tailieunhanh - Thử nghiệm ứng dụng radar kết hợp mô hình số trị trong dự báo mưa hạn cực ngắn

Bài viết giới thiệu một hệ thống kết hợp (blending) đang được triển khai nghiệp vụ tại Trung tâm Dự báo khí tượng thủy văn quốc gia (NCHMF) nhằm tăng cường khả năng dự báo nói chung và trong lĩnh vực dự báo cực ngắn nói riêng. | VNU Journal of Science Earth and Environmental Sciences Vol. 37 No. 3 2021 63-72 Original Article Case Study of Using the Blending Radar-Numerical Weather Prediction Product in Nowcasting Mai Van Khiem1 Woo Wang-chun2 Wong Wai-kin2 Yeung Siu Lun2 Du Duc Tien1 Mai Khanh Hung1 Dang Dinh Quan1 Pham Thi Phuong Dung1 Nguyen Thi Nga1 Chu Thi Huyen Trang1 1 National Center for Hydro-Meteorological Forecastings NCHMF No. 8 Phao Dai Lang Lang Thuong Dong Da Hanoi Vietnam 2 Hong Kong Observatory HKO 132 Nathan Road Tsim Sha Tsui Kowloon Hongkong Received 15 September 2020 Revised 26 January 2021 Accepted 10 Feburary 2021 Abstract In the very short-range forecast less than 12 hours of heavy rain and severe thunderstorms the main challenge is primarily on tracking and predictingng of growth and decay of significant convective systems. With the advances in nowcasting techniques analysis and nowcasting of heavy rain and thunderstorm from radars and satellites have been improving progressively. The intensity changes of convective weather from direct output of numerical weather prediction NWP models would contain substantial errors due to data assimilation and spin-up issues of model physical processes. In this paper we would introduce a new blending system being implemented in NCHMF to enhance nowcasting and forecasting services. The system includes i extrapolations of rain convective systems of Vietnam s radar mosaic or single radars for first 1-6 hours based on radar based on the Short-range Warning of Intense Rainstorms in Localized Systems SWIRLS developed by the Hong Kong Observatory HKO and ii a rapidly update convective- permitting NWP system based on WRF-ARW in the convective scale to provide the forecast up to the next 12 hours. Firstly 1-6 hours forecast from models are calibrated with radar-based quantitative precipitation estimates and nowcasts. The study will present some initial results of the blending system and discuss verification of the quantitative .