Relativistic electrons are an important space weather hazard, being a major source of radiation damage to satellites and posing a risk to humans in space. We use approximately 20 years of data from the US Global Positioning System (GPS) satellite NS41 to determine the characteristics of the geomagnetic storms that lead to the largest relativistic electron fluxes in GPS orbit. The largest CME-driven events are associated with the solar wind having negative excursions of the IMF Bz with minimum values of ~-14 nT two hours prior to zero epoch, defined as the time of the minimum in the Dst index and strong Dst minima, reaching ~-130 nT at zero epoch. In contrast, events driven by high speed solar wind streams (HSSs) are associated with smaller negative excursions of IMF Bz with minimum values of ~-4 nT two hours prior to zero epoch and moderate Dst minima, reaching ~-60 nT at zero epoch. Compared with HSS-driven events, peak E = 2.0 MeV fluxes associated with CME-driven events are larger by factors of 1.3 at L=4.5 and 2.4 at L=6.5. Both the CME- and HSS-driven events are associated with enhancements in the solar wind number density and pressure prior to zero epoch. Following zero epoch the solar wind number density and pressure become low and substorm activity is enhanced for several days.

Relativistic electrons cause internal charging on satellites and are a significant space weather hazard. In this study we analyse approximately 20 years of data from the US Global Positioning System (GPS) satellite NS41 to determine the conditions associated with the largest daily averaged fluxes of E = 2.0 MeV relativistic electrons. The largest flux events at L = 4.5 and L = 6.5 were associated with moderate to strong coronal mass ejection (CME)-driven geomagnetic storms. However, the majority of the fifty largest flux events at L = 4.5 (30 out of 50) and L = 6.5 (37 out of 50) were associated with high speed solar wind streams from coronal holes. Both solar drivers are thus very important for relativistic electron flux enhancements in GPS orbit. The 1 in 3 year flux level was not exceeded following any of the fifteen largest geomagnetic storms as monitored by the Dst index (Disturbance storm time index), showing that the largest geomagnetic storms, most often associated with extreme space weather, do not result in significantly larger relativistic electron flux events in GPS orbit. The datasets include a summary plot of the month associated with the largest flux of 2.0 MeV electrons in GPS orbit during the study period (Figure 1) and a summary plot of the month associated with the largest geomagnetic storm during the study period (Figure 2). The fifty largest 2.0 MeV flux events at L = 4.5 as a function of the minimum Dst of the associated storm are provided in Figure 3.csv, the peak 2.0 MeV electron fluxes associated with the fifteen largest geomagnetic storms at L = 4.5 as a function of the minimum Dst of each of the storms are provided in Figure 4.csv, and the fifty largest 2.0 MeV flux events at L = 4.5 and the sunspot number are provided as a function of time in Figure 5_events.csv and Figure_5_sunspots.csv respectively. The characteristic widths of the fifty largest flux enhancements at L = 4.5 and L = 6.5 are provided in Figure 6.csv. Finally, the fifty largest 2.0 MeV flux events at L = 6.5 as a function of the minimum Dst of the associated storm are provided in Figure 7.csv, the peak 2.0 MeV electron fluxes associated with the fifteen largest geomagnetic storms at L = 6.5 as a function of the minimum Dst of each of the storms are provided in Figure 8.csv, and the fifty largest 2.0 MeV flux events at L = 6.5 and the sunspot number as a function of time are provided in in Figure 9_events.csv and Figure_9_sunspots.csv respectively. The research leading to these results has received funding from the Natural Environment Research Council (NERC) grants NE/V00249X/1 (Sat-Risk), NE/X000389/1 and NE/R016038/1.