A multivariate balance operator for variational ocean data assimilation

• DOI: 10.1256/qj.05.119
• Publication Date: 2006-08-17T22:42:57Z
• Authors: A. T. Weaver; C. Deltel; E. Machu; S. Ricci; N. Daget
• Keywords: Operator (biology); Assimilation (phonology); Uncorrelated

Abstract It is common in meteorological applications of variational assimilation to specify the error covariances model background state implicitly via a transformation from space where variables are highly correlated control can be considered approximately uncorrelated. An important part this balance operator which effectively establishes multivariate component covariances. The use technique ocean data less common. This paper describes that used variable for oceanographic three‐ and four‐dimensional assimilation. proposed has been implemented an incremental system global general‐circulation model. Evidence explain significant percentage background‐error variance presented. analysis structures implied by illustrated using single‐observation experiments. Copyright © 2005 Royal Meteorological Society