Targeted observation to improve tropical cyclone (TC) predictability is among one of the most important research and forecasting issues for TCs. In addition to the hurricane surveillance program conducted by the Hurricane Research Division (Aberson 2003) in the U.S., since 2003 the DOTSTAR (Dropwindsonde Observations for Typhoon Surveillance near the TAiwan Region) program has also successfully performed targeted observations for typhoons in the western North Pacific (Wu et al. 2005). The Japanese team has also planned to launch its targeted observation program (Typhoon Hunting 2008; TH08) in 2008. The joint research collaboration between DOTSTAR and TH08, along with TCS-08 supported by Office of Naval Research, is designed to play an important role in improving the TC predictability in the THORPEX-PARC (T-PARC) experiment in 2008.

To optimize the aircraft surveillance observations using dropwindsondes, targeted observing strategies have been developed and examined. The primary consideration in devising such strategies is to identify the sensitive areas in which the assimilation of targeted observations is expected to have the greatest influence in improving the numerical forecast, or minimizing the forecast error. Since 2003, four objective methods have been tested for operational surveillance missions in the environment of Atlantic hurricanes conducted by NOAA (Aberson 2003) and DOTSTAR (Wu et al. 2005). These products are derived from four distinct techniques, the ensemble Deep-Layer Mean (DLM) wind variance (Aberson et al. 2003), the ensemble-transform Kalman-filter (ETKF, Majumdar et al. 2002), the total-energy singular vector (TESV) technique (Peng and Reynolds 2006), and the Adjoint-Derived Sensitivity Steering Vector (ADSSV) (Wu et al. 2007). The above techniques have been applied in a limited capacity to identify locations for aircraft-borne dropwinsondes to be collected in the environment of the tropical cyclones.

To gain more physical insights into these targeted techniques, studies to compare and evaluate the techniques have been conducted by Majumdar et al. (2006), Etherton et al. (2006), and Reynolds et al. (2007). As a follow-up work, and to highlight the unique dynamic features that affect TC tracks in different regions, in this paper we compare six different targeted techniques based on 92 cases of two-day forecasts of the Northwest Pacific tropical cyclones in 2006. The six targeted methods are TESVs form ECMWF and NOGAPS, the TESV by EPS (Ensemble Prediction System) of JMA (Japan Meteorological Agency), the ETKF based on the multi-model ensemble members [ECMWF, NCEP and CMC (Canadian Meteorological Centre)], the DLM wind variance, and the ADSSV as described above. We note that, compared to the Atlantic region, the Northwest Pacific region has more numerous and more complicated dynamical systems affecting TC motion, such as mid-latitude troughs, the subtropical jet, the southwesterly monsoon and binary interactions. More analyses are ongoing to identify the similarity and the difference of all these different targeted methods and to interpret their meaning dynamically. Results from this work would not only provide better insights into the physics of the targeted techniques, but also offer very useful information to assist the future targeted observations, especially for the DOTSTAR and TH08 in T-PARC, 2008.