A well known bias in satellite feature-tracked winds is their underestimate of wind speed maxima, especially in the upper troposphere. On the other hand, aircraft wind observations provide more accurate wind measurements and have been found to be of great importance in data assimilation in regions where they are available in sufficient numbers. The objective of this research is to optimize the use of both satellite-derived winds and aircraft winds in the NRL Atmospheric Variational Data Assimilation System (NAVDAS) in order to improve forecasts by the Navy Operational Global Atmospheric Prediction System (NOGAPS).

NOGAPS has at times suffered from relatively large forecast errors over North America that can be traced back to analysis errors in sensitive areas of the atmosphere over the North Pacific. Under these conditions, even very small analysis errors rapidly amplify and expand. The research, therefore, focuses on case studies of events in the North Pacific that exhibit sensitivity to the initial conditions. Results will be presented from experiments in which NAVDAS (1) uses all available data, (2) uses aircraft winds but not satellite winds, (3) uses satellite winds but not aircraft winds, and (4) uses both aircraft and satellite winds, but exclude the latter if aircraft winds are available in sufficient number. The fourth strategy is similar to that implemented by the Japan Meteorological Agency in an effort to better depict strong wind speed maxima.