Testing a four-dimensional variational data assimilation method using an improved intermediate coupled model for ENSO analysis and prediction

A four-dimensional variational (4D-Var) data assimilation method is implemented in an improved intermediate coupled model (ICM) of the tropical Pacific. A twin experiment is designed to evaluate the impact of the 4D-Var data assimilation algorithm on ENSO analysis and prediction based on the ICM. The model error is assumed to arise only from the parameter uncertainty. The “observation” of the SST anomaly, which is sampled from a “truth” model simulation that takes default parameter values and has Gaussian noise added, is directly assimilated into the assimilation model with its parameters set erroneously. Results show that 4D-Var effectively reduces the error of ENSO analysis and therefore improves the prediction skill of ENSO events compared with the non-assimilation case. These results provide a promising way for the ICM to achieve better real-time ENSO prediction.     

Tropical storms: representation and diagnosis in climate models and the impacts of climate change

Tropical storms are located and tracked in an experiment in which a high-resolution atmosphere only model is forced with observed sea surface temperatures (SSTs) and sea ice. The structure, geographic distribution and seasonal variability of the model tropical storms show some similarities with observations. The simulation of tropical storms is better in this high-resolution experiment than in a parallel standard resolution experiment. In an anomaly experiment, sea ice, SSTs and greenhouse-gas forcing are changed to mimic the changes that occur in a coupled model as greenhouse-gases are increased. There are more tropical storms in this experiment than in the control experiment in the Northeast Pacific and Indian Ocean basins and fewer in the North Atlantic, Northwest Pacific and Southwest Pacific region. The changes in the North Atlantic and Northwest Pacific can be linked to El Niño-like behaviour. A comparison of the tracking results with two empirically derived tropical storm genesis parameters is carried out. The tracking technique and a convective genesis parameter give similar results, both in the global distribution and in the changes in the individual basins. The convective genesis parameter is also applied to parallel coupled model experiments that have a lower horizontal resolution. The changes in the global distribution of tropical storms in the coupled model experiments are consistent with the changes seen at higher resolution. This indicates that the convective genesis parameter may still provide useful information about tropical storm changes in experiments carried out with models that cannot resolve tropical storms.     

Impact of the Time Scale of Model Sensitivity Response on Coupled Model Parameter Estimation

That a model has sensitivity responses to parameter uncertainties is a key concept in implementing model parameter estimation using filtering theory and methodology.Depending on the nature of associated physics and characteristic variability of the fluid in a coupled system,the response time scales of a model to parameters can be different,from hourly to decadal.Unlike state estimation,where the update frequency is usually linked with observational frequency,the update frequency for parameter estimation must be associated with the time scale of the model sensitivity response to the parameter being estimated.Here,with a simple coupled model,the impact of model sensitivity response time scales on coupled model parameter estimation is studied.The model includes characteristic synoptic to decadal scales by coupling a long-term varying deep ocean with a slow-varying upper ocean forced by a chaotic atmosphere.Results show that,using the update frequency determined by the model sensitivity response time scale,both the reliability and quality of parameter estimation can be improved significantly,and thus the estimated parameters make the model more consistent with the observation.These simple model results provide a guideline for when real observations are used to optimize the parameters in a coupled general circulation model for improving climate analysis and prediction initialization.     

Accounting for CO<Subscript>2</Subscript> variability over East Asia with a regional joint inversion system and its preliminary evaluation

A regional surface carbon dioxide (CO₂) flux inversion system, the Tan-Tracker-Region, was developed by incorporating an assimilation scheme into the Community Multiscale Air Quality (CMAQ) regional chemical transport model to resolve fine-scale CO₂ variability over East Asia. The proper orthogonal decomposition-based ensemble four-dimensional variational data assimilation approach (POD-4DVar) is the core algorithm for the joint assimilation framework, and simultaneous assimilations of CO₂ concentrations and surface CO₂ fluxes are applied to help reduce the uncertainty in initial CO₂ concentrations. A persistence dynamical model was developed to describe the evolution of the surface CO₂ fluxes and help avoid the “signal-to-noise” problem; thus, CO₂ fluxes could be estimated as a whole at the model grid scale, with better use of observation information. The performance of the regional inversion system was evaluated through a group of single-observation-based observing system simulation experiments (OSSEs). The results of the experiments suggest that a reliable performance of Tan-Tracker-Region is dependent on certain assimilation parameter choices, for example, an optimized window length of approximately 3 h, an ensemble size of approximately 100, and a covariance localization radius of approximately 320 km. This is probably due to the strong diurnal variation and spatial heterogeneity in the fine-scale CMAQ simulation, which could affect the performance of the regional inversion system. In addition, because all observations can be artificially obtained in OSSEs, the performance of Tan-Tracker-Region was further evaluated through different densities of the artificial observation network in different CO₂ flux situations. The results indicate that more observation sites would be useful to systematically improve the estimation of CO₂ concentration and flux in large areas over the model domain. The work presented here forms a foundation for future research in which a thorough estimation of CO₂ flux variability over East Asia could be performed with the regional inversion system.     

Accounting for CO_2 Variability over East Asia with a Regional Joint Inversion System and Its Preliminary Evaluation

A regional surface carbon dioxide(CO_2) flux inversion system, the Tan-Tracker-Region, was developed by incorporating an assimilation scheme into the Community Multiscale Air Quality(CMAQ) regional chemical transport model to resolve fine-scale CO_2 variability over East Asia. The proper orthogonal decomposition-based ensemble four-dimensional variational data assimilation approach(POD-4 DVar) is the core algorithm for the joint assimilation framework, and simultaneous assimilations of CO_2 concentrations and surface CO_2 fluxes are applied to help reduce the uncertainty in initial CO_2 concentrations. A persistence dynamical model was developed to describe the evolution of the surface CO_2 fluxes and help avoid the "signal-to-noise" problem; thus, CO_2 fluxes could be estimated as a whole at the model grid scale, with better use of observation information. The performance of the regional inversion system was evaluated through a group of single-observation-based observing system simulation experiments(OSSEs). The results of the experiments suggest that a reliable performance of Tan-Tracker-Region is dependent on certain assimilation parameter choices, for example, an optimized window length of approximately 3 h, an ensemble size of approximately 100, and a covariance localization radius of approximately 320 km. This is probably due to the strong diurnal variation and spatial heterogeneity in the fine-scale CMAQ simulation, which could affect the performance of the regional inversion system. In addition, because all observations can be artificially obtained in OSSEs, the performance of Tan-Tracker-Region was further evaluated through different densities of the artificial observation network in different CO_2 flux situations. The results indicate that more observation sites would be useful to systematically improve the estimation of CO_2 concentration and flux in large areas over the model domain. The work presented here forms a foundation for future research in which a thorough estimation of CO_2 flux variability over East Asia could be performed with the regional inversion system.     

Estimation methods for monthly humidity from dynamical downscaling data for quantitative assessments of climate change impacts

Methods are proposed to estimate the monthly relative humidity and wet bulb temperature based on observations from a dynamical downscaling coupled general circulation model with a regional climate model (RCM) for a quantitative assessment of climate change impacts. The water vapor pressure estimation model developed was a regression model with a monthly saturated water vapor pressure that used minimum air temperature as a variable. The monthly minimum air temperature correction model for RCM bias was developed by stepwise multiple regression analysis using the difference in monthly minimum air temperatures between observations and RCM output as a dependent variable and geographic factors as independent variables. The wet bulb temperature was estimated using the estimated water vapor pressure, air temperature, and atmospheric pressure at ground level both corrected for RCM bias. Root mean square errors of the data decreased considerably in August.     

集合卡尔曼滤波同化多普勒雷达资料的观测系统模拟试验

本文将集合卡尔曼滤波同化技术应用到对流尺度系统中,实施了基于WRF模式的同化单部多普勒雷达径向风和反射率因子的观测系统模拟试验,验证了其在对流尺度中应用的可行性和有效性,并对同化系统的特性进行了探讨。试验表明:WRF-EnKF雷达资料同化系统能较准确分析模式风暴的流场、热力场、微物理量场的细致特征;几乎所有变量的预报和分析误差经过同化循环后都能显著下降,同化分析基本上能使预报场在各层上都有所改进,对预报场误差较大层次的更正更为显著;约8个同化循环后,EnKF能在雷达反射率、径向风观测与背景场间建立较可靠的相关关系,使模式各变量场能被准确分析更新,背景场误差协方差在水平方向和垂直方向都有着复杂的结构,是高度非均匀、各项异性和流依赖的;集合平均分析场做的确定性预报在短时间内能较好保持真值场风暴的细节结构,但预报误差增长较快。     

Implementation of a one-dimensional enthalpy sea-ice model in a simple pycnocline prediction model for sea-ice data assimilation studies

To further explore enthalpy-based sea-ice assimilation, a one-dimensional(1D) enthalpy sea-ice model is implemented into a simple pycnocline prediction model. The 1D enthalpy sea-ice model includes the physical processes such as brine expulsion, flushing, and salt diffusion. After being coupled with the atmosphere and ocean components, the enthalpy sea-ice model can be integrated stably and serves as an important modulator of model variability. Results from a twin experiment show that the sea-ice data assimilation in the enthalpy space can produce smaller root-mean-square errors of model variables than the traditional scheme that assimilates the observations of ice concentration, especially for slow-varying states. This study provides some insights into the improvement of sea-ice data assimilation in a coupled general circulation model.     

ENSO Predictions in an Intermediate Coupled Model Influenced by Removing Initial Condition Errors in Sensitive Areas: A Target Observation Perspective

Previous studies indicate that ENSO predictions are particularly sensitive to the initial conditions in some key areas (so-called "sensitive areas"). And yet, few studies have quantified improvements in prediction skill in the context of an optimal observing system. In this study, the impact on prediction skill is explored using an intermediate coupled model in which errors in initial conditions formed to make ENSO predictions are removed in certain areas. Based on ideal observing system simulation experiments, the importance of various observational networks on improvement of El Niño prediction skill is examined. The results indicate that the initial states in the central and eastern equatorial Pacific are important to improve El Niño prediction skill effectively. When removing the initial condition errors in the central equatorial Pacific, ENSO prediction errors can be reduced by 25%. Furthermore, combinations of various subregions are considered to demonstrate the efficiency on ENSO prediction skill. Particularly, seasonally varying observational networks are suggested to improve the prediction skill more effectively. For example, in addition to observing in the central equatorial Pacific and its north throughout the year, increasing observations in the eastern equatorial Pacific during April to October is crucially important, which can improve the prediction accuracy by 62%. These results also demonstrate the effectiveness of the conditional nonlinear optimal perturbation approach on detecting sensitive areas for target observations.     

Field significance of performance measures in the context of regional climate model evaluation. Part 1: temperature

A new approach for rigorous spatial analysis of the downscaling performance of regional climate model (RCM) simulations is introduced. It is based on a multiple comparison of the local tests at the grid cells and is also known as “field” or “global” significance. New performance measures for estimating the added value of downscaled data relative to the large-scale forcing fields are developed. The methodology is exemplarily applied to a standard EURO-CORDEX hindcast simulation with the Weather Research and Forecasting (WRF) model coupled with the land surface model NOAH at 0.11 ^° grid resolution. Monthly temperature climatology for the 1990–2009 period is analysed for Germany for winter and summer in comparison with high-resolution gridded observations from the German Weather Service. The field significance test controls the proportion of falsely rejected local tests in a meaningful way and is robust to spatial dependence. Hence, the spatial patterns of the statistically significant local tests are also meaningful. We interpret them from a process-oriented perspective. In winter and in most regions in summer, the downscaled distributions are statistically indistinguishable from the observed ones. A systematic cold summer bias occurs in deep river valleys due to overestimated elevations, in coastal areas due probably to enhanced sea breeze circulation, and over large lakes due to the interpolation of water temperatures. Urban areas in concave topography forms have a warm summer bias due to the strong heat islands, not reflected in the observations. WRF-NOAH generates appropriate fine-scale features in the monthly temperature field over regions of complex topography, but over spatially homogeneous areas even small biases can lead to significant deteriorations relative to the driving reanalysis. As the added value of global climate model (GCM)-driven simulations cannot be smaller than this perfect-boundary estimate, this work demonstrates in a rigorous manner the clear additional value of dynamical downscaling over global climate simulations. The evaluation methodology has a broad spectrum of applicability as it is distribution-free, robust to spatial dependence, and accounts for time series structure.     

Idealized Experiments for Optimizing Model Parameters Using a 4D-Variational Method in an Intermediate Coupled Model of ENSO

Large biases exist in real-time ENSO prediction, which can be attributed to uncertainties in initial conditions and model parameters. Previously, a 4 D variational(4 D-Var) data assimilation system was developed for an intermediate coupled model(ICM) and used to improve ENSO modeling through optimized initial conditions. In this paper, this system is further applied to optimize model parameters. In the ICM used, one important process for ENSO is related to the anomalous temperature of subsurface water entrained into the mixed layer(T_e), which is empirically and explicitly related to sea level(SL) variation.The strength of the thermocline effect on SST(referred to simply as "the thermocline effect") is represented by an introduced parameter, αT_e. A numerical procedure is developed to optimize this model parameter through the 4 D-Var assimilation of SST data in a twin experiment context with an idealized setting. Experiments having their initial condition optimized only,and having their initial condition plus this additional model parameter optimized, are compared. It is shown that ENSO evolution can be more effectively recovered by including the additional optimization of this parameter in ENSO modeling.The demonstrated feasibility of optimizing model parameters and initial conditions together through the 4 D-Var method provides a modeling platform for ENSO studies. Further applications of the 4 D-Var data assimilation system implemented in the ICM are also discussed.     

Multi-criteria spatialization of soil organic carbon sequestration potential from agricultural intensification in Senegal

On the eve of the 15th climate negotiations conference in Copenhagen, the pressure to assess all climate mitigation options is mounting. In this study, a bio-physic model and a socio-economic model were designed and coupled to assess the carbon sequestration potential of agricultural intensification in Senegal. The biophysical model is a multiple linear regression, calibrated and tested on a dataset of long-term agricultural trials established in West Africa. The socio-economic model integrates both financial and environmental costs related to considered practice changes. Both models are spatially explicit and the resulting spatial patterns were computed and displayed over Senegal with a geographic information system. The national potential from large-scale intensification was assessed at 0.65–0.83 MtC. With regards to local-scaled intensification as local projects, the most profitable areas were identified in agricultural expansion regions (especially Casamance), while the areas that meet the current financial additionality criteria of the Clean Development Mechanism were located in the northern part of the Peanut Basin. Using the current relevant mode of carbon valuation (Certified Emission Reductions), environmental benefits are small compared to financial benefits. This picture is radically changed if “avoided deforestation”, a likely consequence of agricultural intensification, is accounted for as the greenhouse gases sink capacity of projects increases by an average of a hundred-fold over Senegal.     

Coupled atmosphere–ocean data assimilation experiments with a low-order climate model

A simple idealized atmosphere–ocean climate model and an ensemble Kalman filter are used to explore different coupled ensemble data assimilation strategies. The model is a low-dimensional analogue of the North Atlantic climate system, involving interactions between large-scale atmospheric circulation and ocean states driven by the variability of the Atlantic meridional overturning circulation (MOC). Initialization of the MOC is assessed in a range of experiments, from the simplest configuration consisting of forcing the ocean with a known atmosphere to performing fully coupled ensemble data assimilation. “Daily” assimilation (that is, at the temporal frequency of the atmospheric observations) is contrasted with less frequent assimilation of time-averaged observations. Performance is also evaluated under scenarios in which ocean observations are limited to the upper ocean or are non-existent. Results show that forcing the idealized ocean model with atmospheric analyses is inefficient at recovering the slowly evolving MOC. On the other hand, daily assimilation rapidly leads to accurate MOC analyses, provided a comprehensive set of oceanic observations is available for assimilation. In the absence of sufficient observations in the ocean, the assimilation of time-averaged atmospheric observations proves to be more effective for MOC initialization, including the case where only atmospheric observations are available.     

The impact of observation systems on medium-range weather forecasting in a global forecast system

To investigate the impact of various types of data on medium-range forecasts, observing system experiments are performed using an assimilation algorithm based on the National Centers for Environmental Prediction (NCEP)/Department of Energy (DOE) reanalysis system. Data-denial experiments for radiosonde, satellite, aircraft, and sea surface observations, and selected data experiments for radiosonde and surface data, are conducted for the boreal summer of 1997 and the boreal winter of 1997/1998. The data assimilation system used in this study is remarkably dependent on radiosonde data, which provides information about the three-dimensional structure of the atmosphere. As expected, the impact of radiosonde observations on medium-range forecasts is strongly positive over the Northern Hemisphere and tropics, whereas the satellite system is most beneficial over the Southern Hemisphere. These results are also found in experiments simulating historical changes in observation systems. Over the tropics, assimilation without radiosonde observations generates unbalanced analyses resulting in unrealistic forecasts that must be corrected by the forecast model. Forecasts based on analysis from the observation data before the era of radiosonde observation are found to be less meaningful. In addition, the impacts on forecasts are closely related to the geographical distribution of observation data. The memory of observation data embedded in the analysis tends to persist throughout forecasts. However, cases exist where the effect of forecast error growth is more dominant than that of analysis error, e.g., over East Asia in summer, and where the deficiency in observations is supplemented or the imbalance in analysis is adjusted by the forecast model during the period of forecasts. Forecast error growth may be related to the synoptic correction performed by the data assimilation system. Over data-rich areas, analysis fields are corrected to a greater extent by the data assimilation system than are those over data-poor areas, which can cause the forecast model to produce more forecast errors in medium-range forecasts. It is found that even one month per season is sufficient for forecast skill verification in data impact experiments. Additionally, the use of upper-air observations is found to benefit areas that are downstream of observation data-rich areas.     

A comparison of climatological observing windows and their impact on detecting daily temperature extrema

Climatological observing window (COW) is defined as a time frame over which continuous or extreme air temperature measurements are collected. A 24-h time interval, ending at 00UTC or shifted to end at 06UTC, has been associated with difficulties in characterizing daily temperature extrema. A fixed 24-h COW used to obtain the temperature minima leads to potential misidentification due to fragmentation of “nighttime” into two subsequent nighttime periods due to the time discretization interval. The correct identification of air temperature extrema is achievable using a COW that identifies daily minimum over a single nighttime period and maximum over a single daytime period, as determined by sunrise and sunset. Due to a common absence of hourly air temperature observations, the accuracy of the mean temperature estimation is dependent on the accuracy of determination of diurnal air temperature extrema. Qualitative and quantitative criteria were used to examine the impact of the COW on detecting daily air temperature extrema. The timing of the 24-h observing window occasionally affects the determination of daily extrema through a mischaracterization of the diurnal minima and by extension can lead to errors in determining daily mean temperature. Hourly air temperature data for the time period from year 1987 to 2014, obtained from Toronto Buttonville Municipal Airport weather station, were used in analysis of COW impacts on detection of daily temperature extrema and calculation of annual temperature averages based on such extrema.     

On the use of marine radar imagery for estimation of properties of the directional spectrum of the sea surface

Abstract

The “sea clutter” observable on a standard marine navigation radar has long been recognized as a potential source of information about sea state. In the last decade a number of researchers have published “directional wave spectra” calculated from marine radar images. Our group has continued this line of research using a unique radar system that digitizes and stores radar images eight bits deep directly related to the strength of the radar backscatter.

Our system was deployed on the CSS Hudson during the Grand Banks ERS‐1 SAR Wave Spectrum Validation Experiment cruise in November 1991. We collected in excess of 3000 sea surface backscatter images. From this dataset we have produced a number of directional spectra in an effort to understand the performance of the sensor and to compare it with other wave determining instruments and models.

Analysis has shown a strong azimuthal asymmetry both in the strength of the backscattered signal and in the relative strength of spectral peaks. This asymmetry is similar inform to that observed in scatterometer data. Unbiased estimation of the “true” image spectrum requires removal of these asymmetries. This estimation has been accomplished through calculation and removal of a non‐linear multi‐parameter least‐squares model of the backscatter from each image, and averaging of spectra from many look directions. The resulting spectra compare favourably with those calculated from directional wave buoy data, satellite and aircraft SARs and other directional wave measurements and models.     

A further assessment of vegetation feedback on decadal Sahel rainfall variability

The effect of vegetation feedback on decadal-scale Sahel rainfall variability is analyzed using an ensemble of climate model simulations in which the atmospheric general circulation model ICTPAGCM (“SPEEDY”) is coupled to the dynamic vegetation model VEGAS to represent feedbacks from surface albedo change and evapotranspiration, forced externally by observed sea surface temperature (SST) changes. In the control experiment, where the full vegetation feedback is included, the ensemble is consistent with the observed decadal rainfall variability, with a forced component 60 % of the observed variability. In a sensitivity experiment where climatological vegetation cover and albedo are prescribed from the control experiment, the ensemble of simulations is not consistent with the observations because of strongly reduced amplitude of decadal rainfall variability, and the forced component drops to 35 % of the observed variability. The decadal rainfall variability is driven by SST forcing, but significantly enhanced by land-surface feedbacks. Both, local evaporation and moisture flux convergence changes are important for the total rainfall response. Also the internal decadal variability across the ensemble members (not SST-forced) is much stronger in the control experiment compared with the one where vegetation cover and albedo are prescribed. It is further shown that this positive vegetation feedback is physically related to the albedo feedback, supporting the Charney hypothesis.     

Tropical Pacific internal atmospheric dynamics and resolution in a coupled GCM

A noise reduction technique, namely the interactive ensemble (IE) approach is adopted to reduce noise at the air–sea interface due to internal atmospheric dynamics in a state-of-the-art coupled general circulation model (CGCM). The IE technique uses multiple realization of atmospheric general circulation models coupled to a single ocean general circulation model. The ensembles mean fluxes from the atmospheric simulations are communicated to the ocean component. Each atmospheric simulation receives the same SST coming from the ocean component. The only difference among the atmospheric simulations comes from perturbed initial conditions, thus the atmospheric states are, in principle synoptically independent. The IE technique can be used to better understand the importance of weather noise forcing of natural variability such as El Niño Southern Oscillation (ENSO). To study the impact of weather noise and resolution in the context of a CGCM, two IE experiments are performed at different resolutions. Atmospheric resolution is an important issue since the noise statistics will depend on the spatial scales resolved. A simple formulation to extract atmospheric internal variability is presented. The results are compared to their respective control cases where internal atmospheric variability is left unchanged. The noise reduction has a major impact on the coupled simulation and the magnitude of this effect strongly depends on the horizontal resolution of the atmospheric component model. Specifically, applying the noise reduction technique reduces the overall climate variability more effectively at higher resolution. This suggests that “weather noise” is more important in sustaining climate variability as resolution increases. ENSO statistics, dynamics, and phase asymmetry are all modified by the noise reduction, in particular ENSO becomes more regular with less phase asymmetry when noise is reduced. All these effects are more marked for the higher resolution case. In contrast, ENSO frequency is unchanged by the reduction in the weather noise, but its phase-locking to the annual cycle is strongly dependent on noise and resolution. At low resolution the noise structure is similar to the signal, whereas the spatial structure of the noise deviates from the spatial structure of the signal as resolution increases. It is also suggested that event-to-event differences are largely driven by atmospheric noise as opposed to chaotic dynamics within the context of the large-scale coupled system, suggesting that there is a well-defined “canonical” event.     

Improved ENSO forecasts by assimilating sea surface temperature observations into an intermediate coupled model

A simple method for initializing intermediate coupled models (ICMs) using only sea surface temperature (SST) anomaly data is comprehensively tested in two sets of hindcasts with a new ICM. In the initialization scheme, both the magnitude of the nudging parameter and the duration of the assimilation are considered, and initial conditions for both atmosphere and ocean are generated by running the coupled model with SST anomalies nudged to the observations. A comparison with the observations indicates that the scheme can generate realistic thermal fields and surface dynamic fields in the equatorial Pacific through hindcast experiments. An ideal experiment is performed to get the optimal nudging parameters which include the nudging intensity and nudging time length. Twelve-month-long hindcast experiments are performed with the model over the period 1984–2003 and the period 1997–2003. Compared with the original prediction results, the model prediction skills are significantly improved by the nudging method especially beyond a 6-month lead time during the two different periods. Potential problems and further improvements are discussed regarding the new coupled assimilation system.     

Ensemble data assimilation in a simple coupled climate model: The role of ocean-atmosphere interaction

A conceptual coupled ocean-atmosphere model was used to study coupled ensemble data assimilation schemes with a focus on the role of ocean-atmosphere interaction in the assimilation. The optimal scheme was the fully coupled data assimilation scheme that employs the coupled covariance matrix and assimilates observations in both the atmosphere and ocean. The assimilation of synoptic atmospheric variability that captures the temporal fluctuation of the weather noise was found to be critical for the estimation of not only the atmospheric, but also oceanic states. The synoptic atmosphere observation was especially important in the mid-latitude system, where oceanic variability is driven by weather noise. The assimilation of synoptic atmospheric variability in the coupled model improved the atmospheric variability in the analysis and the subsequent forecasts, reducing error in the surface forcing and, in turn, in the ocean state. Atmospheric observation was able to further improve the oceanic state estimation directly through the coupled covariance between the atmosphere and ocean states. Relative to the mid-latitude system, the tropical system was influenced more by ocean-atmosphere interaction and, thus, the assimilation of oceanic observation becomes more important for the estimation of the ocean and atmosphere.