A pseudo-ensemble hybrid 3DVar data assimilation system for tropical cyclone initialization

Ensemble-based data assimilation methods implement flow-dependent background error covariance during the state estimation, and are particularly useful for mesoscale weather systems with fast evolving forecast uncertainty. In the case of tropical cyclones, recent studies suggest the background error covariance estimated from short-term ensembles is determined primarily by position and intensity uncertainty. Making use of the first-order forecast uncertainty mechanisms, a pseudo-ensemble 3DVar (PE3DVar) system has been developed and implemented in the Weather Research and Forecast model for tropical cyclone implementation. The PE3DVar estimates the background error covariance matrix using a pseudo-ensemble with samples selected from a library of idealized tropical cyclone vortices, which is introduced into the hybrid 3DVar minimization algorithm. The library is comprised of a large quantity of previously spun-up tropical cyclone vortices that vary in size, intensity, and structure, which are binned according to maximum wind speed. In this study, the PE3DVar is then used to assimilate dual Doppler radar wind observations for Hurricane Earl (2010) in comparison to the performance of the standard WRF-3DVar which has static and isotropic background error covariance. Preliminary results show clear advantages of PE3DVar in improving the inner-core temperature and moisture fields during analysis, as well as in the subsequent short- and medium-range intensity forecast. The PE3DVar method benefits from the multivariate, anisotropic covariance typically used in ensemble approaches to data assimilation, while maintaining the same computational cost as the standard 3DVar. The new system also provides the necessary framework for a possible PE4DVar data assimilation algorithm for tropical cyclone initialization.