Electrons of energy several MeV or greater have been implicated in the failure and malfunction of geostationary spacecraft. It is therefore important to be able to specify and even forecast the fluxes of these particles during and following geomagnetic storms. A first step is the understanding of their relationship to lower-energy electrons that can already be well modeled. It is therefore the goal of this paper to examine the relative time, spatial, and spectral relationships between 1.5 MeV electrons and intermediate energy electrons down to about 100 keV. For the November 1993 geomagnetic storm we find that electrons from about 100 keV to 1.5 MeV at GEO can be conveniently characterized by a power law spectrum and that the slope and intercept of this spectrum vary in systematic ways during the storm. This suggests the possibility of developing prediction filters or artificial neural networks, driven by a storm activity indicator (such as Dst), local time and a lower-energy electron flux, to specify the energetic electron spectral characteristics. We further find that local time diurnal effects are an important contributor to the apparent time delay of the recovery of energetic electrons and when these effects are considered the recovery phase enhancement is nearly uniform across the spectrum. This paper will report the spatial and temporal morphology of these intermediate to energetic electrons, their characterization by a power law and the variations of the power law slope and intercept throughout the November 1993 storm. These temporal, spatial, and spectral properties suggest that the recovery phase enhancement is due to the entry of the intermediate energy electrons from the geomagnetic tail as part of the storm injection process. We also discuss our success at building an Artificial Neural Network system to specify the storm time energetic electron flux spectra. ¿ 1998 American Geophysical Union