The results in this paper were obtained with SAPPHIRE, a new auroral Doppler radar experiment designed to study meter-scale E region irregularities. SAPPHIRE is a dual 50-MHz continuous wave, phased array, multibeam, bistatic system which is capable of performing cross-beam measurements from two widely different directions. There are two transmitters, each of which probes the auroral electrojet plasma over a large spatial target grid area of multiple intersections that determine 16 scattering regions or cells. Initial observations using untapered antenna arrays showed a class of scatter characterized by a narrow power spectrum peaking at the same Doppler shift in all, or several, observing cells simultaneously. These are strong echoes ranging in lifetime from a few tens of seconds to a few minutes and occurring preferentially in the midnight and morning magnetic time sectors. The analysis showed that this scatter is strongly anisotropic in azimuth and comes from localized regions of spatially coherent large-amplitude plasma waves that produce mostly type III, but also type I and the rare type IV, radar auroras. By using many events and analyzing a large number of Doppler spectra, we found that type III echoes are the strongest observed, having on the average relative intensities at least 15 dB higher than the type I echoes. The observations relate to the ''short discrete radar auroras'' which are known to originate in spatially confined, dynamic plasma regions.

The possibility exists that the large free energy for instability in these active regions is provided from intense electric fields and/or very sharp electron density gradients, both expected to occur at times near the edges of discrete auroral arcs. Finally, the present results confirm that, because of the large dynamic range of radio auroral echoes, strong scattering regions lead to the complete domination, at times, by backscatter through antenna sidelobes. For the localized regions of strong type III and type I echoes, this means that the convectional 3-dB antenna beam width scale size of the scattering region is unrealistic. Obviously, this has important implications for the radar auroral experiments and the interpretation of observations. ¿ American Geophysical Union 1995