Abstract:Using the forecasting radar echo data from GRAPES_3 km (Global/Regional Assimilation and Prediction System) and WRFRUC (Weather Research and Forecast Rapid Update Cycle) of July and August in 2018, the real radar echo data from SWAN (MCR), and FSS (fractions skill score), the radar echo forecast abilities of two high-resolution models for typhoon and subtropical anticyclone were analyzed. The results showed that the two models have better skill for weak radar echo forecast, with increase of radar echo, the forecasting skill score decreases. When radar echo reaches 55 dBz, FSS even is equal to zero. When spatial neighborhood radius is 3, the forecast is less skillfull in the typhoon rainstorm than subtropical anticyclone rainstorm with radar echo lower than 35 dBz. If radar echo is bigger than 35 dBz, it is in contrast. The forecasting of radar echo by WRFRUC model is always better for typhoon rainstorm than for subtropical anticyclone rainstorm. When spatial neighborhood radius increases to 9, the echo forecasting FSS of WRFRUC model is higher than that of GRAPES model in typhoon rain, but in subtropical anticyclone rainstorm, GRAPES model has the higher skill than WRFRUC model. The biggest FSSs of GRAPES and WRFRUC models appear at the spatial neighborhood radius of 11, and the values are 0.239 and 0.195, respectively. In the first three hours of the forecasting, FSS is bigger than in other hours for the GRAPES model. In the middle hours of the echo forecasting of WRFRUC model, FSS is higher than in other hours.