Prediction of AU, AL, and AE indices using solar wind parameters

An empirical model that predicts the AU index, a measure of the Earth's east electrojet, derived from magnetometers in the Northern hemisphere, is introduced. In addition, we have improved the previous AL model (Li et al., 2007) and have combined it with the AU model to produce an AE model. All models are based on upstream solar wind and interplanetary magnetic field parameters that have been propagated to the magnetopause by a simple ballistic propagation scheme for the years 1995 to 2001. The AU model predicts the 10-min averaged AU index for the seven years 1995-2001 with a prediction efficiency (PE) of 0.716, a linear correlation coefficient (LC) between the AU index and the model of 0.846, and a root mean square (RMS) error of 39.3 nT. We have updated the AL model introduced in Li et al. [2007] using the same prediction functions used to predict AU but with different parameters. The new AL model predicts the seven year AL index with a PE of 0.715, an LC of 0.846, and an RMS error of 81.6 nT. Using AE = AU-AL, the AE index is predicted with a PE of 0.788, an LC of 0.888, and an RMS error of 95.7 nT. The better PE and LC of the AE model over AU and AL models is because AU and AL are better correlated then their prediction errors. It is also found that: (1) The F10.7 index modulates the growth of auroral electrojet indices; (2) AU and AL behave differently during geomagnetic storm main phases. AU can drop to a low level while the magnitude of AL does not drop as much; (3) the longer-averaged auroral electrojets indices can be predicted very well but shorter timescale variations are much less predictable; (4) auroral electrojet activity is strongly dependent on the upstream solar wind velocity and the interplanetary magnetic field but is only weakly dependent on the solar wind density.