Application Of Radar Radial Wind Data Assimilation In GRAPES High Resolution Numerical Model

Doppler radar data has high spatio-temporal resolution and can match with small-and medium-scale numerical models.Therefore,radar data assimilation has great potential for improving the quality of small-and medium-scale information of initial field and improving the level of numerical weather prediction.However,there are many challenges in applying it to GRAPES high-resolution numerical prediction model.In this paper,based on the idea of Doppler radar radial velocity de-blurring scheme,a quality control scheme of Doppler radar radial velocity is developed for the GRAPES assimilation and prediction system.The results show that the proposed scheme can effectively correct the velocity ambiguity and improve the quality of radar radial wind observation data.The observation error of radar data in GRAPES assimilation prediction system is reestimated by using the update vector method.The statistical results show that the observation error of radar radial velocity increases with the increase of altitude.The frequency statistics of the innovation vector are close to the normal distribution,which reduces the deviation of the data statistically,and further verifies the rationality of the three-dimensional variational assimilation scheme of the radar radial wind data.Based on the three-dimensional variational assimilation system and the GRAPES high-resolution model,the radial velocity data assimilation analysis and numerical simulation tests for the precipitation process of Typhoon "Skylark" occurred on August 3,2018 are carried out.The results of assimilation test show that the effective assimilation of Doppler radar data can adjust the initial field of the model.The assimilation of Doppler radial velocity can enhance the convergence intensity of typhoon cyclone in the analytical field,enhance the mesoscale information,and improve the initial wind field.The direct assimilation of radial velocity has a positive contribution to precipitation forecast.The continuous cycle assimilation experiment was carried out for 10 days.The prediction effect of batch assimilation experiment shows that the reasonable radial wind assimilation scheme can effectively improve the forecasting effect of high-resolution GRAPES model for heavy precipitation.