HR: 0830h
AN: SM51A-04    [Abstracts]
TI: Prediction of Auroral Electrojet AL Index
AU: * Tan, A
EM: arjun.tan@email.aamu.edu
AF: Alabama A&M University, 4900 Meridian St., Normal, AL 35762 United States
AU: Lyatsky, W
EM: lyatsky@cspar.uah.edu
AF: Alabama A&M University, 4900 Meridian St., Normal, AL 35762 United States
AB: We studied the possibility for improving the prediction of the auroral electrojet AL index that shows substorm activity. We used 15-min and hourly data for 1995 when high-resolution data for the solar wind and AL indices are available. For prediction of AL indices we used the Akasofu coupling function averaged for a previous time interval and corrected for non-linearity, and season/UT variations of the AL index. For 15-min mean AL indices our method is able to account for ~ 88% of variance (R ¯ 0.94) in AL index for 15-min prediction and for ~ 77.5% of variance (R ¯ 0.88) for 30-min prediction. For hourly AL indices and one-hour prediction, our method is able to account for up to 79% of the variance (R ¯ 0.89) in AL index. These results are higher than earlier results for predicted AL and AE indices. Main causes for high correlation between predicted and actual AL indices are accounting for both previous solar wind history and strong seasonal/UT variations.
DE: 2134 Interplanetary magnetic fields
DE: 2784 Solar wind/magnetosphere interactions
DE: 2788 Storms and substorms
SC: SPA-Magnetospheric Physics [SM]
MN: 2005 Joint Assembly