Spatio-Temporal Dynamics of the Magnetosphere during Geospace Storms

Abstract

Nonlinear dynamical models have became powerful tools for studying and forecasting magnetospheric dynamics driven by solar wind inputs. In this thesis, the techniques of phase space reconstruction from time series data are used to develop new methods for modeling and predicting the spatio-temporal dynamics of the magnetosphere. For these studies, new databases covering the solar maximum period were compiled to enable accurate modeling of the magnetosphere during intense geospace storms. The main contributions of the thesis are: Weighted Mean Field Model and Its Application to the Intense Storms. The nonlinear dynamical models of the coupled solar wind-magnetosphere system derived from observational data yield efficient forecasts of space weather. An improved version of the mean field model, derived from a set of nearest neighbors in the phase space reconstructed from the data, was developed by assigning weights to the nearest neighbors. A new correlated database was compiled and used to model and forecast the geospace storms of October-November 2003 and April 2002, and resulted in improved forecasts of the intense storms. Mutual Information Analysis of Spatio-Temporal Dynamics. The mutual information functions enable studies of the nonlinear correlations of dynamical systems. A high resolution database for a six month period of solar wind and ground-based magnetometer data from 12 high latitude stations was used to compute the mutual information functions representing the correlations inherent in the system. Using two different window lengths of 6 and 24 hr, the spatio-temporal dynamics was analyzed using these functions for the different stations. The spreads in the average mutual information show strong correlations with the solar wind changes and the time evolution of mutual information yields a westward expansion of the disturbed region, starting from the near midnight sectors. Modeling and Predictions of Spatio-Temporal Dynamics of the Magnetosphere. The spatial structure of the magnetospheric dynamics is crucial to space weather forecasting. The database of the magnetic field perturbations at 39 magnetometers belonging to the IMAGE and CANOPUS during year 2002 was used to study the spatio-temporal structure. A longitudinal sampling process utilizing the daily rotation of Earth was used to construct a two-dimensional representation of the high latitude magnetic perturbations. The nonlinear model was used to predict the spatial structure of geomagnetic disturbances during geospace storms. Results presented in this dissetation provide a comprehensive study of the magnetosphere using nonlinear data derived models. The new weighted mean field model, mutual information analysis and spatio-temporal dynamics advance our understanding of the solar wind-magnetosphere coupling. These results can be used to develop new and more detailed space weather forecasting tools.