Nonlinear dynamics of the magnetosphere and space weather

Summary form only given. The solar wind-magnetosphere-ionosphere system exhibits coherence on the global scale and such behaviour can arise from nonlinearity in the dynamics. The observational time series data have been used extensively to analyze the magnetospheric dynamics by using the techniques of phase space reconstruction. Analyses of the solar wind and auroral electrojet and Dst indices have shown low dimensionality of the dynamics and accurate prediction can be made with an input-output model. The predictability of the magnetosphere in spite of the apparent complexity arises from its being synchronized, in the dynamical sense, to the solar wind. The AE and Dst data are used to analyze the storm-substorm relationship based on the input-output model. This shows differences between the storm-time and non-storm substorms, and is interpreted in terms of loading-unloading and directly driven processes. The strong electrodynamic coupling between the different regions of the magnetosphere yields its coherent and thus low dimensional behaviour. The data from multiple satellites and ground stations are used to develop a spatio-temporal model that identifies the coupling between the different regions. These nonlinear dynamical models provide forecasting tools for space weather.