GRAPES-MESO模式不同空间分辨率对中国夏季降水预报的影响分析

国家气象中心业务运行的中尺度数值预报系统GRAPES-MESO（Global/Regional Assimilation and Prediction System mesoscale model）在升级到4.0版本后采用了与以往版本不同的三维空间分辨率设置，本文通过计算精度分析、个例分析及统计分析的方法详细阐述了两者水平分辨率和不等距垂直分层的差异，并由此深入分析了不同模式三维空间分辨率对中国夏季汛期降水预报的影响。主要结论表明，GRAPES-MESO预报系统4.0版本在水平分辨率提高到10 km并同时使用更为合理的加密垂直分层设置后，不仅提高了计算精度和计算稳定性，同时仍能满足业务预报的时效要求。对个例降水特征的分析结果表明，提高模式空间分辨率可以在一定程度上改善对降水中心的预报，但对降水落区的预报改进较为有限。对2012年7月整月批量试验的统计检验结果表明，月平均技巧评分总体变化不大，但对逐日大到暴雨评分提高较大，通过各气象要素统计检验分析可以认为，模式空间分辨率提高的主要作用是通过降低了中低层高度场、温度场和水平风场的误差，改进了对流层中层环流背景场以及对流层低层降水直接触发系统的强度预报，从而能够提高大到暴雨的降水评分。

Impacts of Different GRAPES-MESO model Spatial Resolutions on Summer Rainfall Forecast in China

The Global/Regional Assimilation and Prediction System (GRAPES) mesoscale model(GRAPES-MESO) V4.0 with a new spatial resolution has been put into formal operation in the Numerical Weather Prediction Centre (NWPC) of China Meteorological Administration (CMA). This paper examines the computational accuracy of GRAPES-MESO with different spatial resolutions. The results show that the refined spatial resolution can be applied for operational forecast. Several sensitivity experiments are then performed to determine the impacts of different spatial resolutions on the forecast skill of summer rainfall in July 2012 in China. The simulation results indicate that GRAPES-MESO V4.0 with a refined spatial resolution can better improve the precipitation maxima forecast compared to precipitation location forecast. The equitable threat score (ETS) of 24 hours accumulated precipitation from the batch experiments of the whole July 2012 shows that heavy rainfall forecast is significantly improved. The verification of geopotential height at the middle and lower levels of the troposphere illustrates that the model with a refined spatial resolution reduces the prediction error of the geopotential height and improves the predictions of synoptic background circulation at the middle levels and systems at lower levels that can trigger precipitation. Finally, the model can increase the accuracy rate of prediction of convective systems at lower levels and improve the ETS of heavy rainfall forecast.