Report on the THORPEX Pacific Asian Regional Campaign and Affiliated Programs

Patrick A. Harr and Russell L. Elsberry, Naval Postgraduate School, Monterey, CA USA

Tetsuo Nakazawa, Japan Meteorological Agency-Meteorology Research Institute, Tsukuba, Japan

The THORPEX Pacific Asian Regional Campaign (T-PARC) is a multi-national field campaign that addresses the shorter-range dynamics and forecast skill of high-impact weather events in one region (Eastern Asian and the western North Pacific) and the downstream impact on the medium-range dynamics and forecast skill of another region (in particular, the eastern North Pacific and North America). Although many significant weather events occur over eastern Asia and the western North Pacific, the focus of T-PARC is on various aspects of typhoon activity, which include formation, intensification, structure change, motion, and extratropical transition. Because of the significant impact of typhoon activity on the region of eastern Asia and the western North Pacific, T-PARC is comprised of several affiliated programs. These programs and their national sponsor include:

Tropical Cyclone Structure-2008 (TCS-08) [United States];
Typhoon Hunter-2008 (TH-08) [Japan];
Predictability and Observation Experiment (PROBEX) [South Korea];
Tibetan Plateau Experiment [China];
The South China Sea Experiment [China];
Dropsonde Observations for Typhoon Surveillance near the Taiwan Region (DOTSTAR) [Taiwan].

The combination of observational platforms and collaborative experiments is set such that the experimental design for T-PARC addresses three primary components: (1) A tropical measurement strategy is designed to examine circulations of the tropical western North Pacific monsoon environment as they relate to tropical cyclone formation, tropical cyclone intensification, and tropical cyclone structure change. (2) The measurement strategy for the extratropical transition (ET) and downstream impacts is based on the poleward movement of a decaying tropical cyclone and the resulting intense cyclogenesis that results from its interaction with the midlatitude circulation. (3) The third measurement strategy focuses on identification of regions in which extra observations may reduce numerical forecast error growth. In T-PARC, the targeted observations were aimed primarily at reducing errors associated with forecasts of tropical cyclone track over the western North Pacific.

Field Program Summary

During the field program period of 1 August – 4 October 2008, four typhoon, four tropical storms, and four tropical depressions occurred over the western North Pacific. In addition to these circulations for which official forecasts were issued by the Regional Specialized Meteorological Center (RSMC, Tokyo), a separate set of 51 tropical circulation systems were identified by the T-PARC/TCS-08 research team. These systems were defined as tropical cloud clusters, low-level waves, or subtropical circulations of interest to the project.

During the period, 500 aircraft mission hours were flown during a total of 76 missions. The missions were flown into 4 typhoons, one tropical depression, and 5 tropical cloud clusters. Overall, 72% of the flight missions were flown into the four typhoons. A significant number of flight hours were dedicated to each objective listed above. The final quality-controlled T-PARC data set is nearly complete. Observation data that had never before been obtained in the environment of the western North Pacific include dropwindsondes from 13 driftsondes, the ELectra DOppler RAdar (ELDORA) radar, Twin Otter Doppler Wind Lidar (TWODL) on the NRL P-3, the Stepped Frequency Microwave Radiometer (SFMR) on the USAF 53rd Weather Squadron WC-130J, the DLR Doppler Wind Lidar, and the DLR Differential Absorption Lidar (DIAL). In addition to the in situ observations, comprehensive sets of remotely-sensed observations have been archived. Finally, a complete set of operational and experimental deterministic and ensemble numerical analyses and forecasts have been archived from a variety of operational forecast centers.

Early Stage Tropical Cyclone Objectives

An early TCS08 result is the validation of two subjective and four objective satellite-based tropical cyclone estimation methods with in situ aircraft observations in four typhoons (provided by C. Velden): (i) Significant spread was found in the subjective estimates among the three operational agencies and five independent analysts. (ii) Objective techniques based on microwave and on infrared imagery are very competitive with the subjective techniques. (iii) The satellite consensus (SATCON) improves the accuracy over all methods.

A unique set of oceanographic observations was obtained in TCS-08 from two deployments of drifting buoys in advance of two typhoons—including the first-ever deployment in advance of a Category 5 tropical cyclone. Since the second typhoon followed the first typhoon within a week, the second deployment resulted in about 20 drifters to observe the ocean response. Several research groups are studying this data set as part of the planning for a large ocean response to typhoons field experiment during 2010 called ITOP. The follow-on TCS08 field experiment called TCS10 will also be conducted in conjunction with the ITOP.

As one example of several research modeling efforts related to the TCS08/T-PARC experiment, the NRL—Monterey regional model COAMPS-TC has been revised and improved based on the TCS08/T-PARC cases. An improved vortex initialization technique, plus revisions in the shallow convection, microphysics, and boundary layer representations, have led to a new COAMPS-TC version that is in beta test for western North Pacific implementation during 2009, and participation in the Hurricane Forecast Improvement Program in the Atlantic.

The performance of the ECMWF, UKMO, GFS, and NOGAPS models in predicting tropical cyclone formation has been evaluated (i) An experienced analyst is able to define individual model characteristics and tendencies related to the pre-tropical cyclone seedling to tropical cyclone transition. (ii) When all four global model forecasts are in agreement as to position and evolution, high confidence can be given to the prediction scenario with few false alarms. (iii) The consensus technique was most successful for those seedlings that will later become tropical storms or typhoons, but was not successful for weaker systems.

For the first-time, the WC-130J deployed from 31,000 feet dropwindsondes and air-deployed expendable bathythermographs, and still could make surface wind estimates from the SFMR. This set of aircraft observations in combination with geostationary and polar-orbiting satellite observations will be used by many research teams for years.

Targeted observations during T-PARC and TCS08

As part of the T-PARC/TCS08 objective to investigate the entire life cycle of a tropical cyclone, improvements to the forecasts of tropical cyclone motion were sought by obtaining targeted observations in regions with high sensitivity. Each targeted observation case was selected two days prior to the targeted observation time, and the final flight paths were chosen one day earlier, based on the targeted observation guidance. Several common verification regions, such as for the Guam, Taiwan, and Japan areas, had been agreed prior to the experiment among the nine targeting groups. During the field phase of T-PARC/TCS08, a targeted observation team identified potential opportunities to collect targeted observations. The entire team (members not in Monterey participated via a web-based telecommunication system) then met to discuss these potential targeting opportunities, timelines, and flight tracks.

Although it is still too early to discuss in detail the impacts of these targeted observations on typhoon track forecasts some preliminary results are becoming available. For example, the JMA group is performing observing system experiments for Typhoons Sinlaku and Jangmi. For Sinlaku, the forecast track with special observations has a northward movement during an earlier period, and then northwestward later. However, the track without special observations has westward movement. When bogus data are removed, the JMA best track and the track without special observations are very similar before the recurvature.

Opposite results are obtained for Sinlaku by researchers using the NCEP GFS model. In their cases, the track with dropsondes is closer to the best track. The track without dropsondes tends to have a northward orientation. One of the reasons for such a difference in the track forecast may come from the intensification of the sub-tropical ridge by adding the special targeted observations north of the tropical cyclone. The quality of the dropwindsonde data near the center needs to be examined for observation representativeness in high-wind conditions.

Preliminary results indicate that targeted observations during the field phase of T-PARC demonstrated positive impacts on typhoon track forecasts, although in some cases the overall forecast show some degradation. Further analysis is required to better understand how the targeted observations impact the track forecasts in the different modeling systems.

Extratropical Transition and Downstream Impacts

During the field phase of T-PARC, several tropical disturbances moved poleward to undergo a transition into the midlatitudes. The character of these disturbances included a weak circulation associated with widespread deep convection, a midget tropical cyclone, a typhoon, and a super typhoon. Corresponding to the variety of tropical disturbances was a wide range of forecast and actual structural changes and downstream developments, which provide a broad spectrum of forcing and downstream impacts to be investigated.

A three-aircraft mission was conducted into what was thought to be a weakening TY Sinlaku. However, during the mission convection increased rapidly and the ELDORA observations identify some of the deepest convection observed during the T-PARC/TCS-08 period. Sinlaku re-intensified to a typhoon and then began ET 24 h later. This period represents a source of a large amount of variability in deterministic and ensemble forecast fields from a variety of operational forecast centers. The measurement strategy associated with the case of TY Sinlaku was defined to observe the important physical mechanisms associated with the ET process that influences predictability downstream. TY Jangmi also provided a source of reduced predictability, but this was related to modification of the midlatitude jet stream due to outflow from Jangmi rather than increased convection and downstream ridge building.

During the field phase, a variety of poleward-moving tropical circulation were observed by a wide range of instruments designed to observe the complex physical processes associated with the poleward movement of a tropical cyclone. These observations will provide a basis for examining the impact of the tropical circulations on the region of the midlatitudes into which they are moving and their downstream impacts through a variety of physical mechanisms.