定量降水预报技术研究进展
The Research Progress of Quantitative Precipitation Forecast
中文摘要：
定量降水预报（QPF）是天气预报最重要的业务之一。对21世纪以来QPF的国内外业务现状、集合概率预报技术、短时临近预报技术、检验技术以及订正技术等做了简要综述，初步总结了各类研究所取得的主要进展，包括：QPF准确率的稳步提高得益于数值模式的持续发展，以及预报员对于模式产品的应用订正能力的不断提高；以数值集合预报为基础的概率QPF（PQPF）为用户提供预报不确定性信息，是一种更为科学的天气预报形式；另外，数值模式的实时天气学检验及订正技术的发展为“增加预报员的附加价值”提供有力支撑。最后，提出了目前QPF研究存在的主要问题和需要加强的几个研究方向，分别是：（1）进一步加强数值模式对于大气水汽场的同化和模拟，采用新的模型来描述下垫面与边界层之间的水汽交换，以及大气中真实云和降水物理过程；（2）降水观测和预报的随机特点还没有被充分考虑，需要进一步研究不同时空尺度上的模式预报能力，发展有效的QPF订正技术方法；（3）对于PQPF还存在着理解上的困难和误区，如何将预报不确定性信息传递给用户需要进一步研究；（4）短时临近QPF应由雷达回波外推方法向结合数值模式预报的混合外推技术转变，提高对对流降水系统的预报能力；（5）针对传统的统计评分检验方法的不足，应引入新的QPF检验技术方法，但新方法的解释应用还需不断地积累经验。
中文关键词：
定量降水预报，集合预报，检验，预报业务
KeyWords：
quantitative precipitation forecast (QPF), ensemble forecast, verification, operational forecast
Abstract：
Quantitative precipitation forecast (QPF) is one of the most important operations in weather forecast. A brief overview is made for current domestic and foreign operational situation and techniques of ensemble probability forecast, short-term forecasting and nowcasting, verification methods and error corrections. Preliminary conclusions have been made for some research progress, the improvement of QPF accuracy comes from the development of numerical models and the improvement of correction capability for the application of numerical model productions. Based on the numerical ensemble forecast, the probability QPF (PQPF), as a more scientific weather forecast method, can provide us with uncertainty forecasting information. Moreover, the improvement of real-time synoptic verification and error correction method of numerical model products provide strong support for 'adding forecasters additional value'. Finally, based on this summary of progress, main problems and research fields of QPF research have been introduced. (1) Further strengthening the capability of moisture data assimilation and simulation, using a new model to describe the moisture exchange between land surface and boundary layer and cloud and precipitation physical Processes. (2) Precipitation observation and forecast’s random characteristics have not been considered sufficiently yet, the numerical forecast capability of different spatial and temporal scales and QPF correction methods should be made a close study of. (3) It should be considered how to deliver forecast uncertain information to users due to misunderstanding the meaning of PQPF. (4) The techniques of short-term and nowcasting QPF should be exchanged from radar extrapolation to mixed extrapolation combined with numerical model forecast for improving convective precipitation forecast capability. (5) Based on the shortcoming of traditional statistical methods, new verification methods for QPF should be introduced, but objective explaination and application of the new method takes time to gather experiences.