Stormtime Dynamics of the Relativistic Electron Flux in Earth's Radiation Belts

A state-vector representation is a powerful technique for describing complex plasma systems. Its framework can be adapted for classification methods which can be used to analyze the system's history and for prediction methods which can serve to forecast its future activity. A state-vector description is developed for the electron flux dynamics in Earth's radiation belts, based on an 11-year (1993-2003) dataset of high-cadence flux measurements from a low-Earth (SAMPEX) orbit over a wide L range and at a fixed energy (2-6 MeV). A clustering algorithm is used to divide the state space into regions, or clusters of vectors, and it becomes evident that flux intensifications during storms correspond to characteristic transitions in state space following geoeffective interplanetary disturbances (such as interplanetary coronal mass ejections and high-speed streams). Examples are discussed to show that the classification is valid for medium-term (several-days) and long-term (solar-cycle-phase) timescales. The state-vector representation is then used as the basis of a predictive model of the flux distribution given upstream solar wind measurements. It is found that model accuracy of storm prediction is maximized if the model is tuned at a highly nonlinear regime. The relation to earlier state representations and models of the radiation belt flux is discussed.