Tropical Cyclone Ensemble Forecasting at the Met Office: Upgrades to the MOGREPS Model and TC Products, and an Evaluation of the Benefit of Multi-model Ensembles

Thursday, 21 April 2016: 1:45 PM
Ponce de Leon C (The Condado Hilton Plaza)
Helen A. Titley, Met Office, Exeter, United Kingdom; and R. Stretton
Manuscript (500.3 kB)

For several years the Met Office have been producing ensemble tropical cyclone forecasts and disseminating them to collaborators worldwide. Until recently these have come from the Met Office MOGREPS-15 ensemble, which has a resolution of 60km. In July 2014 we extended the runs of our higher resolution (33km) global ensemble, MOGREPS-G, out to seven days, allowing it to be used in our tropical cyclone products for the first time. At the same time, a new set of products has been developed, including track and intensity forecasts for both named and forming storms. These MOGREPS-G based products are now produced operationally and are available for real-time dissemination.

In order to compare the performance of the MOGREPS-G tropical cyclone forecasts with those from other global ensembles, and produce multi-model ensemble tropical cyclone forecasts, the forecasts from ECMWF ENS and NCEP GEFS are also run through the Met Office tropical cyclone tracker (MOTCTracker). Routine verification is now carried out on the forecasts from each global ensemble, and the various multi-model combinations. Recent verification has shown that the upgrade from MOGREPS-15 to the higher resolution MOGREPS-G model has resulted in significant improvement in the skill and value in the strike probability forecasts. Verification also confirms that there is additional skill and value to be gained by combining output from the individual centre's ensemble forecast models in to multi-model ensemble tropical cyclone strike probability forecasts.

In this talk we will demonstrate the new ensemble tropical cyclone products and show verification results demonstrating the added value from both the MOGREPS model upgrade and from combining several global ensembles in to multi-model ensemble tropical cyclone forecasts.

- Indicates paper has been withdrawn from meeting
- Submission entered in competition