Application of Autoregressive Moving Average Linear Prediction Filters to theCharacterization of Solar Wind-Magnetosphere Coupling

摘要

Linear prediction filtering techniques have been used in studying the couplingprocesses between the solar wind and magnetosphere. Linear models were built and tested on the Bargatze data set, consisting of over 70 days of geomagnetic indices and solar wind data ordered in 34 intervals of increasing geomagnetic activity. Linear filtering techniques employing single-and multiple-input, autoregressive models predicted values of the magnetic index AL from solar wind data. The impulse response curves of the AL-coupling function groups showed amplitude peaks at 25 and 70 minutes, confirming results in previous studies. The separate peaks indicate responses corresponding to the driven and unloading time scales.