An ensemble three-dimensional variational data assimilaiton system for the global ocean: sensitivity to the observation- and background-error variance formulation.

TitleAn ensemble three-dimensional variational data assimilaiton system for the global ocean: sensitivity to the observation- and background-error variance formulation.
Publication TypeMiscellaneous
Year of Publication2008
AuthorsDaget, N, Weaver, AT, M. Balmaseda, A
Secondary TitleTechnical Memorandum
Number562
Abstract

This paper presents an ensemble 3D-Var system that has been developed for global analysis with the OPA ocean general circulation model. The ensembles are created by perturbing the surface forcing fields and the observations used in the assimilation process. Cycled 3D-Var experiments over the period 1993-2000 are presented to test the sensitivity of the analyses to two flow-dependent formulations of the background-error standard deviations sigma^b for temperature and salinity. The first formulation is based on an empirical parameterization of sigma^b in terms of the vertical gradients of the background temperature and salinity fields, while the second formulation derives sigma^b from the spread of the ensemble of analyses. Comparing innovation statistics from the two sigma^b formulations shows that both formulations produce similar results below approximately 150m but the parameterized sigma^b produce slightly better results above this depth where statistical consistency checks indicate that the ensemble sigma^b are underestimated. The rate at which observational information is lost between cycles, however, is shown to be much reduced with the ensemble sigma^b, suggesting that the analyses produced with the ensemble sigma^b are in better balance than those produced with the parameterized sigma^b. Sea surface height anomalies in the northwest Atlantic and zonal velocities in the equatorial Pacific, which are fields not directly constrained by the observations, are clearly better with the ensemble sigma^b than with the parameterized sigma^b when compared to independent data. Areas for improving the ensemble method and for making better use of the ensemble information are discussed.