The generation and propagation of atmospheric gravity waves observed during the Worldwide Atmospheric Gravity-wave Study (WAGS)

During the Worldwide Atmospheric Gravity-wave Study (WAGS) in October 1985, the EISCAT incoherent scatter radar was used to observe the generation of atmospheric gravity waves in the auroral zone in conjunction with a network of magnetometers and riometers. At the same time a chain of five ionosondes, an HF-Doppler system, a meteor radar and a radio telescope array were used to monitor any waves propagating southwards over the U.K.The EISCAT measurements indicated that in the evening sector both Joule heating and Lorentz forcing were sufficiently strong to generate waves, and both frequently showed an intrinsic periodicity caused by periodic variation in the magnetospheric electric field.Two occasions have been examined in detail where the onset of a source with intrinsic periodicity was followed by a propagating wave of the same period which was detected about an hour later, travelling southwards at speeds of over 300 m s⁻¹, by the ionosondes and the HF-Doppler radar. In both cases the delay in arrival was consistent with the observed velocity, which suggests a direct relationship between a source in the auroral zone and a wave observed at mid-latitude.     

Regions of strongly enhanced perpendicular electric fields adjacent to auroral arcs

In a joint campaign involving EISCAT, the Cornell University Portable Radar Interferometer (CUPRI), and sounding rockets, we have observed short-lived elevations of E-region electron temperatures, indicating the presence of strong electric fields. The use of a new pulse-code technique has considerably improved our EISCAT data in regions of low ionospheric electron densities. It has been found that strong and apparently short-lived enhancements of electric fields and associated E-region electron temperatures occur more commonly than long-lived ones. However, earlier EISCAT data with simultaneous optical recordings (and also some CUPRI radar data from the ERRRIS campaign) indicate that many of these events are, in fact, not short-lived, but occur in localized regions and are associated with drifting auroral forms. We show that the observed elevations of electron temperatures are created by very intense electric fields which can be found within narrow regions adjacent to auroral arcs. We discuss our observations against the background of models for electric field suppression or enhancement in the vicinity of auroral precipitation.     

Middle atmosphere and lower thermosphere processes at high latitudes studied with the EISCAT radars

The middle and upper atmosphere and the ionosphere at high latitudes are studied with the EISCAT incoherent scatter radars in northern Scandinavia. We describe here the investigations of the lower thermosphere and the E-region, and the mesosphere and the D-region. In the auroral zone both these altitude regions are influenced by magnetospheric processes, such as charged particle precipitation and electric fields, which are measured with the incoherent scatter technique. Electron density, neutral density, temperature and composition are determined from the EISCAT data. By measuring the ion drifts, electric fields, mean winds, tides and gravity waves are deduced. Sporadic E-layers and their relation to gravity waves, electric fields and sudden sodium layers are also investigated with EISCAT. In the mesosphere coherent scatter occurs from unique ionization irregularities. This scatter causes the polar mesosphere summer echoes (PMSE), which are examined in detail with the EISCAT radars. We describe the dynamics of the PMSE, as well as the combination with aeronomical processes, which could give rise to the irregularities. We finally outline the future direction which is to construct the EISCAT Svalbard Radar for studying the ionosphere and the upper, middle and lower atmosphere in the polar cap region.     

Worldwide atmospheric gravity-wave study in the European sector 1985–1990

Four campaigns of the Worldwide Atmospheric Gravity-wave Study (WAGS) have taken place in the European sector. On many occasions the onset of auroral activity in the evening and midnight sector, as indicated by EISCAT measurements of the electric field, was associated after a suitable delay with the detection of periodic ionospheric disturbance travelling southward over the U.K. at speeds between 500 and 1000 m s⁻¹. The velocity and wavelength of the TIDs corresponded to large-scale atmospheric gravity-waves. The characteristic periods of the travelling disturbances were similar to the intrinsic time scales of the auroral activity for periods of 40 min or more, but variations on a time scale of 20 min or less were strongly attenuated. The r.m.s. amplitude of the auroral electric field was proportional to the r.m.s. amplitude of the HF Doppler-shift associated with the gravity-wave. The time-lag between the onset of strong auroral activity and the arrival of the travelling disturbance over the U.K. was usually about an hour, suggesting a source region about 2000 km north. Similar levels of activity in the afternoon did not appear to produce strong waves in the far field. This is possibly due to ion-drag in the daytime ionosphere although the effects of the lower sensitivity of the HF Doppler-network during daytime must also be considered.     

Sondrestrom and EISCAT radar observations of poleward-moving auroral forms

During many magnetospheric substorms, the auroral oval near midnight is observed to expand poleward in association with strong negative perturbations measured by local ground magnetometers. We show Sondrestrom and EISCAT incoherent scatter radar measurements during three such events. In each of the events, enhanced ionization produced by the precipitation moved northward by several degrees of latitude within 10–20 min. The electric fields measured during the three events were significantly different. In one event the electric field was southward everywhere within the precipitation region. In the other two events a reversal in the meridional component of the field was observed. In one case the reversal occurred within the precipitation region, while in the other case the reversal was at the poleward boundary of the precipitation. The westward electrojet that produces the negative H-perturbation in the ground magnetic field has Hall and Pedersen components to varying degrees. In one case the Hall component was eastward and the Pedersen component was westward, but the net magnetic H-deflection on the ground was negative. Simultaneous EISCAT measurements made near the dawn meridian during one of the events show that the polar cap boundary moved northward at the same time as the aurora expanded northward at Sondrestrom. Most of the differences in the electrodynamic configuration in the three events can be accounted for in terms of the location at which the measurements were made relative to the center of the auroral bulge.     

Interference of tidal and gravity waves in the ionosphere and an associated sporadic E-layer

Observations made on 10 July 1987 with the EISCAT UHF radar are presented. The F-region measurements of both electron density and field-aligned ion velocity show that an upward propagating gravity wave with a period of about 1 h is present. The origin of the gravity wave is probably auroral. The E-region ion velocities show a tidal wave and both upward and downward propagating gravity waves. The gravity waves have three dominant periods with a possible harmonic relationship and similar vertical wavelengths. These waves are either reflected at a single reflection level, ducted between two levels, or they are generated in a non-linear interaction between gravity and tidal waves. The E-region electron density is dominated by particle precipitation. After a short burst of more intense precipitation, a sporadic E-layer forms at 105km and then disappears 40min later. Within this time, the layer rises and falls by a few kilometres, following closely the motion of a convergent null in the velocity profile. We suggest that the formation and destruction of this layer is controlled by both the precipitation, which indirectly provides a source of metal ions through charge exchange, and the superposition of gravity waves and the tidal wave.     

The E-region Rocket/Radar Instability Study (ERRRIS): scientific objectives and campaign overview

The E-region Rocket/Radar Instability Study (Project ERRRIS) investigated in detail the plasma instabilities in the low altitude (E-region) auroral ionosphere and the sources of free energy that drive these waves. Three independent sets of experiments were launched on NASA sounding rockets from Esrange, Sweden, in 1988 and 1989, attaining apogees of 124, 129 and 176km. The lower apogee rockets were flown into the unstable auroral electrojet and encountered intense two-stream waves driven by d.c. electric fields that ranged from 35 to 115 mV/m. The higher apogee rocket returned fields and particle data from an active auroral arc, yet observed a remarkably quiescent electrojet region as the weak d.c. electric fields (~ 10–15 mV/m) there were below the threshold required to excite two-stream waves. The rocket instrumentation included electric field instruments (d.c. and wave), plasma density fluctuation (δn/n) receivers, d.c. fluxgate magnetometers, energetic particle detectors (ions and electrons), ion drift meters, and swept Langmuir probes to determine absolute plasma density and temperature. The wave experiments included spatially separated sensors to provide wave vector and phase velocity information. All three rockets were flown in conjunction with radar backscatter measurements taken by the 50MHz CUPRI system, which was the primary tool used to determine the launch conditions. Two of the rockets were flown in conjunction with plasma drift, density, and temperature measurements taken by the EISCAT incoherent scattar radar. The STARE radar also made measurements during this campaign. This paper describes the scientific objectives of these rocket/radar experiments, provides a summary of the geophysical conditions during each launch, and gives an overview of the principal rocket and radar observations.     

Observations of auroral E-region plasma waves and electron heating with EISCAT and a VHF radar interferometer

Two radars were used simultaneously to study naturally occurring electron heating events in the auroral E-region ionosphere. During a joint campaign in March 1986 the Cornell University Portable Radar Interferometer (CUPRI) was positioned to look perpendicular to the magnetic field to observe unstable plasma waves over Tromsø, Norway, while EISCAT measured the ambient conditions in the unstable region. On two nights EISCAT detected intense but short lived (< 1 min) electron heating events during which the temperature suddenly increased by a factor of 2–4 at altitudes near 108 km and the electron densities were less than 7 × 10⁴ cm⁻³. On the second of these nights CUPRI was operating and detected strong plasma waves with very large phase velocities at precisely the altitudes and times at which the heating was observed. The altitudes, as well as one component of the irregularity drift velocity, were determined by interferometric techniques. From the observations and our analysis, we conclude that the electron temperature increases were caused by plasma wave heating and not by either Joule heating or particle precipitation.     

On the development of folds in auroral arcs

The development of an auroral arc in the midnight sector, from diffuse to discrete with subsequent large scale folding, is studied with the aid of several ground-based observations, including incoherent scatter radar, and data from a HILAT satellite pass. Ion drift velocities in the F-region, as measured by EISCAT, were consistently eastward throughout and after the whole period of development, whilst the ion temperature showed two large enhancements just prior to the appearance of the main auroral fold. The fold moved eastwards and crossed the EISCAT antenna beam, appearing as a short-lived spike in electron density at altitudes between about 100 km and 400 km. The spike in electron density came progressively later at higher altitudes. The observations are interpreted as the result of enhanced convection in the ionosphere and in the magnetosphere. The auroral arc folding is suggested to be caused by the Kelvin-Helmholtz instability in a velocity shear zone in the magnetosphere.     

Model predictions of the occurrence of non-Maxwellian plasmas,and analysis of their effects on EISCAT data

The recent identification of non-thermal plasmas using EISCAT data has been made possible by their occurrence during large, short-lived flow bursts. For steady, yet rapid, ion convection the only available signature is the shape of the spectrum, which is unreliable because it is open to distortion by noise and sampling uncertainty and can be mimicked by other phenomena. Nevertheless, spectral shape does give an indication of the presence of non-thermal plasma, and the characteristic shape has been observed for long periods (of the order of an hour or more) in some experiments. To evaluate this type of event properly one needs to compare it to what would be expected theoretically. Predictions have been made using the coupled thermosphere-ionosphere model developed at University College London and the University of Sheffield to show where and when non-Maxwellian plasmas would be expected in the auroral zone. Geometrical and other factors then govern whether these are detectable by radar. The results are applicable to any incoherent scatter radar in this area, but the work presented here concentrates on predictions with regard to experiments on the EISCAT facility.     

Current density models of the eastward electrojet derived from ground-based magnetic field and radar measurements

Correlated studies of the eastward auroral electrojet using EISCAT radar data and groundbased magnetic field observations from a meridional chain of five stations have been performed during the years 1987 and 1988. Three different models of current distributions—the line-type, the current sheet and the parabolic model—have been tested for their applicability in estimating the current density of the electrojet. The model employing a parabolic cross-section of the current density provides the best results, both from the magnetic profile and from the comparison of magnetic field and radar current density estimates. Current estimates from magnetic field observations are systematically 15% higher than those from EISCAT readings. This discrepancy has been attributed to the induction effect.     

A study of the vertical motion field near the high-latitude summer mesopause during MAC/SINE

We present the results of high-resolution observations of the vertical velocity field obtained with the EISCAT and SOUSY VHF radars near the high-latitude summer mesopause during the MAC/SINE campaign in northern Norway in 1987. The data reveal an energetic motion field with maximum amplitudes of ~ 10 m/s and characteristic periods of ~5–30 min. Motions exhibit a high degree of vertical coherence and a quasi-periodic structure, with typical durations of 5–10 cycles. Estimates of the mean vertical velocity are downward at lower levels and are near zero or positive at greater heights. The mean vertical velocity variance is found to be ~5 m²/s², consistent with other high-latitude measurements. Frequency spectra computed for each radar are found to exhibit considerable variability, while vertical wavenumber spectra are seen to be somewhat variable in amplitude and to have slopes approaching −3 at lower wavenumbers. These results are suggestive of an energetic spectrum of gravity wave motions near the mesopause that has a large vertical flux of wave energy, that may have observed wave frequencies differing significantly from intrinsic frequencies due to Doppler shifting by large horizontal winds, and that is consistent with the separability of the frequency and wavenumber dependence of the motion spectrum and with gravity wave saturation at sufficiently small vertical scales.     

Magnetometer and incoherent scatter observations of an intense Ps 6 pulsation event

In the morning sector of 21 April 1985, during the recovery phase of a geomagnetic storm, a Ps 6 pulsation event was recorded by the EISCAT magnetometer cross in northern Scandinavia. Simultaneously, the EISCAT incoherent scatter radar measured E- and F-region plasma parameters with a latitudinal scanning program. Electric fields and height-integrated Hall and Pedersen conductivities are derived. Two-dimensional patterns of these quantities are constructed for one Ps 6 period. The conductance patterns closely resemble the typical auroral forms of eastward drifting Ω bands with low and high conductances at the northern and southern edges of the scanned area, respectively. From the equatorward region a tongue of high ionization extends poleward into the dark area. The location of the maximum southward current is slightly displaced towards the west from the centre of the conductance tongue. The east-west disturbance electric field points towards the tongue; the north-south fields are enhanced outside and reduced inside the high conductance region. As has been previously suggested, the observations can be explained with a model which superposes currents caused by conductance variations and electric fields. Both effects need to be taken into account for this event. The current structures move within a few degrees in the direction of the background E×B drift, but their speed is about 15% lower than the average F-region plasma drift.     

Formation of ionisation layers responsible for blanketing Es during daytime counter-electrojet over magnetic equator

A model using photochemistry and transport due to electric fields and gravity wave winds has been used to explain the formation of ionisation layers observed over an equatorial station Thumba (dip 0°47′S) with rocket-borne Langmuir probes during two daytime counter-electrojet periods. These layers were seen as blanketing E_s-layers with an ionosonde at Thumba. Convergence of the metallic ions due to three-dimensional gravity wave winds and a westward electric field appears to be mainly responsible for the observed ionisation layer over the equator.     

Thermodynamics and electrodynamics of the auroral E-region during the D salvo of the MAP/WINE campaign

With the help of incoherent scatter (EISCAT) data the thermodynamics and electrodynamics of the auroral E-region north east of Andøya Rocket Range has been investigated between 1740 UT and 2040 UT on 31 January 1984. This time period covers the D salvo of MAP/WINE and the EISCAT incoherent scatter data comprise a useful supplement to interpret the rocket data. Good agreement has been found between the EISCAT temperatures and those derived from mass spectrometer data. Neutral wind velocity estimates from EISCAT and from a falling sphere rocket experiment are in satisfactory agreement for the zonal wind component, but disagree for the meridional component.     

Plasma convection and auroral precipitation processes associated with the main ionospheric trough at high latitudes

Intervals of F-region electron density depletions associated with the main (mid-latitude) ionospheric trough have been studied using latitude scanning experiments with the EISCAT UHF radar. From 450 h of measurements over a one year period at solar minimum (April 1986–April 1987) the local time of appearance of the trough at a given latitude is observed to vary by up to about 8 h. No seasonal dependence of location is apparent, but troughs are absent in the data from summertime experiments. A weak dependence of trough location on K_p is found, and an empirical model predicting the latitude of the trough is proposed. The model is shown to be more appropriate than other available quantitative models for the latitudes covered by EISCAT. Detailed studies of four individual days show no relationship between local magnetic activity and time of observation of the trough. On all four of these days, however, the edge of the auroral oval, evidenced by enhanced electron densities in the E-region, is found to be approximately co-located with, or up to 1° poleward of, the F-region density minimum. Simultaneous ion drift velocity measurements show that the main trough is a region of strong (> several hundred metres per second) westward flow, with its boundary located approximately 1°–2° equatorward of the density minimum. Within the accuracy of the observations this relationship between the convection boundary, the trough minimum and the precipitation boundary is independent of local time and latitude. The relevance of these results is discussed in relation to theoretical models of the F-reregion at high latitudes.     

Comparison of plasma flow and thermospheric circulation over northern Scandinavia using EISCAT and a Fabry-Perot interferometer

The EISCAT incoherent scatter radar, operating in a full tristatic mode, provided data on the ionospheric plasma drift above northern Scandinavia, during the 24 h period, 11 UT 25 November to 11 UT 26 November 1982. For the hours of darkness, 14 UT until 05 UT, observations of thermospheric winds were made by means of a ground-based Fabry-Perot interferometer (FPI) operated at Kiruna Geophysical Institute (21° E, 68° N). During this period, the radar observations describe well the ebbing and flowing of regions of strong convective ion flow associated with the auroral oval. As individual geomagnetic disturbances occur, the overall ion flow pattern intensifies and moves equatorward. The zonal thermospheric wind observed by the FPI responds rapidly to surges of the local ionospheric convection, while the meridional wind response is slower and apparently to much larger-scale features of the geomagnetic input to the high latitude thermosphere. From the data base, periods of strong heating of the ionospheric ions and of the thermospheric gas can be identified, which can be compared with Joule and particle heating rates deduced from the observations of ionospheric drifts, neutral winds, electron densities and auroral emission rates. A three-dimensional, time-dependent global thermospheric model is used to distinguish local and global features of the thermospheric wind field. Meridional and zonal wind components at 312 km may be theoretically derived from the EISCAT data using an appropriate model (MSIS) for neutral temperature. The EISCAT-derived meridional wind is within about 50 m s⁻¹ of the FPI observations throughout the period of joint observations. The EISCAT-derived zonal wind is systematically larger (by about 50%) than the FPI measurement, but the two independent measurements follow closely the same fluctuations in response to geophysical events until 03 UT, when the EISCAT solution is driven away from the FPI measurement by a sharp increase in both neutral and ion temperatures. Between 03 and 05 UT the EISCAT-derived zonal wind is 200–400 m s⁻¹ westward. Allowance for the neutral temperature rise would reduce the EISCAT values towards the very small zonal winds shown by the FPI during this period. We describe the relatively straightforward analysis required to derive the meridional wind from the radar data and the limitations inherent in the derivation of zonal wind, using the ion energy equation, due to the lack of precise knowledge of the background neutral temperature from the EISCAT data alone. For analysis of EISCAT ion drift observations at 312 km, the ground-based FPI temperature measurements do not improve the accuracy of the analysis, since the median altitude of the FPI measurement is probably in the range 180–240 km throughout the observation period. This median altitude and the temperature gradient both fluctuate in response to local geomagnetic events, while the temperature gradient may be considerably greater than that predicted by standard atmospheric models. When the neutral temperature is well known, or when there is a large enhancement of the ion temperature, the EISCAT-derived zonal wind exceeds the FPI measurement, but the consistency with which they correlate and follow ion-drag accelerations suggests that the differences are purely due to the considerable altitude gradients which are predicted by theoretical models.     

Nocturnal high-latitude E-region in winter during extremely quiet conditions

The quiet night-time E-region at high latitudes has been studied using the EISCAT UHF radar. Data from three subsequent nights during a long period of low magnetic activity are shown and typical features of electron density are described. The background electron density is observed to be 5·10⁹ m⁻³ or smaller. Two types of enhancements above this level are observed ; one is due to charged particle precipitation associated with the F-region trough and the other is composed of sporadic-E layers due to waves in the neutral atmosphere. The sporadic-E is observed to exist almost continuously and to exhibit a regular diurnal behaviour. In addition to the typical afternoon and morning sequential layers, a third major descending layer is formed at night after the passage of the F-region trough The afternoon layer disappears simultaneously with the enhancement of the northward trough-associated electric field and the night-time layer appears at high altitudes after the field has again been reduced to a small value. It is suggested that metal ions from low altitudes are swept by the electric field to the upper E-region where they are again compressed to the night-time layer. A set of steeply descending weaker layers, merging to the main night-time layer are also observed. These layers are most probably caused by atmospheric gravity waves. Theoretical profiles for molecular ions indicate that the strongest layers are necessarily composed of metal ions but, during times when the layers are at their weakest, they may be mainly composed of molecular ions.     

The amplitude of auroral backscatter—I. Model estimates of the dependence on electron density

A model of the auroral backscatter amplitude, in the form discussed by Uspensky and Oksman et al., has been derived for the radar geometry appropriate to joint observations by the PGI auroral radars at Karmaselga and Essoyla and the EISCAT incoherent scatter radar. The model shows how refraction effects cause a strongly non-linear dependence of backscatter amplitude on electron density in the E-region. It also explains why the macro aspect sensitivity for auroral radar operating at a frequency of about 45 MHz is only 1–2 dB per degree for aspect angles greater than 5°.     

D-region signatures of substorm growth phase and onset observed by EISCAT

High resolution electron density measurements by EISCAT during the pre-onset phase and onset of an auroral absorption substorm are used to investigate the characteristics of electron precipitation during these substorm phases. The development of the pre-onset phase is the result of a uniform increase of electron fluxes with energies of a few tens of keV, with no particularly hard component. The absorption spike observed at substorm onset contains fine structure when investigated at 10 s resolution, indicating a rapid hardening of the precipitating spectrum at the onset.