Impact of Global Hawk Dropsonde Data Assimilated in the NCEP GFS Model during SHOUT: Hurricane Matthew and an Extratropical Storm over Alaska in 2016

The impact of observations from the Global Hawk (GH) on short-range weather forecasts is investigated using the 2015 operational version of the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS) model. Experiments are performed on two high-impact weather events, which took place as part of the Sensing Hazards with Operational Unmanned Technology (SHOUT) campaign in 2016: (1) Hurricane Matthew in the western Atlantic Ocean on 5 October 2016 during the Hurricane Rapid Response (HRR) mission and (2) an extratropical storm in the central North Pacific on 21-22 February 2016 during the El Niño Rapid Response (ENRR) mission. Additional experiments are performed to examine the benefits of GH observations under a satellite data gap scenario. Overall, the assimilation of GH dropsondes increases weather forecast skill and reduces forecast root-mean-squared error in the region of interest for both storms. For Hurricane Matthew, the reduction in root-mean-squared error is ~ 8% for the 500-hPa geopotential heights and sea-level pressure and this impact is roughly double than that in the extratropical storm. Slight positive improvements are also found for vector winds, temperature, and relative humidity. Of great importance is the influence of the dropsondes on storm track, wind, intensity, and precipitation errors. For Hurricane Matthew, the assimilation of dropsondes leads to a northward shift of the accumulated precipitation, as well as an improvement in the Equitable Threat Score between 16 and 22% in the 72 to 96-hour lead time. These benefits are tied to an improved storm track for Matthew, with the assimilation of GH dropsondes. Furthermore, GH dropsondes are found to partially mitigate a possible gap in satellite data.