JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 98, NO. A7, PAGES 11,299-11,318, JULY 1, 1993
The Galileo Earth Encounter: Magnetometer and Allied Measurements
M. G. KIVELSON,1,2 C. F. KENNEL,1,3 R. L. MCPHERRON,1,2 C. T. RUSSELL,1,2 D. J. SOUTHWOOD,1,4 R. J. WALKER,1 K K. KHURANA,1 P. J. COLEMAN,1,2 C. M. HAMMOND,1,2 V. ANGELOPOULOS,1,3 A. J. LAZARUS,5
R. P. LEPPING,6 AND T. J. HUGHES7
Abstract. The Galileo spacecraft flew by Earth on December 8, 1990, at high speed along a trajectory that traversed the magnetotail and the near Earth magnetosphere. Galileo's orbit through a region of the magnetotail from which limited data are available provided a unique opportunity to study a number of substorm-related phenomena. Several groups cooperated in collecting correlative data in order to take advantage of this special opportunity. Fortunately, geomagnetic conditions were rather disturbed during the entire day, and an interplanetary shock passed Earth when the spacecraft was in the magnetotail at about 30 RE geocentric distance. In this first report we provide an overview of the Galileo magnetometer observations from the crossing of the tail magnetopause at an antisolar distance of close to 100 RE through exit into the solar wind on the dayside. We link these measurements with correlative data from ground stations and from IMP 8 which was ideally located to serve as a monitor of the solar wind upstream of the bow shock. Based on our analysis, we present a time line of the important geomagnetic events of the day that we believe provides a framework for the full multi-instrument analysis of the flyby data. In this paper we use the observations to investigate aspects of the relationship between magnetotail dynamics and the separate intensifications of a multiple onset substorm inferred from ground-based data. The spacecraft spent 6 h downstream of lunar orbit, of which more than 4 h were spent outside of the plasma sheet in regions where traveling compressional regions (TCRs) should have been apparent. Although six substorm intensifications were recorded on the ground during this interval, we did not observe a detectable TCR or plasmoid for every intensification. Our interpretation has important implications for the description of substorm dynamics in the tail. We propose that the signatures associated with individual substorm intensifications are localized in dawn-to-dusk extent even at remote locations in the magnetotail, just as they are in the ionosphere, and that the tail disturbances associated with successive substorm intensifications step across the tail towards the dusk flank. This latter interpretation is appealing as it can explain the failure of Galileo to observe a signature associated with each intensification without invalidating the conclusion of ISEE 3 investigators that in the same region of the magnetotail at least one signature can be associated with each substorm viewed as a collection of individual intensifications. Plasmoidlike signatures with strong axial fields along the GSM y axis and parallel to the By of the interplanetary magnetic field (IMF) were present when the spacecraft was embedded close to the center of the plasma sheet. We interpret these signatures as flux ropes, that is, twisted magnetic structures with one end possibly tied to the ionosphere. The modeled structure yields j x B/jB << I which suggests that the flux ropes are magnetically force free to within the limitations of the model. We point out that plasmoids and flux ropes form a continuum of structures distinguished by the magnitude of By. Our observations lend additional support to the view that bipolar Bz signatures in the magnetotail may often be better described as flux ropes than as disconnected plasmoids. Our other principal results are only summarized in this paper; they will be discussed in greater detail elsewhere. They include (I ) additional evidence that the IMF By controls the lobe magnetic field only in the quadrants that are magnetically linked to the solar wind, and (2) evidence that the low-frequency response (the classic "sudden impulse" or SI signature) to a solar wind shock can be absent in the magnetic signature obtained within a high b plasma sheet. We believe that these observations will provide insight useful for improving phenomenological models of substorms.
On December 8, 1990, the Galileo spacecraft flew by Earth as part of a series of gravity-assist maneuvers designed to send the spacecraft on its way to Jupiter. Many spacecraft equipped to measure particles and fields have flown in orbit about Earth, but Galileo was only the second spacecraft to make a rapid flyby of the sort familiar from many planetary missions (the first was Giotto, as reported by Glassmeier et al. ). It is evident from Figure 1a that the encounter trajectory, traversed at a higher speed than typical for an orbiting spacecraft, was directed up the center of the magnetotail. The flyby occurred during a period of moderate geomagnetic activity that lends interest to this unique data set.
Our present knowledge of the magnetospheres of the large outer planets is from encounter type trajectories. In a dynamic system like a planetary magnetosphere, there is no way to distinguish spatial from temporal variations in a single spacecraft planetary flyby. Fortunately, in the Galileo Earth encounter we have been able to place our data in context in space and time using the extensive correlative data available from other spacecraft and from ground-based facilities, and the correlations have proved to be central to the analysis. Thus although the Earth encounter provides a lesson in data interpretation from a planetary flyby, it also underscores the ambiguity of interpretation that remains in examination of planetary flyby data from a single spacecraft unsupported by correlative measurements.
At the time of the Galileo flyby on December 8, 199O, the IMP 8 spacecraft was ideally located as a monitor of the solar wind conditions upstream of the Earth at a distance of about 38 RE from Earth and close to the Earth-Sun line. Mag- netometer data and plasma data (from the Massachusetts Institute of Technology (MIT) instrument) are used. Several ground-based facilities were operated in special modes in order to obtain relevant correlative measurements. In this report we include data from ground-based magnetometers including USGS (United States Geological Survey) stations, the Canopus Maria magnetometer chain (provided by the University of Alberta by courtesy of John Samson and Gordon Rostoker), the EISCAT (European Incoherent Scatter Radar) magnetometer cross in northern Scandinavia (provided by H. Luhr), the Hermanus observatory (provided by P. R. Sutcliffe), the Goose Bay station (provided by K. Hayashi), and stations of NGDC (National Geophysical Data Center) and the RGON network (Remote Geomagnetic Observatory Network). High time resolution magnetometer data from the Kakioka Observatory (provided by K. Yumoto) were available to identify substorm onsets through Pi 2 power. Additional data were provided by a variety of other sources, notably, the European Incoherent Scatter Radar Facility (EISCAT), the PACE (Polar Auroral Conjugate Experiment) conjugate radars run by the Applied Physics Laboratory and the British Antarctic Survey, the SABRE (Sweden and Britain Auroral Radar Experiment) radar run by Leicester University, the SAMNET (Subauroral Magnetometer Network) magnetometers run by York Univer- sity, the Greenland magnetometer chain (courtesy of E. Friis-Christensen), several records from the former USSR (courtesy of A. Zaitzev), and summaries of additional former USSR station observations (courtesy of Y. Galperin). Supplementary spacecraft data are also available from geosynchronous environmental monitoring spacecraft run by the Los Alamos National Laboratory, the Defence Meteorological Satellite Program spacecraft (provided by P. Newell) and from the CRRES (Combined Release and Radiation Effects Satellite) magnetometer (courtesy of H. Singer). These additional data will ultimately play an important part in the full elucidation of the Galileo fields and particle data from the flyby.
SCENE SETTING AN OVERVIEW OF THE MAGNETOMETER DATA
A description of the Galileo magnetometer instrument is given by Kivelson et al. . As the flyby of Earth provided unique opportunities for instrument calibration and testing, the operational modes were selected to fulfill these important objectives. In particular, the data from the flyby were used to estimate the magnitude of contaminating field components along the spin axis which had not previously been determined because of insufficient high time resolution solar wind data.
An overview of the magnetometer data for December 8, 1990, the day of the Earth encounter, is provided in Figure 2 (GSM coordinates) on a highly compressed scale. The 24-hour plot shows that the spacecraft crossed the magnetopause early in the day (near 100 RE downtail) at 0304 UT, as is more or less consistent with the predicted magnetopause location shown in the trajectory plot in Figure 1. The entry was into the southern lobe of the tail; the field was directed away from the Earth (Bx negative), again entirely as expected. During the next 16 hours the spacecraft moved in and out of the plasma sheet and repeatedly crossed the central neutral sheet at varying distances from Earth. Early in the day, following the appearance of magnetic perturbations in ground records at very high latitudes, Galileo observed magnetic activity, but this was not followed by a significant decrease in magnetotail lobe field strength. Slightly later following the onset of magnetic perturbations at auroral zone latitudes and below, the magnetotail lobe field strength at Galileo decreased at the onset of a disturbance. As the spacecraft approached the dipolar region near Earth, it recorded a final substormlike diamagnetic signature at about 1840 UT. Closest approach to Earth was at 2034 UT, and shortly thereafter the spacecraft encountered the magnetopause where apparent multiple partial crossings of the boundary were detected as the spacecraft exited from the compressed terrestrial field (80 nT) into the magnetosheath slightly before 2200 UT. Multiple crossings of the magnetopause by outbound spacecraft are commonplace but the high velocity of the spacecraft (about 14 km s-1) renders the observations difficult to interpret as simple boundary motion. In the magnetosheath, the field was about 30 nT. The bow shock was encountered only once and its crossing was marked by the expected decrease in field strength.
THE MAGNETOTAIL DATA
The spacecraft orbit projected into a cross section of the tail is shown in Figure lb. The plot is in the yz plane of an aber- rated GSM coordinate system. The new coordinates have the x axis rotated from Sun-pointing by a nominal 4.2° to correct for the effects of the orbital motion of the Earth. Relative to the new x axis, a GSM coordinate system is introduced. The plot shows that the spacecraft entered to the south of the tail z = 0 plane. At about 1000 UT it approached the z = 0 plane and thereafter made its way across in y, attaining the tail center relatively near Earth more than 10 hours later. The trajectory led one to expect encounters with the magnetic neutral sheet in the center of the tail. As recent work has confirmed that the sheet has a warped geometry controlled by the dipole tilt angle [Dandouras, 1988; Fairfield, 1980; Walker et al., 1988; Ham- mond et al., 1992], with distortion most extreme near solstice, we have illustrated the predicted positions of the warped neu- tral sheet at three different times (0440, 1040, and 1640 UT). These times were selected to show the shape at maximum, in- termediate and minimum dipole tilt. Furthermore, the flaring of the magnetopause in the antisolar direction implies that the diameter of the magnetopause cross section decreases as the spacecraft moves towards the Earth. The diagram therefore shows nominal magnetopause positions at the down tail distances corresponding to the three times for which the neutral sheet is shown.
The crossing of the tail magnetopause at 0304 is apparent in Figure 2 as a quietening of the fluctuations and the develop- ment of a clear and persistent negative Bx. The slow increase of the field magnitude during the 2 hours after entry into the magnetotail could represent the traversal of a boundary layer or could have resulted from the addition of flux to the tail during an extended growth phase. During the 4 hours following entry, small amplitude undulations in magnitude and direction were seen. Shortly after 0500 UT, the field exhibited a bipolar fluctuation in Bz centered at 0508 UT. The skewing of the field in the negative y direction and the small increase in the magnitude are the signatures that Slavin et al  have pointed to as distortions of the lobe magnetic field associated with a plasma bubble or plasmoid that moves down the tail somewhere near the center of the plasma sheet. This is what they have called a traveling compression region or TCR. In this case the plasma sheet lies above the spacecraft. The signature is clearer in Figure 3a which provides the magnetometer data for the first 6 hours within the magnetotail on a somewhat expanded scale.
The field magnitude in the magnetotail increased fairly steadily until 0700 UT. At this time, and indeed from entry through until 0700 UT, the relatively low level of noise and the relatively high field value suggest that the spacecraft remained in the magnetotail lobe or the plasma sheet boundary layer rather than the plasma sheet. At 0700 UT, the magnitude dropped and over the next hour the field experienced a series of reorientations. Much of this period was spent close to the neutral sheet (see, for example, the interval between 0717 UT and 0723 UT during which the field magnitude repeatedly dropped below 1 nT). A brief crossing of the neutral sheet occurred at 0705 UT; the field dropped to a very low value and the x component of the field changed sign although Figure 1b shows that at this time the spacecraft was far removed from any expected crossing into the northern lobe. After a brief return to the southern lobe the spacecraft encountered the low field region several more times. It emerged from the disturbed region at about 0725 UT briefly and for a longer sojourn at 0740 UT into what seems to be the northern lobe based on the field magnitude, the reduced amplitude of fluctuations, and the positive value of BX.
Throughout the disturbed field region in the neutral sheet vicinity, that is, in regions where the x component was small, there were noticeably large +y components of the field. In contrast, the z component exhibited excursions in both positive and negative directions, the latter being symptomatic of dynamic behavior. The most striking examples of extreme field perturbations occurred as the spacecraft actually crossed the neutral sheet. For example, in the structure at 0714 UT, the field swung into the YGSM direction in the center of a bipolar signature (negative/positive) in BZ. The structure is quite similar to the signatures of flux ropes described by Hughes and Sibeck  and observed closer to Earth by Elphic et al. . From 0740 to 0810 UT, when the spacecraft was in or near the northern lobe, the BY component remained generally 2-3 nT in amplitude.
A final set of disturbances in the field starting at 0800 UT ended with a crossing to the southern half of the tail (Bx negative). As the Bx component reversed, another fluxropelike signature was observed, this one centered at 0816 UT; once again, a large positive BY (of order 10 nT) was coupled with a bipolar (this time positive/negative) Bz perturbation. In fact, throughout the period of large field changes detected by the spacecraft between 0730 and 0830 UT, those uniformly positive spikes in By that were associated with detectable Bz showed variations of both +/- and -/+ polarity.
Between 0820 UT and 0900 UT, the field at the spacecraft settled down to pointing in the antisolar direction. By remained small as did Bz. The variability of the field amplitude between 0820 and 0940 UT suggests that during much of this time, the spacecraft remained near the boundary of the plasma sheet, and that it returned to the lobe at about 0940 UT. The final crossing from south lobe to north lobe or plasma sheet boundary layer occurred over an extended time starting at 1118 UT. The next neutral sheet encounter (see Figure 3b) occurred at 1150 UT.
Figure lb shows that by 1200 UT, the spacecraft was expected to be traveling across the center of the tail and thus passing near the neutral sheet and the central plasma sheet. The data from 1200 to 1600 UT illustrate this well. The field was highly variable. One can identify three characteristic types of field among which the spacecraft moved during the 4 hours from 1200 to 1600 UT. In the highest field region, where B ~ 20 nT, for example, at 1400 UT, the field was relatively smooth and pointed in either the x or -x directions. We identify this as the tail lobe. Southern lobe encounters predominated. Field signatures that we relate to the plasma sheet were also present, that is, intervals when the field was distinctly variable but still oriented predominantly in the positive or negative x direction. Embedded within this type of signature were current sheet crossings where the x component reversed and relatively large By components were observed. In contrast with the earlier crossings of the current sheet, in these crossings the y component took on both negative and positive values. Occasionally there were clear correlations between the x and y fluctuations as, for example, in the large-amplitude disturbances recorded around 1445 UT. The z component was small throughout and its sign fluctuated.
At about 1740 UT (see Figure 3c), there was a distinct change in the characteristic field and the envelope of the field magnitude began to increase rapidly. The field at the spacecraft position remained predominantly x oriented for some time. Then the spacecraft left the distended field lines of the magnetotail and the x and z components increased as it entered the dipolar field regime. However, from 1748 UT until approximately 1920 UT when Galileo reached XGSM = -7.3 RE, the field magnitude significantly exceeded the magnitudes of both the dipole field and the Tsyganenko model field. A strong ring current flowing inside of 7 RE could produce perturbations of the sort observed.
In the post 1800 UT period, the notable feature was the reduction in field strength and subsequent oscillation of the field which occurred at about 1840 UT. At this time Galileo was at a radial distance of 10 RE, slightly outside of geostationary orbit. The 1840 UT signature of the reduction in the x component coupled with an increase in the y component in conjunction with strong fluctuations resembles the signature of a substorm near geostationary orbit.
THE EXIT FROM THE TERRESTRIAL SYSTEM
After passing close to the planet, the spacecraft magnetometer obtained a "snapshot" view of the dayside field. In Figure 3d with data from the outbound pass, a noticeable feature is the apparent multiple partial crossings of the boundary of the mag- netosphere just prior to the final entry into the magnetosheath field at about 8 RE shortly before 2200 UT. The magnetosheath was crossed in just over half an hour. The magnetosheath thickness was thus of order 2 RE, a reasonable value for the local time (midafternoon) and in the compressed conditions pertaining. The shock was crossed abruptly at 2234 UT.
THE INTERPLANETARY DATA
Some of the changes recorded as Galileo moved through the magnetosphere arose merely because of its motion, but even for the high velocity of Galileo, the changes arising from motion through a time-stationary magnetosphere would be significant only on time scales of hours except during crossings of the plasma sheet and the current sheet. The changes near 1200 UT can be attributed to such a crossing. However, the more rapid variations must be interpreted in other ways. Current sheet crossings at anomalous locations can, for example, result from displacement of the tail symmetry axis in response to a shift in the direction of the solar wind velocity or from rotations about the tail axis in response to externally imposed torque [Cowley, 1981]. Therefore we were fortunate to be able to obtain good data on solar wind plasma and field for use in data interpretation. These data also have proved invaluable in our interpretations of geomagnetic activity.
The IMP 8 spacecraft was located in the solar wind, upstream of the bow shock near the Earth-Sun line at a geocentric distance of about 38 RE on December 8, 1990 (GSE coordinates: (38.4, -4.6, -4.9) at 0000 UT and (37.3, 8.7, -4.3) at 2400 UT). Plasma and field data are shown in Figure 4. The solar wind speed remained steady at around 350 km s-1 in the early part of the day until the arrival of an interplanetary shock at 1418 UT (marked in Figure 4b) after which the speed fluctuated around 400 km s-1. The solar wind density was of order 7 cm-3 initially, rising behind the interplanetary shock to a new steady level, 20 cm-3. The delay time for travel of a disturbance moving at the solar wind speed from IMP 8 to Galileo was of order 40 min at the time of the magnetotail entry, diminishing to about 8 min at the time that the spacecraft left the magnetotail.
The IMF (interplanetary magnetic field) magnitude was relatively constant at about 6 nT until the interplanetary shock arrived. The magnitude then rose to about 10 nT and became more variable. In the early part of the day the field had a clear northward component and a negative y. Before the Galileo entry into the magnetotail at about 0220 UT, the Bz component became small, but it swung to a significantly negative value at 0345 UT; thereafter, the field remained generally southward oriented (returning occasionally to near 0ø) until late in hour 10 when it turned northward. The field oscillated between positive and negative values of Bz on time scales of hours for most of the remainder of the day. The By component was the dominant component of the field for much of the day. From entry until 0800 UT, it was positive. Thereafter it became negative, changing sign several times in the next 6 hours. Following the arrival of the interplanetary shock, both Bx and By changed sign periodically.
GROUND-BASED DATA AND IDENTIFICATION OF SUBSTORM PHASES
Central to the interpretation of the Galileo data is its relation to geomagnetic activity observed by ground stations and by spacecraft elsewhere in the magnetosphere. We have used magnetometer records from a large number of stations (see Table 1) to identify substorm and related signatures.
During the flyby, several rather complex multiple onset substorms can be identified. A low-amplitude Pi 2 burst was recorded at the easternmost stations on the Canopus chain (the stations are identified and their locations are listed in Table 1). A weak positive bay at FRD and a small negative bay at GSB with opposing east-west perturbation components enable us to interpret it as evidence of an extremely weak substorm localized east of midnight (0040 ± 1 hour MLT) at 0440 UT. The ground-based activity initiated with this weak signature grew intermittently stronger with onsets at 0527, 0551, 060O, 0614, and 0643 UT. The recovery was complete by 0730 UT. The second substorm included intensifications at 0755, 0807, and 0834 UT and recovery was complete by 0930 UT. The third substorm onset occurred at 1034 UT and recovery was complete at 1140 UT. Subsequent activity was not observed clearly on the Canopus chain. However, pseudo AK, AU, and AL indices (based on available stations) have been constructed from envelopes of the records from the Canopus chain and several USGS stations (Figure 5). The plots show that electrojet activity started at about 0527 UT and continued all day.
The time line in the Appendix identifies the signatures of activity during the full day of December 8, 1990, and relates them to the upstream solar wind conditions and the measurements of the Galileo magnetometer and other spacecraft. Details will be provided by R. L. McPherron et al. (Magnetic activity during the Galileo flyby of Earth on December 8, 1990, UCLA IGPP manuscript, in preparation, 1993). The data of this Appendix will provide a useful framework for the interpretation of measurements made by all of the Galileo fields and particles instruments, so we have provided greater detail than is justified by the phenomena that we single out for detailed discussion. We have also incorporated timing and intensity information on the bursts of auroral kilometric radiation (AKR) identified by the Galileo Plasma Wave System (PWS) (D. Gurnett and A. Keller, personal communication, 1992). These data will be described more fully by the PWS team in future publications.
Although this report does not include all of the data used to develop the time line given in the Appendix, it seems worthwhile to provide an example of the ground signatures that we consider important for interpretation of the Galileo measurements. We focus on the hours 0500 to 0900 UT and the associated activity both on the ground and at the Galileo position. The Galileo magnetometer data for the interval 0600 to 0900 UT is shown on a somewhat more expanded scale in Figure 6. In Figure 7 we show the 1-min resolution data from relevant available magnetometer stations. Figure 8 shows the high pass filtered data from some of these stations and HER and KAK. The Pi 2 activity apparent in these plots is a useful indicator of the onset of intensifications. Locations of selected stations are plotted in Figure 9.
Although the weak signature at 0440 UT is not apparent in the low time resolution Canopus records, the perturbations starting at 0527 UT are clear at FCHU and nearby stations. The z component of the perturbation (see Figure 7b) has different signs at FCHU and ESKI, indicating that the westward electrojet intensified north of FCHU but south of ESKI. We have not been able to determine the east-west localization. A second intensification of the westward electrojet occurred at 0551 UT and from the signs of the z component of the perturbation it is apparent that it was located north of ESKI but south of RANK. From the sign of the y perturbation one can localize the activity to the east of the Churchill chain.
In the ground records there is only very weak activity at the eastern Canopus stations and GSB prior to the time (0508 UT) when Galileo saw evidence of the outward passage of a TCR whose center passed above the spacecraft. The first disturbances on the ground clearly discernible at low time resolution occurred after 0500 UT. In the discussion section we shall consider how the ground activity may be related to the observed TCR. Substantial electrojetlike activity is seen at ESKI at L = 10.2 at about 0530 UT, preceded by a smaller disturbance at FCHU at L = 8.2. The stations of the central chain were approaching local midnight at this time (see Figure 9).
In contrast to the earlier interval, by 0600 UT, the two highest latitude stations of the central Canopus chain (RANK and ESKI), within an hour of midnight, recorded substantial magnetic bays. Although the L values of the RANK and ESKI (12.44 and 10.2, respectively) would normally lead one to classify them as polar cap stations from which field lines should extend far down the tail, it is difficult to identify a response to this substorm intensification in the magnetotail at Galileo unless one wishes to associate the fluctuations in By and Bz at about 0556 UT with this substorm. Although the association with this weak fluctuation cannot be ruled out, we do not expect a signature in the distant tail to precede the signature observed on the ground. At the (aurora! zone) stations, lower in latitude than RANK and ESKI, and the stations to the east, 0600 UT marks the beginning of a slow decrease in the north- south horizontal component. This would be consistent with the start of a substorm growth phase [McPherron, 1970].
Murphy's law seems not to have been operating during the Galileo flyby. IMP 8 was in an ideal location to serve as a monitor of the solar wind input conditions and recorded an interplanetary shock while Galileo was in the magnetotail. The magnetosphere was geomagnetically active and the tail orientation was such that the spacecraft probed its most interesting regions. In this discussion we shall focus on the implications of the data for the dynamics of the distant magnetotail beyond 20 RE.
Encounters With the Plasma Sheet
As Galileo moved towards the Earth, it encountered different plasma regimes. Many of the changes can be ascribed to large-scale displacements of the entire magnetotail in response to solar wind flow changes. For example, Galileo's initial en- try into the magnetotail appears to have resulted from the shift in the east-west flow angle of the solar wind that can be noted at 0215 in Figure 4. The east-west angle is positive for flow from west to east or equivalently, dawn to dusk. Actually, this angle shifted several times between 0130 UT and 0300 UT. At the solar wind speed, the front should have reached the position of Galileo in about 38 min. The first shift (at ~0135 UT) from ~5° to ~0° would have swung the tail towards the Galileo location starting at about 0213 UT but Galileo did not cross the magnetopause near that time. However, somewhat later, at about 0215 UT, the solar wind flow angle shifted from ~ 4° to ~-3°. Shortly after this change propagated to the position of Galileo (estimated as occurring at 0253 UT) the magnetotail symmetry axis seems to have responded by swinging still further towards dawn causing Galileo to enter the magnetotail. Thereafter the east-west angle remained close to ~-3° until 0800 UT when it again swung to a duskward orientation. The entry into the magnetotail was sharp and consistent with a moving boundary. The associated displacement of the axis of the magnetotail allowed Galileo to penetrate deeply enough that there were probably no further crossings of the magnetopause boundary.
As is evident from Figure lb, if the solar wind flow had been in the nominal radial direction Galileo would not have been expected to cross the neutral sheet until after 1200 UT, so the entry into the plasma sheet at 0656 UT and the subsequent crossing of the neutral sheet required substantial displacement of the magnetotail, either by twisting through about 25° about the tail axis or by shift in a southward direction. Twisting of the open magnetotail in the presence of a finite IMF By was described theoretically by Cowley  and simulated by Brecht et al. . The sense of rotation imposed by positive By is in the right direction (anticlockwise about XGSE) to bring the plasma sheet down over the spacecraft. Furthermore, the small positive By observed during hour 6 prior to the entry into the plasma sheet would be consistent with coupling of the IMF into the magnetosphere as discussed by Cowley and Hughes . However, K. K. Khurana ("A model of magnetospheric convection in the presence of IMF By,'' submitted to Geophysical Research Letters, 1992) has pointed out that the penetration of the By component into the plasma sheet that is evident during most of the interval between 0700 and 0820 is not consistent with tail rotation.
An alternative interpretation is that the crossings of the neutral sheet between 0700 and 0900 UT resulted from a northsouth shift of the entire tail axis. Figure 4 shows that the solar wind velocity had a clear southward orientation during much of the first half of the day. The magnitude of the southward flow angle increased in hours 6 and 7. For a 5° southward flow, the tail axis at 60 RE downtail shifts towards negative ZGSM by ~5 RE As the precision of the north-south angle is only ± 2.5°, this shift alone can account for the crossings into the northern hemisphere of the magnetotail. Although the solar wind data have many gaps, the magnitude of the southward flow angle in the solar wind had diminished when Galileo returned to the southern lobe in hour 8. Hence the solar wind flow direction shift is probably the better explanation of the entries into the different lobes between 0700 and 0900 UT.
Later in the day, especially during the interval 1500 to ~1830 UT the spacecraft spent long periods of time in and near the plasma sheet. During this period, Galileo remained within 2 - 3 RE of the expected neutral sheet position as shown in Figure lb, moving slowly in the ZGSM direction as it crossed towards the center of the tail. The field was significantly depressed below the lobe values that can be estimated by joining the peaks of B, in Figure 2. It was also small compared with average lobe magnetic fields as reported by Behannon  and Mihalov et al. , suggesting that the plasma sheet was unusually hot. As well, the field along the spacecraft trajectory was small compared with estimates from the Tsyganenko  model. This suggests that the plasma b = p/(B2mo) was unusually large. The rather large magnetic pressure fluctuations that were recorded during that time interval can be interpreted as diamagnetic responses to motions of the plasma sheet. It is very probable that the multiple entries and exits from a high b plasma sheet correspond to shifts in the plasma flow directions related to the large-amplitude Alfvén waves present in the upstream solar wind and evident following the shock in Figure 4b (see a discussion of the ground signatures of these waves by Saunders et al. ).
Signature of the 1418 UT Interplanetary Shock
The signature of the interplanetary shock that passed over the IMP 8 spacecraft at 1418 UT should have reached Galileo at some time between 1426 UT and 1433 UT. These estimates assume that the shock propagated from IMP 8 to the subsolar point of the magnetopause at about 500 km/s (solar wind velocity + shock velocity) and continued to x = - 30 RE either as a fast mode wave at 2000 km/s or by convection around the magnetopause at 500 km/s. Recall that the time line in the Appendix shows that a SI (sudden impulse with a magnetic signature) propagated past the two geostationary spacecraft at 1425 UT. One might have expected to see the same type of signature at Galileo. However, the Galileo data are very spiky throughout the entire interval, and no clear magnetic signature of the shock passage was evident in the low time resolution data. Our higher time resolution data (2/9 s resolution) show an enhancement in noise level starting at 1430 UT which increases markedly at 1432 UT. It is possible that the entry into the plasma sheet at 1430 UT occurred as the magnetotail geometry was altered by the shock front passing through the surrounding magnetosheath. We suggest that the absence of a low-frequency magnetic signature at Galileo can be explained as a consequence of the large b of the plasma sheet which allowed the plasma to dominate the pressure response without significantly affecting the magnetic field magnitude.
Magnetic Activity and Azimuthally Localized Plasmoids
During the early part of the day, western hemisphere ground magnetometer stations were well situated to provide very good evidence of auroral zone activity. We have been particularly interested in considering whether the Galileo data confirm the recently reported association between substorms and plasmoids in the magnetotail plasma sheet or traveling compression re- gions (TCRs) which are interpreted as the lobe signatures associated with propagating plasmoids. This association was reported by Slavin et al.  based on study of ISEE-3 data taken at downtail distances between XGSM = -76 RE and XGSM = -80 RE over a 36-hour interval that included 14 substorms. Slavin et al. [1992, p. ] identified substorm onsets using various indices including intensification of AKR power levels, Pi 2 pulsations, the injection of electrons at geosynchronous orbit, and the AL index. They found that "all of the substorms produced one or more plasmoid(s) and/or TCR(s)" after expansion phase onset. The typical time delay between substorm onset and the passage of a TCR or plasmoid by ISEE 3 was 11 min, with considerable variance.
Galileo data from 0300 UT to 0700 UT are pertinent to a comparison with the ISEE 3 data as they were acquired in the magnetotail between XGSM = - 87 RE and - 68 RE The first evidence of magnetic activity at Galileo was the TCR observed just earthward of - 80 RE at 0508 UT, following the very weak substorm of 0440 (see the time line in the Appendix). If we follow Slavin et al.  in assuming that the associated plasma sheet disturbance (e.g., plasmoid) started near x = - 40 RE, its mean travel speed was approximately 250 km/s, a bit slower than the 400 km/s reported by Slavin et al.  from in situ measurements.
After the first intensification at 0440 UT, five additional intensifications, listed in the Appendix, occurred at 0527, 0551, 0600, 0614, and 0643 UT. Allowing for delays of 11 to 30 min for the disturbances to reach Galileo at x = - 60 to - 70 RE, we seek evidence of TCRs approximately between 0538 and 0557 UT, 0602 and 0621 UT, 0611 and 0630 UT, 0625 and 0644 UT, and 0654 and 0713 UT. The last interval includes the entry into the plasma sheet and it is not possible to single out any particular disturbance as dominant. During the rest of the intervals, Galileo was in the lobe. A very small-amplitude bipolar (positive/negative) signature in Bz centered at 0558 UT could be interpreted as a very weak TCR moving tailward past Galileo. The amplitude of the signature in Bz was ~ ± 1 nT over an interval of about 5 min, with no magnitude perturbation apparent, which requires some generalization of the definition of a TCR. No bipolar signatures in Bz were observed during the other three intervals. This means that of the five cases in which we could have identified the presence or absence of a signature, at most two can be linked to the passage of a TCR. Strangely, it was the intensification with the weakest ground signature that corresponded to the unambiguous TCR at Galileo.
This analysis leads to the conclusion that there is not a detectable TCR or plasmoid for every intensification of the substorm electrojet. Our methodology differs enough from that of Slavin et al.  that comparison of the results requires some care. Their indices were used to identify a single onset time for a substorm, whereas we consider separately the multiple onsets that collectively release tension in the magnetotail. If we take the six onsets identified in the Appendix as collectively comprising a multiple onset substorm, then we do associate at least one TCR with the substorm. However, one would be hard put to believe that the very weak first intensification is the only one of the six component parts of the substorm that created a plasmoid.
We are faced with a dilemma. Why did Galileo not see the signatures of a downtail propagating plasmoid following at least three of the intensifications? To us, the most compelling interpretation is that plasmoids are localized in azimuthal extension, and that their signature is observable only across a portion of the tail. This follows because a plasma "bubble" of limited azimuthal extent would bend the field lines only above and below it in the lobes. The field bending is an intermediate mode perturbation that does not propagate across the field, and therefore cannot produce a bipolar signature clear across the tail. The magnetic pressure perturbation associated with the localized disturbance is small and is carried by the fast compressional mode which propagates fairly isotropically; thus the amplitude of the pressure perturbation should be largest in the vicinity of the localized plasmoid and should decay rapidly with distance.
The azimuthal localization of a plasmoid in the tail can be regarded as an extension into the tail of the localized substorm current loop that is associated with the substorm intensification in the auroral ionosphere (see, for example, McPherron ). These current loops tend to be centered increasingly far to the west for successive intensifications within a substorm. Correspondingly, one can argue that the localized plasmoids corresponding to the successive intensifications appear at positions that step across the tail from dawn towards dusk. This would explain why Galileo, which remained relatively close to the dawn flank of the magnetotail between 0530 UT and 0800 UT, clearly observed a TCR after the first and easternmost of the multiple intensifications but saw no clear signatures after the stronger intensifications that followed. We should note that Richardson et al.  did not find plasmoids associated with each substorm disturbance in their CDAW-8 study of ISEE 3 distant geotail (100 to 240 RE) data. They suggested that in these cases, the neutral line remained near Earth and the plasma sheet in the distant tail merely thickened. However, azimuthal localization of plasmoid signatures could provide an alternative explanation. In the near-Earth magnetotail, azimuthal localization of transport has also been inferred for the convection enhancements (referred to as "bursty bulk flows") identified by Angelopoulos et al. . These localized flow bursts may be linked to release of azimuthally confined plasmoids.
Our interpretation is particularly appealing because it explains not only our failure to find a systematic correspondence between tail events and individual substorm intensifications but also the success of Slavin et al.  in finding at least one event associated with every substorm. If a multiple onset substorm consists of successive intensifications, each linked to azimuthally localized flux tubes, and if during the course of the substorm all parts of the plasma sheet eventually undergo reconnection, then a spacecraft in the translunar tail will observe an event associated with at least one of the intensifications. However, the spacecraft will not observe a tail event associated with each intensification.
Our model of multiple azimuthally limited plasmoids released successively during the course of a substorm has important implications for the description of substorm dynamics in the magnetotail, but it should be further tested using other spacecraft data sets. The only other interpretations that appear consistent with both our data and the results of the Slavin et al. study (for example, the plasmoid generation region was beyond 80 RE for the Galileo pass, or, TCRs are only coincidentally linked to substorm intensifications as observed on the ground and their previously reported direct correspondence to substorm intensifications occur only under special circumstances) appear to us to be more forced.
The Plasmoids/Flux Ropes Continuum
Following the passage of the TCR, the magnetic field maintained a slightly southward tilt consistent with flaring of the tail. The first marked turning to a more southward orientation of the field at Galileo occurred at 0656 UT. Concurrent By perturbations were dominated by positive excursions. Another sharp southward turning, this time with the characteristic bipolar signature of a plasmoid, started at 0713 and lasted for about 7 min. The negative/positive polarity is consistent with earthward propagation. In this case, the By perturbation was larger than the Bz perturbation, and again it was predominantly positive. Several additional positive By perturbations follow, with the largest one associated with a southward turning at about 0816 UT.
It appears to us significant that the southward turnings during the active interval are systematically associated with By perturbations of the same sign as the solar wind By. The form of these perturbations corresponds to what would be antici- pated for flux ropes oriented in a dawn-dusk direction such as those found closer to the Earth by Elphic et al.  and Moldwin and Hughes . They differ from the events between 90 and 110 RE in the ISEE 3 data that Sibeck et al.  identified as flux ropes oriented with their axes parallel to the magnetotail axis. The flux ropes that we encountered must lie near the neutral sheet, for they are observed in conjunction with reversals of the Bx component, usually centered near the time of vanishing Bx.
In an attempt to understand the structure of the perturbations, we have analyzed the structure of the perturbation centered at 0714 UT. In Figure 10 we show a high time resolution plot of the data for 0714 ± 0010 UT. The lobe field prior to entry into Bx became very small. The bipolar signature in Bz could be modeled as either a plasmoid or a flux rope moving earthward but, close to the center of the bipolar signature, the field rotated almost fully into the +y direction. As noted by Sibeck et al.  and Moldwin and Hughes , this is expected near the center of a flux rope but not in a plasmoid. the plasma sheet was about 12 nT. The bipolar perturbations developed as
If we regard the perturbation at 0714 UT as that of a helical flux tube embedded in the center of the plasma sheet with its axis along the YGSM direction, that is, parallel to the y component of the interplanetary magnetic field, we can follow Elphic et al.  and model the perturbation field in the four-parameter form
Br = 0
Bf = B2(r/a2 )/[1+ (r/a2)2]
Bz = B1/ [1 + (r/a2)]
where the perturbation is given in cylindrical coordinates (r, f, z) and z is distance along the axis of the flux tube.
The Bx panel of Figure 10 indicates that the spacecraft approached the structure from the southern lobe, but emerged into a region of low Bt, which appears to have been near the center of the plasma sheet. This means that the signature of the passage out from the center of the flux rope is not obscured by the variations of the ambient field, so we model only the signature starting from 0714 UT and assume symmetry to obtain the first half of the perturbation. We assume that the outbound spacecraft velocity was along the XGSM direction, thus accounting for the prolonged interval spent near the center of the plasma sheet. We do not know z0, the north-south distance between the flux rope axis and the spacecraft trajectory, but we assume that at the point where By (or in the model, Bz, ) is maximum, p = z0 and elsewhere p2 = Zo2 + (X2 - Xo2). Then
BXGSM = -Bfsinf
BZGSM = Bfcosf
where f = tan-1 [z0/(x-x0)] and f is measured from the XGSM direction.
As the plasma is in motion relative to the spacecraft (along the x direction, by assumption), the relation between time in the measured profile and distance is not known. Thus we give the spatial scale in undefined "distance units." We find that B1 = B2 = 11 nT, a1 = a2 = 50 distance units and z0 = 20 units gives a good fit. Figure 10 shows the fit superimposed on data for the 20 min surrounding the peak perturbation. The fit to Bx is not good because this component is dominated by the signature of the current sheet in which the flux rope is embedded. The associated Bx variations reflect motion through this current sheet which was not modeled. However, we think that the fit to the dominant low-frequency variation of the other two components is satisfactory for a very simple model.
Our model is one that has been used previously [Elphic and Russell, 1983] to represent putative flux ropes, and although it is not unique, it is useful in characterizing the relationship between the radial and axial scale lengths of the structure. It has a nonvanishing field component along its axis in the z (or YGSM) direction throughout the flux rope, but the field becomes aligned with the axis as the spacecraft approaches the center of the structure. The corresponding current structure is found from m0j = Ñ x B
jp = 0
jf = 2B1r/[mo(a1)2(l + (p/al)2 )2]
jz = 2B2/ [moa2 (1 + (p/a2)2 )2] (3)
Thus the current, like the field, has a characteristic form as the spacecraft approaches the flux rope axis. At the center of the structure, the current flows purely along the axis of the flux rope.
Note that if B1 = B2 = B and a1 = a2 = a, as it is in the fit to the data, then j= (2B/moa) [1 + (r/a)2]-1 and with j and B everywhere parallel, the configuration is magnetically force free in the sense that j x B = 0, although we did not impose the force free condition on the flux rope model.
There is an interesting parallel between the magnetic signature of the putative flux ropes that we analyze here and magnetic clouds in the solar wind identified by Burlaga et al , although we cannot yet confirm that the plasma conditions are similar. Burlaga  addresses the issues associated with force free models of the observed helical structure of the field in a magnetic cloud and provides a thoughtful summary of the conditions under which a magnetically force free model is expected to correspond to a stable configuration. He models the signatures of several magnetic clouds using a form with the underlying symmetry that we have adopted but with j and B related by a constant factor rather than the r dependent factor of our model. He remarks that the boundary conditions are not addressed in his model, that a study of the stability of the model is needed, and that the assumption of cylindrical symmetry can be valid only locally. These caveats apply equally to the present analysis. Nonetheless we believe that both his model and ours provide insight into the structure of the observed field perturbations.
We conclude that the magnetic perturbations centered at 0714 UT at XGSM = - 68 RE and analogous perturbations that occurred over the next hour are consistent with the criteria proposed for magnetically force free flux ropes embedded within the magnetotail plasma sheet. The core field, oriented along the y direction with the same sense as the y component of the solar wind magnetic field, has a magnitude comparable with the lobe field magnitude; there is a bipolar signature in the transverse component of the field. The perturbations appear within the plasma sheet and are centered very near to the neutral sheet crossings. They last between 6 and 12 min.
This is somewhat longer than the duration of the event reported by Elphic et al.  but comparable in duration with the events of Moldwin and Hughes . The fit to the flux rope observed by Galileo at about 0714 UT suggests that the spacecraft flew through rather near the center of the structure. The axis was found to lie near to the GSM y axis. This draping of the flux rope across the tail is similar to the geometry that emerged from the Ogino et al.  magnetohydrodynamic simulations of a reconnecting magnetosphere in the presence of a southward IMF with a finite y component. Their flux ropes are connected to the ionosphere at one end, drape across the tail near the center of the plasma sheet, and may go out through the magnetopause into the solar wind or return to the ionosphere. In the simulation, at some large downtail distance, these structures disconnect from the ionosphere. We speculate that they may unwind, and reconnected portions may start to move earthward in response to field line tension.
The 0714 UT event had -/+ polarity in Bz; a second clear event, centered at about 0816 UT, had +/- polarity. These polarities and the ground activity indices of the Appendix suggest that the 0714 UT flux rope was convecting earthward in the recovery phase of a substorm and the 0816 UT flux rope was convecting tailward in the expansion phase of a substorm.
Strong spiky unipolar signatures in By are not unique to the plasmoidlike disturbances observed by Galileo. Examples have been noted by Moldwin and Hughes . As well, the substorm-associated signatures observed by ISEE 3 in the data published by Slavin et al.  provide numerous examples that are called plasmoids although they are accompanied by very evident By spikes. This supports the proposals by Elphic et al. , Hughes and Sibeck , Sibeck , Moldwin and Hughes  and others that the substorm-associated structures, referred to as plasmoids, that move tailward in the substorm expansion phase may be flux ropes, probably force free, connected at the ends to the ionosphere or the solar wind. The plasmoids that are found in two-dimensional analysis (whether schematic diagrams or simulations) appear to be unconnected to the ionosphere or the solar wind. The structure is inevitable in two dimensions where field-aligned currents vanish. In a three-dimensional situation, the plasmoidlike structures are likely to connect to other parts of the magnetosphere through field-aligned currents and thus to display to some extent the character of flux ropes. It is possible that plasma or field signatures that have been called plasmoids or flux ropes form a continuum of structures that differ only in the size of the axial field (By) that they contain. We cannot reach any firm conclusions based on our limited data set, but we point out that a large axial field in the center of what could easily be taken as a plasmoid requires explanation. A more complete analysis of the putative flux ropes observed by Galileo will be published elsewhere (K K Khurana et al., "Flux ropes in the magnetotail of Earth and their relation to magnetospheric substorms", manuscript in preparation 1993). Ultimately, the use of correlative plasma data is expected to illuminate further the nature of the magnetic structures discussed here.
This first report on the magnetometer observations made during the Earth I flyby of the Galileo spacecraft and interpreted in conjunction with the extensive correlative measurements acquired during the flyby is intended to reveal the richness of this unique data set. Spacecraft data from a complement of fields and particle instruments were acquired during a well-documented, geomagnetically active interval. Our initial report, based principally on the magnetometer data, have revealed the complexity of the links between substorm signatures at the ground and the signatures found in the translunar tail. In particular, we are intrigued with the suggestion that plasmoids and TCRs may occasionally be highly localized in cross-tail dimensions, a possibility that requires further investigation. We have also found additional evidence that some or all of the signatures that have been identified as plasmoids in the tail have the structure of flux ropes and that they may become magnetically force free as they travel down the tail.
APPENDIX: TIME LINE FOR GALILEO EARTH ENCOUNTER I
Interval including first substorm (with intensifications at 0440, 0527, 0551, 0600, 0614, 0643, ending at 0730)
0212 IMF rotates to Bz = 0.
0303 IMF becomes weakly southward.
0304 Galileo enters the magnetosphere's southern lobe from the magnetosheath.
0345 IMF more strongly southward.
0400 IMP 8 solar wind becomes very steady at 350 km s-1 from 5° east of Sun line; density is 7 cm3.
0407 IMF Bz suddenly turns north with near zero magnitude.
0440 IMF Bz becomes strong, and steady southward. Pi pulsation activity begins at the south edge of the Churchill chain, with onset at HER at 0442. Very weak positive H and positive D bay at FRD, weak perturbation at LRV possibly linked to pulsation activity, very small negative bay at GSB with negative D perturbation.
0442 Onset of a weak AKR burst at Galileo followed by intermittent weak bursts until 0551.
0508 TCR at Galileo.
0515 Field at GOES-7 (-2200 LT) begins to decrease.
Expansion Phase: Onset
0551 Intensification of WEJ north of ESKI but south of RANK, to east of Churchill chain and Pi 2 onset at HER. No effect at mid-latitude.
0557 Positive D perturbations at BOU and NEW, but central meridian uncertain.
0600 Intensification of WEJ on Churchill chain which may be due to surge from previous onset passing overhead, WEJ reaches RANK. Positive D spike to the west at RABB. Galileo footprint is far north and east of location of first substorm.
0614 Intensification of WEJ at high latitudes north of RANK and west of Churchill chain, probably not as far as RABB. No effect at mid-latitude.
0620 Beginning of auroral zone recovery phase.
0632 AKR of moderate intensity. Intermittent until after 0700.
0635 Northward turning of IMF.
0640 LANL spacecraft 046 records electron drop out at 1940 LT.
Recovery Phase: Onset 5
0643 Intensification of WEJ at high latitude east of Churchill chain, moves west.
0655 IMF Bz turns weakly southward. Positive D at NEW, but nothing else to RABB.
0656 Galileo MAG encounters edge of plasma sheet in south lobe; B: negative and By is strongly positive. Impulsive changes continue intermittently until after the flux rope signature centered at 0816.
0705 Galileo MAG first encounters neutral sheet, then returns to southern boundary.
0715 IMF turns suddenly northward and reverses to south.
0730 End of substorm.
0733 Weak AKR continuing until 0755.
0740 To 0810. Galileo appears to be in northern lobe.
Interval including second substorm (with intensifications at 0755, 0807, 0834, ending at 0930)
0755 WEJ onset west of the Churchill chain. The WEJ was north of MCMU and south of FSMI. The WTS may have been far to the west beyond DAW! CMO sees a weak WEJ onset. Weak positive H bay begins at BOU, TUC, and NEW. There are negative D perturbations at FRN and NEW which implies that the central meridian was west of NEW, consistent with auroral zone indication, despite being far west of midnight. Small amplitude Pi 2 onsets at HER and KAK. Intense AKR which continues to second intensification at 0805.
0800 The flow at IMP 8 suddenly switches from 5° east to 5° west, IMF Bz ends gradual increase to zero. The By component suddenly reverses; this is a rotational discontinuity. Field goes from east of Sun (spiral angle) to west (antispiral). 0800 To 0810. Galileo reenters the plasma sheet and crosses towards the southern lobe which it reenters at 0820.
0805 Intense AKR continuing until 0829.
0807 Major intensification of WEJ west of RABB and east of FSMI almost exactly at magnetic midnight. Both FSM and later DAW see WTS. The WEJ flows at latitude below all stations in Churchill chain. Strong positive H bay onset, central meridian around TUC or FRN, but east of NEW.
0811 Pi 2 onset at HER but not at KAK.
0812 Field dipolarization begins at GOES-7 (at 0100 LT). A Pi 2 starts at GOES-6 (2300 LT).
0815 Galileo crosses the neutral sheet returning to the south.
0816 Pi 2 activity at GOES-7 Galileo sees a flux rope with strong positive By and bipolar Bz0820 Pi 2 activity at GOES-6. LANL spacecraft 046 at 2120 LT observes a very large spike of protons. Electrons recover to normal but no heating.
0834 Possible intensification but no effects at mid-latitudes.
0835 Intense burst of AKR for 3 min.
0840 Start of data gap in IMP 8 data.
0843 AKR intensification until 0852.
0855 Intense AKR for 3 min Moderate AKR for 10 min. Intense for three more minutes before dying out at 0922.
0900 Drifting electrons begin to arrive at LANL spacecraft 046 from the east.
0900 To 1100. Continuous tailward flow gradually weakening as Galileo moves through south boundary layer towards the southern lobe.
0910 End data gap in IMP 8 data.
0925 Moderate AKR until 1036, weak until end at 1110.
0930 End of substorm.
Third Substorm (1034 - End 1140)
1020 IMP 8 flow shifts towards south. 1034 WEJ onset to west of DAW, south of CONT. Flow at IMP 8 shifts to eastward. 1100 On ground, auroral zone recovery begins. 1117 To 1132. Rapid crossing of Galileo from southern lobe to northern plasma sheet boundary layer, strong earthward flow in the central plasma sheet. 1140 End of substorm. Dispersive electron injection at LANL spacecraft 046. 1150 Pi 2 onset at KAK and at DAW. Negative bay at CMO.
Additional Classification for Which Details Are Not Available As Cround Magnetograms Not Yet Available
WEJ and WTS are abbreviations for westward electrojet and west- ward traveling surge, respectively.
Acknowledgments. We owe special thanks to Lee F. Bargatze for creating the pseudo-AE index used for Figure 5, and to Jim Slavin for calling our attention to the presence of TCR signatures in our data. The support of the Galileo Magnetometer programming staff at UCLA, including Todd King, Steven Joy, and Joe Mafi, was critical to this research and is gratefully acknowledged. The proficient clerical support provided by Rose Silva was greatly appreciated. We are grateful to John Samson and Gordon Rostoker of the University of Alberta for providing data from the Canopus magnetometer chain supported by the Canadian Space Agency. Thanks are due to Hermann Luhr of Technische Universitat, Braunschweig, P. R. Sutcliffe of the Hermanus Magnetic Observatory, K. Hayashi of the University of Tokyo, Y. Yumoto of STElab, Nagoya University, Mark Lester of Leicester University, E. Friis-Christensen of the Meteorological Institute, Copenhagen, A. Zaitzev of IZMIRAN, Y. Galperin of the Space Research Institute, Moscow, H. Singer of the Air Force Geophysical Laboratory, P. Newell of the Johns Hopkins Applied Physics Laboratory, and S. Cowley of Imperial College for providing correlative data. We appreciate that D. Gurnett and A. Keller of the University of lowa characterized AKR activity with measurements from the Galileo Plasma Wave System and allowed us to include the information in the time line given in the Appendix. Finally, we wish to thank the members of the JPL staff whose professional excellence were critical to the achievement of a successful Earth flyby. In particular, we take this opportunity to recognize the invaluable contributions of Carol Polanskey, our JPL Science Coordinator, through whose careful attention to detail in planning and executing the command sequences used during the Earth flyby we were able to obtain an outstandingly useful data set. This research was supported by the Jet Propulsion Laboratory, Pasadena, under contract JPL 958694. UCLA Institute of Geophysics and Planetary Physics Publication 3744.
The Editor thanks J. S. Dandouras and A. Otto for their assistance in evaluating this paper.
T. J. Hughes, Herzberg Institute of Astrophysics, National Researeh Council, Ottawa, Ontario, Canada KIA OR6.
A. J. Lazarus, Center for Space Research, Massachusetts Institute of Technology, Cambridge, MA 02140.
R. P. Lepping, Laboratory for Extraterrestrial Physics, Goddard Space Flight Center, Greenbelt, MD 20771.
D. J. Southwood, Blackett Laboratory, Imperial College of Scienee, Technology and Medieine, London, SW7 2BZ. U.K.
(Received June 3, 1992;
revised December 1, 1992;
accepted December 3, 1992.)
Copyright 1993 by the American Geophysical Union.
Paper number 92JA03001.
0148 0227/93/92JA-03001$05.00 1 1,299
1Institute of Geophysics and Planetary Physics, University of California, Los Angeles.
2Department of Earth and Space Sciences, University of California, Los Angeles.
3Department of Physics, University of California, Los Angeles.
4Blackett Laboratory, Imperial College of Science, Technology and Medicine, London.
5Center for Space Research, Massachusetts Institute of Technology, Cambridge, Massachusetts.
6Laboratory for Extraterrestrial Physics, Goddard Space Flight Center, Greenbelt, Maryland.
7Herzberg Institute of Astrophysics, National Research Council, Ottawa, Ontario, Canada.
Fig. 1. The encounter trajectory for the flyby of Earth on December 8, 1990. (a) The trajectory in aberrated GSM coordinates. Model bow shock and magnetopause traces have been provided. Filled circles along the trajectory are separated by 12 hours and open circles are separated by 2 hours. (b) A cross section of the magnetotail bounded by a nominal magnetopause. Ticks along the spacecraft trajectory indicate time as in Figure 1a. The neutral sheet model is from Hammond et al.  for 0440 UT (approximately the time of maximum dipole tilt), 1040 UT, and 1640 UT (approximately the time of minimum dipole tilt). Magnetopause cross sections are nominal for the spacecraft XGSM positions at these times.
Fig. 2. Compressed day plot of the magnetic field measured by the Galileo magnetometer on December 8, 1990. The panels show the components in GSM coordinates and the total field in nanoteslas plotted versus universal time (UT). The abscissa also is marked with the XGSM position of the spacecraft.
Fig. 3. Plots of the magnetic field in GSM coordinates as in Figure 2 but Figures 3a, 3b, 3c are for 6-hour intervals and Figure 3d is for a 3-hour interval on the outbound pass.
Fig. 4. Interplanetary parameters for December 8, 1990, measured by the IMP 8 spacecrafl located upstream of the bow shock in the solar wind. Figure 4a provides data from 0000 to 1200 UT. Figure 4b provides data from 1200 to 2400 UT. Plotted from top to bottom are the north-south flow angle (degree), positive for flows from south to north, that is, towards positive ZGSE, the east-west flow angle (degree), positive for flows from west to east, that is, towards positive YGSE, the bulk flow speed in kilometers per second, the ion thermal speed in kilometers per second, the density in number/cm3, all from the MIT plasma instrument in a geocentric ecliptic coordinate system; and the magnetic field components in GSM coordinates from the Goddard magnetometer. The aberration angle of 4.25°, based on a nominal 400 km/s solar wind velocity appropriate to this data set, has been removed from the east-west flow angle. The data are plotted versus UT. A vertical line at 1418 UT identifies the interplanetary shock.
Fig. 5. Fourteen-station AU, AL, and AE indices created from the envelopes to the records of the Canopus chain supplemented by the USGS stations at College and Barrow.
Fig. 6. Plot of the Galileo magnetometer data in GSM in coordinates from 0600 to 0900 UT.
Fig. 7. Ground-based magnetometer data for the 6 hours from 0400 to 1200 UT. (a) Records from the Canopus Maria east-west chain of magnetometer stations (FCHU, RARE, FSMI, DAWS) supplemented by CMO and BRW. (b) Records from the Canopus Maria north-south chain of magnetometer stations RANK, ESKI, FCHU, BACK, GILL, and PINA. (c) Records from the mid-latitude stations SJG, FRD, PINA, BOU, TUC, NEW, HON, and MMB. Station positions are given in Table 1. The stations extend over a range of latitudes from L values of 12.44 to 4.3. The CGM (corrected geomagnetic) coordinate system has x towards IGRF (International Geomagnetic Reference Field) north, y towards IGRF east, and z vertical and positive downwards and the plots are ordered with x, y, and z components from left to right. Arrows indicate the time of local midnight at each station.
Fig. 8. High pass filtered magnetometer data (IGRF north component) from the listed stations with a filter of 2 min. The data from all stations are at 5-s resolution apart from CMO and BRW which are at l-min resolution. The auroral stations are those in Figure 7a and the Canopus meridional chain stations are those in Figure 7b. The third panel provides data from low-latitude stations. The Pi2 bursts (irregular pulsations of periods between 40 and 150 s which are systematically associated with substorm onsets) provide the start times of the different events.
Fig. 9. A plot of the locations of the Canopus Maria station array and of other stations used in the analysis of the Galileo flyby data in CGM coordinates and the Galileo footprint. The different panels provide orientations of the arrays at times labeled on the plots. These are Lamber conformal conic views of the nightside polar cap. Snapshot time at (a) 0500 UT and (b) 0900 UT. A corrected geomagnetic coordinate grid is superposed on the plots. Also indicated is the expected position of the auroral oval based on the Kp activity. The footprint of the Galileo spacecraft based on the Tsyganenko  model is also shown for Figure 9b. The model does not extend far enough down the tail to provide a footprint for the earlier interval.
Fig. 10. Galileo magnetometer data (solid curves) for 20 min surrounding the center of a possible flux rope centered at 0714 UT on December 8, 1990. The dashed curve is a plot of the model of the flux rope as described in the text.
Table 1. Magnetometer Stations