Fluctuations in the direction of propagation of intermittent low-frequency ionospheric waves

Abstract

Low-frequency (8-28 Hz), long-wavelength electrostatic waves in the ionospheric E region over northern Scandinavia are studied by using data obtained from an instrumented rocket having four probes mounted on two perpendicular booms. Two data sets are available, one for upleg and one for downleg conditions with somewhat different ionospheric parameters. The ionospheric plasma is unstable with respect to the electrostatic Farley-Buneman instability in both cases, but the DC electric field is somewhat enhanced during the downleg part of the flight. We find that the direction of wave propagation as given by the local normalized fluctuating electrostatic field vector varies randomly within an interval of aspect angles. The distribution of the directional change per time unit is determined. The waves propagate predominantly in the electrojet direction, but significant variations in directions can be found, both with respect to the magnetic field (the aspect angle) and with respect to the electrojet direction. Some of our results are in variance with related radar observations in the electrojet near the equator. Indications of significant spatial intermittency of the signal is demonstrated. Large-amplitude electrostatic fluctuations are confined to spatially localized regions and have a narrower aspect angle distribution with reduced directional fluctuations. We introduce an intermittency measure based on average excess time statistics for the record for the absolute value of the detected time-varying electric fields. We thus determine the average of time intervals spent above a prescribed amplitude threshold level. The results are compared with an analytical expression obtained for a reference nonintermittent Gaussian signal. The general analysis requires the joint probability density of signal amplitude and its time derivative to be known. The analytical models for quantifying the intermittency effects were tested by synthetic time series allowing study of the transition from non-Gaussian to Gaussian random signals.