偏波ドップラーレーダの同化によるメソ対流系の降水予測精度向上に関する研究

Abstract

The short lead time rainfall prediction by Numerical Weather Prediction model has some difficulties in the spin-up problem. Therefore, data assimilation (DA) is expected to improve the initial condition in the model. In this study, our developed ensemble DA system, CReSS-LETKF, and the method of estimation of ice-water mixing ratios are employed. DA of rain, graupel, ice crystal, snowflake and Doppler velocity estimated by polarimetric Doppler radar are carried out after the first convective cloud in mesoscale convective systems is generated. As a result, the first convective clouds formed in initial condition have effective influence on the short lead time rainfall prediction. As the next challenging step, DA is carried out before the first convective cloud. As a result, convective clouds are not generated although the atmosphere conditions, such as potential temperature change