The possibility is examined of obtaining information about the orientation and state of motion of the magnetopause current layer from the convection electric field, Ec=-v¿B, calculated from measured plasma velocities, v, and magnetic fields, B. Using AMPTE/IRM data from one particular traversal of a magnetopause rotational discontinuity as a guide, a new method has been developed for predicting the magnetopause normal vector, n. In its simplest form, the method consists of performing variance analysis on Ec data, or modifications thereof which take into account the motion of the magnetopause. The procedure is similar to the minimum variance analysis commonly performed on B data, except that the direction of maximum variance of Ec, rather than minimum variance of B is used as a predictor for n. In its most developed form, the method depends on the existence of a deHoffmann-Teller (HT) frame in which the electric field vanishes or nearly vanishes. Procedures have been developed for the determination of the initial velocity, vH To, and acceleration aHT, assumed constant over the analysis interval, of that frame.

When these effects are taken into account, an accurate n vector may often be obtained; in the trial case, error analysis and consistency tests indicate uncertainties of the order of 2¿--3¿, even though the variance in the normal magnetic field, B ⋅ n, in this crossing was large enough to render the minimum variance direction of B useless as a predictor of n. In general, it is not possible to determine the initial magnetopause normal speed, uno, from the electric field data but the analysis places certain, sometimes narrow, bounds on uno and on the normal magnetic field, B ⋅ n, the tangential electric field, Ec¿n, and the distance, h, along n traversed during the data interval, all of which are found to depend linearly on uno. For the trial case, the following results were obtained: -97<uno<-88 km/s; 0<B ⋅ n<2.4 nT; 0<‖ Ec¿n ‖<1.1 mV/m; 3036<h<3643 km (the actual magnetopause thickness was much less than h). The acceleration aHT in general has both a normal and a tangential component: for the test case (which had a duration of 69 s) both were of the order of 1 km/s2, the former corresponding to a deceleration of the inward magnetopause motion, the latter to a decrease and slight rotation of the tangential component of vHT, an effect that is shown to have a significant influence on the tangential stress balance.