HR: 0800h
AN: SM41A-1129    [Abstracts]
TI: Electron Flux Prediction in the Radiation Belt via Autoregressive Models With Solar Wind Drivers
AU: * Presicci, M R
EM: presicci@colorado.edu
AF: Lasp/University of Colorado, 1234 Discovery Drive, Boulder, CO 80303-7814 United States
AU: Baker, D N
EM: Daniel.Baker@lasp.colorado.edu
AF: Lasp/University of Colorado, 1234 Discovery Drive, Boulder, CO 80303-7814 United States
AU: Rigler, E J
EM: jrigler@colorado.edu
AF: Lasp/University of Colorado, 1234 Discovery Drive, Boulder, CO 80303-7814 United States
AU: Weigel, R S
EM: Robert.Weigel@lasp.colorado.edu
AF: Lasp/University of Colorado, 1234 Discovery Drive, Boulder, CO 80303-7814 United States
AB: Several prior studies have suggested the solar wind velocity acts as a driver for the relativistic electron flux in the radiation belts. In this study, time series of ACE solar wind velocity measurements are included along with past electron flux measurements from SAMPEX in order to predict future electron flux. Since previous flux measurements are used to predict future flux, the autoregressive model is chosen. The inclusion of solar wind velocity measurements can also be modeled by an autoregressive equation. Since the entire model remains autoregressive, the Kalman filter state and measurement vectors are augmented to accommodate previous solar wind velocity along with the previous electron flux measurements. The filter processes these measurements as they become available at different rates. ACE measurements are updated more frequently than SAMPEX. The filter identifies the model coefficients of electron flux and solar wind velocity recursively, so that the coefficient estimates are optimal up to the time of the most recent measurement. The improvement in prediction accuracy over coefficient identification using Kalman filters without solar wind velocity is assessed. As a self consistent check, the autocorrelation function for the solar wind velocity with time, using actual ACE solar wind measurement data is computed justifying the modeling of solar wind velocity by an autoregressive equation.
DE: 2720 Energetic particles, trapped
DE: 2784 Solar wind/magnetosphere interactions
DE: 2788 Storms and substorms
SC: SPA-Magnetospheric Physics [SM]
MN: 2004 AGU Fall Meeting