The impact of an observation upon analysis is defined as the difference between the analysis using the observation and the analysis not using the observation. The statistical interpolation method of objective analysis enables the efficient computation of such observational impacts. Impact calculations provide a rigorous quantitative method of identifying the most influential observations, both upon individual analyses, and in an ensemble sense. They also enable the identification of data that, although rejected by quality control, would have had a large impact if they had been used. This report describes the computational algorithm for impact assessment, and presents examples of such assessments made within the framework of the Australian Bureau of Meteorology's global assimilation system. The examples include the identification of the most influential Southern Hemisphere sea-level pressure data, a ranking of Australian upper air stations with respect to analysis impact, and the identification of important quality-control decisions.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below