Meso-scale analysis and predictability of Taiwan heavy rainfall using VDRAS and WRF

Taiwan Island is frequently inundated with heavy rainfall that can lead to substantial flooding and loss of life. Understanding the evolution of heavy rainfall events and their predictability is a primary concern for the Taiwan Central Weather Bureau. Particularly critical is providing advance warning of potential for flash floods and rainfall exceeding 100 mm over a 1-2 hr period which is not a rare event for Taiwan. During 2008, an enhanced set of data were collected over Taiwan as part of the Southwest Monsoon Experiment/Terrain-influenced Monsoon Rainfall Experiment (SoWMEX/TiMREX) to improve understanding of the physical processes associated with terrain-induced heavy precipitation systems and the monsoonal environment in which they are embedded. A dense set of instrumentation, including radar, rawinsonde, dropsonde and surface stations, were located over southern Taiwan to collect high resolution observations needed for data assimilation and numerical modeling studies. The NCAR dual-polarization radar (S-Pol) was specially sited in southern Taiwan to provide comprehensive observations for studying the dynamical and microphysical processes and environmental conditions leading to the development of heavy rainfall.

This paper provides a comparison of the 4-D Variational Doppler Radar Analysis System (VDRAS) radar data assimilated fields with Weather Research and Forecasting (WRF) Model output fields for two southwest monsoonal periods during SoWMEX/TiMREX, 2-6 June and 14-16 June and for a heavy rainfall event on 31 May associated with the Mei Yu front. Radar reflectivity and radial velocity data from S-Pol and the Taiwan operational radars are assimilated into VDRAS, providing a suite of fields (wind, thermodynamic, shear and stability) that are compared with WRF analysis and forecast fields over Taiwan. A comparison of VDRAS analysis fields with the Taiwan Local Analysis Prediction System (LAPS) analysis and forecasts fields for 31 May will also be shown. The strength of using the VDRAS fields for understanding the evolution of and analyzing the predictability of the heavy rainfall events lies in its ability to assimilate the high resolution radar observations every 10 min and provide more detailed results on the meso-scale than is possible with the current resolution of the WRF analysis and forecast fields. In addition, the LAPS analysis fields are updated every 60 min in real-time and may be an ideal dataset to use as a background field for VDRAS instead of WRF. Those fields that lend toward greater predictability of heavy rainfall will be discussed.