Development and Implementation of a 4Dvar Data Assimilation system for the Hindcast/Forecast of Circulation in the Gulf of Alaska, Bering, Chukchi, East Siberian, and Beaufort Seas, and the Arctic Ocean

Researcher: Gleb Panteleev
Funding Source: JAMSTEC-70%, NSF-15%, NPRB -15%
Collaborators: D. Nechaev (University of Southern Mississippi), A. Proshutinsky (Woods Hole Oceanographic Institute), T. Kikuchi (Jamstec, Japan), M. Yaremchuk (International Pacific Research Center), N. Maximenko (International Pacific Research Center), R. Woodgate (Applied Physical Laboratory, University of Washington), J. Zhang (Applied Physical Laboratory, University of Washington), P. Stabeno (Pacific Marine Environmental Laboratory, NOAA), M. Ikeda (University of Hokkaido, Sapporo, Japan), E. Carmack (IOS, Canada)

The project goals are:

  1. Development of new 4Dvar data assimilation tools: ensemble-based 4Dvar, conventional (adjoint) 4Dvar based on Semi Implicit-Ocean Model (SIOM) and Princeton Ocean Model (POM).
  2. Reanalysis, hindcast, and forecast (Model-data synthesis) of circulation in the Bering, Chukchi, and East Siberian seas.

The approach is to use conventional adjoint-based 4Dvar data assimilation, ensemble-based 4Dvar.

The following results are anticipated:

  1. Development of a hydrophysical climatological atlas of the Bering Sea, and a historical reanalysis of circulation in the Arctic Ocean for several periods. These products will include dynamically balanced ocean states (temperature, salinity, velocity, and ext.).
  2. Reconstruction of the Reference Sea Surface Height (RSSH) in the Bering, Chukchi, and East Siberian seas. Obtaining a reliable estimate of the RSSH will allow us to use the satellite altimeter data from ERS-2 and Envisat satellite missions. These data will allow reconstruction and analysis of circulation from 1992 to the present.
  3. Development of the optimal sampling strategy in the Bering and Chukchi seas.
  4. Hindcast/Forecast of flow through the major Aleutian passes.

Development and application of the 4Dvar data assimilation system is the natural and the more effective way of model-data synthesis. It allows us to combine all available data with dynamical constraints of the existing physical models and obtain an optimal and dynamically balanced estimate of the ocean state.