千米尺度模式降水的检验评估进展及展望
A Review and Outlook of Verification and Evaluation of Precipitation Forecast at Convection-Permitting Resolution
中文摘要：
数值模式检验评估是理解模式偏差、完善数值模式和提升预报服务产品质量的有效方法，是现代数值模式发展的基础环节之一。传统的降水业务检验主要基于累积降水量的评估，较少考虑降水的发展演变过程。随着模式分辨率和降水精细化预报需求的提升，仅基于降水量的检验评估已不能满足定时、定点、定量的降水预报服务需求。在回顾已有降水检验评估方法的基础上，重点介绍基于降水频次、强度等小时尺度特征量及其日变化峰值时间的新评估方法，并示例给出此类评估结果对于深入认识日内降水预报偏差的应用价值。初步结果表明，相关方法和结论为丰富和深化数值模式评估，改进客观预报方法具有有益的启发作用。随着千米尺度模式在降水精细化预报业务中的广泛应用，未来仍需进一步发展和完善针对不同区域降水过程分类、分型的细致评估方法，并进一步研发面向快速更新循环系统和对流可分辨集合预报模式应用的检验评估指标，为千米尺度数值模式的改进和模式产品的合理应用提供科技支撑。
中文关键词：
检验评估，对流可分辨模式，精细化预报，小时尺度降水特征
KeyWords：
verification, convective permitting model, fine-scale forecast, hourly precipitation characteristics
Abstract：
Evaluation and verification of numerical model provides an effective method to understand the model deviation, improve the numerical model as well as improve the quality of forecast service products. The traditional operational verification of precipitation is mainly based on the evaluation of accumulated rainfall amount, and less attention has been paid to the evolution of rainfall systems. With the increasing of the model resolution and the requirement of fine precipitation prediction, the verification and evaluation based on amount only will not satisfy the demand of fine-scale quantitative precipitation forecast service. On the basis of reviewing the existing methods of precipitation verification and evaluation, this paper mainly introduces a new evaluation method based on hourly scale characteristics such as frequency and intensity of precipitation and the peak time of diurnal variation. Examples are given to show the application of such evaluation methods for in-depth understanding of the subdaily precipitation forecast deviation. The preliminary results show that the results can enrich the numerical model evaluation and provide more useful information to improve the objective prediction method. With the widely application of convective permitting models in operational precipitation forecast, it is necessary to further develop and improve the detailed evaluation methods for different regions through classification, and further develop the verification and evaluation metrics for the application of rapid updating system and convective permitting ensemble forecasting model, so as to improve the performance of kilometer scale numerical model and post-possess the model products to provide scientific and technological support.