The impact of T‐TREC‐retrieved wind and radial velocity data assimilation using EnKF and effects of assimilation window on the analysis and prediction of Typhoon Jangmi (2008)

Abstract

AbstractThis study examines the relative impact of assimilating T‐TREC‐retrieved winds (VTREC) versus radial velocity (Vr) on the analysis and forecast of Typhoon Jangmi (2008) using an ensemble Kalman filter (EnKF). The VTREC and Vr data at 30 min intervals are assimilated into the ARPS model at 3 km grid spacing over four different assimilation windows that cover, respectively, 0000–0200, 0200–0400, 0400–0600, and 0000–0600 UTC, 28 September 2008. The assimilation of VTREC data produces better analyses of the typhoon structure and intensity than the assimilation of Vr data during the earlier assimilation windows, but during the later assimilation windows when the coverage of Vr data on the typhoon from four Doppler radars is much improved, the assimilation of Vr outperforms VTREC data. The combination of VTREC and Vr data, either by assimilating both VTREC and Vr data in all cycles or by assimilating VTREC in the first cycle and Vr in the remaining cycles (labeled VTFVR), further improves the analyses of the typhoon structure and intensity compared to assimilating VTREC or Vr data alone. Quantitative verifications of 24 h forecasts of the typhoon show that the VTFVR assimilation experiments produces forecasts that best match the best track data and also have the highest precipitation prediction skills. The track forecast errors in experiment that assimilate VTREC data through the later cycles are the largest. The behaviors are discussed based on the coverage, information content, and accuracy of the various forms of data.