Improving Vortex Initialization and Hurricane Forecasting Through 3dEnVar and 4dEnVar Hybrid Data Assimilation Methods

Prof. Zhaoxia Pu
University of Utah, USA

10:00, 2 July, 2019
3#1118

Abstract

This presentation summarizes the recent research efforts from the speaker’s research on improving the vortex initialization using the state-of-the-art numerical weather prediction model and NCEP GSI-based ensemble-variational hybrid data assimilation systems (e.g, both 3dEnVar and 4dEnVar). A series of results will be presented to demonstrate the effectiveness of various data assimilation configurations in improving vortex initialization and numerical prediction of track and intensity of hurricanes. Specifically, results are summarized in the following aspects: 1) the influence of enhanced background error covariance terms; 2) 3dEnVar versus 4dEnVar; 3) assimilation of the inner core tail Doppler radar (TDR) radial velocity observations; 4) assimilation of GPM satellite radiances and satellite-derived atmospheric motion vectors (AMVs); Problems, challenges, recent progress in improving hurricane vortex initialization will be discussed.