GODAE is sponsored by
List of poster abstracts
[A-B] [C-E] [F-G] [H] [I-K] [L] [M-O] [P-R] [S] [T-W] [X-Z]
Authors X - Z
Number 181 - Session 2
Development of a 3DVAR ocean data assimilation system for the South China Sea
Xianjun Xiao1, 2, Dongxiao Wang2, Jiang Zhu3, Changxiang Yan3
1 Beijing Climate Centre,Beijing,China
2 South China Sea institute, Chinese Academy of Sceince, Guangzhou, China
3 Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
Abstract
There is in great demand of the circulation analysis in South China Sea (SCS), but it is difficult to improve the quality of prediction only depending on the ocean circulation numerical model. So A South China Sea 3D variation assimilation system is set up to research SCS circulation.
Three dimension variation method based on Zhu and Yan (2006) and POM model are connected to construct the South China Sea 3D variation assimilation system to assimilating along-track Altimeter data. The assimilation method is improved in the system. There are two primary modifications: I) The assimilation along a satellite track and retrieval a two dimension temperature and salinity section. Compare with the original method that retrieval a T-S diagram in every single point, new method make the analysis harmonized and smooth, which avoided the singular value and ensure the average error along every track is minimum. II) A recursive filter algorithm is introduced into cost function as a preconditioning. There is no need to the inverse of background error covariance matrix and simplify the calculation.
The verification making use of the in situ data proved, the analytical temperature and salinity from the system is improved relative to the simply numerical simulation only using POM whenever the abnormal phase (1998) or the normal phase (2000) of SCS because the information included in the sea surface height is efficiently transformed into the subsurface ocean, which manifest the system is steady. In particular, the fresh water cover from the continental runoff and precipitation over SCS is represented in the result of the system. The error statistical analysis provide us more information about the system, such as the temperature and salinity is more closed to a Gauss distribution pattern compare to the simulation, that is to say the distribution of error is more reasonable. The standard error variance of temperature is reduced at every cruise with an average value 1.5 and the standard error variance of salinity is 0.2 and 0.04 lower than simulation. The circulation of assimilation is also inconsistent with other observation or numerical diagnose and some mesocale eddy is resolved.
At present, this assimilation system is applied in a 10 years reanalysis of SCS and has become a part of the Chinese regional assimilation system of GODAE.
Back to list of session 2 abstract (Operational Oceanography Infrastructure)
Number 72 - Session 4
A DATASET OF GLOBAL OCEAN SURFACE CURRENTS DERIVED FROM ARGO FLOAT TRAJECTORIES
Jiping Xie, Jiang Zhu
Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
Abstract
We estimate the surface current vectors from the Argo float surface trajectories over the global ocean during the period from January 1999 to December 2007. The estimation is done by processing each surface trajectory using the Kalman filter. The estimation errors are also provided as a part of the dataset. The errors of the estimated surface currents are about 4.7 cm s-1 which is equivalent with the precise of the currents from the surface drifting buoys. Also the estimated surface currents are compared to that directly estimated from the surface drifters. The results show good agreement both on the current amplitude and on the direction of velocities. Based on this surface current data, the averaging characteristics and seasonal variability of the main currents in global oceans could be illustrated as more detailed. This dataset of the global surface currents provides complementary information and fill many gapes to the existing datasets such as the surface currents from the surface drifters. We also made the dataset available to users of interest for many types of applications.
Keywords: ARGO floats; Surface Currents; Kalman Filter; Global Ocean
Number 48 - Session 4
SIMULATING MOW IN A BASIN-SCALE MODEL
Xiaobiao Xu1, Eric P. Chassignet2
1 DMS,University of Southern Mississippi, Stennis Space Center, USA
2 COAPS, Florida State University, Tallahassee, USA
Abstract
A high resolution, basin-scale model is used to simulate the Mediterranean outflow water (MOW) in North Atlantic Ocean. The model is based on the hybrid-coordinate ocean model (HYCOM) and includes an algebric parameterization for entrainment process.
The simulation reproduced about 0.7Sv Mediterranean outflow water exiting the Strait of Gibraltar, with salinity of 38.2~38.3. The volume transport contains a seasonal cyle (~0.1Sv) similar to the observation shown by Candela (2001). CTD data from two sections across the Gulf of Cadiz at 8.5W and across the North Atlantic Ocean at 36N is then used to evaluate the performance of entrainment paraterization. The comparison shows that the simulated MOW is equilibrated at observed depth in the Gulf after entrainment. The MOW core is however slightly warmer and more saline than observed. Beyond the Cape of Vincent, the upper part of the simulated MOW spread northward, while the lower (major) part spreads toward west, agrees with the observed pattern in general. The spreading is not via a well-defined current, but via eddies.
Number 57 - Session 5
METEOROLOGICAL IMPACT OF REGIONAL OCEAN DATA ASSIMILATION: BENEFIT OF EDDY-RESOLVING SST ESTIMATES
M. Yamamoto1, and N. Hirose1
1 Research Institute for Applied Mechanics, Kyushu University, Kasuga, Japan
Abstract
It is difficult to detect high-resolution sea surface temperature (SST) in wintertime marginal seas under bad weather conditions. In order to overcome this problem, the SST is estimated by combining satellite measurements (SST and SSH) into an eddy-resolving ocean general circulation model (OGCM) (Hirose et al., J. Oceanogr., 2007). Recently, we applied the high-resolution SST to atmospheric mesoscale simulations in the wintertime Japan Sea area, and demonstrated the benefit of the eddy-resolving assimilated SST estimates on the basis of comparative experiments using the OGCM-assimilated SST (Exp. R) and optimum-interpolated SST (Exp. N). In Exp. R, mesoscale SST features resulting from ocean dynamics significantly influence the surface turbulent heat flux, precipitable water, and cloud water for a developing extratropical cyclone (Yamamoto & Hirose, Geophys. Res. Lett., 2007) and cold-air outbreak (Yamamoto & Hirose, Atmos. Sci. Lett., 2008). The results in Exp. R agree well with the observations in comparison with those in Exp. N.
In the present study, we investigate an impact of the high-resolution OGCM-assimilated SST on the long-term (monthly) mesoscale simulation for January, 2005. The persistent oceanic mesoscale eddies found in the assimilated SST play an important role as the heat and moisture sources in the regional weather system. In particular, the southeastward cold tongue around 39°N, 137°E influences the precipitation in the coastal area. The precipitation in Exp. R is smaller than that in Exp. N in a large part of the coastal area. Since the difference between the observed and modeled precipitations in Exp. R is 20 % smaller than that in Exp. N, the assimilated SST improves the meteorological simulation for the monthly precipitation in comparison with the interpolated SST.
Finally, we emphasize that our approach could be commonly applied to many marginal seas, not only the Japan Sea. The abovementioned meteorological applications of ocean data assimilation are expected to elucidate the influence of ocean mesoscale eddies on the weather system in the coastal areas, and to contribute to the improvement of regional weather forecasting.
Number 40 - Session 4
NON-STERIC EFFECT FROM THE GRACE MISSION ON THE ALTIMETRY DATA ASSIMILATION
C. Yan1, J. Zhu1
1 Institute of atmospheric Physics, Chinese Academy of Sciences, Beijing, 10029, China
Abstract
Up to now, the altimetric satellites have provided high-precision, high-resolution, and quasi-synoptic observations of sea surface height (SSH), representing the height of the sea level. The SSH includes the signal caused by temperature and salinity fluctuations (the steric effect) and non-steric barotropic and mass variations. The residual is interpreted as the non-steric variation after subtracting the steric variation from the altimeter SSH measurements. The relatively accurate non-steric variation is difficult to be measured until the launch of GRACE. Some studies have shown that the non-steric variation is significant from the view of the relative contribution to the whole sea level rise, with annual amplitudes between 7 and 9mm of water, and a maximum in late-summer, early-fall. However, in some studies about the altimteric data assimilation, the non-steric variation is neglected. Then what is the relative contribution of non-steric variation to the data assimilation? This study aims to evaluate the impact of non-steric part from the GRACE mission in the sea level rise on the altimetric data assimilation. An altimetric data assimilation scheme, which may retrieve vertical temperature and salinity profiles from one SSH measurement, is used by combining satellite altimetry (which measure steric + non-steric vatiation) with GRACE (non-stetric) in the data assimilation. The results are compared to that obtained using usual approach, i.e. the assimilation of satellite altimetry alone. The ARGO data is used to validate the impact of non-steric change. The RMSE (root-mean-square-error) of the retrieved temperature profile relative to ARGO is reduced by comparison with the usual approach. The heat content obtained by assimilating the combination of satellite altimetry and GRACE is closer to ARGO. The results show that in the area of large non-steric variation, the non-steric change impacts greatly on the altimetric data assimilation. That implies that it is necessary and important to consider the GRACE-derived mass variation in the large non-steric change area whether in the data assimilation or in the sea level rise study.
Number 74 - Session 4
AN OPTIMAL ENSEMBLE INTERPOLATION SCHEME FOR ASSIMILATION OF ARGO PROFILES INTO HYCOM
J. Zhu, J. Xie, C. Yan
Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
Abstract
An ensemble optimal interpolation scheme is presented for assimilating Argo temperature and salinity profile data into the HYCOM, a hybrid vertical coordinate ocean circulation model. Two related problems are addressed. The first problem is caused by the nature of the “layer model”. The ocean stratification is represented by the layer thickness and temperature (or salinity) in the HYCOM. The straightforward way to assimilate a pair of Argo temperature and salinity profiles T(z), S(z) is via an observation operator that transforms the model layer thicknesses, layer temperatures and layer salinities to T(z) and S(z) using some vertical interpolations. By doing so, the observation operator is a nonlinear function of model variables, and thus the linear Kalman update equation cannot deliver the optimal solution. As a consequence, after one assimilation step, the temperature and salinity profiles from the updated layer thicknesses, layer temperatures and layer salinities may not get closer to the observations than that before the assimilation. In this paper, the method proposed is based on Thacker and Esenkov (2002). First the Argo temperature and salinity profiles are used to obtain the “observations” of layer thicknesses and they are assimilated to adjust the model layer thickness and model velocity fields based on the covariance matrix that is formed from a set of ensemble. In the second step, the Argo temperature profiles are assimilated to adjust the model layer temperatures. In this step, the observation operator is based on the previously adjusted layer thicknesses. At the third step the model layer salinities are derived from the equation of state of sea water.
The second problem is caused by the sampling errors. The background error covariance matrix is calculated from a set of stationary ensembles that are model snapshots. The fake long-distance correlations in the vertical direction can be effectively reduced by a simple vertical-localization scheme in the vertical direction. Another sampling error source is the underestimate of the background error in the thermocline and below. An inflation scheme is applied to increase the ensemble spread in the thermocline and below based on the information of average “ensemble mean minus observation”. We also proved that the assimilation scheme keeps the conservation of volume, or the vertical sum of layer thicknesses is not changed by assimilating Argo observations. The assimilation scheme is illustrated and tested using Argo data from 2005 in the Pacific.
Number 4 - Session 4
A VARIATIONAL METHOD FOR PROCESSING RADIAL VELOCITY DATA ACQUIRED BY COASTAL RADARS
Max Yaremchuk1, Alexei Sentchev2
1 University of Hawaii, Honolulu, USA
2 Universit´e du Littoral, Wimereux, France
Abstract
High-frequency radars measure projections of surface velocity vectors on the directions of the radar beams. We consider retrieval of 2d velocity field from such observations using the variational approach. The interpolation problem is regularized by penalizing high-frequency variability of the vorticity and divergence fields. Regularization terms are inversely proportional to the local density of HFR observations, whereas their magnitude depends on the three scalar parameters retrieved from the statistical analysis of the radial data. Twin-data experiments are used to assess the method's skill and compare its performance with the open-boundary modal analysis (OMA) technique. It is shown that the variational method is more accurate when observations have gaps in spatial coverage and/or signal-to-noise ratio is larger than 2.5. At noise levels higher than 40% and with homogeneous data coverage both methods produce similar results.
Number 5 - Session 4
A method of successive corrections of the control subspace in the reduced-order 4dVar
Max Yaremchuk1, Dmitri Nechaev2 and Gleb Panteleev3
1 University of Hawaii, Honolulu, USA
2 University of Southern Mississippi, USA
3 University of Alaska, Fairbanks, USA
Abstract
A version of the reduced control space four-dimensional variational method (R4dVar) of data assimilation into numerical models is proposed. In contrast to the conventional 4dVar schemes, the method does not require development of the tangent linear and adjoint codes for implementation. The proposed R4dVar technique is based on minimization of the cost function in a sequence of low-dimensional subspaces of the control space. Performance of the method is demonstrated in a series of twin-data assimilation experiments into a highly nonlinear quasigeostrophic model utilized as a strong constraint. When the adjoint code is stable R4dVar's convergence rate is comparable to that of the standard 4dVar algorithm. In the presence of strong instabilities in the direct model, R4dVar works better than 4dVar whose performance is deteriorated due to the breakdown of the tangent linear approximation. Comparison of the 4dVar and R4dVar also shows that R4dVar becomes advantageous when observations are sparse and noisy.
Number 141 - Session 4
CHARACTERISTICS AND FORCING MECHANISMS OF EDDY VARIABILITY AROUND THE HAWAIIAN ISLANDS
S. Yoshida, B. Qiu, R. Lukas, Y. Jia, P. Hacker
University of Hawaii, Honolulu, USA
Abstract
The Navy Coastal Ocean Model (NCOM) and Hybrid Coordinate Ocean Model (HYCOM) are evaluated by comparing eddy variability with satellite sea surface height (SSH) anomaly data near the Hawaiian Islands. On the large scale, NCOM SSH anomaly produces patterns similar to those calculated from altimetric SSH anomaly data with two strong variability regions, one north of the island chain and one in the lee of the islands. The two high variability regions are located over the broad eastward mean flow of the North Pacific Current (NPC) and the narrow eastward flowing Hawaiian Lee Counter Current (HLCC). This study also examines forcing mechanisms of the dominant 90-day HLCC eddy variability, which extends westward from 158°W and has been observed in tide gauge records at Wake Island (Mitchum, 1995). The HLCC region has a strong north-south horizontal shear centered on the HLCC at 19°N, with the north-westward flowing Hawaiian Lee Current (HLC) to the north and westward flowing North Equatorial Current (NEC) to the south. Eddies are dominant features in the lee of the Big Islands of Hawaii, and are caused mostly by local wind stress curl and intrinsic oceanic instability processes.
The mean eddy kinetic energy (EKE) calculated from altimetric data has a maximum to the west of the Big Island. This pattern weakens to the west. The pattern is quantitatively different in the NCOM EKE or SSH variability structures where the high values are seen further to the west in the region of the HLCC. Wind stress curl has high values in the wake of the island of Maui and Hawaii (Calil et al, 2008). In order to estimate SSH anomalies resulting from the convergence/divergence of the anomalous surface Ekman pumping, we apply a simple Ekman pumping model over the strong wind stress curl region in the wake of islands using,
Where is g’ the reduced gravity (~0.03m/s2), ρ0 is the reference density and is the wind stress curl. The SSH anomaly calculated from the Ekman pumping model explains about 10 to 15 percent of the altimeter or NCOM SSH variability with no clear correlation pattern. 
Since the current system in the lee of island is horizontally sheared, it possesses the mean kinetic energy necessary for barotropic instability. In order to clarify the energy sources from mean to eddy variability, energetic analysis with the convergence of the Reynolds stress in the momentum equations is conducted. In addition to use of altimeter data, we also use shipboard ADCP and 3 years of mooring data from the Hawaii Ocean Timeseries (HOT) to study near shore mean current and eddy variability and its representation in NCOM and HYCOM. The figure shows NCOM SSH standard deviation with two high variability regions, one north of the island chain and one in the lee of the islands (unit is cm).
Number 106 - Session 4
North-Western Mediterranean circulation study: flow interaction between the Northern Current and the coastal circulation in the Gulf of Lion and the Côte d’Azur Region
Y. Ourmières1, A. Molcard1, B. Zakardjian1, P. Forget1, P. Fraunié1, Y. Barbin1, J. Gaggelli1, A.
Duchez1,2, C. Langlais3
1 LSEET, Université du Sud Toulon Var, 83957 La Garde, France
yann.ourmieres@lseet.univ-tln.fr
2 LEGI, BP53X, 38041 Grenoble, France
3 CSIRO, University of Tasmania, Australia
Abstract
Coastal areas are transition zones between the continents and the open sea. Interest in these areas comes along with the rapid evolution of environmental issues. Among them, the key role played by interactions between coastal and open sea circulation is pointed out. In this context, a North-Western Mediterranean circulation integrated study is developed, focusing on the interaction between the Northern Current (NC), a major actor of the circulation dynamics in this region, and coastal areas namely the Côte d'Azur and the Gulf of Lion. Short to medium scale processes are studied, mostly transient coastal eddies impact and NC meanders instabilities both affecting coast offshore water mass exchanges. This project makes use of high resolution modelling, HF-VHF radar surface currents monitoring, as well as in situ data and satellite images, for model assessment and eventually data assimilation. The numerical modelling is conducted using the NEMO-OPA9 code configured for the Gulf of Lion and the Côte d'Azur region, with a high-resolution of 1/64°.
The model is forced with realistic atmospheric fluxes (REMO, Majewski, D., 1991) and the open boundary conditions are provided by the NEMO-MED16 model (Alhammoud et al. 2005), configured for the Mediterranean basin at a 1/16° resolution. HF-VHF radars providing surface currents data are planned to be installed at key locations along the Mediterranean coast, in order to monitor the currents variability in the studied area matching the model domain. Long term monitoring is targeted, allowing comparisons, model validation and potentially model-data integration. Additional numerical work is also conducted in order to assess the impact of high-resolution altimetry (ALTIKA project) data assimilation into high-resolution models.
References:
Alhammoud B, Béranger K, Mortier L, Crépon M, Dekeyser Y (2005) Surface circulation of the Levantine Basin: comparison of model results with observations, PiO 66(2-4), 299-320.
Majewski, D., 1991: The Europa-Model of DW, ECMWF Seminar on numerical methods in Atmospheric Science, 2, 147-191.
Number 122 - Session 3
ONE-WAY NESTING OF HYCOM TO HYCOM
Luis Zamudio1, Patrick Hogan2, & Joseph Metzger2
1 Center for Ocean-Atmospheric Prediction Studies, Florida State University, USA
2 Naval Research Laboratory, Stennis Space Center, Mississippi, USA
Abstract
The existence of high-resolution regional nested numerical ocean models is in part due to the ability of these models to simulate short time and scale processes at relative low computational cost. These models are nested inside a lower-resolution global or basin scale models (i.e. outer models) that supply the boundary conditions for the nested inner models. The larger domain models have the ability to simulate a wide range of spatial and temporal signals that originate far away from the nested region and these may play a key role in the modulation of the circulation and sea level variability in the nested regions (e.g. coastally trapped waves and eddies).
Since HYCOM is becoming a community ocean model, then the publicly accessible (http://www.hycom.org/dataserver) daily outputs of operational global HYCOM can be used as boundary conditions for regional models. The coupling of global (or basin scale) to regional HYCOM is a one-way robust routine, which has been used to isolate dynamical processes at relative low computational cost. Hence, this study documents a systematic approach to nest HYCOM inside HYCOM, which can be used for potential HYCOM users, to select a priori (without repeating the sensitivity experimentation) the characteristics of the HYCOM nesting parameters.
Number 45 - Session 5
IMPACT OF INITIAL OCEAN SURFACE AND SUBSURFACE STATES ON 1997, 2002, AND 2006 EL NIÑO PREDICTIONS
Fei Zheng 1, and Jiang Zhu 2
1 International Center for Climate and Environment Science (ICCES), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
2 State Key Laboratory of Atmospheric Boundary Layer Physics and Atmospheric Chemistry (LAPC), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
Abstract
Ocean initial states play a crucial role on dominating the onset and developments of the El Niño events, but there are distinct differences for the development of El Niño due to the differing effects of the ocean surface and subsurface conditions. In this study, based on a ensemble prediction system (EPS), roles of the initial ocean surface and subsurface states on these three El Niño predictions are examined in a set of retrospective forecast experiments, respectively. Three sets of experiments were carried out. In the first set of experiments, only the SST data was assimilated to provide accurate surface initial conditions, while only sea level (SL) data was assimilated to provide the updated information of the temperature of the subsurface water entrained into the mixed-layer in the model. In the third set of experiments both SST and SL data were assimilated.
All of the hindcast experiments with three different initial conditions can successfully predict the three El Niño events before 12 months, and the hindcasts with assimilating both SST and SL data into the initial fields performed best (referred as the best hindcast hereafter). For the 1997 El Niño, the hindcast with only assimilating SST data was much closer to the best hindcast than that with only assimilating SL data. It indicated that a more accurate initial zonal SST gradient across the Pacific could effectively alter the westerly wind event, and induce a subsequent strong El Niño, the initial surface ocean state mainly contributed to promote favorable conditions for the explosive growth, while the initial subsurface state had little impact on the growth of the 1997 event. For the 2002 and 2006 El Niño events, the hindcast results with only assimilating SST data were more similar to the best hindcast results than those with only assimilating SL data in the first half of the two events, while the hindcast results with only assimilating SL data were more similar to the best hindcast results in the second half of the two events. These indicated that the accurate initial surface state could trigger the onset of the both El Niño events, and the more reasonable initial subsurface state contributed to improve the growth of the two events in the second half.
Keywords: El Niño, ensemble prediction system, initial surface state, initial subsurface state
(Last Updated: 30-10-2008)