Application of the variational data-assimilation technique for the measurements of the transport through the Bering Strait.
G. Panteleev, M.Yaremchuk, D. Nechaev,
Collaborative project between IARC, IPRC and USM.
Goals
1) To optimized the sampling strategy in the Northern Bering Sea (NBS).
2) To build data assimilation system for operational hindcast/forecast of the local circulation and the Bering Strait transport during the summer period through the assimilation of the satellite altimetry data.
Optimization of the mooring array in the Bering Strait region.
Final results: manuscript was accepted for publication in Atmosphere Ocean
The *.fdf file can be douwloaded here
Final results: manuscript was accepted for publication in Atmosphere Ocean
Manuscript can be viewed here
Abstract
The problem of the optimal sampling strategy for moored current velocity
observations in the Northern Bering Sea (NBS) is addressed. We
analyze dynamically induced correlations in the NBS currents
and conduct their sensitivity analysis to optimize positions of a limited
number of moorings. Optimization of the sampling strategy is performed with
respect to robustness of the reconstruction
of the NBS circulation with a particular emphasis on the accurate monitoring
of the mean Bering Strait transport (BST). Computations reveal four major regions
in the NBS basin that are highly correlated with the BST. Apart from
the regions within the Bering Strait itself, they include the Anadyr Strait
and a region 100 km south of the Cape of Prince of Wales (CPW). Results of
the sensitivity analysis are tested in the framework of twin data experiments
with the quasi-stationary and oscillatory background circulations.
Data-assimilation technique
To find the optimal solution of the model we performed strong constraint minimization of the cost
function measuring the distance between the model solution and data on the space of the control variables
[Le Dimet and Talagrand}, 1986]. Control variables include the initial conditions, the model
field values required to specify the open boundary conditions, and the surface fluxes of momentum,
heat, and salt [Nechaev et al.}, 2005].
Forward and adjoint models
The primitive equation model utilized in this study was successfully used for the reconstruction of
the evolution of summer climatological circulation in the Barents Sea [Panteleev et al., 2006] and
for the hindcast of the circulation in the Tsushima Strait [Nechaev et al., 2005] for a period
of 10 months. The forward model is a modification of the C-grid, z-coordinate OGCM designed by [Madec
et al., 1999]. The model is implicit both for barotropic and baroclinic modes permitting model
runs with relatively large time steps. The adjoint code was built analytically.
Preliminary results
The primitive equation model utilized in this study was successfully used for the reconstruction of the evolution of summer climatological circulation in the Barents Sea [Panteleev et al., 2006] and for the hindcast of the circulation in the Tsushima Strait [Nechaev et al., 2005] for a period of 10 months. The forward model is a modification of the C-grid, z-coordinate OGCM designed by [Madec et al., 1999]. The model is implicit both for barotropic and baroclinic modes permitting model runs with relatively large time steps. The adjoint code was built analytically.
We completed several twin experiments and assimilated data from four-five moorings. The data assimilation model with horizontal resolution of 6km was configured for the domain shown by the white line in Figure 4. The gridded SSH data were linearly interpolated into the model grid and assimilated. Because in situ temperature/salinity observations are absent, we utilized the climatological temperature/ salinity distributions in Amukta Strait.
Figure 1. Slowly variable climatological circulation in the Bering Strait region. Blue arrows – model results.
In the first experiment we assimilated the data from four moorings deployed in the eastern and western parts of the Bering Strait. The assimilation of the velocity data from these moorings allows to reconstruct the Bering Strait transport with the error of 0.002-0.005. However, the circulation pattern (Figure 2) differs significantly from the "true climatological" pattern in the region (Figure 1)
Figure 2. Climatological circulation in the Bering Strait reconstructed through the assimilation of velocity data from four moorings in the Bering Strait (red asterisk).
In the second experiment velocity data from two moorings deployed in the eastern part of the Bering Strait and three additional moorings deployed in the northern part of the Bering Sea (Figure 3). The assimilation of the velocity data from these moorings allows to reconstruct the Bering Strait transport with the error of 0.03-0.05. The circulation pattern in the northern part of the Bering Sea was also reconstructed with very high accuracy (Figure 3).
Figure 3. Climatological circulation in the Bering Strait reconstructed through the assimilation of velocity data from five moorings in the Bering Strait (red asterisk).
Operational estimates of the Bering Strait transport through the 4Dvar data assimilation. Preliminary results
Climatological circulation in the Bering Sea.
Reconstruction of the seasonally mean climatological circulations has been performed using the technique developed for the reconstruction of the mean BS climatological circulation (see for example http://www.frontier.iarc.uaf.edu/~gleb/bering_sea/bering_sea.html). We assimilated drifter, temperature/salinity, and other observations available for a particular season of the year. This approach, at a resolution of approximately 25 km, was applied to the study of BS summer circulation by Panteleev et al. (2006b; see also http://people.iarc.uaf.edu/~gleb/bering_sea/bering_sea.html). Recently, we decreased the model resolution to 18km and reconstructed the circulation in the entire BS. That allowed improved representation of the Kamchatka, Near, and Builder passes in the model. The derived estimates of the mean climatological transports through these passes are -28Sv, +13Sv, and +3Sv, respectively. The transports through other Aleutian passes also differ from previous estimates, but insufficient model resolution does not allow us to make reliable estimates at this time (Panteleev et al., 2007a). The preliminary results are shown at Figure 4.
Figure 4. Optimized mean climatological SSH at a resolution of 18 km.
Figure 5. Mean 1992-2001 dynamic ocean topography in the BS according to Rio and Hernandez (2004).
Satellite altimetry data.
In this project we will use sea level anomaly data distributed by Aviso (http://www.aviso.oceanobs.com/) and originally measured along the satellite tracks (Figure 1). Along-track altimeter records were processed to eliminate such errors as inverse atmospheric pressure, wet troposphere, and tides, and to retain the signal that is due to the geostrophic ocean dynamic. An "anomaly" is a deviation from the seven-year mean (nominally, 1992-1999). Along-track data of TOPEX/Poseidon (main and secondary orbits), Jason-1, ERS-1 and -2, Envisat, and GFO missions are available to the public at http://www.aviso.oceanobs.com/html/donnees/produits/hauteurs/global/
Figure 6. TOPEX/Poseidon (TP) altimeter data coverage in the Bering Sea. The density of the satellite tracks in the Bering Sea significantly increases (Cherniawsky et al., 2005) moving northward