Adjoint-Free 4D Variational Data Assimilation Into Regional Models
Document Type
Book Chapter
Publication Date
1-1-2016
Department
Marine Science
School
Ocean Science and Engineering
Abstract
The ongoing trend towards parallelization in computer technologies propels ensemble methods toward the forefront of data assimilation studies in geophysics. Of particular interest are ensemble techniques which do not require the development of tangent linear numerical models and their adjoints for optimization. These “adjoint-free” methods detect effective search directions for optimization through direct perturbation of the numerical model across carefully chosen sets of states.Optimization proceeds by minimizing the cost functionwithin the sequence of subspaces spanned by these perturbations. In this chapter, an adjoint-free variational technique (a4dVar) is described and demonstrated in an application estimating initial conditions of two numerical models: the Navy Coastal Ocean Model (NCOM), and the surface wave model (WAM). It is shown that a4dVar is capable of providing forecast skill similar to that of conventional 4dVar at comparable computational expense while being less susceptible to excitation of ageostrophic modes that are not supported by observations. Prospects of further development of the a4dVar methods are discussed.
Publication Title
Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications (Vol. III)
First Page
83
Last Page
114
Recommended Citation
Yaremchuk, M.,
Martin, P.,
Panteleev, G.,
Beattie, C.,
Koch, A.
(2016). Adjoint-Free 4D Variational Data Assimilation Into Regional Models. Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications (Vol. III), 83-114.
Available at: https://aquila.usm.edu/fac_pubs/19570
COinS