Impact of UAS Global Hawk Dropsonde Data on Tropical and Extratropical Cyclone Forecasts in 2016

• Published in: Weather and forecasting Vol. 33; no. 5; pp. 1121 - 1141
• Main Authors: Kren, A. C., Cucurull, L., Wang, H.
• Format: Journal Article
• Language: English
• Published: Boston American Meteorological Society 01.10.2018
• Subjects: Aircraft; Cyclones; Data; Data assimilation; Data collection; Dropsonde; Dropsonde observations; Dropsondes; Errors; Experiments; Extratropical cyclones; Hurricanes; Laboratories; Meteorological satellites; Operational hazards; Precipitation; Relative humidity; Satellite data; Satellite observation; Satellites; Sea level; Sea level pressure; Severe weather; Steering; Storms; Tropical climate; Tropical cyclones; Unmanned aerial vehicles; Weather; Weather forecasting; Florida; United States > US; North America
• ISSN: 0882-8156; 1520-0434
• DOI: 10.1175/WAF-D-18-0029.1

Abstract A preliminary investigation into the impact of dropsonde observations from the Global Hawk (GH) on tropical and extratropical forecasts is performed using the National Centers for Environmental Prediction (NCEP) Global Data Assimilation System (GDAS). Experiments are performed during high-impact weather events that were sampled as part of the NOAA Unmanned Aerial Systems (UAS) Sensing Hazards with Operational Unmanned Technology (SHOUT) field campaigns in 2016: 1) three extratropical systems in February 2016 and 2) Hurricanes Matthew and Nicole in the western Atlantic. For these events, the benefits of GH observations under a satellite data gap scenario are also investigated. It is found that the assimilation of GH dropsondes reduces the track error for both Matthew and Nicole; the improvements are as high as 20% beyond 60 h. Additionally, the localized dropsondes reduce global forecast track error for four tropical cyclones by up to 9%. Results are mixed under a satellite gap scenario, where only Hurricane Matthew is improved from assimilated dropsondes. The improved storm track is attributed to a better representation of the steering flow and atmospheric midlevel pattern. For all cases, dropsondes reduce the root-mean-square error in temperature, relative humidity, wind, and sea level pressure by 3%–8% out to 96 h. Additional benefits from GH dropsondes are obtained for precipitation, with higher skill scores over the southeastern United States versus control forecasts of up to 8%, as well as for low-level parameters important for severe weather prediction. The findings from this study are preliminary and, therefore, more cases are needed for statistical significance.