Argo Data Assimilation Research Articles

Observing system experiments over the Atlantic Ocean with the REMO ocean data assimilation system (RODAS) into HYCOM

The Oceanographic Modeling and Observation Network (REMO) focuses on scientific and technological development of operational oceanography in Brazil considering both numerical forecasting and observational systems. A key component of the forecasting system is the recently constructed REMO Ocean Data Assimilation System (RODAS) into the HYbrid Coordinate Ocean Model (HYCOM). Here, RODAS is presented for the first time with its full capability. RODAS employs a multivariate ensemble optimal interpolation scheme that is able to assimilate Argo temperature (T) and salinity (S) profiles, sea surface temperature (SST) analyses, and satellite along-track or gridded sea-level anomaly (SLA) data. RODAS is presented together with a series of Observing System Experiments (OSEs), in which components of the Global Ocean Observing System (GOOS) were systematically withheld from 3-year assimilation runs over the Atlantic Ocean. Using the same initial condition from a full assimilation run, OSEs were performed from 1 January 2010 to 31 December 2012 withholding (i) only Argo data; (ii) only UK MetOffice OSTIA SST analyses; (iii) only satellite along-track altimetry data; and (iv) all observation types. These runs were also compared with the full assimilation run and the model free run to evaluate the impact of different observations into the model state. The results show that each observation type brings complementary information into the analyses. Assimilation of SST is needed to better constrain the mixed layer temperature, while assimilation of SLA mainly improves the representation of circulation by adding mesoscale-like features, such as those found in the Gulf Stream and the Brazil-Malvinas Confluence. In the subsurface, only Argo observations are able to constrain the thermohaline state. When Argo data are withheld, the quality of S is seriously compromised and becomes worse than the free run in the upper ocean. Additionally, the run withholding all observation types shows that the model state in the surface almost reaches the free run state by the end of the third year. However, below 300 m, the memory of the Argo data assimilation is longer and the quality of T and S is only degraded by 35% in comparison with the full assimilation run.

Evaluation of the Impact of Argo Data on Ocean Reanalysis in the Pacific Region

Observing System Simulation Experiments (OSSEs) have been conducted to evaluate the effect of Argo data assimilation on ocean reanalysis in the Pacific region. The “truth” is obtained from a 5-year model integration from 2003 to 2007 based on the MIT general circulation model with the truly varying atmospheric forcing. The “observations” are the projections of the truth onto the observational network including ocean station data, CTD, and various BTs and Argo, by adding white noise to simulate observational errors. The data assimilation method employed is a sequential three-dimensional variational (3D-Var) scheme within a multigrid framework. Results show the interannual variability of temperature, salinity, and current fields can be reconstructed fairly well. The spread of temperature anomalies in the tropical Pacific region is also able to be reflected accurately when Argo data is assimilated, which may provide a reliable initial field for the forecast of temperature and currents for the subsurface in the tropical Pacific region. The adjustment of salinity by using T-S relationship is vital in the tropical Pacific region. However, the adjustment of salinity is almost meaningless in the northwest Pacific if Argo data is included during the reanalysis.

ARGO data assimilation into the ocean dynamics model with high spatial resolution using Ensemble Optimal Interpolation (EnOI)

The article proposes parallel implementation of the Ensemble Optimal Interpolation (EnOI) data assimilation (DA) method in eddy-resolving World Ocean circulation model. The results of DA experiments in North Atlantic with ARGO drifters are compared with the multivariate optimal interpolation (MVOI) DA scheme. The sensitivity of the model error, i.e., the difference between the model and observations depending on the number of ensemble elements, is also assessed and presented. The effectiveness of this method over the MVOI scheme is confirmed. The model outputs with and without assimilation are also compared with independent sea surface temperature data from ARMOR 3d.

How essential are Argo observations to constrain a global ocean data assimilation system?

Abstract. Observing system experiments (OSEs) are carried out over a 1-year period to quantify the impact of Argo observations on the Mercator Ocean 0.25° global ocean analysis and forecasting system. The reference simulation assimilates sea surface temperature (SST), SSALTO/DUACS (Segment Sol multi-missions dALTimetrie, d'orbitographie et de localisation précise/Data unification and Altimeter combination system) altimeter data and Argo and other in situ observations from the Coriolis data center. Two other simulations are carried out where all Argo and half of the Argo data are withheld. Assimilating Argo observations has a significant impact on analyzed and forecast temperature and salinity fields at different depths. Without Argo data assimilation, large errors occur in analyzed fields as estimated from the differences when compared with in situ observations. For example, in the 0–300 m layer RMS (root mean square) differences between analyzed fields and observations reach 0.25 psu and 1.25 °C in the western boundary currents and 0.1 psu and 0.75 °C in the open ocean. The impact of the Argo data in reducing observation–model forecast differences is also significant from the surface down to a depth of 2000 m. Differences between in situ observations and forecast fields are thus reduced by 20 % in the upper layers and by up to 40 % at a depth of 2000 m when Argo data are assimilated. At depth, the most impacted regions in the global ocean are the Mediterranean outflow, the Gulf Stream region and the Labrador Sea. A significant degradation can be observed when only half of the data are assimilated. Therefore, Argo observations matter to constrain the model solution, even for an eddy-permitting model configuration. The impact of the Argo floats' data assimilation on other model variables is briefly assessed: the improvement of the fit to Argo profiles do not lead globally to unphysical corrections on the sea surface temperature and sea surface height. The main conclusion is that the performance of the Mercator Ocean 0.25° global data assimilation system is heavily dependent on the availability of Argo data.

Argo data assimilation and its effect on climate state estimation and forecasting in the western North Pacific using a coupled model

AbstractIn this study, we investigated the effects of ocean subsurface data (Argo data) on climate state estimation and forecasting, focusing on the reproduction of North Pacific subtropical mode water (STMW) using a four‐dimensional variational data assimilation system with a coupled model. We produced two reanalysis plus forecast data sets for the ocean and atmosphere in 2010 using a 3 month assimilation period: the first including Argo data (Argo case) and the second did not include Argo data (control case). In the control case, the Kuroshio, Kuroshio Extension front, and recirculation gyres along the front were not adequately reproduced. Consequently, there were large biases in temperature and salinity in the western North Pacific. The assimilation of Argo data effectively corrected these biases and significantly improved reproduction of the Kuroshio fronts and recirculation gyres, resulting in a more realistic reproduction of the winter mixed layer and STMW. The correction of these biases is critical to the 1–3 year predictions of the STMW core properties, and the assimilation of Argo data enables prediction of these properties for more than a year. We showed that assimilation of Argo data affects the surface atmospheric temperature above the STMW formation region.

Argo data assimilation into HYCOM with an EnOI method in the Atlantic Ocean

Abstract. An ocean data assimilation system to assimilate Argo temperature (T) and salinity (S) profiles into the HYbrid Coordinate Ocean Model (HYCOM) was constructed, implemented and evaluated for the first time in the Atlantic Ocean (78° S to 50° N and 98° W to 20° E). The system is based on the ensemble optimal interpolation (EnOI) algorithm proposed by Xie and Zhu (2010), especially made to deal with the hybrid nature of the HYCOM vertical coordinate system with multiple steps. The Argo T–S profiles were projected to the model vertical space to create pseudo-observed layer thicknesses (Δ pobs), which correspond to the model target densities. The first step was to assimilate Δ pobs considering the sub-state vector composed by the model layer thickness (Δ p) and the baroclinic velocity components. After that, T and S were assimilated separately. Finally, T was diagnosed below the mixed layer to preserve the density of the model isopycnal layers. Five experiments were performed from 1 January 2010 to 31 December 2012: a control run without assimilation, and four assimilation runs considering the different vertical localizations of T, S and Δ p. The assimilation experiments were able to significantly improve the thermohaline structure produced by the control run. They reduced the root mean square deviation (RMSD) of T and S calculated with respect to Argo independent data in 34 and 44%, respectively, in comparison to the control run. In some regions, such as the western North Atlantic, substantial corrections in the 20 °C isotherm depth and the upper ocean heat content towards climatological states were achieved. The runs with a vertical localization of Δ p showed positive impacts in the correction of the thermohaline structure and reduced the RMSD of T (S) from 0.993 °C (0.149 psu) to 0.905 °C (0.138 psu) for the whole domain with respect to the other assimilation runs.

Improvement of short-term forecasting in the northwest Pacific through assimilating Argo data into initial fields

The impact of assimilating Argo data into an initial field on the short-term forecasting accuracy of temperature and salinity is quantitatively estimated by using a forecasting system of the western North Pacific, on the base of the Princeton ocean model with a generalized coordinate system (POMgcs). This system uses a sequential multigrid three-dimensional variational (3DVAR) analysis scheme to assimilate observation data. Two numericalexperimentswereconductedwith andwithoutArgotemperatureandsalinityprofiledata besides conventional temperature and salinity profile data and sea surface height anomaly (SSHa) and sea surface temperature (SST) in the process of assimilating data into the initial fields. The forecast errors are estimated by using independent temperature and salinity profiles during the forecasting period, including theverticaldistributionsofthehorizontallyaveragedrootmeansquareerrors(H-RMSEs)andthehorizontal distributions of the vertically averaged mean errors (MEs) and the temporal variation of spatially averaged root mean square errors (S-RMSEs). Comparison between the two experiments shows that the assimilation of Argo data significantly improves the forecast accuracy, with 24% reduction of H-RMSE maximum for the temperature, and the salinity forecasts are improved more obviously, averagely dropping of 50% for H-RMSEs in depth shallower than 300 m. Such improvement is caused by relatively uniform sampling of both temperature and salinity from the Argo drifters in time and space.

A Time-Averaged Covariance Method in the EnKF for Argo Data Assimilation

In the ensemble Kalman filter (EnKF), ensemble size is one of the key factors that significantly affects the performance of a data assimilation system. A relatively small ensemble size often must be chosen because of the limitations of computational resources, which often biases the estimation of the background error covariance matrix. This is an issue of particular concern in Argo data assimilation, where the most complex state-of-the-art models are often used. In this study, we propose a time-averaged covariance method to estimate the background error covariance matrix. This method assumes that the statistical properties of the background errors do not change significantly at neighbouring analysis steps during a short time window, allowing the ensembles generated at previous steps to be used in present steps. As such, a joint ensemble matrix combining ensembles of previous and present steps can be constructed to form a larger ensemble for estimating the background error covariance. This method can enlarge the ensemble size without increasing the number of model integrations, and this method is equivalent to estimating the background error covariance matrix using the mean ensemble covariance averaged over several assimilation steps. We apply this method to the assimilation of Argo and altimetry datasets with an oceanic general circulation model. Experiments show that the use of this time-averaged covariance can improve the performance of the EnKF by reducing the root mean square error (RMSE) and improving the estimation of error covariance structure as well as the relationship between ensemble spread and RMSE. RÉSUMÉ [Traduit par la rédaction] Dans le filtre de Kalman d'ensemble (EnKF), la taille de l'ensemble est l'un des facteurs clés qui ont une influence importante sur la performance d'un système d'assimilation de données. Il faut souvent choisir une taille d'ensemble assez petite à cause des limites des ressources informatiques, ce qui biaise souvent l'estimation de la matrice de covariance de l'erreur de fond. Cette question revêt une importance particulière pour l'assimilation des données Argo, qui fait souvent appel à des modèles de pointe très complexes. Dans cette étude, nous proposons une méthode de covariance moyennée dans le temps pour estimer la matrice de covariance de l'erreur de fond. Cette méthode suppose que les propriétés statistiques des erreurs de fond ne changent pas de façon importante d'une étape d'analyse à la suivante durant un court laps de temps, ce qui permet d'utiliser dans les étapes courantes les ensembles générés aux étapes précédentes. Ainsi, on peut construire une matrice d'ensembles conjoints combinant les ensembles des étapes précédentes et courantes pour former un plus grand ensemble dans le but d'estimer la covariance de l'erreur de fond. Cette méthode peut accroître la taille de l'ensemble sans augmenter le nombre d'intégrations du modèle; elle équivaut à estimer la matrice de covariance de l'erreur de fond en utilisant la covariance moyenne de l'ensemble calculée sur plusieurs étapes d'assimilation. Nous appliquons cette méthode à l'assimilation des ensembles de données Argo et d'altimétrie avec un modèle de circulation océanique générale. Des essais montrent que l'emploi de cette covariance moyennée dans le temps peut améliorer la performance de l'EnKF en réduisant l’écart-type et en améliorant l'estimation de la structure de la covariance de l'erreur de même que la relation entre l'étalement et l'écart-type l'ensemble.

Argo data assimilation in ocean general circulation model of Northwest Pacific Ocean

The Argo temperature and salinity profiles in 2005–2009 are assimilated into a coastal ocean general circulation model of the Northwest Pacific Ocean using the ensemble adjustment Kalman filter (EAKF). Three numerical tests, including the control run (CTL) (without data assimilation, which serves as the reference experiment), ensemble free run (EnFR) (without data assimilation), and EAKF experiment (with Argo data assimilation using EAKF), are carried out to examine the performance of this system. Using the restarts of different years as the initial conditions of the ensemble integrations, the ensemble spreads from EnFR and EAKF are all kept at a finite value after a sharp decreasing in the first few months because of the sensitive of the model to the initial conditions, and the reducing of the ensemble spread due to Argo data assimilation is not much. The ensemble samples obtained in this way can well represent the probabilities of the real ocean states, and no ensemble inflation is necessary for this EAKF experiment. Different experiment results are compared with satellite sea surface temperature (SST) data and the Global Temperature-Salinity Profile Program (GTSPP) data. The comparison of SST shows that modeled SST errors are reduced after data assimilation; the error reduction percentage after assimilating the Argo profiles is about 10 % on average. The comparison against the GTSPP profiles, which are independent of the Argo profiles, shows improvements in both temperature and salinity. The comparison results indicated a great error reduction in all vertical layers relative to CTL and the ensemble mean of EnFR; the maximum value for temperature and salinity reaches to 85 % and 80 %, respectively. The standard deviations of sea surface height are employed to examine the simulation ability, and it is shown that the mesoscale variability is improved after Argo data assimilation, especially in the Kuroshio extension area and along the section of 10°N. All these results suggest that this system is potentially useful for improving the simulation ability of oceanic numerical models.

Importance of the assimilation of Argo float measurements on the Meridional Overturning Circulation in the South Atlantic

[1] The meridional overturning circulation (MOC) and meridional heat transport (MHT) obtained from two GFDL coupled models, with and without data assimilation, are examined and compared with observations collected at nominally 34°S in the South Atlantic. The results demonstrate that the performance of the Geophysical Fluid Dynamic Laboratory (GFDL) coupled data assimilation (CDA) model is quite different between the two periods, 1979–2002 and 2003–2007, due to the assimilation of Argo data in later period. The MOC components from the GFDL CDA during 1979–2002 are similar to those from GFDL CM2.1 IPCC simulation, both give weak boundary currents and strong interior overturning transport compared to observations. However, after assimilating temperature and salinity profiles from the Argo floats, the performance of the GFDL CDA is greatly improved in terms of representing the observed MOC and MHT structure: the transports of boundary currents are twice as strong as those during pre-Argo period, and the overturning flow in the interior region is reduced. Possible causes for the changes in model performance are discussed.

Using ensemble adjustment Kalman filter to assimilate Argo profiles in a global OGCM

An ensemble adjustment Kalman filter (EAKF) is used to assimilate Argo profiles of 2008 in a global version of the Modular Ocean Model version 4. Four assimilation experiments are carried out to compare with the simulation without data assimilation, which serves as the control experiment. All experiment results are compared with dataset of Global Temperature–Salinity Profile Program and satellite sea surface temperature (SST). The first experiment (Exp 1) is implemented by perturbing temperature of upper layers in the initial conditions (ICs) with an amplitude of 1.0°C and no ensemble inflation. The results from Exp 1 show that the simulated temperature (salinity) deviation in the upper 400 m (500 m) is reduced through Argo data assimilation; however, these deviations are increased in deeper layers. The error reduction in SST is much greater during January to June than during the rest of the year. Three more experiments are designed to understand the responses in different layers and months. Two of them test model sensitivities to ICs by perturbing them vertically: one over the vertical extent of the whole water column (Exp 2) and the other employs smaller perturbation amplitude of 0.1°C (Exp 3). Exp 2 shows that the simulated temperature and salinity deviations are systematically improved in the whole water column. Comparison between Exps 2 and 3 suggests that perturbation amplitude is important. Exp 4 tests the influence of the optimal inflation factor of 5%, which is determined by other set of numerical tests. Exp 4 improves assimilation performance much more than the other three experiments without inflation. Therefore, we conclude that the perturbation should be introduced to all model layers, proper perturbation amplitude is important for Ocean data assimilation using EAKF, and the ensemble inflation by an optimal inflation is critical to improve the skill of the EAKF analysis.

Observing System Evaluations Using GODAE Systems

Global ocean forecast systems, developed under the Global Ocean Data Assimilation Experiment (GODAE), are a powerful means of assessing the impact of different components of the Global Ocean Observing System (GOOS). Using a range of analysis tools and approaches, GODAE systems are useful for quantifying the impact of different observation types on the quality of analyses and forecasts. This assessment includes both existing and future observation platforms. Many important conclusions can be drawn from these studies. It is clear that altimeter data are extremely important for constraining mesoscale variability in ocean forecast systems. The number of altimeters is also important. For example, near-real-time applications need data from four altimeters to achieve skill that is similar to systems using data from two altimeters in delayed mode. Another important result is that sea surface temperature is the only observation parameter that adequately monitors ocean properties in coastal regions and shallow seas. Assimilation of Argo data provides a significant, measurable improvement to GODAE systems, and is the only observation platform that provides global-scale information for constraining salinity. The complementary nature of different components of GOOS is now clear and the emergence of new assimilation techniques for observing system evaluation provides the GODAE community with a practical path toward routine GOOS monitoring.