More than 85% of the observations routinely assimilated at the major operational global NWP centers around the world are satellite observations. Many recent improvements in the NWP forecast skills are due in part to the steadily increasing in use of the satellite observations in data assimilation. However, it is interesting to note that not all satellite observations have the same impact on reducing the 24-hr forecast errors - even for the same sensor within the same data assimilation window. For example, results from recent 4D-Var observation impact studies suggest that satellite observations near end of data assimilation window generally have a larger beneficial impact on the reduction of 24-hr forecast errors than the ones near the beginning of the data assimilation window. It is important to understand why this is the case, so that the operational data delivery and assimilation systems may be improved. In this study, we use the US Navy's operational 4D-Var global atmospheric data assimilation system, NAVDAS-AR* to examine various possible explanations for these recent findings. We first present the overall impact of assimilated satellite observations on the 24-hr operational NOGAPS forecasts. The temporal impact of assimilated individual observation on the analysis is then shown by results from assimilating a “single” AMSU-A observation using NAVDAS-AR. Results from assimilating “single” observations from other sensors are also presented, followed by concluding remarks.

*NAVDAS-AR: NRL Atmospheric Variational Data Assimilation System - Accelerated Representer