Cyclone Structure Characterization and Evolution Forecasting Using Cluster Analysis and Statistical-Dynamical Modeling

PIs试图从三个方面深入了解气旋结构演变的动力学：一是通过将最近发展的气旋相空间诊断扩展到所有气旋结构演变来检测和表示低维特征；第二种方法使用聚类分析来帮助确定旋风相空间内结构演变的首选路径；在第三章中，将建立一个基于回归的统计动态预测模型，该模型基于一个简化的预测集。此外，将根据旋风相空间诊断和利用卫星观测资料进行旋风相空间诊断，通过对这些模式进行验证，寻求对目前大气旋风操作模式的不确定性的某种度量。

The PIs seek to gain insights into the dynamics of cyclone structure evolution in three ways: one is the detection and representation of lower-dimensional features by extending the recently developed cyclone phase space diagnostic to all cyclone structure evolutions; the second uses cluster analysis as an aid in defining preferred paths of structural evolution within the cyclone phase space; in the third a regression based statistical dynamical forecast model based on a reduced predictor set will be developed. In addition, some measure of the uncertainty in present operational models for atmospheric cyclones will be sought through a validation of these models based on cyclone phase space diagnostics against their own analyses and a cyclone phase space diagnostic using satellite-derived observations.