C. T. Russell, J. Berchem and J. G. Luhmann

  Institute of Geophysics and Planetary Physics, University of California, Los Angeles, CA 90024, U.S.A.

Originally published in:
Adv. Space Res., Vol. 5, No. 4, pp. 363-368, 1985.



We have examined the region of occurrence of flux transfer events for three distinct orientations of the interplanetary magnetic field: nearly horizontal in the solar magnetospheric equator, diagonally southward at 45o to the magnetospheric equator and nearly due south. For horizontal IMF conditions the FTE's occur in a horizontal band about + 6 RE wide. For diagonally southward IMF conditions, the FTE's occur in a diagonal swath about + 6 RE wide passing through the subsolar point. For duskward but nearly due southward IMF conditions, our observations reveal FTE's throughout the northern morning quadrant. These observations are consistent with a near equatorial source for flux transfer events and hence with component merging and not anti-parallel merging. These observations also help understand the energetic ion anisotropies seen in these events.



Reconnection is the process by which the solar wind transfers energy into the magnetosphere through the interconnection of the interplanetary and magnetospheric magnetic fields /1/. Reconnection is generally believed to control magnetospheric dynamics both through variations in the dayside reconnection rate at the magnetopause and on the nightside in the geomagnetic tail /2/. While the evidence for the reconnection prior to the launch of the ISEE spacecraft was overwhelming /3/, there had not been a direct observation of the reconnecting plasma. ISEE-1 and -2 provided direct evidence for two apparently distinct forms of reconnection: patchy reconnection, also called Flux Transfer Events (FTE's) /4,5/ and quasi-steady state reconnection /6,7,8/. However, this quasi-steady state reconnection is a rare or elusive phenomenon. Most often, essentially whenever the interplanetary magnetic field is southward, fluctuations appear in the magnetic field normal to the surface of the magnetopause /9/. About half the time these events have the classical signature that has been interpreted as the formation of a flux tube connected between the terrestrial magnetosphere and the shocked solar wind plasma in the magnetosheath /10/. Although it is clear that reconnection has taken place in these events, it is not immediately obvious where reconnection was initiated /3/. It may have been initiated at high latitudes as postulated by Haerendel et al. /11/ or at low latitudes as advocated by Cowley /12/. We note that it has been recently shown that the events treated by Haerendel et al. /11/ did in fact have all the characteristics of FTE's in the boundary layer just inside the magnetopause /13/.

We might expect reconnection to occur primarily at the subsolar or subflow point because here the flow speed of the reconnecting plasma is slowest and the pressure of the reconnecting fields the largest. Reconnection has also been postulated to occur along a tilted separator, whose orientation rotated with changes in the Y-Z GSM component of the IMF so that the separator lay along the direction of the magnetopause current /14/. However, this hypothesis has been challenged /15/. Finally, reconnection may occur preferentially at locations in which the magnetosheath magnetic field is nearly exactly antiparallel to the magnetospheric magnetic field /16,17/. Recently Luhmann et al. /18/ have used a realistic magnetosheath and magnetospheric field to determine the locations of these antiparallel fields on the magnetopause. Thus, we can have some confidence in the location of the merging sites predicted by this model. As illustrated in the lower left-hand panel of Figure 1, when the IMF is due southward, the expected merging line is along the equator but when the IMF is more duskward, as in the other three panels, the separator splits into two halves which move toward the southern polar cusp on the dawnside and toward the northern polar cusp on the duskside, for a duskward directed IMF. The reverse occurs for a dawnward IMF. It is the purpose of this paper to attempt to determine from the available evidence whether reconnection tends to occur near the subsolar point or away from it as predicted by antiparallel merging by examining the known properties of flux transfer events.

Fig. 1. The merging sites predicted by the antiparallel merging hypothesis for due southward interplanetary fields (O, O, -1), a horizontal duskward field (O, 1, O) and two intermediate southward and duskward fields (O, 1, -1) and (O, 1, -3) (from /18/). The strength of the interplanetary field is arbitrary in the gas dynamic model of the magnetosheath used. The contours show the regions in which the magnetospheric and magnetosheath fields have an anti-parallel component in steps of cos-1 = 0.1.



Fig. 2. Locations of flux transfer events projected in the Y-Z GSM plane (from /10/). The arrow gives the simultaneously measured direction of the IMF. Dashed arrows show the location of the reverse FTE's.

The orientation of the interplanetary magnetic field at the time of flux transfer events has been examined by Berchem and Russell /10/. Figure 2 shows the locations of the FTE's and the projected interplanetary magnetic field in solar magnetospheric coordinates from this study. When the actual distribution of the IMF during the study period is used to normalize the occurrence rate, there is a strong maximum for southward fields, but since the IMF generally lies close to the ecliptic plane, there are many FTE's with almost horizontal magnetic fields. The pattern in Figure 2 is confusing, but there is much order in the distribution when we separate these data further by the orientation of the IMF. Figure 3 shows the resulting distribution for fields pointing above -30o to the magnetospheric equator in the Y-Z plane (top panel), from -30o to -60o and from -60o to -90o. In constructing this figure, symmetry of the effects of magnetic fields pointing to dawn and dusk has been assumed. Thus a point on the dawnside with a dawnward pointing IMF has been plotted at the equivalent location on the duskside with a duskward field. In other words we have reflected the field direction and the FTE locations in the noon-midnight meridian.

Fig. 3. The locations of flux transfer events across the dayside magnetosphere projected on the Y-Z GSM plane together with the Y-Z projection of the interplanetary magnetic field. (Top panel). The locations for IMF directions above -30 to the GSM equator. (Middle panel). The location for IMF directions from -30o to -60o. (Bottom panel). The locations for IMF directions from -60o to -90o. All FTE's for dawnward IMF directions have been moved to equivalent positions for duskward fields by reflection across the Z-axis.

The top panel, which treats nearly horizontal magnetic fields, shows some very interesting properties. First, there are no events within a circle of 4 RE of the subsolar point. Second, there is a fairly uniform occurrence of FTE's outside this circle within a band about 10 RE wide centered on the equator. We can interpret this observation with a very simple model. Consider an interplanetary flux tube convecting to the subsolar point, lying parallel to the GSM equator. When it gets to the subsolar point, it can convect north or south. Let us assume it convects northward and begins to reconnect. The two halves of the interplanetary tube are pulled apart by convection and field tension. The part connected to the northern hemisphere is pulled to dawn. The part connected to the southern hemisphere is pulled to dusk. Had the tube been initially convected southward, the same result but in the southern hemisphere would have ensued. The absence of events near the subsolar point merely reflects their lack of mature development in this region so that the classical signature is not identifiable. We note that there appears to be no separation into a northern dusk and southern dawn merging region as predicted by anti-parallel merging. Rather there is a broad band of FTE's across the central equatorial region of the magnetosphere as expected from a subsolar merging line extending at maximum about + 4 RE along the noon meridian. The length of the merging region could be smaller than this. If so then the width of the FTE band would be due to the subsequent convection of the structures once formed.

The middle panel is equally as interesting. This shows the FTE occurrence pattern for diagonal or ~ 45o fields. The events occur in a broad diagonal band extending across the middle of the magnetosphere, roughly about 12 RE wide. This distribution is also consistent with a very simple model in which a diagonal flux tube is convected against the magnetopause and drifts either north or south, reconnecting beginning near the subsolar point.

The observed FTE's show no evidence for connection to the postulated antiparallel merging sites which as Figure 1 shows, lie to the upper right and lower left of the FTE band. It is somewhat difficult to gauge the length of the merging line here but it is probably about + 4 RE wide perpendicular to the FTE band. A possible reason for the scarcity of FTE's along the center of the diagonal band is that this is a region of steady reconnection. A narrow band of steady reconnection for diagonally southward fields could supply the potential drop that is apparently present in the magnetosphere for such IMF orientations but which is not supplied by FTE's alone.

The bottom panel shows the FTE's for nearly southward (but slightly duskward) fields. This slight tilt in the field direction apparently is enough to cause an asymmetry in the occurrence pattern, for most events were observed in the northern dawn quadrant. There is a circle of avoidance within about 4 RE of the subsolar point where there is but one event. Again we suspect that the events in this region are present, but that the events there do not meet our strict criteria for classification. Again we interpret this distribution as one in which the flux tube is convected to the subsolar region and propagates either dawnward or duskward. Because these fields are not precisely north-south they are pulled toward dawn in the north and toward dusk in the south thus causing some asymmetry. However, the observations are clearly consistent with a near equatorial source possibly about 8 RE across. Of course, the antiparallel merging mechanism also predicts a near equatorial merging site for this field orientation. If the scarcity of FTE's in the afternoon quadrant is not simply due to the statistics of small numbers, then a possible explanation is that in this region steady-state reconnection is occurring. Since southward directed IMF's lead to the largest drops in potential across the magnetosphere it would seem that such an occurrence of simultaneous steady-state merging is essential.

In summary then, the control of the location of occurrence of FTE's by the interplanetary magnetic field orientation clearly favors the origin of FTE's in a region about 4 RE radius around the subsolar point. Another set of data, that of energetic ion anisotropies, however first appears to confuse this picture /19/. Let us now examine those data to see how they are consistent with subsolar reconnection.



Strong energetic ion flows have been detected in flux transfer events /20/. These have been interpreted as being due to the leakage of energetic particles from the magnetosphere. The sign of the anisotropy, whether parallel or antiparallel to the magnetic field then determines the hemisphere to which the flux tube is connected /19/. Antiparallel flows indicate connection to the northern hemisphere and parallel flows connection to the southern hemisphere. The observed anisotropies are very confusing. For duskward pointing fields, the field lines in FTE's appear to be almost always connected to the northern hemisphere north of the GSM equator. For dawnward directed fields the FTE's appear to be almost always connected to the southern hemisphere on the dawnside and predominantly connected to the northern hemisphere on the duskside.

In order to make some sense out of this pattern let us treat the anisotropy data in much the same way as the field data in Figure 3. Let us combine the anisotropy occurrence diagrams for dawnward and duskward fields by reversing both sign of the Y-component of the field and the location of the observation. The sign of the anisotropy remains unchanged in this switch. We cannot combine the few data in the southern hemisphere with the northern hemisphere data because the anisotropy should reverse across the merging line. As Figure 3 suggests the merging line should be vertical along the noon-midnight meridian for horizontal fields and along the equator for vertical fields. The compacted data are shown in Figure 4 in GSM coordinates.

Fig. 4. The location of flux transfer events and the associated ion anisotropies. The locations of FTE's for dawnward IMF directions have been combined with the locations for duskward IMF's by reflection across the Z-axis. The sign of the anisotropy and the order of the BN signature denoted by the tail on the points was not altered. The tails on the data points indicate normal (direct) or reverse FTE signatures for upward and downward pointing tails respectively.

The most striking feature of Figure 4 is a broad diagonal swath of open circles across the center of the diagram indicating connection to the northern hemisphere, with a circle of avoidance around the subsolar point. This immediately invokes association with the middle panel of Figure 3. However, we do not know the times of the FTE's used by Daly et al. /19/, so we have not yet been able to check this association. We would expect northern hemisphere reconnection events to be connected principally to the northern hemisphere north of the reconnection line. Based on our earlier discussion of Figure 3, we would expect the region in which merging is occurring to be roughly perpendicular to the incoming field lines and passing through the subsolar region. The southerly terminus of this broad swath of open circles is consistent with such conjecture.

The puzzle in this distribution is not the broad swath along the diagonal but rather the region of southerly connection in the north toward dusk. This appears to be counter-balanced by a similar region of northerly connection (below the diagonal swath) near dawn but there are few statistics here. Returning to Figure 3, we see that the only IMF orientation that has events at high latitudes near dusk for duskward directed fields is a nearly horizontal orientation. Thus we expect that these FTE's with the anomalous southerly connections are simply FTE's occurring when the IMF was nearly in the equatorial plane which became connected in the north and were convected to the north. As mentioned earlier we expect that the dusk part of such flux tubes would be connected to the southern hemisphere. The "counter-balancing" dawn region of course should be connected to the northern hemisphere. We note that, in the antiparallel merging model, the flux tubes to dawnward and south of the northern dusk merging region would be connected to the northern hemisphere.

This is just the opposite of what is observed. These "anomalous" points are connected to the south and must therefore be on the dusk side, i.e., "below" the merging line which is what a north-south subsolar merging line would predict.

We note the existence of several exceptions to these general patterns. Some of these are probably due to situations in which the IMF was nearly north-south and the merging line was parallel to the equator and possibly slightly shifted from it. There seems to be no evidence for reconnection sites away from the subsolar region in this set of data.



We have examined the ordering of the occurrence of FTE's according to the directions of the IMF as well as the ordering of energetic ion anisotropies by the dawn-dusk component of the IMF. Comparing the spatial distribution of the ion anisotropies and FTE occurrences in the two studies we can make some inferences about the reconnection site and the origin of the complexities of the ion distribution. First, we deduce that the reconnection site is in the equatorial region regardless of IMF orientation as long as the IMF is southward. Second, we infer that the anomalous connection to the southern hemisphere of duskside (dawnside) northern hemisphere FTE's when the IMF is duskward (dawnward) occurs for mainly horizontal IMF conditions. This prediction must be checked in the near future.

It is tempting to speculate on the interpretation of the twist /21/ seen in the magnetic field of the FTE's, given in Figure 4. The twist of the FTE's is inferred from the signature in the normal component together with motion of the FTE. As an FTE rolls across the magnetopause, it will twist. The faster it rolls the more it will twist. The ion anisotropies in Figure 4 can be thought of as different samples of the same FTE. If we do so, then the signature in the normal component indicated by the tails on the points appears to vary along the FTE as if some points on the flux tube were rotating (rolling) faster than others. Again the data are scarce and thus this inference is not conclusive. A better study would be to quantify the degree of twist and the inferred motion (from 2 satellite studies) of all these, and other events. Until that is done there will be much room for speculation on the source of the twist.

Finally, the estimates of the contribution of FTE's to the potential drop of the magneto- sphere /9,21/ appears to fall far short of what is observed at active times. The distributions of FTE's shown in Figure 3 allow room for a band of simultaneous steady reconnection. We suspect that this band of steady reconnection grows in width as the IMF becomes more southward.



The authors would like to acknowledge many discussions and close collaboration with their various colleagues in their study of flux transfer events especially, S. W. H. Cowley, N. U. Crooker, P. W. Daly, R. C. Elphic, R. P. Rijnbeek, M. A. Saunders and D. J. Southwood. This work was supported by the National Aeronautics and Space Administration under NASA contract NAS 5-25772.



J. W. Dungey, Interplanetary field and the auroral zones, Phys. Res. Lett., 6, 47 (1961).

C. T. Russell and R. L. McPherron, The magnetotail and substorms, Space Sci. Rev. 15, 205 (1973a).

C. T. Russell, Reconnexion, in Physics of Solar Planetary Environments, (edited by D. J. Williams), p 526-540, American Geophysical Union Washington, D. C. (1976).

C. T. Russell and R. C. Elphic, Initial ISEE magnetometer results: Magnetopause observations, Space Sci. Rev., 22, 681-715 (1978).

C. T. Russell and R. C. Elphic, ISEE observations of flux transfer events at the dayside magnetopause, Geophys. Res. Lett., 6, 33-36 (1979).

G. Paschmann, B. U. 0. Sonnerup, I. Papamastorakis, N. Sckopke, G. Haerendel, S. J. Bame, J. R. Asbridge, J. T. Gosling, C. T. Russell and R. C. Elphic, Plasma acceleration at the earth's magnetopause: Evidence for reconnection, Nature, 282, 243-246 (1979).

B. U. 0. Sonnerup, G. Paschmann, I. Papamastorakis, N. Sckopke, G. Haerendel, S. J. Bame, J. R. Asbridge, J. T. Gosling and C. T. Russell, Evidence for magnetic field reconnection at the Earth's magnetopause J. Geophys. Res., 86, 10,049-10,067 (1981).

J. T. Gosling, J. R. Asbridge, S. J. Bame, W. C. Feldman, G. Paschmann, N. Sckopke and C. T. Russell, Evidence for quasi-stationary reconnection at the dayside magnetopause, J. Geophys. Res., 87, 2147-2158 (1982).

R. P. Rijnbeek, S. W. H. Cowley, D. J. Southwood and C. T. Russell, A survey of dayside flux transfer events, observed by the ISEE-1 and -2 magnetometers, J. Geophys. Res., 89, 786 (1984).

J. Berchem and C. T. Russell, Flux transfer events on the dayside magnetopause: Spatial distribution and controlling factors, J. Geophys. Res., in press (1984).

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

S. W. H. Cowley, The causes of convection in the Earth's magnetosphere: A review of developments during the IMS, Rev. Geophys. Space Phys., 20, 531-565 (1982).

R. P. Rijnbeek and S. W. H. Cowley, Magnetospheric flux erosion events are flux transfer events, Nature, 309, 135 (1984).

B. U. 0. Sonnerup, Magnetopause reconnection rate, J. Geophys. Res., 79, 1546 (1974).

S. W. H. Cowley, Comments on the merging of non-antiparallel magnetic fields, J. Geophys. Res., 81, 3455-3458 (1976).

N. U. Crooker, Dayside merging and cusp geometry, J. Geophys. Res., 84, 951-959 (1979).

N. U. Crooker The half-wave rectifier response of the magnetosphere and antiparallel merging, J. Geophys. Res., 85, 575~578 (1980).

J G Luhmann, R. J. Walker, C. T. Russell, N. U. Crooker, J. R. Spreiter and S S Stahara, Patterns of potential magnetic field merging sites on the dayside magnetopause, J. Geophys. Res., 89, 1739 (1984).

P. W. Daly, M. A. Saunders, R. P. Rijnbeek, N. Sckopke and C. T. Russell, The distribution of reconnection geometry in flux transfer events using energetic ion, plasma and magnetic data, J. Geophys. Res., 89, 3843 (1984).

P. W. Daly, D. J. Williams, C. T. Russell and E. Keppler, Particle signature of magnetic flux transfer events at the magnetopause J. Geophys. Res., 86, 1628 (1981).

M. A. Saunders, C. T. Russell and N. Sckopke, Flux transfer events: Scale size and interior structure, Geophys. Res. Lett., 11, 131-134 (1984).

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