C.T. Russell and R.C. Elphic

Institute of Geophysics and Planetary Physics, University of California

Los Angeles, California 90024

Originally Published In: Geophys. Res. Lett., 6, 33-36, 1979.


Magnetic field measurements from the ISEE 1 and 2 spacecraft are examined in the vicinity of the magnetopause near local noon on a typical pass when the magnetosheath field is southward. The data clearly show evidence for patchy impulsive reconnection. The flux transfer rate for these events is at least of the order of 1 to 2 x 1012 Maxwells per second, and possibly greater. This rate is similar to rates deduced for magnetopause erosion events. Not only are these observations relevant to the substorm process, but the impulsive nature of the flux transfer events leads to boundary oscillations that could also be the source of long period magnetic pulsations in the outer magnetosphere.


The earth's magnetopause is constantly in motion. Part of this motion is an oscillation, sometimes regular, sometimes irregular, about its average position (Holzer et al., 1966; Heppner et al., 1967; Cummings and Coleman, 1968; Kaufmann and Konradi, 1969, 1973; Aubry et al., 1970, 1971; Ledley, 1971). The velocities of this motion usually far exceed that of the spacecraft traversing the boundary. Estimates of the velocity of these motions usually range in the tens of kilometers per second. To measure these velocities on an individual and more or less routine basis, was one of the prime objectives of the ISEE 1 and 2 spacecraft, which were launched into nearly identical orbits on October 22, 1977 with apogees of 23 Re initially near 10 LT. These spacecraft carry a sophisticated complement of particles and fields instrumentation, but for the purposes of this discussion we will concentrate on the magnetic fields measurements performed by the UCLA fluxgate magnetometer (Russell, 1978).

The average position of the magnetopause is often found to move in response to a southward interplanetary field even when the dynamic pressure of the solar wind remains constant. This motion has been termed erosion of the magnetopause (Aubry et al., 1970) because the flux is transferred to the earth's magnetotail. The erosion of flux from the magnetopause and its appearance in the tail prior to substorms plus its eventual release from the tail provides a very consistent model for the substorms process (cf. Russell and McPherron, 1973). However, despite the overall consistency of the model with the flux transfer process, many have refused to accept it because certain "predictions" of the model were not confirmed. In particular the expected "classic" signature of a rotational discontinuity with a significant normal component was rare (cf. Sonnerup and Ledley, 1978) and the flows and heating of the plasma were not clearly evident (cf. Heikkila, 1978).

Haerendel et al. (1978) have made an exhaustive study of the dayside magnetopause using low resolution magnetic field and particle data from the HEOS-2 satellite and have come to the conclusion that reconnection of the magnetosheath and magnetospheric fields occurs at the polar cusps, not the equator, and impulsively, not in a steady state. Such behavior goes a long way in explaining previous difficulties in reconciling the observed and expected structure of the dayside magnetopause.

The ISEE spacecraft complement the HEOS-2 observations by providing measurements near the magnetospheric equator, and represents an improvement over the HEOS measurements in both angular and temporal resolution, in addition to its dual satellite nature. We have examined in detail four passes of ISEE 1 and 2 data at the subsolar magnetopause: one for northward magnetosheath fields and three with varying degrees of southward fields. While all three of these latter crossings exhibit the behavior of the type to be discussed below, we discuss only the crossing with the intermediate southward field, because its structure was most amenable to analysis, in that the events to be analyzed were regular and long-lived.


Vector measurements of the magnetic field are obtained at a rate ranging from 4 to 32 times per second depending on instrument state. The amplitude resolution of the instrument depends on gain and instrument mode but is ± 8 mg for the data used in this analysis. The magnetometer zero levels are checked periodically by flipping the spinaxis aligned sensor into the spin plane. Spacecraft associated fields were determined during a period of crossed spin axis operation by comparing data in the spin plane of one spacecraft with that along the spin axis of the other. Otherwise, the data processing of the two spacecraft has been completely independent. In regions where the fields are believed to be uniform the instruments agree to within 1/10g or 0.1%. The differences between ISEE 1 and 2 measurements visible in Figure 1 are entirely due to the ambient field.

To display the data we use a boundary normal coordinate system. The boundary normal system has its N-direction outward along the boundary normal. For the boundary normal we chose a model normal to a conic section with eccentricity 0.4 and focus on the earth. The L-direction is chosen along the projection of the solar magnetospheric Z-direction perpendicular to the magnetopause normal. The M-direction completes the right hand orthogonal system and points roughly opposite the earth's direction of rotation. We will also use 12-second averages of the data overlapped by 2/3 to restrict the bandwidth of the data to highlight the features of interest. Overlapping the averages eliminates signals outside the nominal bandwidth of the data, i.e. aliasing.

Fig. 1. Twelve-second averages of the magnetic field at ISEE-1 (heavy line) and ISEE-2 (light line) on November 8, 1977 in boundary normal coordinates. The spacecraft were at (10.16, -1.77, 5.07) Re in solar magnetospheric coordinates and separated by 299 km along the model boundary normal. Boundary normal coordinates are defined so that BN is outward along the boundary normal, BL is along the projection of Z GSM and BM completes the right-handed set. BL stands for boundary layer as identified by the LASL/MPI fast plasma analyzer.


Figure 1 shows ISEE 1 and 2 magnetic field measurements from 0200 to 0306 UT on November 8, 1977, orbit 5 inbound, in boundary normal coordinates. ISEE 1 measurements are shown with a heavy line; ISEE 2 with a light line. The location of the ISEE spacecraft was (10.16, -1.77; 5.07) Re in solar magnetospheric coordinates, and their sepa ration was (231, -181, 192) km. ISEE 1 was leading ISEE 2 by 299 km measured along the model magnetopause normal. The average magnetosheath field locally made an angle from 115° to 145° with the magnetospheric field during this period.

If one were to examine solely the BL and |B| traces, they would interpret this section of data to contain two partial entries into the magnetosphere: one at 0212 and one at 0236 UT together with a full entry, albeit with much back and forth motion, from 0244 to 0252. The BL-traces are nested as expected for an in-and-out boundary crossing at 0212 and 0236, and the field strength reaches magnetospheric levels. However, the BM and BN traces exhibit strange behavior at these times. The BM component increases in strength here while in the magnetosphere it is close to zero. The BN component has a very peculiar pat tern of an increase to large values, then a de crease to rather large negative values and then a return to zero. This behavior of BN is not restricted solely to these pseudo-magnetopause crossings. There are similar events at 0241 and in the magnetosphere at 0257 and 0300 UT.

When one sees large normal components as are evident here, one is tempted to conclude that reconnection is present. However, this explanation is not unambiguous for BN is not a measure of the instantaneous normal component. The true boundary normal direction can and, in fact, did change during this event. If the boundary develops a bulge, magnetosheath field lines must wrap around the bulge. The signature in the normal component, as the bulge passed by the spacecraft would be first an increase in the normal component reaching a maximum at the steepest slope of the bulge; then a return to zero as the bulge reached its maximum displacement. Then as the bulge receded away from the location of the spacecraft, the BN perturbation would go negative and eventually return to zero. This sequence of events is exactly what occurred here. The coherence of the variation at the two spacecraft indicates that the scale length of the bulge is much larger than the separation of the two spacecraft. We note that the variation in Bn in the magnetosheath is sensitive to oscillations in the N-M plane since the field is mainly along the M-direction. In the magnetosphere it is sensitive to L-N plane oscillations.

The signature in BL indicates a magnetopause entry at 0212 and 0236 UT. The difference between the traces is due to the different penetration depths of ISEE 1 and 2. The motion of the boundary has carried the magnetosphere part way across the satellite. This is confirmed by the occurrence of magnetospheric particle fluxes in both the Berkeley/Toulouse energetic particle experiment (G. Parks, R. Lin and K. Anderson, personal communication, 1978) and in the MPI/LASL fast plasma analyzer (G. Paschmann and S. Bame, personal communication, 1978). However, BM is increasing, rather than decreasing, at these times. Thus, the horizontal field exerts the greater stress and the "magnetospheric flux" is not vertical but rather is bent in the direction of the magnetosheath field.

Our explanation of these events is illustrated in Figure 2. The magnetosphere has undergone patchy, impulsive reconnection upstream of the satellites. After the reconnection ceased the flux tube was pulled out of the magnetosphere. As the stressed field lines attempted to relax (or shorten) themselves they swept up field lines ahead of them and made the bulge in the sheath field seen before flux tube entry. We cannot determine from a single event whether reconnection was initiated in the equatorial regions or near the southern polar cusp.

Fig. 2. Qualitative sketch of a flux transfer event. Magnetosheath field lines, slanted arrows, have connected with magnetospheric field lines, vertical arrows, possibly off the lower edge of the figure. As the connected flux tube is carried by the magnetosheath flow in the direction of the large arrow, the stressed field condition at the "bend" tends to relax, effectively shortening the flux tube and straightening the bend. Magnetosheath field lines not connected to the magnetosphere drape over the connected flux tube and are swept up by its motion relative to the magnetosheath flow.

Although the magnetic signature of a flux transfer event is, as a whole, three-dimensional, separate variance analyses of the individual flux tube boundary entries and exits reveal a two-dimensional structure. Such an analysis for the entry to the 0236 event, for example, yields variance directions are (0.976, -0.134, -0.174) and (0.205, 0.268, 0.941) respectively, in boundary normal coordinates. The latter direction is 20° from the undisturbed normal; it is tilted 12° up from the M-N plane, and 16° back (anti-sunward) from the L-N plane. This direction is completely consistent with the picture of a flux tube being lifted out and back by reconnection with the magnetosheath field. A similar analysis for the flux tube exit gives variance eigenvalues of (451, 16, 1)g 2 and the eigenvector in the minimum variance direction is (-0.435, -0.692, 0.576) in boundary normal coordinates. This is 55° from the undisturbed normal: down 260 from the M-N plane, and 44° forward (sunward) from the L-N plane, which is consistent with passage through the trailing edge of the flux tube depicted in Figure 2. It is noteworthy that the exit from the flux tube has a well-defined normal component of 18 ± 4g , which suggests the flux tube is connected to the surrounding magnetosheath. Thus, the trailing edge of the flux tube is not as well defined as shown in Figure 2, and further reconnection may be taking place as the flux tube moves.

As noted in the introduction impulsive reconnection has been postulated before to explain high latitude observations (Haerendel et al., 1978). This is the first observation of impulsive reconnection at low latitudes, but the reason for the lack of previous reports of this phenomenon is not that it is rare at low latitudes. Thus far we have seen evidence for flux transfer events on every orbit when the interplanetary field was southward, and high-resolution data was examined. We attribute our successful identification in art to having two satellites, in part to the use of boundary normal coordinates, and in part to the display of the data in cartesian coordinates.

In Figure 1 we have indicated the region where hot electrons in both the Berkeley/Toulouse and MPI/IASL instruments were seen. Energetic protons were also seen but they are not as ideal tracers of field lines. We note also that flowing magnetosheath plasma was seen throughout the 0212 and 0236 events.

The magnetopause at 0250 UT was not essentially different on this pass than on other passes. A plasma boundary layer was seen inside the magnetosphere, detached from the magnetopause from 0254:30 to 0258. For further details about this magnetopause crossing, and about these and other flux transfer events the interested reader is referred to Russell and Elphic (1978).


To approximate the amount of magnetic flux eroded during one of these flux transfer events, we first estimate the cross-section of the reconnected flux tube and the flux contained in this cross-section. The duration of the flux tube crossing provides an estimate of one dimension assuming it is convecting with the speed of the plasma. The size of the BN perturbation indicates the flux tube has similar dimensions in the orthogonal direction. Using the velocity measurements of Frank et al. (1978) we obtain an area for the flux tube of approximately 12 square Re and a total flux of approximately 2.9 x 1015 Maxwells for the first flux transfer event and 2.2 x 1015 for the second. At a rate of 3 per hour, these events represent an erosion rate of about 2 x 1012 Mx/sec. This rate is somewhat greater than that found by Aubry et al. (1970) when the magneto sheath had a southward component of about 30g . The velocity measurements of the MPI/LASL fast plasma analyzer are 0.70 of those of the Iowa detector during this event, but the directions agree within 10° (G. Paschmann, personal communication, 1978). If we use the MPI/LASL measurements our flux transfer rate is a factor of 2 less. Finally, since these events have finite dimensions they could be missed if the satellite were poorly situated. Thus, our estimate should be treated as a lower limit to the true rate of erosion at this time.


In the data examined to date which represent data obtained in a limited region of space and over limited range of magnetosheath field directions, we cannot comment on the absolute occurrence rate of these flux transfer events, nor pinpoint their exact location of initiation. However, the difficulties of past researchers in finding the expected "classical" signature of reconnection (cf. Sonnerup And Ledley, 1978) combined with the preponderance of evidence for reconnection (cf. Russell, 1976) suggests that flux transfer events provide much of the flux transport associated with magnetospheric dynamics. Further, our interpretation, based principally on the morphology of the temporal variations in field direction, agrees with those of Haerendel et al. (1978) based on the variations in field strength. The existence of multiple flux transfer events offer an attractive explanation of impulsive ion injection into the polar cusp, as observed on rocket flights by Torbert and Carlson (1976). Finally, it should be noted that the events on November 8 are exemplary structures and flux transfer events may, in general, be less obtrusive and well-defined than those on this day.

Flux transfer events lead naturally to boundary oscillations as they propagate down the flanks of the magnetopause, very much mimicking the behavior of the Kelvin-Helmholtz instability (cf. Dungey and Southwood, 1970). Since the amplitude of flux transfer events may also depend on the solar wind velocity, the simple dependence of the amplitude of magnetospheric magnetic pulsations on the solar wind velocity (Singer et al. 1977) should not necessarily be taken as evidence for the efficacy of the Kelvin-Helmholtz instability. Further work investigating the role of the north-south component of the interplanetary field must be undertaken.


We are grateful for receipt of data from the MPI/IASL fast plasma instrument of G. Paschmann and S. Bame and from the Berkeley/ Toulouse energetic particle experiment of K. Anderson. Discussions of these data with G. Paschmann, G. Parks and R. Lin have been extremely helpful. This research was supported by the National Aeronautics and Space Administration under contract NAS 5-20064.


Aubry, M.P., C.T. Russell and M.G. Kivelson, Inward motion of the magnetopause before a substorm, J. Geophys. Res., 75, 7018, 1970.

Aubry, M.P., M.G, Kivelson and C.T. Russell, Motion and structure of the magnetopause, J. Geophys. Res., 76, 1673, 1971.

Cummings, W.D. and P.J. Coleman, Jr., Magnetic fields in the magnetopause and vicinity at synchronous altitude, J. Geophys. Res., 73, 5699-5717, 1968.

Dungey, J.W. and D.J. Southwood, Ultra low frequency waves in the magnetosphere, Space Science Rev., 10, 672-688, 1970.

Frank, LA., K.L. Ackerson, R.J. De Coster and B.G. Burek, Three-dimensional plasma measurements within the earth's magnetosphere, Space Science Rev., in press, 1978.

Haerendel, G., G. Paschmann, N. Sckopke, H. Rosenbauer and P.C. Hedgecock, The frontside boundary layer of the magnetosphere and the problem of reconnection, J. Geophys. Res., 83, 3195, 1978.

Heikkila, W.J., Criticism of reconnection models of the magnetosphere, Planet. Space Sci., 26, 121, 1978.

Heppner, J.P., M. Sugiura, T.L. Skillman, B.G. Ledley and M. Campbell, OGO-A magnetic field observations, J. Geophys. Res., 72, 5417-5471, 1967.

Holzer, R.E., M.G. McLeod and E.J. Smith, Preliminary results from the OGO-1 search coil magnetometer: Boundary positions and magnetic noise spectra, J. Geophys. Res., 21, 1481, 1966.

Kaufmann, R.L. and A. Konradi, Explorer 12 magnetopause observations: Large-scale non-uniform motion, J. Geophys. Res., 74, 36099 1969.

Kaufmann, R.L. and A. Konradi, Speed and thickness of the magnetopause, J. Geophys. Res., 78(28), 6549-6568, 1973.

Ledley, B.G., Magnetopause attitudes during OGO-5 crossings, J. Geophys. Res., 76(10), 6736, 1971.

Russell, C.T., Reconnexion, in Physics of Solar Planetary Environments, (edited by D.J. Williams), p. 526, American Geophysical Union, Washington, D.C., 1976.

Russell, C.T., The ISEE-1 and 2 fluxgate magnetometers, ISEE Trans. Geoscience Electronics, in press, 1978.

Russell, C.T. and R.C. Elphic, Initial ISEE magnetometer results: Magnetopause observations, Space Science Rev., in press, 1978.

Russell, C.T. and R.L. McPherron, The magnetotail and substorms, Space Science Rev., 15, 205,

Singer, H.J., C.T. Russell, M.G. Kive-1-son and E.W. Greenstadt, Evidence for the control of Pc 3,4 magnetic pulsations by the solar wind velocity, Geophys. Res. Letters, 4, 377, 1977.

Sonnerup, B.U.0. and B.G. Ledley, CGO-5 magnetopause structure and classical reconnection, J. Geophys. Res., submitted, 1978.

Torbert, R.B. and C.W. Carlson, Impulsive ion injection into the polar cusp, in Magnetospheric particles and fields, (B.M. McCormac, editor), p. 47, 1976.

(Received August 31, 1978;

accepted November 3, 1978.)

Back to CT Russell's page More On-line Resources
Back to the SSC Home Page