Based on the wind lidar observation data and 12 wind shear events reported by pilots at Lanzhou Zhongchuan airport from July to November 2020, the performance and effectivity of two lidarbased lowlevel wind shear alerting algorithms were investigated. One algorithm is the regional divergence algorithm (RDA), and another is the combinational algorithm of single ramp and double ramps detection. The IRIS algorithm, the default algorithm in the lidar, was used as a reference. The conclusions are as follows: (1) RDA and ramp detection algorithm had better ability in alerting wind shear events under the convection weather or downward momentum weather conditions. However, they had a poor performance on identification of wind shear events under turbulent conditions. (2) Compared with ramp detection algorithm, RDA uses multiradial directions to reconstruct headwind profile, and consequently it can provide some temporal and spatial evolution information of radial velocities along the runway and its extension line, which is conducive to identification and alerting of wind shear events ahead of time. (3) The alerting results on 12 wind shear events indicated that RDA was superior to ramp detection algorithm and IRIS algorithm.

The precipitation forecast performance of GRAPES_Meso model with 3 km spatial resolution and its cloud analysis system was investigated and compared through numerical experiments for 13 heavy rainfall cases in summer and batch forecast in July 2018 in Northwest China. The results are as follows: (1) GRAPES_Meso model with 3 km spatial resolution had a good forecasting skill and stable performance for precipitation forecast in Northwest China, it could provide favourable outputs for short-term forecasting and nowcasting business. The average threat score of light rain and above for 13 heavy rainfall cases was between 0.5 and 0.6, while the threat score of batch experiments in July was slightly worse than that of 13 heavy rainfall cases. (2) By introducing observed data of radar reflectivity, satellite TBB and CTA, the cloud analysis system could reasonably adjust cloud water, rain water and other hydrometeor content, consequently reduce spin-up time in the mesoscale model and improve the forecast ability of precipitation with different magnitudes. However, the forecast results of cyclic assimilations in cloud analysis system weren’t stable. (3) The predicted coverage of radar echo by GRAESP_Meso model and cloud analysis system was consistent with observations, while the intensity of composite reflectivity was higher than the observations.