Application of the Multi-scale Blending Scheme on Continuous Cycling Radar Data Assimilation
The torrential rains result from the short duration extreme rainfall system is the most critical issues for the disaster prevention. However, the predictability of the short duration extreme rainfall system is very limited due to the fast involved and strong nonlinearity nature. The assimilation of the radar observation with rapid update cycle frequency and the high resolution model is a key to level up the predictability of such a system.
The continuous rapid update cycle is able to capture and keep convective-scale structure and avoid the model spin-up problems. However, it has some disadvantages, for example, the limited-area regional model analyses and forecasts suffer a general de?ciency in effective representation of large-scale features and thus there are obvious large-scale forecast errors. In addition, the model is easy to accumulate the model errors in the areas with sparse observations.
In this study, a multi-scale blending scheme (Hsiao et al. 2015) apply to a continuous hourly cycling 3DVAR and LETKF radar DA system. This scheme combines the global model analysis and the convective scale model 1-hr forecast. Case studies show that the blending scheme is able to correct the bias of the large-scale flow from the global model and keep the convective rainfall structure from the convective scale radar DA system.
Ms. Siou-Ying Jiang Central Weather Bureau (CWB), Taiwan