from the ECCO model.
The 1/6° run does not show a significant improvement in the model performance over the 1/3° run, despite the use of lower viscosity and diffusivity values. In contrast, the increase in vertical resolution from 23 to 39 levels and the associated decrease in the vertical viscosity have significant effects on the tropical currents in the 1° model runs. However, recent results from a 1/3° run with 50 layers suggest no significant changes compared to the same run but with 39 layers.
A 1/3° horizontal resolution seems just sufficient for a skillful simulation of the tropical Pacific circulation and its momentum balances. With this resolution, the ECCO wind stress produces a NECC trough that is too deep, resulting in overly large instability driven by eddy variability in the western tropical Pacific. The estimated time variability of the ECCO wind stress curl is important for stimulating eddy variability in the Gulf of Tehuantepec, which is nearly absent in the NCEP-forced runs.
Because the 1/3° solution is sensitive to details of the diffusivity and viscosity values, it may become important to include those parameters in the set of control parameters in an optimization of the tropical Pacific circulation. The present study, however, cannot conclusively determine the real effects of (parameterized) mesoscale eddies in large-scale simulations of the tropical Pacific Ocean. This problem has been addressed previously by Stockdale et al. (1993) and Megann and New (2001).
The conclusions suggest that the optimization of the model parameters, such as the forcing fields and initial conditions, could improve the model behavior by making it more consistent with the available data. Our next step will therefore be to use the adjoint method to assimilate all available data in the tropical Pacific Ocean.
Acknowledgments
Reanalysis surface forcing fields from the National Centers for Environmental Prediction/National Center for Atmospheric Research were obtained through a computational grant at NCAR. Computational support from the Center of Modeling and Prediction at Scripps and the National Center for Atmospheric Research is acknowledged. This work was supported by NOAA through the Consortium for the Ocean Role in Climate and by the Consortium for Estimating the Circulation and Climate of the Ocean funded by the National Oceanographic Partnership Program (ONR ECCO Grants N00014-99-1-1049).
REFERENCES
Auad, G., Miller A. , Roads J. , and Cayan D. , 2001: Pacific Ocean wind stress and surface heat flux anomalies from NCEP reanalysis and observations: Cross-statistics and ocean model responses. J. Geophys. Res., 106 , 22249–22265.
Behringer, D. W., Ji M. , and Leetmaa A. , 1998: An improved coupled model for ENSO prediction and implications for ocean initialization. Part I: The ocean data assimilation system. Mon. Wea. Rev., 126 , 1013–1021.
Bennett, A. F., Chua B. S. , Harrison D. E. , and McPhaden M. J. , 2000: Generalized inversion of Tropical Atmosphere–Ocean (TAO) data using a coupled model of the tropical Pacific. J. Climate, 13 , 2770–2785.
Bjerknes, J., 1966: A possible response of the atmospheric Hadley circulation to equatorial anomalies of ocean temperature. Tellus, 18 , 820–829.
Blanke, B., and Raynaud S. , 1997: Kinematics of the Pacific equatorial undercurrent: An Eulerian and Lagrangian approach from GCM results. J. Phys. Oceanogr., 27 , 1038–1053.
Bonekamp, H., van Oldenborgh G. J. , and Burgers G. , 2001: Variational assimilation of TAO and XBT data in the HOPE OGCM, adjusting the surface fluxes in the tropical ocean. J. Geophys. Res., 106 , 16693–16709.
Bonjean, F., and Lagerloef G. S. E. , 2002: Diagnostic model and analysis of the surface currents in the tropical Pacific Ocean. J. Phys. Oceanogr., 32 , 2938–2954.
Chelton, D. B., and Coauthors, 2001: Observations of coupling between surface wind stress and sea surface temperature in the eastern tropical Pacific. J. Climate, 14 , 1479–1498.
Courtier, P., Thepaut J-N. , and Hollingsworth A. , 1994: A strategy for operational implementation of 4D-Var, using an incremental approach. Quart. J. Roy. Meteor. Soc., 120 , 1367–1388.
Durand, F., and Delcroix T. , 2000: On the variability of the tropical Pacific thermal structure during the 1979–96 period, as deduced from XBT sections. J. Phys. Oceanogr., 30 , 3261–3269.
Fu, L., Christensen E. , and Yamarone C. , 1994: TOPEX/POSEIDON mission overview. J. Geophys. Res., 99 , 24369–24381.
Gent, P. R., and Cane M. A. , 1989: A reduced gravity, primitive equation model of the upper equatorial ocean. J. Comput. Phys., 81 , 444–480.
Giese, B. S., and Carton J. A. , 1999: Interannual and decadal variability in the tropical and midlatitude Pacific Ocean. J. Climate, 12 , 3402–3418.
Gould, J., and Coauthors, 2004: Argo profiling floats bring new era of in situ ocean observations. Eos, Trans. Amer. Geophys. Union, 85 , 185–190.
Grima, N., Bentamy A. , Katsaros K. , and Quilfen Y. , 1999: Sensitivity of an ocean circulation model forced with satellite wind stress fields. J. Geophys. Res., 104 , 7967–7989.
Halpern, D., Chao Y. , Ma C-C. , and Mechoso C. , 1995: Comparison of tropical Pacific temperature and current simulations with two vertical mixing schemes embedded in an ocean general circulation model and reference to observations. J. Geophys. Res., 100 , 2515–2522.
Hoteit, I., Cornuelle B. , Köhl A. , and Stammer D. , 2005: Treating strong adjoint sensitivities in tropical eddy-permitting variational data assimilation. Quart. J. Roy. Meteor. Soc., 131 , 3659–3682.
Johnson, G., Sloyan B. , Kessler W. , and McTaggart K. , 2002: Direct measurements of upper ocean currents and water properties across the tropical Pacific Ocean during the 1990’s. Prog. Oceanogr., 52 , 31–61.
Josey, S., Kent E. , and Taylor P. , 2002: Wind stress forcing of the ocean in the SOC climatology: Comparisons with the NCEP–NCAR, ECMWF, UWM/COADS, and Hellerman and Rosenstein datasets. J. Phys. Oceanogr., 32 , 1993–2019.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Kessler, W. S., 2002: Mean three-dimensional circulation in the northeast tropical Pacific. J. Phys. Oceanogr., 32 , 2457–2471.
Kessler, W. S., Johnson G. C. , and Moore D. W. , 2003: Sverdrup and nonlinear dynamics of the Pacific equatorial currents. J. Phys. Oceanogr., 33 , 994–1008.
Köhl, A., Stammer D. , and Cornuelle B. , 2007: Interannual to decadal changes in the ECCO global synthesis. J. Phys. Oceanogr., 37 , 313–337.
Lagerloef, G., Mitchum G. , Lukas R. , and Niiler P. , 1999: Tropical Pacific near-surface currents estimated from altimeter, wind, and drifter data. J. Geophys. Res., 104 , 23313–23326.
Large, W., McWilliams J. , and Doney S. , 1994: Oceanic vertical mixing: A review and a model with a nonlocal boundary layer parameterization. Rev. Geophys., 32 , 363–403.
Levitus, S., and Boyer T. P. , 1994: Temperature. Vol. 4. World Ocean Atlas 1994, NOAA Atlas NESDIS 4, 117 pp.
Levitus, S., Burgett R. , and Boyer T. P. , 1994: Salinity. Vol. 3, World Ocean Atlas 1994, NOAA Atlas NESDIS 3, 99 pp.
Marshall, J., Adcroft A. , Hill C. , Perelman L. , and Heisey C. , 1997: A finite-volume, incompressible Navier Stokes model for studies of the ocean on parallel computers. J. Geophys. Res., 102 , 5753–5766.
McCreary, J. P., Lu P. , and Yu Z-J. , 2002: Dynamics of the Pacific subsurface countercurrents. J. Phys. Oceanogr., 32 , 2379–2404.
McPhaden, M. J., and Coauthors, 1998: The Tropical Ocean-Global Atmosphere (TOGA) observing system: A decade of progress. J. Geophys. Res., 103 , 14169–14260.
Megann, A., and New A. , 2001: The effects of resolution and viscosity in an isopycnic-coordinate model of the equatorial Pacific. J. Phys. Oceanogr., 31 , 1993–2018.
Pedlosky, J., 1996: Ocean Circulation Theory. Springer-Verlag, 453 pp.
Putman, W., Legler D. , and O’Brien J. , 2000: Interannual variability of synthesized FSU and NCEP–NCAR reanalysis pseudo-stress products over the Pacific Ocean. J. Climate, 13 , 3003–3016.
Qiao, L., and Weisberg R. H. , 1997: The zonal momentum balance of the equatorial undercurrent in the central Pacific. J. Phys. Oceanogr., 27 , 1094–1119.
Qiu, B., and Chen S. , 2004: Seasonal modulations in the eddy field of the South Pacific Ocean. J. Phys. Oceanogr., 34 , 1515–1527.
Reynolds, R. W., and Smith T. M. , 1994: Improved global sea surface temperature analysis using optimum interpolation. J. Climate, 7 , 929–948.
Smith, W., and Sandwell D. , 1997: Global seafloor topography from satellite altimetry and ship depth soundings. Science, 277 , 1956–1962.
Stammer, D., Wunsch C. , Fukumori I. , and Marshall J. , 2002a: State estimation improves prospects for ocean research. Eos, Trans. Amer. Geophys. Union, 83 , 289.
Stammer, D., and Coauthors, 2002b: The global ocean circulation during 1992–1997, estimated from ocean observations and a general circulation model. J. Geophys. Res., 107 .3118, doi:10.1029/2001JC000888.
Stammer, D., Ueyoshi K. , Köhl A. , Large W. G. , Josey S. A. , and Wunsch C. , 2004: Estimating air-sea fluxes of heat, freshwater, and momentum through global ocean data assimilation. J. Geophys. Res., 109 .C05023, doi:10.1029/2003JC002082.
Stockdale, T. D., Anderson D. , Davey M. , Delecleuse P. , Kattenberg A. , Kitamura Y. , Latif M. , and Yamagata T. , 1993: Intercomparison of tropical ocean GCMs. World Climate Research Programme, WCRP-79, WMO/TD-No. 545, 90 pp.
Sun, D-Z., Fasullo J. , Zhang T. , and Roubicek A. , 2003: On the radiative and dynamical feedbacks over the equatorial Pacific cold tongue. J. Climate, 16 , 2425–2432.
Swenson, M. S., and Hansen D. V. , 1999: Tropical Pacific Ocean mixed layer heat budget: The Pacific cold tongue. J. Phys. Oceanogr., 29 , 69–81.
Vialard, J., Menkes C. , Boulanger J-P. , Delecluse P. , and Guilyardi E. , 2001: A model study of oceanic mechanisms affecting equatorial Pacific sea surface temperature during the 1997–98 El Niño. J. Phys. Oceanogr., 31 , 1649–1675.
Weaver, A. T., Vialard J. , and Anderson D. L. T. , 2003: Three- and four-dimensional variational assimilation with a general circulation model of the tropical Pacific Ocean. Part I: Formulation, internal diagnostics, and consistency checks. Mon. Wea. Rev., 131 , 1360–1378.
Yang, S-K., Hou Y-T. , Miller A. J. , and Campana K. A. , 1999: Evaluation of the earth radiation budget in NCEP–NCAR reanalysis with ERBE. J. Climate, 12 , 477–492.
Yu, Y., Emery W. , and Leben R. , 1995: Satellite altimeter derived geostrophic currents in the western tropical Pacific during 1992–1993 and their validation with drifting buoy trajectories. J. Geophys. Res., 100 , 25069–25085.
Zhang, K., and Marotzke J. , 1999: The importance of open boundary estimation for an Indian Ocean GCM data synthesis. J. Mar. Res., 57 , 305–334.
Mean of the adjustments imposed by the ECCO optimization procedure on the NCEP wind stress (differences between ECCO and NCEP mean wind stress) for the (a) zonal and (b) meridional components. Standard deviation of the (c) NCEP and (d) ECCO wind stress magnitude. Units are in N m−2.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Mean SSH on the equator from T/P data and the different model runs (see Table 1 for list of runs).
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
RMS difference with the spatial mean removed between T/P anomalies and the different model runs (see Table 1 for list of runs). Units: cm.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Standard deviation of monthly averaged SSH fields from (a) T/P data, (b) run with NCEP forcing, (c) 1/3° run with ECCO forcing, (d) 1° × 1° run, (e) ECCOM forcing, and (f) NCEPW.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
RMS differences (°C) between monthly means of the different model runs and Reynolds SST with temporal means removed.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
(left) Mean of the SST from (a) Reynolds data, (b) run with NCEP forcing, (c) run ECCO forcing, and (d) 1° × 1° run. (right) Longitude–time plots of the SST along the equator from (e) TAO data, (f) run with NCEP forcing, (g) run with ECCO forcing, and (h) 1° × 1° run.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Correlations between TAO and model temperature from the different model runs (see Table 1 for list of runs). Correlations at (left) 5-m, (middle) 100-m, and (right) 250-m depths for (top) 8°N, (middle) the equator, and (bottom) 8°S. Insignificant correlations at the 95% confidence level are marked with symbols of the same colors.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Mean zonal velocity at 15-m depth from (a) drifters data, (b) run with NCEP forcing, (c) run with ECCO forcing, and (d) 1° × 1° run.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Vertical distribution of the mean zonal current along the equator from (a) Johnson ADCP data, (b) run with NCEP forcing, (c) run with ECCO forcing, and (d) 1° × 1° run. Positive values indicate eastward current.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Profiles of the standard deviation for the (top) zonal and (bottom) meridional velocity along the equator at 110°W, 165°E, and 147°E as estimated from the model ADCP measurements and the different model runs (see Table 1 for list of runs).
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Vertical distribution of the correlations between TAO data and model zonal velocity from the different model runs (see Table 1 for list of runs). Insignificant correlations at the 95% confidence level are marked with symbols of the same colors.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
(left) Zonal and (right) meridional acceleration terms (m s−1) at 140-m depth for 1° × 1°, 1/3° × 1/3°, and 1/6° × 1/6° numerical models forced by ECCO forcing. The sections are shown at 189°E. Balances for the (top) 1° resolution, (middle) 1/3° resolution, and (bottom) 1/6° resolution.
Citation: Journal of Atmospheric and Oceanic Technology 25, 1; 10.1175/2007JTECHO528.1
Experiments for horizontal resolution sensitivity study and for forcing sensitivity study. “NCEPF” means original NCEP forcing, “ECCOF” means optimized forcing from the ECCO estimation, “NCEPW” means NCEP winds and ECCO heat and salinity fluxes, and “ECCOM” means NCEP forcing plus mean ECCO adjustments. In all cases, the model runs on a 1/3° × 1/3° grid. ECCOF and NCEPF will be used to assess the quality of the adjustments produced by the ECCO optimization procedure relative to the NCEP forcing fields. ECCOF and NCEPW only differ in wind stress, to assess the quality of the wind forcing. ECCOM was performed to test the skills of the mean ECCO adjustments. Finally, the comparison between NCEPF and NCEPW, which are forced with different heat (HF) and freshwater (E − P) fluxes but identical wind stress, will be used to assess the effects of these forcings. LP stands for Laplacian operator and BH is for biharmonic operator.