Application of surface pressure measurements of O-2-band differential absorption radar system in three-dimensional data assimilation on hurricane: Part II - A quasi-observational study

摘要

This is the second part on assessing the impacts of assimilating various distributions of sea-level pressure (SLP) on hurricane simulations, using the Weather and Research Forecast (WRF) three dimensional variational data assimilation system (3DVAR). One key purpose of this series of study is to explore the potential of using remotely sensed sea surface barometric data from O-2-band differential absorption radar system currently under development for server weather including hurricane forecasts. In this part II we further validate the conclusions of observational system simulation experiments (OSSEs) in the part I using observed SLP for three hurricanes that passed over the Florida peninsula. Three SLP patterns are tested again, including all available data near the Florida peninsula, and a band of observations either through the center or tangent to the hurricane position. Before the assimilation, a vortex SLP reconstruction technique is employed for the use of observed SLP as discussed in the part I. In agreement with the results from OSSEs, the performance of assimilating SLP is enhanced for the two hurricanes with stronger initial minimum SLP, leading to a significant improvement in the track and position relative to the control where no data are assimilated. On the other hand, however, the improvement in the hurricane intensity is generally limited to the first 24-48 h of integration, while a high resolution nested domain simulation, along with assimilation of SLP in the coarse domain, shows more profound improvement in the intensity. A diagnostic analysis of the potential vorticity suggests that the improved track forecasts are attributed to the combined effects of adjusting the steering wind fields in a consistent manner with having a deeper vortex, and the associated changes in the convective activity. (C) 2014 Published by Elsevier Ltd.