江淮梅雨锋暴雨对流尺度模拟偏差特征分析及可预报性研究

业务对流尺度数值预报系统（水平分辨率1~4km）能够解析γ到α中尺度的多尺度天气系统，各尺度误差在模拟过程中相互作用并快速增长最终限制了模式可预报性。梅雨锋暴雨是江淮地区最具代表性的高影响天气，具有典型的多尺度特征，对其开展对流尺度模拟的偏差分析和可预报性研究可以解释模拟偏差成因并为模式优化提供依据。本项目针对近年来江淮地区的梅雨锋暴雨实例，分析业务对流尺度数值预报系统降水模拟偏差特征，构建典型偏差案例库，加深对模式应用的认知；基于对流尺度集合预报分析典型降水偏差的天气学不确定性来源，揭示产生各类典型降水偏差的关键影响因子；研究造成典型降水偏差的预报误差增长机制及对实际和内在可预报性的影响，揭示误差增长及其影响降水偏差的机理。研究结论不仅是对对流尺度模拟可预报性理论的完善，还能提高对流尺度模式的业务应用价值，并为相应集合预报技术改进提供科学依据。

Operational convection-allowing model (CAM) with a resolution between 1 to 4km resolves multiscale weather systems from meso-γ- to meso-α-scales, rapid forecast error growth from different sources and different scales strongly limits the predictability of CAM. Convective precipitation forecasting of Meiyu frontal extreme rainfall events is an inherently multi-scale challenge because of broad range of spatial scales impacting the initiation and evolution of convective systems. Therefore, the study on the predictability of CAM of Meiyu frontal rainfall event helps explain and potentially aids the improvement of CAM. In this project, we aim to understand the application of CAM on frontal rainfall events through systematical forecast evaluation and precipitation error analysis. Through CAM ensemble forecasts, we use ensemble sensitivity analysis to reveal the uncertainty source of representative precipitation forecast error. We investigate error growth dynamics within CAM ensemble forecasts and associated impacts on both practical and intrinsic predictability, and further build a conceptual model to reveal the relationship between error growth and representative precipitation forecast error. In summary, the project has important scientific significance not only for the predictability of CAM, but also aids the improvement of CAM-based ensemble generation.