Heavy rainfall prediction during the Meiyu seasons in Taiwan with the WRF-LETKF system: what we have learned from the SoWMEX IOP8 event

Shu-Chih Yang* and Shu-Hua Chen, Ching-Yuan Huang, Keichii Kondo, Takemasa Miyoshi and Yu-Ching Liou
Department of Atmospheric Sciences, National Central University, Univ. of California at Davis USA, National Central University Taiwan, RIKEN Advanced Institute for Computational Science Japan, RIKEN Advanced Institute for Computational Science Japan, and National Central University Taiwan

Strong convections associated with Meiyu fronts often result in heavy precipitation in Taiwan and can bring disasters like flood or mudflow. Meiyu front is a multi-scale interactive weather system, which includes small-scale convection, a mesoscale front, and large-scale southwesterly monsoonal flow. A regional Ensemble Kalman Filter system, which couples Local Ensemble Transform Kalman Filter with the Weather Research and forecasting model (WRF-LETKF), has been used to study the heavy rainfall prediction during Meiyu season.

Based on a case study of the heavy precipitated SoWMEX IOP8 event on 16 June 2008, we have several findings with the WRF-LETKF system.
(1) Assimilation of the GPS-radio occultation (RO) data helps better represent the moisture transport from South China Sea toward southwestern Taiwan. Particularly, the use of RO local bending angles is superior to the use of RO local reflectivity due to its capability of recovering a deeper moist layer, extending from the surface to the mid-troposphere, which leads to a rainfall forecast closer to the observed.
(2) Assimilation of the radar data improves the rainfall nowcasting. However, the characteristics of the rainbands are greatly influenced by how we initialize the ensemble. A moist bias inherited from the initial ensemble can be alleviated by adjusting the ensemble with additive inflation or a mean-recentering method.
(3) The dual-resolution and dual-scale localizations WRF-LETKF framework can further improve the short-term heavy rainfall prediction, in terms of its intensity and timing.

More details will be presented in the talk.


*email: shuchih.yang@gmail.com
*Preference: Oral