Each SuperDARN radar operates on a fixed frequency selected between 8 and 20 MHz, and is therefore sensitive to decametre-scale ionospheric irregularities. They form a narrow beam which is stepped in 3.25 increments through 16 adjacent beam positions to provide an azimuth scan of 52 . The operating system of the radars is very flexible, allowing a wide range of operating modes. However, the usual mode chosen is to integrate the backscatter return in 70 range cells simultaneously for six seconds in each beam position in turn, therefore a scan of the entire field of view is completed in under 2 minutes. The first range and range cell size can be changed under software control, but typical values would be 270 km and 45 km respectively, giving a maximum range of 3420 km in this instance. For each range and beam cell combination, a 22 lag complex auto-correlation function is determined, from which the radial Doppler spectrum is deduced.
The backscatter observed with the radars arises from small-scale ionospheric irregularities in the E and F regions. Such irregularities are normally present over large regions of the auroral oval and polar cap, and the F-region irregularities move with the background plasma flow (Ruohoniemi et al., 1987). They thus act as scattering targets from which to determine high latitude plasma convection. It is likely that they are generated by cascading from larger scale irregularities, or from a number of plasma instability processes (Kelley et al., 1982; Tsunoda, 1988). Statistical studies (Leonard, 1991) of the backscatter returns observed using the Halley radar have established that irregularities are normally present throughout the high latitude ionosphere for all magnetic local times. The radars are thus highly effective tools for studying plasma motions at ionospheric altitudes.
The radars provide a complete description of the Doppler spectrum in the radial direction for each range/beam cell for which backscatter is observed. The parameters normally derived are the backscatter power, the line-of-sight Doppler shift and the Doppler spectral width. The latter is a measure of the velocity shear and velocity turbulence of the plasma. Ruohoniemi et al. (1989) have demonstrated that, for many occasions, it is possible to convert the resulting maps of the radial component of velocity into maps of horizontal vector velocity. The technique assumes a measure of spatial uniformity. Freeman et al. (1991) and Yeoman et al. (1992) have identified the conditions under which this assumption may cause errors to be significant.
Maps of true horizontal velocity can be obtained using two radars with overlapping fovs, and this is one of the major reasons for establishing SuperDARN. In the Antarctic, a second HF radar is planned to be deployed and commence operation at SANAE station in early 1995 as part of a tri-national (UK/USA/South Africa) collaboration known as the Southern Hemisphere Auroral Radar Experiment (SHARE, see Dudeney et al., 1994). In addition, Japanese scientists plan to deploy a further radar at Syowa station in January 1995, with a fov which overlaps that of SHARE. Thus we expect to be able to provide maps of horizontal vector flow from the Southern Hemisphere by the mid 1990s.
The Halley radar is managed and operated as part of the SESAME programme, whilst its data enter GGS as part of the DARN investigation, along with the other SuperDARN radars. Sanae data will be processed at BAS to produce merged vectors prior to submission of SHARE key parameters via DARN. The overlap of dual radar operation is in excess of 3 × 106 km2, extends from about 65 to 85 PACE geomagnetic latitude and over 3 hours of magnetic local time at 75 . The magnetic coordinate system used for conjugate comparison of radar data is based on the International Reference Geomagnetic Field for 1985, extrapolated to 1988, and a revision of the corrected geomagnetic coordinate system described by Gustafsson (1984) to produce well-conditioned mapping in the vicinity of the singularity at the South Pole (see Baker and Wing, 1989, for details).
The PACE radars have demonstrated their capabilities for geospace research by establishing the manner in which the ionospheric convection responds to IMF By changes in the vicinity of the cusp (Greenwald et al., 1990), and on the nightside (Dudeney et al., 1991). These studies have shown that for quasi-stationary IMF conditions the dayside convection is well represented by the Heppner and Maynard (1987) empirical patterns, but that on the nightside there are By related hemispherical asymmetries which are not contained in the empirical patterns. The dayside ionosphere starts to respond, on average, within 8 minutes of a change in By at 30 RE upstream; on the nightside the response is delayed by about 30 minutes. Baker et al. (1990) have used a combination of Halley PACE radar and the DMSP F9 spacecraft to identify the signature of the cusp in the radar data. More recently Baker et al. (1991) have shown that the cleft (low latitude boundary layer) exhibits a different signature. This opens up for the first time the possibility of studying, in detail, the spatial and temporal variability of the cusp/cleft. Pinnock et al. (1991, 1993) have provided excellent 2-D images of the localised regions of enhanced plasma flow. These bursts appear to be the ionospheric signatures of changes in the rate of magnetic merging in the vicinity of the cusp, and may be responsible in part for spawning polar patches - regions of high F-region plasma concentration which form in the cusp and then drift anti-sunward into the polar cap (Rodger et al., 1994a; 1994b; Rosenberg et al., 1993).