Journa/ofAtmo.~pheric and Terrcsrrial Phyncs, Vol. 50, Nos 4/S, pp. 323-338, 1988. Printed in Great Britain. 0021-9169/X8 $3X0+ .W Pergamon Press plc The generation and propagation of atmospheric gravity waves observed during the Worldwide Atmospheric Gravity-wave Study (WAGS) P. J. S. WILLIAMS,’ G. CROWLEY, ‘* K. SCHLEGEL, 3 T. S. VIRDI, I. MCCREA,’ G. WATKINS, ’ N. WADE,’ J. K. HARGREAVES,~ T. LACHLAN-COPE,’ H. MULLER, 5 J. E. BALDWIN,~ P. WARNGR,~ A. P. VAN EYKEN,‘~ M. A. HAPGOOD’ and A. S. RODGER’ Department of Physics, University College of Wales, Aberystwyth, U.K. ‘Department of Physics, Leicester University, U.K. 3 Max-Planck-Institut fur Aeronomie, Katlenburg-Lindau, F.R.G. 4Department of Environmental Sciences, University of Lancaster, U.K. ‘Department of Physics, Sheffield University, U.K. 6Cavendish Laboratory, University of Cambridge, U.K. ‘Rutherford Appleton Laboratory, Didcot, Oxon, U.K. a British Antarctic Survey, Cambridge, U.K. zyxwvutsrqponmlkji (Receivedfor publication 8 December 1987) Abstract-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 aurora1 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- ‘, 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 aurora1 zone and a wave observed at mid-latitude. INTRODUCTION Since the seminal work by Hines over 25 years ago, atmospheric gravity waves in the thermosphere (AGWs) have been the subject of many theoretical and experimental studies (e.g. FRANCIS, 1975 ; RICH- MOND, 1978, 1979; HUNSUCKER, 1982). Many of the general properties of AGWs are well known, They are often conveniently divided into two classes: large scale and medium scale. Large scale waves have horizontal wavelengths of 500 km or more, and periods in the range 3&180 min. They are generated in the aurora1 zone and propagate over long distances with horizontal phase velocities of at least 300mss’. Medium scale waves, on the other hand, have wavelengths of a few hundred kilometres and periods less than 30 min. They are thought to originate both in aurora1 and in tropospheric sources, and propagate for shorter distances than the large scale waves. * Present address: NCAR, Boulder, Colorado, U.S.A. t Present address: Physics Department, Southampton University, U.K. AGWs of aurora1 origin are very important in transporting energy, initially of solar wind origin, from the high latitude ionosphere to other locations. For example, in the results of RISHBETHet al. (1985) it is seen that following a geomagnetic disturbance the first increase in neutral temperature observed at mid- latitude is associated with the equatorward propa- gation of wave activity. However, despite extensive study, the generation of these waves and their propa- gation characteristics are not fully understood. Three basic processes for generating aurora1 AGWs have been proposed, but the relative importance of the three is not certain. (a) Joule heating occurs when magnetospheric elec- tric fields drive a large scale current, the aurora1 elec- trojet, through the high latitude ionosphere. The heat- ing rate q5 = aPEf2, where rrp is the ionospheric Pedersen conductivity (which depends on electron concentration and on the ion-neutral and electron- neutral collision frequencies) and E’ is the effective electric field. (b) The Lorentz force follows from the interaction between the electrojet current and the geomagnetic field. 323