Data taken during joint observations of the European incoherent scatter (EISCAT) radar and the Karmaselga auroral coherent radar (46.5 MHz) are combined to investigate the dependence of the coherent echo power upon the maximum electron density in the E region. For electron densities in the range (1–3)×1011 m−3, an increase in the echo power with the electron density was found. A few coherent echoes were observed for electron densities larger than 3×1011 m−3, but when such echoes occurred, their power was small even though the ionospheric electric field was strong, more than 30 mV/m. These measurements agree reasonably well with the model of Uspensky and Williams [1988] which takes into account ionospheric refraction and assumes that auroral coherent echo is a superposition of scattering from a thick layer of electrojet irregularities. Results of the comparison indicate that electron density at altitudes around 110 km can be roughly estimated solely from coherent radar data.

EISCAT measurements of the electric field in the auroral electrojet are compared with the signature of TIDs propagating equatorward as observed by an HF-Doppler network. At night-time the onset of auroral activity is usually followed by the arrival of a TID at lower latitude. Cross-correlation of the time variations of the electric field measured by EISCAT and the frequency offset recorded by the HF-Doppler system confirms a relationship between the auroral activity and the gravity wave, indicating both the travel time and the periodicity of the wave. The relationship is especially close under quiet conditions when the cross-correlation coefficient is typically 60%, significant at 0.1%. When the observed electric field is used as input to a thermosphere-ionosphere coupled global model it predicts the time signature of the observed HF-Doppler variation reasonably well but seriously underestimates the amplitude of the disturbance. Examination of this discrepancy may lead to a better understanding of the mechanisms involved in the generation and propagation of atmospheric gravity waves.

The case for a multi-antenna capability to serve the EISCAT Svalbard radar is presented under three headings: a) the need to make reliable measurements of plasma velocity, b) the ability to operate with increased sensitivity when electron concentrations are low and c) the flexibility to employ several different modes during an experiment.

Observations of the signals from distant short-wave transmitters, made by Granville Beynon in 1942 and 1943, suggested movements of or within the ionosphere. This work led directly to research in Travelling Ionospheric Disturbances and tidal modes in the neutral atmosphere. It is particularly appropriate that EISCAT has made such a substantial contribution in both fields.

The theory of auroral coherent echoes developed for VHF scattering by Uspensky et al. (1988, 1989) is applied to the interpretation of intensity and Doppler velocity slant range profiles of HF radar aurora. The theoretical model includes the effects of irregularity aspect sensitivity, ionospheric refraction of the radar beam, and the reception of signals from different heights. The predicted profiles of HF radar aurora are compared with Schefferville HF radar observations in the frequency interval of 9-18 MHz. Satisfactory agreement is found between theory and experiment for the intensity profiles. However, there are significant discrepancies for the Doppler velocity profiles. We discuss this lack of agreement in light of other recent observations.

The EISCAT Common Programme can be used in three ways to monitor tidal oscillations in the lower thermosphere. In Common Programme One (CPI) tristatic observations provide measurements of the ion-velocity vector at several heights in the E-region and one height in the F-region. In Common Programme Two (CP2) monostatic measurements give profiles of ion velocity in the E-region while tristatic measurements give continuous measurements of ion velocity in the F-region. From the ion velocities and the ion-neutral collision frequency, the vector of the E-region neutral wind can be determined and both east-west and north-south components of the diurnal, semi-diurnal and ter-diurnal oscillations can be identified. CP1 and CP2 also provide profiles of the field-aligned ion velocity, and these can be used to calculate the north-south component of the neutral wind without knowing the ion-neutral collision frequency, but the result is affected by any vertical component of neutral velocity. The three methods are compared and the advantages of CP2 demonstrated.

This paper studies the effect of ionospheric refraction upon auroral radar backscatter under conditions where the aspect angle appears far from ideal, i.e., when the unrefracted ray path trajectory is at least a few degrees from the perpendicular to the Earth’s magnetic field. It is found that wave trapping by curved electron density layers can cause ionospheric refraction as large as 20o, even at 150 MHz. This suggests that many so‐called off‐orthogonal VHF echoes are in reality due to backscattering at near‐orthogonal aspect angles, the discrepancy arising from increased ionospheric refraction by curved or tilted layers.

Measurement of the observed anti-correlation between the field-perpendicular component of F-region plasma velocity in the north-south plane and the downward, field-parallel component has been proposed as a way to determine the value of the O-O+ collision frequency. However, random noise errors in measurements of plasma velocity made at EISCAT may combine in analysis to induce a spurious anti-correlation between the derived values of these components which is hard to distinguish from any genuine anti-correlation.

Measurements of ion velocity from EISCAT Common Programme Two (CP-2) have been used to determine the profile of neutral winds in the high latitude E-region, following the method proposed by Brekkeetal. [(1974) J. geophys. Res. 78, 8235]. CP-2 has a number of advantages over the previously-used Common Programme One (CP-1) : in particular it provided a complete profile of neutral wind velocity whereas CP-1 only determined values at four separate heights. The time variation of neutral wind velocity at each height in turn was analysed into diurnal, semi-diurnal and ter-diurnal components. The height profile of amplitude and phase for each component was interpreted in terms of tidal modes and the results were compared with previous results from EISCAT and the Chatanika incoherent-scatter radar and the theoretical model derived by forbes [(1982) J. geophys. Res. 87, 5222 and 5241].

The effect of tilting of the E-layer on auroral backscatter is considered for the conditions applying in the vicinity of an auroral arc. Tilting modifies the degree of refraction, sometimes causing internal reflection, and this can explain a number of previously anomalous phenomena observed in auroral backscatter, including strong backscatter at apparently large aspect angles. As a result of tilts, refraction effects may be relevant over a wider range of frequencies than considered in the past, and the apparent aspect sensitivity and altitude of backscattered signals should be re-examined carefully.

Simultaneous observations of the Polar Geophysical Institute 83‐MHz auroral radar and the European Incoherent Scatter facility were used to study the coherent echo threshold phenomenon. It is shown that, for electric fields not far from the Farley‐Buneman instability threshold of ∼20 mV/m, the onset of echo occurrence was due to an electron density increase above some threshold value. This value was estimated to be ∼2.5 × 1011 m−3. For smaller electric fields of 5‐10 mV/m, echoes were not observed even for electron densities double the above threshold value. The importance of both factors, electric field and electron density, and some other effects are discussed in light of recent studies of coherent echo intensities and their dependence on plasma conditions at electrojet altitudes.

We examine the nature of flows in the nightside auroral zone ionosphere and their relationship with concurrent geomagnetic activity by studying a three-hour interval of ground-based radar, magnetometer array and spacecraft data. We find that the flows are bursty in nature, a characteristic previously reported, and that the bursts are related to a series of substorm electrojet intensifications initiated in the pre-midnight sector.

The F-region ion velocity measured by EISCAT in the auroral electrojets shows rapid variations, characterised by intervals of greatly increased velocity, usually in the same directions as the prevailing velocity associated with magnetospheric convection. These bursts of enhanced plasma velocity may be either short-lived but spread over several degrees of latitude and longitude, or relatively long-lived but narrow in extent and associated with auroral features drifting through the EISCAT beam. It is suggested that the short-lived, widespread events are driven directly by magnetic reconnection at the dayside magnetopause or they are associated with reconnection on the nightside during the expansion phase of a substorm. Several examples are shown where the interplanetary magnetic field remains southward for several hours but the plasma velocity measured by EISCAT is dominated by a series of bursts.

Incoherent scatter measurements at high latitudes show that in the height range between 100 and 130 km the velocity of the neutral atmosphere shows a prominent sinusoidal component with a 12-hour period. The variation of phase with height for this component is approximately constant from day to day, and corresponds to the predicted variation for the (2,4) mode of the global tide; at greater heights, the variation of phase with height is much slower and corresponds to the (2,2) mode. The amplitude of this semi-diurnal mode as measured at EISCAT shows a deep modulation with a period of about 53 hours, and this is interpreted as the result of a non-linear interaction in the mesosphere between the semi-diurnal tide and the most prominent planetary wave at these latitudes. A diurnal mode is also observed which shows even greater day-to-day variability. There is good qualitative agreement between the measurements made at EISCAT, Chatanika and Søndrestrøm, especially for the semidiurnal component. There is also reasonable agreement with the predictions made using the various models of the lower thermosphere, but the effect of ion on-drag driven by magnetospheric convection is difficult to model in detail, especially at Søndrestrøm. Tidal modes are also observed in the semi-diurnal variation of ion temperature.

When strong electric fields are applied to the auroral E-region they cause a modified two-stream instability which heats the electron population. The theory of electron-plasmon collisions [Robinson (1986) J. atmos. terr. Phys. 48, 417] predicts the relationship between the strength of the applied field and the corresponding increase in electron temperature at different heights. This relationship is non-linear and to make a valid comparison of the predicted increases in Te with the values observed accurate measurements must be made with a time resolution which matches the rapid variations in field strength. The measurements must also be corrected for the effects of neutral velocity, and care must be taken to minimise statistical bias when the results are averaged, including the effect of small errors in the assumed direction of the magnetic field. Results from EISCAT using alternating codes show a very good agreement between theory and observation. However, the significance of this comparison is limited by uncertainties in the assumed values of the electron-neutral collision frequency and the electron cooling rate.

Three campaigns of the Worldwide Atmospheric Gravity-wave Study 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, was followed after an hour or so by the arrival of a large-scale atmospheric gravity wave over the UK. Similar levels of activity in the early afternoon did not appear to produce waves in the far field, possibly as a result of ion-drag in the daytime ionosphere. The intrinsic time variation of the auroral activity determined the original time variation of the travelling disturbance, but waves with a period of about 30 min were strongly attenuated whereas waves with a period of about 60 min propagated to very great distances.

Measurements of plasma velocity along the magnetic field line by EISCAT provide an estimate of the meridional neutral wind in the E-region. During March and April 1988 such observations were made during the pre-midnight period on fourteen separate evenings and the results were used to fit the amplitude and phase of the semi-diurnal component of neutral velocity between 100 and 160 km altitude. Large variations in the dominating tidal modes were observed from day to day, during geomagnetically active periods. Smaller changes were observed during five successive days of quiet geomagnetic conditions. The most stable feature was found to be the phase at 110 km; on 10 out of 14 cases studied the maximum southward velocity was encountered at 1800 LT to within 30 minutes. On three successive days a phase-height variation corresponding to the (2, 4) tidal mode dominated below 130 km with phase varying less than ten minutes from day-to-day, and the phase-height variation of the (2, 2) mode was observed above 130 km. These observations indicate the presence of a stable tidal oscillation. However, even then the amplitude changed from day to day, which is explained as a modulation produced by a planetary wave having a two day period.

During the first Lower Thermosphere Coupling Study (LTCS), September 21–25 1987, data were recorded from the incoherent scatter radar sites at EISCAT, Millstone Hill, Sondrestrom, and Arecibo. These experimental facilities measured ionospheric parameters (Ne, Te, Ti, and plasma velocity) in the E and the F regions which have been used to determine the E region neutral wind and infer the neutral temperature in the height range 100–150 km. Propagating tides are clearly visible in some of the parameters, and the latitude structure and phase variations with height indicate the presence of at least the (2,2) and (2,4) global tidal Hough modes. The influence of geomagnetic forcing is also clearly present at high latitudes. The University College London‐Sheffield University three‐dimensional coupled thermosphere‐ionosphere model has been used to simulate this period of observation, by imposing tidal forcing at the lower boundary and magnetospheric forcing at high latitudes, in an attempt to interpret and understand the experimental data. Model simulations are able to predict where the signature of a particular tidal mode is likely to be observed in the respective responses of the temperature and wind structure. The numerical simulations predict the range of observed tidal amplitudes at mid and high latitudes, provided the tidal forcing functions imposed near the lower boundary of the model are larger (400 m geopotential height variation) than those inferred from linear tidal models. At high latitudes, the semi‐diurnal ion temperature response is driven by geomagnetic heating, rather than lower atmospheric tides and, as low as 110 km altitude, there is a large difference between ion and neutral temperatures during active conditions. Temperature and zonal wind tidal amplitudes observed at the low‐latitude site of Arecibo are consistently larger than those produced by any of the simulations.