中文摘要：

Super-parameterization (SP ) 试图明确地在粗糙的分辨率以内代表深传送对流由嵌入在母亲模型的每列解决模型(CRM ) 的云的全球模型。第一次，我们在 mesoscale 实现了 SP 地区性的天气模型，全球 / 地区性的吸收和预言系统(葡萄) 。构造 SP 葡萄使用一二维(2D ) 在每格子列的 CRM。控制和二 SP 模拟被进行让北京 7.21 重降雨事件用 SP 在葡萄评估改进。SP-run-I 是仅仅交付微视物理学反馈的基本 SP 跑，而 SP-run-II 交付 microphysical 和云部分反馈。跑的比较显示 SP-run-I 比控制跑在降水预报上有稍微积极的影响。然而，云部分反馈的包括导致明显的全面改进，特别地以云部分和 24-h 累积降水。尽管这仅仅是用 SP 葡萄的初步的研究，我们相信它将在中国用 SP 为后续研究提供可观的指导。

英文摘要：

Super-parameterization（SP） aims to explicitly represent deep convection within a coarse resolution global model by embedding a cloud resolving model（CRM） in each column of the mother model. For the first time, we implemented the SP in a mesoscale regional weather model, the Global/Regional Assimilation and Pr Ediction System（GRAPES）. The constructed SP-GRAPES uses a two-dimensional（2D） CRM in each grid column. A control and two SP simulations are conducted for the Beijing ＂7.21＂ heavy rainfall event to evaluate improvements in GRAPES using SP. The SP-run-I is a basic SP run delivering microphysics feedback only, whereas the SP-run-II delivers both microphysical and cloud fraction feedbacks. A comparison of the runs indicates that the SP-run-I has a slightly positive impact on the precipitation forecast than the control run. However, the inclusion of cloud fraction feedback leads to an evident overall improvement, particularly in terms of cloud fraction and 24-h cumulative precipitation. Although this is only a preliminary study using SP-GRAPES, we believe that it will provide considerable guidance for follow-up studies using SP in China.