The Impact of Assimilating Enhanced AMVS, Dropsonding Observations, and HIRAD Data on Analyses and Forecasts of Hurricanes Edouard and Gonzalo (2014)

Thursday, 21 April 2016
Plaza Grand Ballroom (The Condado Hilton Plaza)
Shixuan Zhang, University of Utah, Salt Lake City, UT; and Z. Pu, C. S. Velden, and D. J. Cecil

Atmospheric wind data is essential for hurricane prediction. In this study, three types of the wind data products, including the enhanced atmospheric motion vectors (AMVs), derived from multiple satellites by CIMSS, University of Wisconsin, the dropsonding observations, and the wind retrievals from the Hurricane Imaging Radiometer (HIRAD), obtained from the NASA Hurricane and Severe Storm Sentinel (HS3) Global Hawk, are assimilated into the NCEP Hurricane Weather Research and Forecasting (HWRF) model to examine the impacts of wind data on analyses and forecasts of hurricanes. The NCEP gridpoint statistical interpolation (GSI) data assimilation system at a configuration of the ensemble Kalman filter – variational hybrid is employed. Two major hurricanes, Edouard and Gonzalo during the 2014 hurricane season are used for case studies. First, the CIMSS AMVs data products have been assimilated into the HWRF model for Hurricanes Edouard and Gonzalo (2014) during their rapid intensification and mature stages. Various sensitivity studies have been conducted. Results show positive impacts of CIMSS AMVs on the intensity forecasts in the mature phase of Edouard and rapid intensification phase of Gonzalo. Specifically, the intensity (in terms of minimum sea level pressure and maximum surface wind speed) errors have been reduced significantly in the whole period of 72 h forecast during the rapid intensification of Hurricane Gonzalo, while a moderate reduction of intensity errors have been shown in the first 24 h forecast during the mature stage of hurricane Edouard. Further analyses show that those improvements in the intensity forecasts can be partially attributed to the reduction of imbalances in the initial conditions due to the assimilation of CIMSS AMVs. However, the lack of mass field observations (e.g., surface pressure, temperature and specific humidity) can degrade the benefit of the CIMSS AMV data assimilation. For instance, the impact of CIMSS AMVS is nearly neutral during the rapid intensification stage of hurricane Edouard and mature stage of hurricane Gonzalo. In addition, the HS3 dropsonding data have been assimilated into the HWRF model for Hurricane Edouard (2014) during its mature stages. Results from several sensitivity experiments show moderate positive impacts on both track and intensity forecasts. Meanwhile, compared with non-cycled data assimilation, the cycled data assimilation results in improved analyses and forecasts. In addition, assimilation of wind, temperature and moisture all together leads to better track and intensity forecasts than the assimilation of wind and temperature only does. Additional experiments are being performed to examine the impact of HIRAD data products on HWRF forecasts of Hurricanes Edouard and Gonzalo. An integrated data assimilation experiments is also being performed. Relative impacts of different wind data will be compared and the results will be presented during the conference.
- Indicates paper has been withdrawn from meeting
- Submission entered in competition