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

PI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;二是使用聚类分析作为在气旋相空间内确定结构演变的优选路径的辅助手段;第三，将开发基于减少的预测器集合的基于回归的统计动态预测模型。此外，将通过对这些模型进行验证来确定目前大气气旋运行模型中的不确定性，验证的基础是气旋相空间诊断与其自身分析的对比，以及利用卫星观测结果进行气旋相空间诊断。

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.