Convection of closed magnetic tubes in the central part of the Earth's magnetospheric tail reduces their length and compresses the plasma contained in them. In a reasonable approximation the problem of plasma compression in a shrinking tube reduces to a classical hydrodynamical problem of a piston being pushed into a cylinder which has contained the gas at rest. In this case the top of the flux tube plays the role of the piston. As the piston is pushed in, a shock wave is known to separate from it. Between the piston and the shock front the gas is compressed and heated, while the gas ahead of the front is undisturbed. Similar processes occur in the plasma contained in the contracting flux tube and are accompanied by the formation in the tail of two shock waves which have separated from the neutral plane in the northward and southward directions, between which the plasma is hotter and denser than in the tail lobes. In the case of stationary magnetospheric convection, shock waves are also stationary. The distinctive plasma characteristics in the region between the shock waves and also the spatial position of this region suggest that in the real magnetosphere this region corresponds to the plasma sheet. We assess the validity of this hypothesis and its effectiveness in describing geophysical effects. We show that the plasma particle energy expected from this model corresponds to experimentally derived values. At the same time, within the framework of the MHD theory with isotropic pressure it is impossible to obtain the required high degree of compression during the transition of the plasma across the shock front (100 times or more).

By analyzing the constraints imposed by the laws of conservation of mass, energy, and momentum on plasma parameter jumps in a collisionless shock wave, we find that in order to achieve the observed degree of compression it is sufficient to have a small excess of transverse pressure in the plasma sheet over longitudinal pressure, which is consistent with observations. We have numerically simulated the motion of shock waves in convecting magnetic flux tubes. By postulating the relationship between the shock wave velocity and plasma parameters, we have calculated the form of the shock surface, which we interpret as the plasma sheet boundary. The line of intersection of the shock surface with the ionosphere is taken here as the polar boundary of the auroral oval. Results of calculations agree with the present-day understanding of the structure of the plasma sheet and of the auroral zone. The application we consider of this hypothesis lead us to conclude that this approach to describing phenomena associated with the plasma sheet boundary and with the auroral region is promising. ¿ American Geophysical Union 1993