Proposal of North Western Pacific Tropical Cyclone Track

PROPOSAL OF “NORTH WESTERN PACIFIC TROPICAL CYCLONE TRACK

ENSEMBLE FORECAST RESEARCH PROJECT”

Project Summary

Ensemble and multi-ensemble forecasts are effective ways for operational communities to reduce uncertainty and improve accuracy. A WWRP-RDP project “North Western Pacific Tropical Cyclone (TC) Track Ensemble Forecast Research (NWP-TCTEFRP) Project” intends to build on the TIGGE concept and take advantage of the TIGGE CXML data provided by multiple organizations for improving TC track forecast skill over the North western Pacific, starting from 2009, as a five-year project.

The objectives of the NWP-TCTFRP Project are

- to explore and develop effective ways of obtaining and utilizing the track forecast data from TIGGE data providers to improve medium range forecast of TC track forecast

- to develop software for a real time multi-model tropical cyclone forecasting system from data transfer to verification

- to evaluate the utility of multi-model forecasts of tropical cyclones track predictions and provide recommendations on future multi-model ensemble systems and on future GIFS-RDPs

- to encourage forecasters of involved Members to utilize the information which the Project will deliver on a Web and to exchange their experience in use of TIGGE ensembles data for tropical cyclone forecast via internet forum.

The implementation of the Project should under the participation of GIF TIGGE WG, TIGGE data providers, TIGGE archive centers, TC Warning Centers, TC RSMCs etc.

The Project will first show its visibility and preliminary evaluation of the data during the Shanghai EXPO 2010, May 1 to October 31, 2010. It is followed by the enhancement of the data utilization, which may includes the extension of the CXML database (intensity, wind speed, and precipitation) and better guidance for operational purposes in the rest of the period through examining the value of the CXML database and training activities for the member countries in the Typhoon Committee.

Background of the Proposal

In the EC-LX1, The Council reaffirmed the use of ensemble techniques including multi-model consensus forecasting by the national and regional tropical cyclone warning centres to further improve the application of NWP to tropical cyclone forecasting. It also underlined the dissemination of ensemble-based probabilistic guidance to improve the representation of forecast uncertainty which will be especially useful for disaster risk management in threatened areas. The Council recognized the need for more research and development for the combined use of deterministic and EPS-based products to help Members in their risk assessment at an early stage in forecasting, and to improve decision-making processes for disaster risk management. (EC-LXI, 3.1.16)

The Council recognized that, while tropical cyclone forecasts have attained increasing accuracy in the track forecasting, they still rely heavily on the research and technology developments for improvement of forecasting tropical cyclone intensities, associated heavy rainfall and storm surge, as well as their seasonal frequency. The Council therefore reiterated that high priority be continuously given to transferring from the research and development domain into operations new scientific advances related to the forecasting of rapid changes of track and intensity of tropical cyclones and the impact of associated hazards during tropical cyclone landfall due to its significance for disaster prevention. An ensemble-based probabilistic approach is especially important in the low-threat/high-risk scenario associated with tropical cyclone forecasting. To focus research and development activities and facilitate the transfer to operations, the Council encouraged active interaction between operational forecasters and researchers as a key to success. (EC-LXI, 3.1.26)

Thus the Council recommended that the GIFS-FDP should begin with the prediction of tropical cyclone tracks and ensemble-based diagnostics, since EC-LX urged continuation of this real-time programme. The Council encouraged the participation of the relevant TIGGE data providers, TIGGE archive centres, Tropical Cyclone Warning Centres (TCWCs), Regional Specialized Meteorological Centres (RSMCs, including RSMCs with activity specialization in Tropical Cyclones) and WMO Members in executing such GIFS-FDP activities, which will require training and the development of a common set of products. The Council also recommended a follow on from the GIFS-FDP should focus on improving prediction of heavy rainfall and other problems of high priority, such as contributing to improving food security. The Council requested support from the WWRP SERA Working Group and THORPEX Regional Committees in exploring various societal application areas (EC-LXI, 3.1.31)

Under the TIGGE activities, the TIGGE Cyclone XML (CXML) datasets, including deterministic and probabilistic forecasts from eight centers, are available. Beginning in 2008, seven of the ten TIGGE providers joined with the Shanghai Typhoon Institute of the CMA to produce typhoon track positions in CXML and made them available in near real-time in support of the THORPEX Pacific Asian Regional Campaign (T-PARC). During the period, the TIGGE CXML datasets showed a distinct capability for the regional campaign and the experimental forecast guidance. Based on this exercise, the CMA proposed the Tropical Cyclone FDP, to enhance the use of the CXML at the WWRP JSC in September 2009. Some operational centers regularly obtain the CXML datasets for experimental use. For example, In KMA, there is an internal web page, to show all CXML track forecasts. The RSMC Tokyo also has a web site for the JMA track forecasts for Typhoon Committee members with user ID and password.

Objective of the Proposal

The objective of the Proposal is to provide the tropical cyclone track forecast information, extracted from the TIGGE CXML datasets to all Typhoon Committee Members and researchers in real-time, based on their requests as an experimental and research purposes. This is the first step and in collaboration with TIGGE WG, the Project will extend to cover more information, such as precipitation and wind forecasts of Tropical Cyclones in further development.

The Project will also assess and quantify the values of ensemble forecasts using TIGGE CXML, while enhancing ability of forecasters to use probabilistic forecasts for tropical cyclone track. The Project will establish links and feedback mechanism with forecasters and end-users for further developments, and will improve our knowledge how to extract useful information from TIGGE dataset, to facilitate the technology transfer to NHMSs.

Content of the Project

-To use the data in real-time

Currently the TIGGE CXML data is available at 48 hour delay. For the Project, the real-time access to the CXML data is required. The Project will request the real-time access to the CXML data during the typhoon period from May to November every year to the TIGGE WG. A ftp server will be set up to collect all CXML data for the final product production and delivery to the Members of the Typhoon Committee and interested researchers through Web interface.

-To develop the forecast methodology

A key issue of the Project is to find scientific methods of applying single EPS, multi-EPS, or even Ensemble-of-Ensemble to TC track and precipitation forecasts and how the Project delivers the final products through Web interface. Based on TIGGE multi-model ensembles data, a few research studies have been carried out to investigate the benefits from multi-model ensembles compared with single ensemble. As summarized by Dr. Phillipe Bougeault and Zoltan Toth in 7th session of International Core Steering Committee for THORPEX (2008), the results from TIGGE have demonstrated only limited benefits of multi-model ensembles for forecasts of 500hpa height and sea level pressure, but give better benefits for 2m temperature and 850hpa temperature. It is also found out that the benefits are more significant when ensembles with comparable skill are combined from different centers, whereas the benefits are less clear when poorer performing ensembles are added to a better performing system. But in their summary, there is no report for tropical cyclone, and it is still uncertain what benefits will be obtained from multi-model tropical cyclone ensembles, especially for track, and precipitation. Sharanya (2009) assesses the ability of global ensemble prediction system to predict tropical cyclone track probabilities in both Atlantic and North western pacific region, and results show that, the combined ECMWF and UKMET ensemble produced more mean forecasts of small error than either individual ensemble, and the combined ensemble also yielded more accurate probabilistic 12-36 hour forecasts in Atlantic Ocean.

It is worth to perform a research on tropical cyclone multi-model ensembles, such like the best way to combine those ensembles information. Based on TIGGE CXML data, it is found out that some tropical cyclone track ensemble forecasts are good for North western Pacific region, but some ensembles perform with large biases. Is it true that multi-model tropical cyclone ensembles with better performance will give more beneficial than combining all ensemble? It is uncertainty and needs to be investigated.

For heavy precipitation induced by tropical cyclone, the location and probability of heavy precipitation is extremely important for the forecaster, because it is strongly connected with landslide, mudslide or flood, and will cause severe damage, and sometimes human being lost. To use the precipitation information in multi-model ensemble forecast data is a bit premature at this point, but the Project will perform the feasibility study on this for future implementation.

Based on above discussion, the tropical cyclone track forecast research project will focus on following items:

i. Based on the TIGGE CXML data, the Project will assess the ensemble forecast capability for tropical cyclone track in the North-western Pacific region.

ii. The combining results from multi-model ensembles will be compared to single model ensemble, and combining results from all multi-model ensembles and selected ensemble members will be compared to decide the best way for combination.

iii. Combining results from multi-model ensembles will be compared to the consensus track of multi-center single deterministic forecasts.

iv. A verification will be done to assess the probabilistic forecast capability of TC track forecast, by comparing methods between ensemble of multi-model ensembles and consensus of multi-model deterministic forecast, and single model EPS.

v. The Project will investigate the feasibility of the extended use of the TIGGE CXML data, including the information on TC intensity, maximum wind speed and precipitation. The Project will perform a feasibility study of the current predictability of TC intensity, maximum wind speed, which is currently available on the CMXL data.

vi. The combined probabilistic result from multi-model ensembles for tropical cyclone track and will be provided in real-time to the Members of the Typhoon Committee and interested researchers through Web interface with id and password.

Approach and Structure

The Project will establish a Committee and the following two working groups (WGs) to accomplish all above objectives.

WG1: to develop how to extract useful information of TC track forecast and techniques to assess the added value of ensemble information

WG2: to implement to create a web-based interface to deliver probabilistic TC

track forecasts

The Project Committee will be co-chaired by Drs. Jiandong Gong (CMA) and Tetsuo Nakazawa (JMA).

The Project Committee will ask the Members of the Typhoon Committee to recommend the candidates of the working groups.

Timelines

-Year 2009

Establishment of the Project

Oct. 2009

-The outcomes of the Shanghai meeting are used to determine if this research demonstration project will require an expansion of the CXML format or possibly even another exchange approach

- WMO sends a letter to all TIGGE CXML data providers, asking their permission to use the CXML dataset in real-time for the project of NWP- TCTEFRP.

- WMO sends a letter to all Typhoon Committee members, asking their interests to receive TC track ensemble forecast products in real-time and notifies the relevant RSMC(s) of this project

-The Chair of the Asian THORPEX Regional Committee sends a letter to all Typhoon Committee Members, asking their recommendations of candidates of two working groups

Nov 2009

-Report on the outcome of the planning process to the THORPEX ICSC

-Letters sent from the SG’s Office to data providers informing them of the project requesting access and any data changes (e.g., CXML format can carry other information such as central pressure, peak winds by quadrant, etc

Dec. 2009

-ARC establishes two working groups

-Year 2010

Execution of Shanghai EXPO 2010

Jan 2010

-Short report of the project description and goals due to THORPEX IPO

Mar. 2010

-Two working groups report to the ARC and THORPEX IPO on the interim status (including prototype of the products)

Apr. 2010

-ARC holds a training workshop of TC track forecasts for forecasters in Typhoon Committee

May 2010

-The web page opens

November 2010

-Typhoon committee reports on the early outcomes of the data set and lessons learned for subsequent TC activities (La Reunion is interested in a subsequent project). The report must include an initial assessment of the added value of ensemble/multi-model ensemble approaches and guidance on how to introduce ensemble products to forecasters in other regions.

-Year 2011

Evaluation of the TC CXML Database in Shanghai EXPO 2010

-Year 2012

Improvement/Enhancement of the Utilization of the TC CXML Database

-Year 2013

Final Assessment and Recommendation of the TC CXML Database

References

Aberson, S. D. and C. R. Sampson, 2003: On the predictability of tropical cyclone tracks in the northwest Pacific basin. Mon. Wea. Rev., 131, 1491-1497.

Buizza, R., 1997: Potential forecast skill of ensemble prediction and spread and skill distributions of the ECMWF ensemble prediction system. Mon. Wea. Rev., 125, 99-119.

Buizza, R., Bidlot, J. R., Wedi, N., Fuentes, M., Hamrud, M., Holt, G. and Vitart, F., 2007: The new ECMWF VAREPS (variable resolution ensemble prediction system). Quart. J. Roy. Meteor. Soc., 133, 681-695.

Ebert E., Z. Toth, M. Charles, and G. Ross, 2008: An XML format for cyclone analyses and forecasts. [online http://www.bom.gov.au/bmrc/projects/THORPEX/CXML/index.html, 27 February 2008]

Rappaport, E.N., J.L. Franklin, L.A. Avila, S.R. Baig, J.L. Beven, E.S. Blake, C.A. Burr, J.G. Jiing, C.A. Juckins, R.D. Knabb, C.W. Landsea, M. Mainelli, M. Mayfield, C.J. McAdie, R.J. Pasch, C. Sisko, S.R. Stewart, and A.N. Tribble, 2009: Advances and Challenges at the National Hurricane Center. Wea. Forecasting, 24, 395–419.

Sharanya, J. M. and M. F. Peter 2009: On the ability of global Ensemble Prediction Systems to predict tropical cyclone track probabilities. Submitted to Weather and Forecasting.

High Impact Weather forecasting for GEOSS: The Global Interactive Forecast System (GIFS), THORPEX GIFS-TIGGE Working Group, Geneva, Switzerland, 18-20 November 2008.