Author(s): Jiyang Tian; Jia Liu; Chuanzhe Li; Fuliang Yu
Linked Author(s):
Keywords: Data assimilation; Doppler radar reflectivity and radial velocity; Numerical rainfall prediction; Observation height; Weather Research and Forecasting model (WRF)
Abstract: Hydrological prediction needs high-resolution and accurate rainfall information, which can be provided by mesoscale Numerical Weather Prediction (NWP) models. However, the predicted rainfall is not always satisfactory for hydrological use. The assimilation of Doppler radar observations is found to be an effective method through correcting the initial and lateral boundary conditions of the NWP model. The aim of this study is to explore an efficient way of Doppler radar data assimilation from different height layers for mesoscale numerical rainfall prediction. The Weather Research and Forecasting (WRF) model is applied to the Zijingguan catchment located in semi-humid and semi-arid area of Northern China. Three-dimensional variational data assimilation (3-DVar) technique is adopted to assimilate the Doppler radar data. Radar reflectivity and radial velocity are assimilated separately and jointly. Each type of radar data are divided into seven data sets according to the observation heights: (1) < 500m; (2) < 1000m; (3) < 2000m; (4) 500～1000m; (5) 1000～2000m; (6) > 2000m; (7) all heights. Results show that the assimilation of radar reflectivity leads to better results than radial velocity. The accuracy of the predicted rainfall deteriorates as the rise of the observation height of the assimilated radar data. Conclusions of this study provide a reference for efficient utilisation of the Doppler radar data in numerical rainfall prediction for hydrological use.
Year: 2018