Singular Vector Ensemble Forecast of Tropical Cyclone

Thawatchai Sirikantisophon, Dusadee Sukawat

Abstract

Singular vector is a method to generate several initial conditions for weather prediction models from the same observation data. In this paper, a singular vector associated with a barotropic model is applied to tropical cyclone track forecasts by creating 50 initial conditions. These initial conditions are used to create 50 track predictions with slightly different initial conditions, called ensemble forecast. The ensemble mean should have more forecast skill than a single forecast. In addition, probability forecast of storm track based on the ensemble is also possible.