Empirical Prediction Models of Relativistic Electrons Following Storms

A number of parameters of the solar wind and magnetosphere are correlated with the production of relativistic electrons following storms. These include the level of relativistic electrons at storm onset, seed electron flux, solar wind velocity and number density, the variation in velocity and number density, IMF Bz, AE and Kp indices, and ULF and VLF wave power. We use multiple regression analyses to determine which are the most predictive of flux when other variables are controlled for. We identified 176 storms (1992-2002) with at least 72 storm free hours after the minimum Dst. We obtained hourly averaged fluxes for relativistic electrons (> 1.5 MeV) and seed electrons (100 keV) from several spacecraft (Los Alamos National Laboratory geosynchronous energetic particle instruments). For each storm, we found the log10 maximum relativistic electron flux for each satellite 48-72 hours after the end of the main phase of each storm. No spacecraft was in operation for this entire period, so we standardized the observations to a common mean and standard deviation and then averaged over all available satellites in each hour. Each predictor variable was averaged over onset, main phase, and the 48 hours following minimum Dst. The rise in relativistic electron flux following storms is best explained by a set of variables rather than by one or two factors. Main phase seed electron flux, ULF, solar wind velocity and its variation, and IMF Bz following the storm were the most significant explanatory variables. We produced predictive models using the coefficients from the regression models, and assessed their effectiveness in predicting novel observations from an additional 44 storms during the same span of years (1992-2002). Correlations between observed values and those predicted by these empirical models ranged from .75-.79.