The Australian Bureau of Meteorology’s Global Assimilation and Prediction system (GASP) has been progressively upgraded in recent years. A major enhancement was the use of multivariate statistical interpolation (MVSI) instead of the previous univariate (UVSI) method of objective analysis. The methodological differences between MVSI and its predecessor, and the corresponding enhanced data monitoring and quality control procedures, are described. The data assimilation properties of the MVSI system are demonstrated to be superior to the UVSI by: (a) better background fields in the assimilation cycle, (b) smaller adjustments to analysed fields during prediction model initialisation, and (c) a better fit of the initialised fields to the observations. Parallel forecast trials over several periods in different seasons indicate a systematic improvement in medium-range forecast skill when using MVSI analyses. The prediction model initialisation has also been upgraded by application of a non-linear normal mode procedure to analysed increments from a background field, rather than to full fields. This incremental implementation is shown to produce: (a) an improved representation of the semidiurnal tidal mode, (b) improved tropical background fields in the assimilation cycle, and (c) a useful reduction in bias in low latitude medium-range forecasts. Impacts of resolution enhancements, forecast verification comparisons with global models from other centres, and forecast case studies are presented in an accompanying paper.
Seaman et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: