Impacts of Adding/removing Satellite and Dropsonde Data On Winter Storm Forecast Accuracy Over The United States

Abstract

In the event of a gap in satellite data, Unmanned Aircraft Systems (UAS) may be useful to partially mitigate a loss in accuracy when forecasting high-impact weather events. We conduct Observing System Simulation Experiments (OSSEs) to investigate the impacts of satellite data and UAS data on the forecast accuracy of two winter storms. The NCEP Global Forecast System (GFS) is initialized with and without simulated data from numerous NOAA and NASA satellites and UAS dropsondes simulated from the ECWMF T511 Nature Run. Forecast energy error at 3-day lead times increases by an average of about 10% when nine satellites are removed, and this increase in error is mitigated when UAS data over a large portion of the Pacific Ocean is added. We investigate the impacts of removing specific satellites, and adding UAS data over smaller regions/flight paths on forecast accuracy as well as the causes of forecast errors.