Experimental study of high-energy electron drift echoes observed on board ATS 5

Abstract

The high-energy electron data (0.7 ≲ E ≲ 2.6 MeV) which were obtained by the University of California at San Diego experiment on board ATS 5 during the period from September 1969 to January 1971 present a large number of drift echo events. A detailed analysis of ∼60 of these events has been performed in order to study their main characteristics. In all cases, the drift period at each energy follows the classical law T = T0(1 + x)/x(2 + x), where x = E/m0c². An event is therefore characterized by T0 (the drift period at E = 320 keV), by the amplitude of the modulation A equal to (Jmax − Jmin)/(Jmax + Jmin), and by its starting time t0. The events may occur at any local time, although there is a noticeable decrease of their occurrence between ∼2300 and ∼0200 LT. There is no clear relationship between the occurrence of these events and the parameters which characterize the magnetic activity (Kp, AE, and Dst); however, in 80% of the cases, ssc's are followed (within 1 min) by drift echoes. Empirical laws have been obtained which relate T0 with the AE index (at constant Kp) or with Kp (for small values of AE). These laws are in good agreement with the theoretical values of T0 which are obtained by using a Mead-Williams magnetospheric model in which the subsolar distance Rb is a function of Kp and the tail field BT is a function of AE. This model is consistent with recent experimental data (Coleman and McPherron, 1976) relating the midnight magnetic field intensity at the geostationary orbit and the AE index, provided one chooses the right coefficients of the Mead-Williams development. There is evidence that the large majority of drift echoes are not the consequence of a direct particle injection. Their quasi-sinusoidal flux variation is consistent with a redistribution of particle drift shells after a compression or an expansion of the magnetosphere. The amplitude of the modulation is of the order of the one which was predicted by Brewer et al. (1969). During the course of this study we found an empirical relation between BT and AE, valid for AE ≳ 40 γ: BT = 2 + 23 log (AE/33), where both BT and AE are expressed in γ (1 γ = 1nT).