Application of nonlinear autoregressive moving average exogenous input models to geospace: advances in understanding and space weather forecasts

Abstract

Abstract. The nonlinear autoregressive moving average with exogenous inputs (NARMAX) system identification technique is applied to various aspects of the magnetospheres dynamics. It is shown, from an example system, how the inputs to a system can be found from the error reduction ratio (ERR) analysis, a key concept of the NARMAX approach. The application of the NARMAX approach to the Dst (disturbance storm time) index and the electron fluxes at geostationary Earth orbit (GEO) are reviewed, revealing new insight into the physics of the system. The review of studies into the Dst index illustrate how the NARMAX approach is able to find a coupling function for the Dst index from data, which was then analytically justified from first principles. While the review of the electron flux demonstrates how NARMAX is able to reveal new insight into the physics of the acceleration and loss processes within the radiation belt.