Enhancing Forecasts of Tropical Cyclone Extratropical Transition By Statistical Pattern Recognition

该项目建立在先前工作的基础上，该工作开发了一种称为投影追踪（PP）的统计模式识别技术，可用于根据可从全球分析、模式预报和卫星观测中收集到的周围大气的大尺度特征来预测热带气旋的温带转变（ET）。通常情况下，热带风暴在离开亚热带温暖水域并转向极地时减弱。ET导致风暴再次增强，使其保持危险。研究人员计划推广和完善PP技术，将其从概念验证变为适用于实时预测的实验预测工具。虽然该项目具有实质性的应用方面，但研究人员还旨在了解区分两类风暴（ET与消散）的潜在物理机制，这两类风暴由风暴特征和更大天气环境的组合所体现，并区分不同的增强速率。最终目标是建立一个自动化系统，可以从模型和遥感信息中识别出可能的外星人候选者。该方法将应用于飓风和台风预报，并将适用于T-PARC的实地活动。将资助两名研究生和一名本科生。

The project builds on previous work which developed a statistical pattern recognition technique called Projection Pursuit (PP) that can be used to predict the extratropical transition (ET) of tropical cyclones based on large-scale characteristics of the surrounding atmosphere that can be gleaned from global analyses, model forecasts, and satellite observations. Normally tropical storms weaken as they leave the warm waters of the subtropics and turn poleward. ET leads to a reintensification of the storm causing it to remain dangerous. The investigators plan to generalize and refine the PP technique, taking it from proof-of-concept to an experimental forecasting tool applicable for real-time forecasting. While the project has a substantial applied aspect, the investigators also aim to understand the underlying physical mechanisms that separate the two categories of storms (ET versus dissipating) embodied by a combination of the storm characteristics and those of the larger synoptic environment, as well as differentiate between various rates of intensification. The ultimate goal is an automated system that can identify likely candidates for ET from model and remote sensing information. The methodology will have application for hurricane and typhoon forecasting and would be applicable in the T-PARC field campaign. Two graduate students and one undergraduate will be supported.