N. Romanova, V Pilipenko, O. Kozyreva, and N. Yagova romanova@ifz.ru, pilipenk@augsburg

1. Relativistic Electrons, Geomagnetic Indices, and ULF wave activity in the Terrestrial Magnetosphere N. Romanova, V Pilipenko, O. Kozyreva, and N. Yagova romanova@ifz.ru,pilipenk@augsburg.edu

2. “Killer” electrons and satellite anomalies Relativistic electrons fluxes are known to be one of the main factors responsible for the high-orbit satellites malfunctions. Thus, theprediction of relativistic electronfluxes is very important for the space industry.

3. Threshold electron fluxes for the increased risk of the GEOanomaly occurrence

4. Mechanism of the acceleration of ~100 keV electrons supplied by substorms is revival of the idea of magnetospheric geosynchrotron. Pumping of energy into seed electrons is provided by large-scale MHD waves in a resonant way, when the wave period matches the multiple of the electron drift period, e.g. Appearance at GEO of relativistic electrons following storms resists definitive explanation. These electron events are not merely a curiosity for scientists, but they can have disruptive consequences for spacecrafts. While it has been known a general association between the solar wind velocity and electron enhancements, the wide variability of the response and the puzzling time delay (~1-2 days) between storm main phase and the response has frustrated the identification of responsible mechanisms. Some intermediary must more directly provide energy to the electrons? Rather surprisingly, ULF waves in the Pc5 band (~few mHz) have been suggested as a possible energy reservoir: the presence of Pc5 wave power after minimum Dst may be a good indicator of relativistic electron response [O’Brien et al., 2001]. Thus, in a laminar, non-turbulent magnetosphere the “killer” electrons would not appear!? Geosynchrotron:ULF waves = intermediary between the solar windand “killer” electrons during magnetic storms?

5. Algorithm of the ULF wave index construction • For any UT, magnetic stations in the MLT sector 05 – 15, and in the latitudinal range 60 - 75 CGM are selected. • Spectra of two detrended (cut-off 0.5 mHz) horizontal components are calculated with Filon’s method in 1-h time window. • The frequency range for the index definition is the Pc5 band (fL=3 mHz, fH=7mHz) – the range of the most intense fluctuations. • In order to discriminate broad-band and narrow-band variations we applied an algorithm based on the determination of “bump” above the linear fit to background “colored-noise” spectra in the range 1-8 mHz. As a result one obtains: • Noise spectral power (N) - the band-integrated area beneath the background spectra; • Signal spectral power (S) - the area of the bump above the background spectra; • Total spectral power(T) - T=S+N The summation is performed with respect to all N stations where the signal amplitude is above K*Bmax Global ULF wave index:

6. Space weather events in 1994 Surprisingly, the sustained intense increase of the relativistic electrons (E>2 MeV) fluxes up to ~104 is observed after the weak storm (Dst~-100nT), whereas the increase after the strong storm (Dst~-200nT) is much shorter and less intense (up to ~103 only). The electron behavior matches well the variations of the global ULF-index: after the first weak storm this index increases much more substantially and for a longer period than after the second strong storm! During the March-April 1994 storms geostationary satellites suffered numerous anomalies from the “killer” electrons. Relativistic electron flux has a time delay ~1-2 days with respect to the ULF-index. Thus, this index could be used as a “precursor” of the risk of GEO satellite anomalies?! ?

7. “Killer” electrons, solar wind velocity and ULF index • The interconnection between the relativistic electron fluxes and ULF wave index is high throughout all phases of solar cycle. However, the relevant correlation coefficient always appears to be lower than the correlation coefficient between the electron flux and solar wind velocity. • The ULF wave index and solar wind velocity are highly correlated

8. Cross-correlation between the electron flux variations, ULF-index, and solar wind velocity The cross-correlation function shows that the electron flux increases about a day after the enhancements of ULF wave activity and solar wind velocity Correlation between the ULF-index and electron flux somewhat increases for the time-integrated over pre-history ULF index values: Increase of correlation probably implies the occurrence of the cumulative effect, that is, the the long-lasting (with characteristic time ) ULF wave activity is important for the electron flux increase, but not just instant values!

9. Relationship between the relativistic electron flux and geomagnetic activity indices The electron flux is also highly correlated with the Dst and AE indices. Sometimesthe correlation with Dst is even higher than that with the ULF index. Geomagnetic disturbances, as well as an elevated level of ULF wave activity, precede the growth of relativistic electron flux for about two days on average

10. The new ULF wave index, analogous to geomagnetic indices, derived from ground-based and satellite observations in the Pc5 frequency band – the range of natural MHD resonators and waveguides, characterizes the turbulent level of the solar wind-magnetosphere-ionosphere system. The database for interval 1994-2001 is freely available to space community via anonymous FTP site for testing and validation:space.augsburg.edufolder:/pub/MACCS/ULF_Index/CD with ULF index database may be requested! We acknowledge the provision of • Noon-reconstructed electron fluxes provided by P. O’Brien. • GOES data from NOAA NSDC; • Ground magnetic data from INTERMAGNET, CPMN, MACCS, and Greenland arrays; • OMNI-2 database from NASA NSSDC;