Two Methods of Global Data Assimilation.
Abstract
An optimum interpolation (OI) method and a method that fits observations to basis functions are described. The selected basis functions are the normal modes of the linearized primitive equations. In both methods the assimilation system consists of the analysis procedure, a non-linear normal mode initialization, and the AFGL global spectral NWP model. Both methods were tested using FGGE analyses for 7 Jan 79 OOGMT as the starting point. Successive 12 h assimilations were conducted for ten cycles for the OI and five cycles for the fitting procedure. For the OI method results showed a slight increase in error with time when compared with observations, partly attributable to the 12 h forecast length, which was excessive for the forecast error standard deviation prescribed. The fitting procedure showed significantly stronger error increases with time, due in part to faulty conversion of heights to temperature with the hydrostatic equation and to possible over smoothing by the procedure. Keywords include: Optimum interpolation; Normal modes; Data assimilation; Objective analysis; Initialization.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1984
- Accession Number
- ADA155981
Entities
People
- Craig J. Johnson
- D. C. Norquist
- I. Halberstam
- S. L. Tung