1 Institute of Geophysics and Planetary Physics,
University of California, Los Angeles 90024.
2 Lockheed Palo Alto Research Laboratory, Palo Alto, California 94304.
3 Los Alamos Scientific Laboratory, University of California, Los Alamos, New Mexico 87544.
4 Jet Propulsion Laboratory, California Institute of Technology, Pasadena 91103.
5 Space Sciences Laboratory, TRW Systems Group, Redondo Beach, California 90278.
Abstract. During the large magnetic storm of November 1, 1968, the Ogo 5 spacecraft encountered the apparent direct penetration of magnetosheath plasma into the dayside magnetosphere at magnetic latitudes as low as 43o. Because this region of magnetosheath plasma occurred on magnetospheric field lines immediately adjacent to the zone of trapped energetic particles, it is interpreted to be the polar cusp. The temperature of the electrons in the polar cusp was 4 times greater than the electron temperature measured simultaneously by Vela 4B in the downstream magnetosheath at local dusk, and the electron energy distribution in the cusp was similar to the distribution observed later when Ogo 5 entered the magnetosheath. During the encounters with the polar cusp the amount of depression of the magnetic field implies the presence of magnetosheath protons together with the directly measured electrons. During this period of time the polar cusp was very turbulent at both ULF (magnetic) and VLF (electric) frequencies. Energetic ( 50 kev) electrons were also observed in the cusp. The cusp moved equatorward and poleward in response to changes in the north-south component of the interplanetary magnetic field.
In a closed magnetospheric model, the confinement of the earth's magnetic field by the solar-wind plasma produces two neutral points on the noon-midnight meridian, one in the northern and one in the southern hemisphere [see Mead, 1967]. Spreiter and Summers  argue that these neutral points fill up with magnetosheath plasma to form a cusp. The plasma can leak from this cusp into the dayside auroral oval at rates estimated to be from 1022 protons/sec [Spreiter and Summers, 1967] to 5 x 1024 protons/sec [Willis, 1969]. On the other hand, Piddington  argues that the existence of the earth's magnetotail requires that two neutral lines rather than neutral points exist on the magnetospheric boundary. These lines have also been called demarcation lines by Walters .
Until recently only indirect evidence was available to support the contention that magnetosheath plasma had access to the dayside auroral oval. Most of this evidence was from various dayside ionospheric observations, which have recently been reviewed by Heikkila and Winningham . Other indirect support of this hypothesis is that the dayside magnetopause maps into the auroral oval [Fairfield, 1968].
Within the past year three measurements of direct influxes of magnetosheath plasma have been reported: two at low altitudes and one at high altitudes. Frank and Ackerson  report that a narrow band ~50 km wide located at the high-latitude boundary of the dayside auroral precipitation pattern is the low-altitude signature of the polar neutral point. Heikkila and Winningham  report a much broader latitude band with a width of several degrees magnetic latitude. However, the latter data were obtained during a magnetic storm.
Imp 5, launched into a highly elliptic, high-altitude, nearly polar orbit, repeatedly encountered the direct influx of magnetosheath plasma over a wide range of altitudes and local times [Frank, 1971]. This direct influx of plasma is a usual feature of the magnetosphere and occurs in a band across the magnetosphere rather than in a single tube of magnetic flux. Frank has termed this region the polar cusp. The observed magnetic latitudes of the polar cusp are lower than those calculated by Spreiter and Briggs  and higher than those calculated by Antonova and Shabansky (1968], agreeing most closely with the calculations of Roederer . The dipole magnetic latitude of the cusp under quiet conditions at 4 RE is about 67o and at 8 RE is about 63o [Frank, 1971]. (Magnetic latitude, the angle between a point of observation and the magnetic equator, as we use it in this paper, should not be confused with invariant latitude, the magnetic latitude of a field line's intersection with the surface of the earth.)
Since the Ogo 5 orbit reaches a maximum dipole magnetic latitude of only about 45o during its first year of operation, it does not usually encounter the polar cusp. However, under disturbed conditions the polar cusp moves equatorward. Nevertheless, the lowest position encountered on Imp 5 was at a magnetic latitude of 53.4o at 6.5 RE, still above the Ogo 5 orbit.
On the other hand, the observations discussed in this paper were obtained during a very unusual period of time. First, as will be discussed below, the solar-wind momentum flux during this storm was exceptionally high. Thus, the equilibrium subsolar radius of the magnetopause was small. Secondly, Aubry et al.  have reported that a southward interplanetary magnetic field erodes the dayside magnetopause from its equilibrium position. During a large fraction of this period the interplanetary field was southward, reaching a maximum of 30 southward. Accordingly, Ogo 5 encountered the magnetopause at only 7.2 RE at 61o from the earth-sun line. Later on the same day the ATS satellite crossed the magnetopause at local noon at 6.6 RE (J. Barfield, private communication, 1971).
It is reasonable, then, that Ogo 5 did encounter the polar cusp on this day. We will show below that the observations are indeed consistent with the known properties of the polar cusp. Further, we will compare the Ogo, 5 suprathermal electron observations with simultaneous observations made on the Vela 4B satellite in the magnetosheath. The advantages of using data from the Ogo 5 satellite to study this region are the high data rate of this satellite and the large complement of sophisticated experiments carried by it. However, these observations will characterize the polar cusp only during a highly disturbed period of time.
The period from October 24 to November 6, 1968, was the most geomagnetically active period during 1968, including at least five geomagnetic storms. Relevant geophysical and solar data for this period have been collected at and published by the World Data Center A . Of the five storms during this period, only the last two had well-developed main phases. These two storms began at 0859 UT on October 31 and 0918 UT on November 1. The period of interest to this paper is from 1130 to 1500 UT on November 1, as Ogo 5 traveled from perigee, through the dayside magnetosphere, and out into the magnetosheath.
Figure 1 shows the Kp and Dst indices during the October 31 and November 1 storms. There are two distinct Dst minimums, one 18 hours before the Ogo 5 pass under consideration and one 7. hours afterwards, of -211. The Kp index ranged from 3- to 8+ during these two storms and in fact was 8+ during the Ogo 5 pass.
Figures 2 and 3 show the H components of the magnetograms from 6 low-latitude and 9 auroral-zone stations for the period from 0800 UT on November 1 to 1600 UT on November 2 [Akasofu and Kawasaki, 1970]. We see that the outbound pass of Ogo 5 occurred during the expansion and recovery phase of a very large substorm. At College the depression in H reached -2500 .
On October 31, 1968, Vela 4B was above the afternoon hemisphere and initially was in the solar wind. Figure 4a shows the Vela and Ogo 5 orbits from 0600 UT of October 31 to 1800 UT of November 1, 1968. These trajectories have been rotated into the game plane about the earth-sun line. The apogee of Ogo 5 was approximately in the dusk meridian on this orbit. The average positions of the earth's bow shock and magnetopause [Fairfield, 1971] are shown for comparison.
The Vela data are not continuous, but rather occur in segments of about 1-2 hours duration with like intervals of no transmission. Nevertheless, these data are sufficient to define the changes in the solar wind that led up to the events under study here. We will describe these data in perhaps more detail than is necessary for the major topic of this paper. However, since these data have not appeared elsewhere and may prove useful in further studies of this storm, we include them here. The plasma detector and analysis procedure has been described by Montgomery et al. [1968, 1970]. Table 1 lists the range of plasma parameters measured on board Vela during each of these data segments. Figure 4b shows the location of each of these data segments on the Vela trajectory and summarizes these observations.
|Fig. 1. The 3-hourly Kp index and the hourly Dst index for the period from 08W UT on October 31 to 1000 UT on November 2, 1968.|
Initially during segment a, from 0600-0800 on October 31, Vela was in the solar wind as expected from Figure 4a. The solar wind was somewhat fast at this time but otherwise typical. Data segment b occurred after the first sudden commencement shown in Figure 1. The solar wind had changed owing to the passage of an interplanetary shock. The velocity increased so that the solar wind was extremely fast (up to 800 km/sec) and the plasma was much denser (up to 22 protons/cm3). Furthermore, the proton temperature had increased an order of magnitude and it was now even greater than the electron temperature.
Substitution of the individual proton velocities and densities during this interval in the formula of Schield  for the subsolar mag netopause radius with an interaction parameter (f2/k) equal to 1.5 [cf. Fairfield, 1971] gives magnetopause distances from 7.1 to 7.7 R.
During data segment c, 1500-1800 UT, Vela encountered the earth's bow shock, spending some time in the shocked interplanetary plasma and some time in the doubly shocked magnetosheath plasma. From examining Figure 4b we see that this shock crossing was close to the normal shock position, despite the fact that from the solar-wind parameters in Table 1 the sub-solar magnetopause radius would be expected to be about 8.2 RE. This position was due to several factors. First, the large Dst index (-183 ) at this time indicates that the magnetosphere was inflated. Schield  states that this should increase the magnetopause radius by about 0.5 RE. Second, use of the hourly average magnetic-field strength of 28.5 given by Ness and Behannon  and the solar-wind density and and velocity from Table 2 gives values of 8, 3.5, and 3.1 for the sonic, Alfven, and magnetosonic Mach numbers, respectively. Use of the magnetosonic Mach number together with the dependence of standoff distance on Mach number of Spreiter and Jones [19631 gives an expected subsolar shock position of 12.7 RE. This is still inside the average shock position, but the lower than normal Mach number would result in a larger than normal Mach angle, which would cause a significant outward displacement of the shock at the Vela, position.
|Fig. 2. The H component of the magnetograms from 6 low-latitude observatories from 0800 UT of November 1 to 1600 UT of November 2, 1968, after Akasofu and Kawasaki [19701. Vertical lines bracket the interval during which Ogo 5 passed through the dayside magnetosphere.|
At the beginning of the data segment d at 2230 UT, Vela was very near the magnetopause as determined from the plasma characteristics 013 board Vela. This was much before the expted time for the magnetopause crossing. This nearness could be due to two causes. First, the magnetosphere had become inflated as the ring current developed. At 2200 UT, Dst was -153 . Second, the solarwind moment flux may have dropped. At 2300 UT, Vela left the magnetosphere and re-entered the magnetosheath. We note that the plasma density had indeedropped from the previous magnetosheath values. We note also that at 2117 UT Ogo 5 entered the magnetosphere, where it remained until 0000 UT. Between 0000 and 0100 UT, Ogo 5 re-encountered the magnetopause several times.
During data segment e, Vela remained in the magnetosheath. But at the start of data segment f on November 1 at 0500 UT Vela was near the magnetopause again, and during the next data segment starting at 1015 it was entirely in the magnetosphere. This was after the sudden commencement, associated with the November 1 stormy had occurred (0918 UT). It was not until 1337 UT that Vela re-entered the magnetosheath. We note that, aside from calculations of the magnetopause position from the solar-wind momentum flux, this is the first direct observation that the magnetosphere was compressed during this storm.
|Fig. 3. The H component of the magnetograms from 9 auroral-zone observatories from 0800 UT on November 11 to 1600 UT on November 2, 1968, after Akasofu and Kawasaki [19701. Vertical lines bracket the interval during which Ogo 5 passed through the dayside magnetosphere.|
After 1337 UT the magnetosheath is very unusual. First, the velocity ranges from 700-1000 km/sec and, since the solar wind in front of the bow shock must have been even faster, the interplanetary flow velocity was certainly exceptionally high. Second, the magnetosheath is moderately dense with 9-33 particles/cm!. Finally, both the proton and electron temperatures continue to be moderately high even for the magnetosheath. During the last data segment i, from 1745-1900, Vela is again in the magnetosphere but near the magnetopause, as determined from the Vela plasma measurements. Since Vela data segments h and i roughly define a magnetopause position nearly simultaneous with the Ogo 5 magnetopause crossing at 1428, it is tempting to reconstruct the magnetopause cross section at this time. However, this cannot be done very accurately for several reasons: the position of the Vela crossing is not lmown precisely, Ogo 5 is near the neutral points where the magnetopause cross section may be quite different from a simple conic section, and the two magnetopause positions are far from the subsolar point, the location of which is the object of such a calculation. Nevertheless, if we do attempt such a fit, we obtain, for a simple conic section centered on the earth, a standoff distance of 5.6 RE and an eccentricity of 0.81. This is sketched in Figure 4b.
|Fig. 4a. The trajectories of Vela 4B and Ogo 5 from 0600 UT on October 31 to 1800 UT on November 1, 1968. The distance perpendicular to the earth-sun line is plotted versus the distance parallel to the earth-sun line for both satellites. The average positions of the magnetopause and shock front in the afternoon hemisphere near the ecliptic plane (Fairfield, 19711 have been sketched for comparison. Both Vela 4B and Ogo 5 (except near perigee) were in the afternoon hemisphere.|
From this brief examination of the Vela data of October 31 and November 1 we see the following facts. Although the solar wind was fast around 15W UT on October 31, its Mach number was quite low. During this sequence of storms the magnetopause position was highly variable. At times the solar-wind pressure was sufficiently intense to severely compress the magnetopause. In particular, the solar-wind velocity in the magnetosheath after 1337 on November 1 corresponds to an interplanetary flow velocity greater than has ever been measured. It is quite likely that the subsolar radius of the magnetopause was less than 6 RE at this time.
|Fig. 4b. The location of the Vela data segments in the same coordinate system as Figure 4a. The nature of the plasma measurement has been schematically shown for each segment. The normal magnetopause and shock positions are also shown [Fairfield, 19711. The dashed curve is an earth-centered ellipse fitted through the Ogo 5 magnetopause crossing at 1428 UT (the cross) and the Vela crossing at 170 UT.|
Ogo 5 is in a highly eccentric orbit with a period of 2.6 days. On November 1, 1968, perigee was 1.73 RE and apogee was 23.35 RE geocentric. Owing to the large inclination of the orbit to the earth's equator, the outbound satellite passes are at much higher magnetic latitudes than the inbound passes.
On the outbound pass on November I the satellite moved from perigee near dawn and passed through the dayside magnetosphere remaining close to 400 magnetic latitude for most of its magnetospheric pass; this is shown in Figure 5. The lower panel shows the plot of magnetic latitude versus radial distance. The upper panel shows the trajectory in L value local time space. Universal times are indicated every 2 RE in radial distance (lower panel) or L value (upper panel). Table 2 gives the L value, magnetic latitude, local time, radial distance, and GSM Cartesian coordinates of the position of Ogo 5 every 15 min from 1130 to 1500 UT.
|Fig. 5. The trajectory of Ogo 5 during its outbound pass on November 1, 1968. The upper panel shows the orbital trace in local time versus L-value space. The universal time of the crossing of even-numbered L shells has been indicated. The lower panel shows the orbital trace in magnetic latitude versus radial distance. The universal time every 2 Rir has been indicated. Since the satellite is at a high magnetic latitude on this pass, it reaches the limit of the upper panel (L = 11) much sooner than the limit of the lower panel (R = 11).|
The Ogo 5 satellite can transmit data in real time at three rates, 1, 8, and 64 kbits/sec, and store data at a rate of 1 kbit/sec. The different telemetry rates govern the sampling capability (if the various experiments so that the data at the lowest telemetry rate will have less resolution in time than at the higher telemetry rates. On this orbit the telemetry rate was 64 kbits sec from 1140 to 1328 UT. From 1328 until 1500 UT the satellite was in data storage mode.
The three panels of Figure 6 show simultaneous measurements from three of the Ogo 5 particle detectors. The lowest panel is the thermal proton density from 0-600 ev as measured by the Lockheed ion mass spectrometer. This instrument has been described in detail by Harris and Sharp . The apparent plasmapause between the points labeled A and B is actually a secondary plasinapause. The main plasmapause occurred at approximately 1138 UT at an L value of 2.1. This density profile has been discussed in more detail by Chappell et al. . No data are available from this experiment for the period from 1330 to 1500 UT.
The middle panel shows the energetic electron flux from 50 to 1100 kev as measured by the UCLA energetic electron detector. The outer-zone maximum occurred at 1149 at an L value of 2.8. Although this was a rather low L value for the outer-zone maximum and in fact occurred where the quiettime slot usually occurs, this position was outside of the primary plasmapause, in agreement with previous studies during the main phase of a geomagnetic storm [Russell and Thorne, 1970]. The limit of trapped energetic electrons was coincident with the beginning of the secondary plasmapause at point A, 1218 UT. The dipole L value here was 5A, corresponding to an invariant latitude of 65o, and the magnetic latitude was 42o. Beyond this point the counting rates were at or near background except near the points labeled C and E and between points F and G.
The upper panel Shows the energy density of suprathermal electrons from 50 to 3200 ev as measured by one of the Jet Propulsion Laboratory solar wind spectrometers. This instrument has been described in detail by Neugebauer . The data used here were obtained by a curved-plate analyzer on the main body ofthe spacecraft facing radially away from the earth. The instrument measures the electron spectrum at 60 logarithmically spaced energies from 50 to 3200 ev every 18 sec at the 64-kbit/sec telemetry rate and every 295 sec at the 1-kbit rate. The energy density shown here was obtained by integrating the individual electron spectra assuming an isotropic pitch-angle distribution.
|Fig. 6. The electron energy density from 50 to 3200 ev measured by the JPL solar-wind experiment; the energetic electron flux from 50-1100 kev, measured by the UCLA energetic electron spectrometer; and the thermal proton density, measured by the Lockheed ion mass spectrometer, during the Ogo 5 outbound pass through the magnetosphere on November 1, 1968.|
The first flux of suprathermal electrons was encountered immediately outside the energetic electron trapping boundary. Later, large fluxes of supratherinal electrons were again encountered at times nearly simultaneous with enhancements of the energetic electrons. Finally between points F and G a large and fairly steady flux was encountered. As will be shown below, point G is the magnetopause. The increase in the energy density after point G was due primarily to an increase in number density and may be due at least in part to temporal changes in the magnetosheath, since the change in the electrons occurred over a much longer period than the duration of the magnetopause crossing.
|Fig. 7. The pitch angle of the curved-plate analyzer of the JPL solar-wind experiment mounted on the main body of Ogo 5, which was measuring electrons during this paw; the energy density of these suprathermal electrons (50-3200 ev) assuming isotropy; the number density of these electrons assuming isotropy; and their average energy.|
From top to bottom, Figure 7 shows the angle between the normal to the suprathermal electron detector and the magnetic field, the supra-thermal electron energy density (same as in Figure 6), the density of 50- to 3200-ev electrons, and the mean energy per electron. With only a few minor exceptions, the pitch angle is always greater than 1200 within the magnetosphere. Since the satellite was in the northern hemisphere during this pass, the electron detector therefore was always looking up the magnetic-field line toward the magnetosheath. In fact near 1300 UT the detector was almost antiparallel to the magnetic field.
The number density during the enhanced fluxes in the magnetosphere ranged from 10 to 40 and reached 100 particles/cm3 in the magnetosheath. However, the absolute values of both the number density and the energy depend on our assumption of isotropy. On the other hand, the average energy and the shape of the energy spectra of these electrons do not. As can be seen from Figure 7, the average energy in the regions of enhanced fluxes in the magnetosphere and the average energy in the magnetosheath are similar.
Figure 8 shows four electron spectra during this pass. The first two spectra were obtained at 64-kbit/sec telemetry rate during the first two encounters with enhanced fluxes while Ogo 5 was deep in the magnetosphere. The interval between the sampling df the first and last points on these spectra is approximately 7 sec. The second two spectra were obtained at the 1-kbit/ sec telemetry rate, for which approximately 2 min elapsed between the sampling of the first and last points. The first of these two low bit rate spectra was obtained between points F and G, and the last in the magnet6sheath after point G.
|Fig. 8. Four representative spectra of the suprathermal electrons detected on board Ogo 5. Each spectrum is displayed as the logarithm of the distribution function versus the square of the velocity. In the absence of temporal or spatial variations during the measurement of a spectrum, this display results in a straight line for a Maxweuian distribution. Best fit straight lines have been drawn for each spectrum and the corresponding temperatures and number densities are indicated on each spectrum. The times shown correspond to the first. and last point plotted. The first two spectra were obtained at the highest sample rate and the last two at the lowest rate. All but the last spectrum were measured within the magnetosphere. The last spectrum was measured ia the magnetosheath.|
These four spectra are displayed as the logarithm of distribution functions versus the square of the velocity so that the slope of the spectrum is proportional to temperature. If the spectra were strictly nonconvective Maxwellian, with no temporal variations during the time in which the spectrum was measured, the spectra would be straight fines in this display. We see that there are in fact deviations from the best fit temperatures displayed on these graphs but that it is not unreasonable to assign a temperature to these spectra. Within a factor of 2 the temperatures were similar deep in the magnetospbere, in the outer magnetosphere, and in the magnetosheath. This similarity in temperature suggests that to lowest order the particles gain entry to the magnetosphere without energization. However, the data also suggest that some energization is taking place.
The average energy in Figure 7 shows that the electrons were significantly hotter near C and E than during the remainder of the pass and significantly cooler immediately after the entry into the magnetosheath at 1428 UT. We cannot rule out the possibility that temporal changes were occurring in the magnetosheath electrons during this period, but we note that the times of high average energy near C and D were also times when energetic electron fluxes were observed. Furthermore, some of the deviations of the electron spectra from a Maxwellian distribution were not simply temporal fluctuations. For example, the large fluctuation near v2 = 150 x 106 km2/sec2 in the first spectrum in Figure 8 persisted throughout this encounter with the cusp.
We can also compare simultaneous electron spectra in the magnetosphere and in the magnetosheath during a portion of this pass. Table I shows that at 1337 UT Vela entered the magnetosheath from the magnetosphere possibly as a result of a change in the solar wind. Simultaneously Ogo 5 reencountered the enhanced suprathermal electron fluxes (point F in Figures 6 and 7). We note that from 1337 to 1415 UT (the end of the Vela data segment) the electron spectrum at Ogo 5 (as can be judged from the number density and average energy plotted in Figure 7) and the spectrum at Vela remained roughly constant. Figure 9 shows two simultaneous spectra. Although the electron temperature is somewhat higher than usual at Vela, the spectral shape is typical (Montgomery et al., 19701. The flat-topped portion at low velocities has been suppressed by the choice of abscissa. Although the number densities are comparable, we note that the temperature of the Vela electrons is about a factor of 4 less than that of the Ogo 5 electrons. Such a difference could be easily due to the fact that the Ogo 5 electrons enter the magnetosphere much closer to the stagnation point than the position of Vela. As the electrons move from the front of the magnetosphere along the flanks of the magnetosphere, they would be expected to cool [Spreiter et al., 1966].
|Fig. 9. Simultaneous electron spectra obtained by Ogo 5 in the polar cusp and Vela 4B in the downstream magnetosheath. Best fit temperatures are shown for both spectra. Vela 4B was behind the dawn-dusk meridian at this time, while Ogo 5 was in the polar cusp within the magnetosphere near noon. The spectra are plotted as the logarithm of the distribution function versus the square of the velocity as in the previous figure.|
Again, it appears that the enhanced fluxes of suprathermal electrons are consistent with the direct entry of magnetosheath plasma into the magnetosphere. Since Ogo 5 first encountered these fluxes just outside the outer zone of energetic electrons, we identify these regions as encounters with the polar cusp. However, if each of these regions of enhanced suprathermal electron fluxes is indeed the polar cusp, then either the polar cusp is quite broad in latitudinal extent and is capable of switching on and off or the polar cusp can move about quite freely, at least during a large magnetic storm. Frank  has shown that at quiet times the polar cusp is a thin region about 1 RE thick near the magnetopause and about 30 km thick near the auroral zone. Furthermore, it has not been observed to switch on and off. Thus the second alternative seems most likely and, in fact, we show below that this motion appears to be controlled by the direction of the interplanetary magnetic field as well as by the solar-wind momentum flux.
Magnetosheath protons are also expected to be present in the polar cusp. The solar-oriented curved-plate analyzer on Ogo 5 was in a positive-ion measurement mode during this pass but did not detect any protons until Ogo 5 entered the magnetosheath at point G. The threshold sensitivity of this analyzer is 2 X 105 normally incident protons/cm2/sec/energy channel. The detector had an acceptance cone of 5o, half-width half-maximum, and between 1220 and 1430 UT the angle between the detector normal and the magnetic field decreased from 1500 to 40o. Thus this null result is consistent with Frank's  observation that the angular distribution of polar-cusp protons is strongly peaked along the local magnetic field and does not imply that such protons were not present.
Figure 10 compares 1-min averages of the magnetic Figure 10 field measured by the UCLA triaxial fluxgate magnetometer with the expected field due to the earth's internal sources. The magnetometer has been described in detail by Aubry et al. . The upper panel shows the difference between the measured field strength and the expected field strength. A negative difference indicates that the measured field is weaker than the reference field. The middle panel shows the inclination of the field, which is the angle between the magnetic field and thelocal horizontal. It is positive in the northern magnetic hemisphere. The lower panel shows the declination of the field, which is the angle of the field in the local horizontal plane. The zero of this angle is the projection of the north dipole on the horizontal plane; positive angles point eastward.
|Fig. 10. The deviation of the field magnitude from the reference field, and the inclination and declination of the field at Ogo 5 derived from 1-min averages of the vector field components for the outbound pass on November 1, 1968. The inclination and declination of the reference field are indicated by thin lines. The letters A to G indicate the times corresponding to the various features labeled on Figures 6 and 7.|
From Figure 10 we see that the many features of the particle data of Figures 6 and 7 have their counterparts in the magnetic field. Specifically, the enhanced suprathermal electron fluxes between points A and B and at points 0, D, and E were accompanied by depressions in the field magnitude. However, the electrons do not have sufficient energy density to account for the 'missing' magnetic energy density in these spikes. The difference between the magnetic energy density in the spikes and in the adjacent regions is of the order of 1(r ev/cm!, which is an order of magnitude greater than the energy density of the suprathermal electrons. Since the electron number densities at these peaks are of the order of 30 cin, it is probable that the supratherinal electrons were accompanied by protons with average energies of about 3 kev. We note that the kinetic energy of a proton with a velocity of 800 kin/sec, which is typical of the velocities measured by Vela from 1337-1415 UT, is 3.3 kev. Thus it is reasonable that 3-kev protons would be present in the polar cusp at this time.
Before 1330 UT the measured field was less and the inclination was greater than that of the reference field. This is to be expected in an inflated magnetosphere. However, the depression at the surface of the earth was close to - 100 , while the average depression at Ogo 5 outside the spikes was about -20 . This difference could be due to the fact that Ogo 5 was outside of the bulk of the ring current or that the ring current was highly asymmetric at this time. We note that Ogo 5 was above the morning hemisphere until 1245 UT. After 1330 the field magnitude increased above the reference field and the inclination decreased below that of the reference field, thus indicating a compressed magnetosphere. This is consistent with the Vela entry into the magnetosheath at 1337. With the exception of the period around 1300 UT, the declination was consistent with the sweeping back of field lines toward the tail: a positive declination in the morning hemisphere and a negative declination in the afternoon. At 1300 UT the declination swept through 360o. This means that the field lines rotated a full turn about the radial direction. At 1428 UT, labeled G, the satellite crossed the magnetopause. This is seen most clearly in the difference field and the declination. It is also quite evident in the rms deviations of the field discussed in the next section.
Figure 11 shows the rms deviations of the magnetic field at ULF frequencies and the maximum and minimum VLF electric field amplitudes during this pass. The rms deviations are 1-min averages of the square root of the sum of the power on each of the 3 vector components for oscillations in the frequency range from 0.07 Hz to the Nyquist frequency. Before 1328 UT the Nyquist frequency was 27.8 Hz, and after this time it was 0.43 Hz. This rather large decrease in bandwidth at 1328 UT did not appreciably affect the values for the rins deviations, since the major fraction of the power was at low frequencies. The major fraction of the decrease in the rms deviations occurring near this time actually occurred while the satellite was still transmitting at the high data rate.
The power at ULF frequencies here is much larger than is usually present in the magnetosphere. Typical wave amplitudes in this frequency band in the dayside magnetosphere are less than 0.25 . We note that, while the field is almost continually more disturbed than on a typical quiet day, the major field disturbances occurred during or near the encounters with the suprathermal electrons. The peak noise levels were commonly between 5 and 10 at these times and occasionally exceeded 10 . At point G (1428 UT) the magnetopause was crossed and the noise level increased sharply to about 20 . Such a large amplitude is unusual even in the magnetosheath. Typical values are from 2 to 5 in this frequency range, but by 1500 UT on November 1 the rms deviations reached an amplitude of 99 .
|Fig. 11. One-min averaged rms deviations of the magnetic field and the VLF electric-field amplitude at 13 kHz on this pass. The rms deviations are the square root of the sum of the power in each of the vector components of the field from 0.07 Hz to the Nyquist frequency. The two electric-field tram are the maximum and minimum electric-field amplitudes over a 30-sec interval once every 3.2 min.|
The lower panel of Figure 11 shows the maximum and minimum electric field amplitudes at 1.3 kHz during a period of 30 sec every 3 1/4 min as measured by the TRW electric-field experiment. This experiment has been described by Crook et al. . As with the ULF magnetic noise, the VLF electric measurements are above typical quiet-day values almost continuously. In addition, noise enhancements occurred at or near the regions of large suprathermal electron fluxes. At ELF frequencies the Ogo 5 search coil magnetometer also detected strong bursts of magnetic noise between points A and B, and at points C, D, and E. The noise near points A and B extended from at least 10 to 1000 Hz, while the noise at C, D, and E extended only up to about 100 Hz (R. K. Burton and R. E. Holzer, private communication, 1971).
Figure 12 shows four power spectra calculated from the fluxgate magnetometer data at times near points B, C, D, and E. The frequency range covered is from 0.06 to 27.7 Hz, which corresponds to the frequency band of the rms deviations shown in Figure 10. The data were first rotated into a field-aligned coordinate system and detrended. Then, using the Fast Fourier Transform algorithm on segments of 8192 points (2.5 min of data), power spectra with 20 degrees of freedom were calculated. The power parallel to the magnetic field, b, and the power in one of the two perpendicular components, b, is shown. The dashed line shows the power spectrum during the quiet period at 1245 UT for reference. At this time all three components had similar powers and were close to the noise level of the magnetometer.
The spectra in Figure 12 have several common features and several differences. First, all spectra have more power transverse to the field than parallel to the field at every frequency. However the ratio of the transverse to compressional power varies from spectrum to spectrum and is a function of frequency. Second, all spectra have approximately a f2 dependence over most of the frequency range. However, the last spectrum has about a f4 dependence above the proton gyrofrequency. The lack of any change in the spectrum at the proton gyrofrequency for the second and third spectra is actually the surprising feature. (Any change in the first spectrum near p, could be masked by the close proximity of the Nyquist frequency and the low power spectral density.) The explanation of this could be either that no waves propagating in the left-hand mode have been generated or that the noise observed is not due to propagating waves but is rather the near field magnetic fluctuations caused by time-varying currents. The second explanation seems most attractive, since Frank  has found that the polar cusp carries current and since large fluctuating field-aligned currents are present in the auroral zone [Russell and Holzer, 1970].
|Fig. 12. Four representative power spectra of the fluctuations observed in the polar cusp. The dashed line is the spectrum observed outside the polar cusp at 1245 UT. The spectrum labeled b, is the power spectral density of the component along the observed magnetic field; the spectrum labeled b is the power spectral density of the component perpendicular to the field in the dipole magnetic meridian. Each power spectrum was computed using 8192 original data points with 20 degrees of freedom, and each section of data was detrended before the power spectrum was computed.|
As stated previously, the presence of strong fluxes of suprathermal electrons occurring first at the termination of the outernzone energetic electrons suggests that these flux enhancements represent the entry into the polar cusp. The average energy of these electrons, the implied presence of low-energy protons from the magnetic field depression, and the possibility of fluctuating currents provide further support for this view. However, questions arise as to why the polar cusp is encountered at such a low magnetic latitude and why the cusp appears to come and go. Figure 13 shows the necessary approximate position of the polar cusp and the normal quiettime location obtained by Frank , together with the trajectory of Ogo, 5. We see that our interpretation requires a dramatic displacement of the polar cusp from its usual position.
|Fig. 13. The inferred configuration of the polar cusp at 1220 and 1315 UT during this pass drawn in the dipole meridian plane. The normal location of the polar cusp [Frank, 19711 is shown for comparison. The arrows indicate the direction of the projection of the field in this plane. The location of the polar cusp has been drawn by requiring the field to be dipolar near the earth and parallel to the observed field at Ogo 5. The field line beyond the orbit of Ogo 5 is highly schematic. The invariant latitude of the polar cusp, A, obtained by this graphical technique is also shown.|
However, it is quite reasonable that such a dramatic change in the magnetosphere occurred. First, the Vela data show that the solar-wind momentum flux must have been very high at this time. Although no data are available in the solar wind, the magnetosheath data indicate that the momentum flux must be similar to or even greater than the flux observed on October 31, which was strong enough to give a calculated magnetopause radius of 7 RE. The magnetopause position observed by Ogo 5 at 1428 UT and the fact that Vela was in the magnetosheath from 1337 to 1415 are consistent with such a severe compression. Second, the interplanetary magnetic field was strongly southward at times during this event. Aubry et al.  have shown that a southward-directed interplanetary magnetic field leads to an erosion of the dayside magnetosphere. Of course, such an erosion will cause the neutral points and therefore the polar cusp to move equatorward. At the time of the first Ogo 5 encounter with the cusp at 1220 UT, the interplanetary magnetic field in GSM coordinates was 25 southward at Explorer 33 (positioned 30 R1, in front of the earth and 30 RE to the dawnside of the earth-sun line). At 1230 UT the field reached 30 southward (D. Colburn and E. Ungstrup, private communication, 1970).
Figure 14 shows the field strength and the north-south component in the GSM coordinate system as measured by the Explorer 33 ARC magnetometer during the Ogo 5 pass, together with two Ogo 5 measurements. Since no velocity measurements were available from Explorer 33 or 35 at this time and the direction of the field in the ecliptic plane varied during this interval, it is impossible to calculate delay times and it is probable that the delay time changed throughout this period. Nevertheless, these data can be used to more fully understand the Ogo 5 observations.
At the first cusp encounter the southward component was increasing. Thus it is probable that the cusp was moving inward at Ogo 5 and equatorward on the magnetopause. Together with the outward motion of Ogo 5 this caused the satellite to penetrate the cusp at time A and pass into the polar-cap field lines at time B. However, just after 1230 the magnetic field turned northward for 10 min, then returned southward for 9 min, and finally returned northward for a long period of time.
If we suppose that the magnetosphere could have responded almost instantaneously to these changes, the cusp should have moved poleward, then equatorward, and then polarward again. This is reasonable because the erosion on the dayside should cease immediately when the field turns northward; moreover, there was a large substorm in progress, as can be seen from Figures 2 and 3, and the flux eroded into the tail was continually being returned from the tail at this time. Therefore our interpretation of the encounter with the suprathermal electrons from 1252 to 1302 UT, the gap from 1302 to 1309 UT, and the second encounter from 1309 to 1321 UT is that when the interplanetary field turned northward the cusp moved poleward, embedding Ogo 5 in the cusp starting at 1252 UT. However, the cusp badnot completely crossed Ogo 5 when the interplanetary field again turned southward, causing the cusp to proceed equatorward again, and Ogo 5 left the cusp and entered the polar cap at 1302 UT. Then, the interplanetary field returned northward and the polar cusp began to drift poleward, initially encountering Ogo 5 at 1309 UT and finally crossing to a position poleward of Ogo 5 at 1321 UT. We note that the duration of the first of these two cusp encounters was equal to the duration of the northward field and the gap between encounters was similar to the duration of the southward field (7 versus 9 min). The time lag between the change in direction of the interplanetary field at Explorer 33 and the resultant appearance and disappearance of the cusp at Ogo 5 was 16 min.
|Fig. 14. The interplanetary field during the polar-cusp encounters. The upper panel shows the deviation of the magnetic field from the reference field at Ogo 5 and the panel below it shows the suprathermal electron energy density. The bottom two panels show the interplanetary field measured by Explorer 33 at 0900 LT and 42 Ri, from the earth. The upper of these two panels is the field strength and the lower panel is the geocentric solar magneospheric (GSM) Z component of the field. The Ogo 5 field data are 1-min averages; the Explorer 33 data are 81.2-sec averages.|
Later, as the satellite proceeded outward it appears that the satellite also encountered the cusp at F and remained in it until the magnetopause. Since the satellite was moving outward through the magnetic field and the interplanetary field remained northward during this interval, it is possible that the satellite entered the cusp simply owing to its outward progress across field lines. However, this entry into the cusp was coincident with the Vela entry from the magnetosphere into the magnetosheath. Thus, it appears that the cusp did move equatorward at this time, probably owing to a change in the solar-wind momentum flux.
Finally, at 1420 UT the solar-wind magnetic field turned southward again. It is probably more than a coincidence that 8 min later, at G, Ogo 5 entered the magnetosheath. Such a time delay, although not identical to the required time delay at 1250 UT, could be due to a change in the field in the ecliptic plane or a different response time of the magnetopause from that of the polar cusp.
During, the November 1 Ogo 5 pass, both a solar-proton and a solar-electron event were in progress. These particles were detected in the magnetosphere, magnetosheath, and interplanetary medium by the Ogo 5 UCLRL energetic electron and proton spectrometer (H. West, personal communication, 1971). Solar electrons, while they can gain access to nonpolar-cap field lines to a limited extent (Vainpola, 1971], provide a good indicator of the polar-cap field lines. Preliminary data from the UCLRL detector showed, in fact, solar electrons everywhere outside of the outer zone and the cusp encounters. However, to use the solar electrons to identify polar-cap field lines unambiguously, i.e., to distinguish open from closed field lines, it would be necessary to check whether there was a double or single loss cone [Vampola, 1971]. However, at the altitude of Og6 5 the loss cone is very small and the range of pitch angles sampled is insufficient for such a test.
Solar protons were also present at these times, but since solar protons have a much wider access to the magnetosphere they are not useful as tracers of the polar cap. However, they do provide support for our identification of the magnetopause at 1428 UT. At this time the solar proton distributions became highly anisotropic about roughly the solar direction and remained anisotropic thereafter through the magnetosheath and into the interplanetary medium.
The observations made on board Ogo 5 on November 1, 1968, provide no information on the normal configuration of and processes occurring in the quiettime polar cusp. However, they do show that certain of the characteristics of the quiettime polar cusp, as explored by Frank , are unchanged at disturbed times. The Ogo 5 data show that both protons and electrons gain entry from the magnetosheath deep into the magnetosphere. They also show that this region is quite extensive in local time. The first encounter occurred at 1100 and the last, at 1400 LT.
Other characteristics of the cusp, not reported by Frank  but revealed by this set of measurements, are that the cusp is a magnetically and electrically turbulent region, that energetic electrons can be present in the cusp although absent in the magnetosheath plasm, and that the cusp location is apparently controlled by the interplanetary magnetic field as well as by the solar-wind momentum flux. The magnetic turbulence was seen both at ULF and ELF frequencies and was much greater on this day than is normally seen in the dayside magnetosphere. In fact, the power observed in the cusp at ULF frequencies exceeded the power observed on Ogo 5 in any other region of the magnetosphere at any time, and the amplitude of the VLF electric-field noise is similar to that commonly seen at the bow shock. Energetic electrons were not observed in the cusp by Frank  at quiettimes. Thus their presence at this time may be an unusual occurrence. These electrons may be related in some way to the energetic electrons observed in a layer near the magnetopause [Meng and Anderson, 19701. The control of the cusp position shown on this day provides further evidence of the importance of the sign and magnitude of the north-south component of the interplanetary magnetic field in controlling the dynamical processes of the magnetosphere. This is not to say that the magnetospheric configuration is independent of the incident solar-wind dynamic pressure but rather that the interplanetary field variations cause changes from the normal equilibrium position.
Finally, we note that we have only briefly surveyed the observations from a small number of the Ogo 5 experiments operating on this orbit. Thus there are many very intriguing detailed observations of the particles and fields on this day that have not been mentioned. While further study of these data is not expected to alter any of the conclusions of this paper, such a study is providing a wealth of information on the microscopic behavior of the magnetosphere on this day, and further papers on these data are planned.
Acknowledgments. Other coinvestigators for the Ogo 5 experiments discussed in this paper were: C. W. Snyder (plasma spectrometer), P. J. Coleman, Jr., T. A. Farley, and D. L. Judge (fluxgate magnetometer and energetic electron spectrometer), G. W. Sharp (light ion mass spectrometer), and G. M. Crook (VLF electric field experiment). The principal investigator for the Vela 4 plasma experiment was S. J. Bame of the Los Alamos Scientific Laboratory. We are indebted to C. P. Sonett and D. S. Colburn for the interplanetary field data from their NASA Ames Research Center magnetometer on Explorer 33, and we acknowledge with thanks the receipt of the hourly Dst values from W. E. Valente of the National Space Science Data Center. We are particularly grateful to L. A. Frank for furnishing us with advance information about the Imp 5 measurements and for many valuable discussions about the cusp morphology. We also thank J. N. Barfield and R. L. McPherron, R. K. Burton and R. E. Holzer, and H. I. West, Jr., for allowing us to examine their data in advance of publication. Useful discussions of these observations have been held with R. K. Burton, T. A. Farley, M. G. Kivelson. R. L. McPherron, V. M. Vasyliunas, and H. I. West, Jr.
This work was supported by the National Aeronautics and Space Administration at UCLA under NASA contract NAS 5-9098, at Lockheed by NASA contract NAS 5-9092, and at TRW by NASA contract NAS 5-9278 and represents one phase of research carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract NAS 7-100 sponsored by the National Aeronautics and Space Administration. The Vela observations were obtained as part of the VeIa Nuclear Test Detection Satellite Program which is jointly sponsored by the Advanced Research Projects Agency of the Department of Defense and the U.S. Atomic Energy Commission.
The Editor thanks S.-I. Akasoftt and L. A. Frank for their assistance in evaluating this paper.
Akasofu, S.-I., and K. Kawasaki, Geomagnetic disturbances (October 26-November 5, 1968) associated with McMath plage region 9740, in Upper Atmos. Geophys. Rep. UAG-8, compiled by J. Virginia Lincoln, p. 243, US. Dep. of Commer., ESSA, Asheville, N. C., 1970.
Antonova, A. E., and V. P. Shabansky, On the particle motion in the magnetosphere disturbed by the solar wind, Moscow State Univ. Res. Rep, Moscow State Univ., Moscow, 1968.
Aubry, M. P., C. T. Russell, and M. G. Kivelson, On inward motion of the magnetopause before a substorm, J. Geophys. Res, 75(34), 7018, 1970.
Aubry, M. P., M. G. Kivelson, and C. T. Russell, Motion and structure of the magnetopause, J. Geophys. Res, 76(7), 1673, 1971.
Chappell, C. R., K. K. Harris, and G. W. Sharp, Ogo 5 measurements of the plasmasphere during observations of stable auroral red arcs, J. Geophys. Res., 76(10), 2357,1971.
Crook, G. M., F. L. Scarf, R. W. Fredricks, I. M. Green, and P. -Lukas, The Ogo 5 plasma wave detector: Instrumentation and in-flight operation, IEEE Trans. Geosci. Electron, GE-7(2), 120, 1969.
Fairfield, D. H., Average magnetic field configuration of the outer magnetosphere, J. Geophys. Res., 73(23), 7329, 1968.
Fairfield, D. H., Average and unusual locations of the earth's magnetopause and bow shock, J. Geophys. Res, 76, this issue, 1971.
Frank, L. A., Plasma in the earth's polar magnetosphere, J. Geophys. Res., 76, 5202, 1971.
Frank, L. A., and K. L. Ackerson, Observations of charged particle precipitation into the auroral oval, J. Geophys. Res, 76(16), 3612, 1971.
Harris, K. K, and G. W. Sharp, Ogo 5 ion mass spectrometer, IEEE Trans. Geosci. Electron., GE-7(2), 93, 1969.
Heikkila, W. J., and J. D. Winningham, Penetration of magnetosheath plasma to low altitudes through the dayside magnetospheric cusps, J. Geophys. Res., 76(4), 883, 1971.
Mead, G. M., Neutral points on the boundary of the closed magnetosphere, Space Sci. Rev., 7, 158, 1967.
Meng, C.-I., and K. A. Anderson, A layer of energetic electrons (>40 kev) near the magnetopause, J. Geophys. Res., 75(10), 1827, 1970.
Montgomery, M. D., S. J. Bame, and A. J. Hundhausen, Solar wind electrons: Vela 4 measurements, J. Geophys. Res., 73(15), 4999, 1968.
Montgomery, M. D., J. R. Asbridge, and S. J. Bame, Vela 4 plasma observations near the earth's bow shock, J. Geophys. Res., 75(7), 1217,1970.
Ness, N. F., and K. W. Behannon, Interplanetary magnetic field measurements October 24, 1968- November 23, 1968, in Upper Atmos. Geophys. Rep. UAG-8, compiled by J. Virgina Lincoln, p. 221, U.S. Dep. Commer., ESSA, Asheville, N. C., 1970.
Neugebauer, M., Initial deceleration of solar wind positive ions in the earth's bow shock, J. Geophys. Res., 75(4), 717, 1970.
Piddington, J. H., The magnetosphere and its environs, Planet. Space Sci., 13(5), 363, 1965.
Roederer, J. G., Quantitative models of the magnetosphere, Rev. Geophys., 7, 77, 1969.
Russell, C. T., and R. E. Holzer, AC magnetic fields, in Particles and Fields in the Magnetosphere, edited by B. M. McCormac, p. 195, D. Reidel, Dordrecht, Netherlands, 1970.
Russell, C. T., and R. M. Thorne, On the structure of the inner magnetosphere, Cosmic Electrodynamics, 1, 67, 1970.
Schield, M. A., Pressure balance between the solar wind and magnetosphere, J. Geophys. Res., 75(5), 1275, 1969.
Spreiter, J. R., and B. R. Briggs, Theoretical determination of the form of the boundary of the solax corpuscular stream produced by interaction with the magnetic dipole field of the earth, J. Geophys. Res., 67(l), 37, 1962.
Spreiter, J. R., and W. P. Jones, On the effect of a weak interplanetary magnetic field on the interaction between the solar wind and the geomagnetic field, J. Geophys. Res, 68(12), 3555, 1963.
Spreiter, J. R., and A. L. Summers, On conditions near the neutral points on the magnetospheric boundary, Planet. Space Sci., 15, 787, 1967.
Spreiter, J. R., A. L. Summers, and A. Y. Alksne, Hydromagnetic flow around the magnetosphere, Planet. Space Sci., 14, 223, 1966.
Vampola, A. L., Access of solar electrons to closed field lines, J. Geophys. Res., 76(l), 36, 1971.
Walters, G. K., On the existence of a second standing shockwave attached to the magnetopause, J. Geophys. Res., 71(5), 1341, 1966.
Willis, D. M., The influx of charged particles at the magnetic cusps on the boundary of the magnetosphere, Planet. Space Sci., 17, 339, 1969.
World Data Center A., Data on solar geophysical activity October 24-November 6, 1968, in Upper Atmos. Geophys. Rep. UAG-8, compiled by J. Virginia Lincoln, U.S. Dep. Commer., ESSA, Asheville, N. C., 1970.