Development of Advanced Guidance for Forecasting High-Impact Weather

9714154 Fritsch美国天气研究计划（USWRP）是一项跨部门活动，旨在执行和实施必要的研究，以改善向全国提供的天气服务。根据该计划，国家科学基金会、国家海洋和大气管理局、国家航空航天局和海军研究办公室将联合评估和支持USWRP的高优先级研究。有充分的文件证明，原始观测和数值模型输出的效用可以通过统计后处理大大增强。最近在数值模拟和安装新的观测系统方面的进展创造了重新设计现有后处理技术和发展新技术的机会和需要。首席研究员将设计、构建和测试新的统计后处理程序，这些程序将利用新兴的集成技术，并将操作数值模型变化对统计后处理程序的影响降至最低。他还将建立一个先进的基于观测的统计预报系统，为高影响天气提供可靠的短期概率指导。基于观测的系统将利用高频观测平台，如WSR-88D多普勒雷达和卫星信息。提出的工作的一个主要工具将是一般的可加性模型的应用，它允许在统计回归关系中的某些非线性。预计由此产生的技术将在敏感天气参数预报的提前期方面取得重大进展，并为高影响天气事件期间的短期决策提供急需的概率指导。＊＊＊

9714154 Fritsch The U.S. Weather Research Program (USWRP) is an interagency activity designed to perform and implement the research necessary to improve the delivery of weather services to the nation. Under this Program, the National Science Foundation, the National Oceanic and Atmospheric Administration, the National Aeronautics and Space Administration and the Office of Naval Research are jointly evaluating and supporting research of high priority to the USWRP. It is well documented that the utility of raw observations and numerical model output can be greatly enhanced by statistical post-processing. Recent advances in numerical modeling and the installation of new observing systems have created an opportunity and a need to redesign existing post-processing techniques and to develop new techniques. The Principal Investigator will design, construct, and test new statistical post-processing procedures that will capitalize on the emerging ensemble technology and will minimize the effects of changes in operational numerical models on statistical post-processing procedures. He will also construct an advanced observations-based statistical forecasting system for providing reliable probabilistic short-term guidance on high-impact weather. The observations- based system will capitalize on high-frequency observing platforms such as WSR-88D Doppler radars and satellite information. A primary tool of the proposed work will be the application of general additive models that allow for certain non- linearities in statistical regression relationships. It is expected that the resulting techniques will yield significant lead time advances in forecasts of sensible weather parameters and provide much-needed probabilistic guidance for short-term decision making during high-impact weather events. ***