Pkmct. Space Sci., Vol. 24, pp. 921 to 935. Pergamon Press, 1976. Printed io Northern Ireland
PC5 PULSATIONS IN THE OUTER MAGNETOSPHERE: A SURVEY OF HEOS-1 DATA P. C. HEDGECOCK The Blackett
SW7 2BZ, England
(Receiued in final form 3 Februnry 1976) Abstrct-Magnetic field measurements from 133 low-latitude transits of the HEOS-1 satellite through the magnetosphere have been used to analyse the low-frequency pulsation activity in the outer regions of the geomagnetic field. Providing full longitude coverage in the sunward hemisphere at geocentric distances larger than -7.5 R,, this survey complements previous low-frequency pulsation data from satellites at smaller geocentric distances. Several giant PC5 events, each being mainly compressional and lasting l-2 hr, are described in detail and it is shown that this phenomenon is relatively common in the 8-12 R, geocentric distance range near dusk. A depression of the ambient field magnitude always accompanied the events, suggesting that they are associated with a region of enhanced plasma pressure. The properties of these wave events are compared with the predictions of current micropulsation theories involving a Kelvin-Helmholtz generation mechanism and field-line resonance. Unlike the PC5 events observed nearer Earth, these events were not obviously related to periods of enhanced geomagnetic activity. 1. INTRODUCTlON
at least for the longer period events, changes appreciably from station to station over a small range of magnetic latitude. The sense of polarisation of the component transverse to the ambient field has been found to reverse in a systematic manner in going from north to south on a given meridian through the latitude of maximum amplitude and a similar reversal is found in going from the morning to the evening sector at constant magnetic latitude (Kaneda et al., 1964; Oberts and Raspopov, 1968, Samson et al., 1971). Raspopov et al. (1968), Chen and Hasegawa (1974), Hasegawa and Chen (1974) and Southwood (1974) discuss the theoretical implications of this behaviour in terms of pulsation source mechanisms. Despite the wealth of ground station data, there are relatively few reports of satellite measurements of long period pulsations in the distant magnetosphere. The early measurements from the Pioneer space probes (Coleman et al., 1960; Sonnett, 1963) and from Explorers 6 (Judge and Coleman, 1962), 12 (Patel, 1966a) and 14 (Pate1 and Cahill, 1964; Patel, 1965), displayed several examples of fluctuations with periodicities in the range 10-200 sec. These were oberved at many widely separated lowlatitude locations in the Sunward hemisphere. Instrumental limitations prevented the observation of detail but it was clear that these events occurred relatively frequently at 5-10 R, geocentric distance and that the wave characteristics ranged from mainly compressional to mainly transverse (with respect to the ambient field). Typically, the waves had a 2-4 y amplitude in an ambient field of -50 y
Following the first reports of long-period, quasisinusoidal pulsations in the geomagnetic surface field (Rolf 1931; Sucksdorff 1939), much effort has been devoted to understanding the morphology and occurrence characteristics of the 150-600 set period waves classified as PC5 by Jacobs et al. (1964). Reviews of the observational and theoretical aspects of this work have been presented by Troitskaya (1967), Campbell (1968), Saito (1969), Dungey and Southwood (1970), Jacobs (1970) and Orr (1973). Recently, the statistical properties of PCS in the aurora1 zones have been discussed by Samson et al. (1971), Raspopov et al. (1972) and Gupta (1973, 1974, 1975). In the context of this paper, the important points to note from the ground station results are that the PC5 event may span some 10-30” of latitude (Saito, 1969, Fig. 25) and at least a comparable range of longitude (e.g. Patel, 1966b; Lanzerotti et al., 1974, Fig. 5), that event durations of l-2 hr are common (e.g. 01’ 1963) and that the waveforms are rather sinusoidal. Statistically, the events have maximum amplitude in a rather small range of latitude in the aurora1 regions (Saito, 1969, Fig. 25) where the most probable amplitude is a few tens of gamma (Gupta, 1975, Figs. 3 and 4). The diurnal variation in their probability of occurrence maximises near dawn and dusk (Kitamura, 1963; Ol’, 1963; Saito, 1964a, 1969; Kato and Utsumi, 1964). Gupta (1975, Fig. 5), however, presents evidence which indicates that the form of this daily variation, 921
(1 y = In T’), displayed a rather irregular waveform, had a duration corresponding to a few periods of the wave and showed a polarisation reversal of the transverse component near the 10:00 LT meridian (Patel, 1965). The latter is in general agreement with the polarisation properties noted at ground stations (Nagata et al., 1963) and correlation with ground station events near the “foot of the field line” was obtained (Patel, 1965, 1966b). Subsequently, several events have been observed in detail. Brown et al. (1968), Sonnerup et al. (1969) and Lanzerotti et al. (1969) have described an event observed with Explorer 26 at -5 R, equatorial geocentric distance near the 13: 30 LT meridian. The magnetic field fluctuations were found to correlate with ground station magnetic pulsations and with changes in the fluxes of relativistic electrons and 100-300 keV protons measured at the satellite. Lanzerotti et al. (1969), Barfield et al. (1971) and Lanzerotti and Tartaglia (1972) have also discussed a similar event observed by ATS-1 in geosynchronous orbit in the 1300-1600LT zone. More recently, Lanzerotti et al. (1974, 1975a,b) have described a storm-time PC5 event observed with Explorer 45 at -5.5 R, equatorial geocentric distance near 20:00 LT. On this occasion the disturbance at the satellite was mainly transverse and the event was recorded at several ground stations near the “foot” of the field line passing through the satellite’s position. Heppner et al. (1970) have discussed the magnetospheric distribution of pulsations in the PCl-5 categories, using data from the satellites OGO-3 and OGO-5. This survey covered the range of local times from 21: 00 to 12: 00 through midnight, at latitudes less than -20”. Their statistical analysis relied on a visual inspection of field magnitude data but simultaneous vector measurements suggested that the majority of wave events had a significant wmpressional component. They quote a detection threshold of 0.5-2 y for events with periods in the range 120-240 set and find that these events were most frequent at 8-12 R. geocentric distance in the local time range 03:OO to 09:OO. Part of the large spectrum of waves observed at 6.6 R, equatorial geocentric distance by ATS-1 (Coleman, 1970) consists of large amplitude, approximately sinusoidal, wmpressional waves with periods in the PC5 range (Barfield and Coleman, 1970). Typically, these have a period -200 set, -10 y and a duration -1 hr. an amplitude They are observed only in the 12:00-19: 00 LT zone during the main-phase minimum of a geomagnetic storm (Barfield and McPherron,
1972). These events have the interesting polarisation property that as the field strength increases the field direction becomes more inclined to the local horizontal. This feature was also noticed in the event observed by Sonnerup et al. (1969) in the same general location. In contrast, the geomagnetic quiet-time, low-frequency pulsations at ATS-1 are mainly transverse with periods in the PC4 range (Cummings et al., 1972, 1975; Kokubun et al., 1975). Similar transverse pulsations have been found near geosynchronous altitude within a few hours of noon in data from the Dodge satellite (Dwarkin et al. 1971). Recently, Kivelson (1975) has used ATS-1 and OGO-5 field and plasma data to show that these transverse pulsations are associated with isolated regions of enhancement in the cold plasma density. Thus, the ATS-1 and Explorer 26 data suggest that in the 5-7 R, equatorial geocentric distance range, the wmpressional PC5 event might be a storm-time phenomenon confined to the afternoon sector. On the other hand, the survey by Heppner et al., which did not wver the afternoon sector, suggests that PC5 events are also a relatively wmmon feature in the outer magnetosphere near dawn. Unfortunately, no other information is available on the occurrence of PC5 pulsations at distances beyond geosynchronous orbit in the afternoon and evening sectors. The present analysis aims to fill this gap by surveying the occurrence of wmpressional PC5 events throughout the Sunward hemisphere from dawn to dusk, using HEOS-1 data obtained during 1969 and 1970. No detailed theoretical interpretation of these events is attempted here: the interested reader is referred to recent papers by Sonnerup et al. (1969), Hasegawa (1969), Southwood (1974, 1975) and Hasegawa and Chen (1974). 2.THEINSTRUMENTATlON The HEOS-1 satellite was launched on 5 December 1968 into a very eccentric Earth orbit with an apogee at a geocentric distance of 35.8 R,, 5 1.5” north of the ecliptic. The mean orbital period was 4.646 days and the plane of the orbit was inclined, initially, at an angle of 52” to the ecliptic. The experiment complement included a three-axis fluxgate magnetometer with a dynamic range of l 64 y. Instantaneous vector measurements were digitised to ~tO.25 y resolution every 1.5 set but the low available telemetry rate limited transmission to 1 vector every 48 sec. In addition to this set of data, the experiment’s 16-kilobit wre memory provided 17 min periods of measurement with samples at
Giant PCS pulsations in outer magnetosphere 1.5 set intervals. These recurred every 5 hr after replay. Full details of the instrument, of the orbit and of the measurement accuracies are recorded elsewhere (Hedgecock, 1975a,b); we recall here only that the field component measurements are thought to be accurate, on average, to *O.l y or *l% of range. Owing to the inclination of the orbital plane, the satellite traversed inbound and outbound orbital segments at rather different latitudes, the high apogee limiting the magnetospheric coverage (i.e. between magnetometer saturation and the magnetopause) to some 3-6 hr per segment in the Sunward hemisphere. In the 8-13 R, geocentric distance range, the inbound segments during 1969 were approximately equatorial while the outbound segments were at -30” latitude. Fortunately, a chance combination of orbital geometry and period resulted in maximum variation between the magnetic latitude of consecutive inbound or outbound segments near dawn and dusk. However, relatively large orbital perturbations increased both the orbital inclination and the perigee altitude to the extent that low-latitude low-inclination transits through the 8-13 R, geocentric range did not occur after 1970. This analysis, therefore, is limited to data obtained during 1969170. 3. DATA
The occasional -6 hr periods of magnetospheric measurement were identified by inspection of daily plots of reduced data. Selected periods were replotted on an expanded time scale with the field represented by its vector magnitude together with the three field components in a Cartesian dipolemeridian system. This coordinate system is righthanded with X radially outwards from Earth at the point of observation, Z towards magnetic north and Y towards magnetic east, the magnetic meridian being defined by the local radius vector and the field vector predicted by the reference field IGRF65 (IAGA, 1969). This coordinate system rotates about all three axes as the satellite moves through the magnetosphere but provides an easy means for identifying the effects of wave motion, field-aligned currents, etc. Owing to the variable orientation of the geomagnetic dipole relative to the solar vector, the shape of the magnetosphere changes considerably from event to event. As a first-order correction for this variability, the satellite positions are quoted in geocentric magnetospheric equatorial (GME) coordinates (Hedgecock and Thomas, 1975). These, in effect, assign an equivalent position in a mag2
from HEOS-1 data
netosphere with the dipole orthogonal to the solar vector. Thus, in the GME system, +X is directed to the Sun, +Z is antiparallel to the geodipole moment vector and +Y completes a right-handed set. When the geodipole is orthogonal to the solar vector, the GME axes coincide with the geocentric solar magnetospheric (GSM) axes (+X directed to the Sun with the dipole moment in the X2 plane) and with the solar magnetic (SM) axes (t-Z antiparallel to the dipole moment with the Sun in the XZ plane). For non-orthogonal dipole orientations, the magnetospheric equator is represented as a multiply-curved sheet, Z GME is the perpendicular distance of the satellite from this sheet and X, Y GME are obtained from X, Y GSM by an approximate curvilinear transformation. The GME system tends towards the SM system near Earth, the GSM system far from Earth (at large IZI, provided [XI< 10) and towards the displaced GSM system of Fairfield and Ness (1972) in the geomagnetic tail at X< -17 R,. Full details of the coordinate transformation to GME, a discussion of the limitations of this representation and an illustration of the improved ordering of field direction measurements when expressed in the GME system are presented elsewhere (Thomas and Hedgecock, 1976). The data recorded tailward of 20:00 and 04:OO local times were discarded from the present analysis: the low-latitude data showed the irregular perturbations characteristic of the presence of the tail plasma sheet while the higher-latitude data obtained in the tail lobes showed no pulsation activity. Sunward of these local times, 64 lowlatitude and 69 higher-latitude orbit segments were found to have sufficient telemetry coverage for useful observation of PC5 events during 1969170. -80% of the total. These orbit This represents segments were distributed approximately uniformly in local time through the selected range. The presence of an enhanced level of lowfrequency fluctuation near the magnetopause was noticed in a large fraction of the intervals selected. These fluctuations had an amplitude -10% of the ambient field strength and displayed an irregular waveform; they could, perhaps, best be described as low frequency “noise.” They occupied a region which extended l-2 R, Earthward of the magnetopause near the subsolar point, expanding to -5 R, Earthward of the magnetopause on some occasions near dawn and dusk. An analysis of the characteristics and distribution of this “noise” will be presented in a future paper. Contrasting with these irregular fluctuations, four good examples of “continuous,” PCS-period fluctuations were found. These events had an appreciably
P. C. HBDGEC~C~K
different character and their selection did not depend in any way on critical selection criteria. They had nearly sinusoidal waveform, were mainly compressional and lasted for l-2 hr. Since the satellite moves radially through a distance -3 R, in this time, the “observed” duration may reflect the satellite’s motion rather than the actual duration of the event. In the following, we describe these events and, by contrasting with data obtained in the same general location with no pulsation present, attempt to define the magnetospheric condition which leads to their production. 4.THEEl’ENl-S
The equivalent magnetospheric locations of the four good events are indicated in Figs. 1 and 2 in GME coordinates. Figure 1 is a plan view of the magnetospheric equator with Earth and a symmetrical form of Fairfield’s (1971) magnetopause shown to scale. The dash curves represent the projections of the satellite orbit segments for time ranges corresponding to the data displayed in Figures 3 and 4. The continuous sections of the curves for 12 October, 21 October, 4 November and 9 February indicate the range of location in which the PC5 events were observed, while the continuous curve for 14 November and the star on the curve for 4 November indicate the locations of additional pulsation data discussed in the following. Figure 2 illustrates the latitudinal configuration of these orbit segments, the coordinate CHI representing the radial distance from the Z GME axis. The latitude difference between the dawn and dusk orbit segments should be noted; the former, on 30 January, 9 February and 13 February are at latitudes. -25” while the latter, generally, are more equatorial. Field lines intersecting the Earth at magnetic latitudes of 70”, 72” and 74” also are indicated. These were derived from a magnetospheric field model based on the MF73 representation (Mead and Fairfield, 1975), with coefficients adjusted to provide a better fit to distant highlatitude magnetic measurements and with the inclusion of Olson’s ring current (Hedgecock et al., 1976, Model A). The field magnitude signatures recorded during the six inbound orbit segments near dusk are displayed in Fig. 3. They are arranged from top to bottom in order of decreasing longitude relative to noon; the field strength is in gamma and the satellite position in GME coordinates is indicated every 30 min above each plot (in R,). Each plot also
OF THE SEGMENTS To
The thick traces represent the positions at which giant pulsation events were observed while the asterisk on the 4 November trace indicates the location of the finer resolution data displayed in Fig. 6. The Earth and the equatorial section of the maanetooause (after Fairfield, 1971) are shown to scale. The &ordinate system is described in the text.
displays two reference fields; the continuous curve is IGRF65 while the dash curve is the MF73 prediction (MF73Q in Mead and Fairfield, 1975). These two curves are stepped because orbit position data were available only at -6 min intervals. The solid sections of the orbit segments plotted in Figs. 1 and 2 correspond to the clear PCS events observed on 12 October between 18:00 and 2O:OOUT, on 2112 October between 23:20 and 01:50 UT and on 4/5 November between 23:25 and 00:30 UT. On all six occasions, the magnetopause was encountered l-2 hr before the start of the plot at geocentric distances of 15-18 R,,
Giant PC5 pulsations in outer magnetosphere
from HEOS-1 data
,’ Feb 9
70”, 72” AND 74”. The coordinate CHI represents the radial distance from the 2 axis in the GME coordinate system in which the geomagnetic dipole is always parallel to Z and normal to the solar vector (see text). It should be noted that the Jan and Feb segments are near dawn while the remainder are near dusk (Fig.
close to the position predicted by Fairfield’s best-fit conic (Fairfield, 1971). The dawn event and comparison data sequences from the two neighbouring orbit segments are displayed in the same way in Fig. 4. Again, the order is that of decreasing angular distance from the subsolar point and the two reference fields have been included. On these occasions the magnetopause was encountered at the positions indicated by the arrows. The solid trace in Figs. 1 and 2 corresponds to the giant pulsation observed on 9 February between 02 : 00 and 03 : 00 UT. 5.
These were investigated by first transforming the field component data to a coordinate system with the Z axis aligned with the mean field direction. This was derived from the dipole meridian system, described previously, by the equivalent of rotating first about the local vertical to remove the residual declination and then about the new Y axis to remove the inclination. The resulting mean-fieldaligned (MFA) coordinates have +Z in the direction of the mean field vector while +X and +Y, respectively, are in the local upward and eastward hemispheres
and radial distance with a velocity of -lo4 km/hr during these observations, the “mean field direction” varies continuously. Therefore, rather than evaluating a long-term average direction, we have derived a “running mean” by numerically filtering the data, using a sine-terminated low-pass filter designed according to the algorithm of Behannon and Ness (1966, equation 31). This filter produces zero phase difference between its input and output at all frequencies and, with 48-set samples, has an amplitude response which falls from 99% at a -2OOOsec to 1% at -1OOOsec. The period filtered sequence was used to define the transformation for each vector in such a way that the Z MFA axis follows the vector direction specified by the filtered data sequence. As an added convenience for observing wave phenomena, the slow variation in the field magnitude caused by the satellite’s motion has been removed from the Z MFA component by subtracting the magnitude of the vector represented by the filtered sequence. The component fluctuations in the MFA coordinate system and the satellite position every 30 min in GME coordinates are displayed in Fig. 5. The individual measurements, occurring at a rate of 75/hr, are resolved and it should be noted that -1 y fluctuations in the magnitude can be caused
or ‘I 2
-3.03 15.17 0.46
: -3.10 14.67 0.66
0 -3.18 14.03 0.66
1 -3.23 13.60 0.77
-3.29 12.62 0.61
3 -3.39 10.79 0.76
-3.40 IO.02 0.67
L -3.40 9.21 0.61
-3.36 8.63 0.39
1969 5 OCT 710
UT -3.36 7.64 0.13
-3.31 6.66 -0.16
1969 21 OCT 12
15 Y ‘I 2
-1.07 IS.45 1.73
16 -1.16 14.96 1.46
-I .26 14.66 I .24
17 -1.40 14.03 0.90
-I .56 13.46 0.51
10 -1.71 13.02 0.17
-I .92 12.41 -0.28
-0.69 Il.91 2.99
-0.63 11.29 2.37
-2.63 10.42 -1 .61
-2.69 9.59 -2.09
-3.13 6.66 -2.56
-2.15 11.75 -0.72
-2.37 II.lB -I .I2
-,.I, 10.62 I .e4
-1.42 9.69 1.31
-1.69 9.26 0.66
-2.02 6.37 0.36
-2.32 7.39 -0.13
-0.62 Il.26 0.33
-0.64 10.36 0.16
-1.06 9.41 -0.03
-1.22 6.63 -0.16
-3.31 7.88 -2.94
-0.01 14.38 4.67
-0.10 13.68 4.61
23 -0.30 12.93 3.74
-0.43 12.37 3.30
30 .. ___ 0
x I 2
-1.41 7.65 -0.,2 RI
x Y 7
3.64 14.66 4.94
IL 3.51 14.14 4.76
3.34 13.66 4.43
17 2.21 I I .29 2.32
I .S2 10.72 1.80
1.49 IO,22 I .39
I .o* 9.65 0.68
0.49 6.61 0.37
lg 0$?6 -0.05 7.97 -0.11
FIG. 3. FIELDMAGNITVDEFLUCNATIONSINGAMMA OBSERVEDDURINGTHESIX NEAR DUSKINDICATED IN FIGS. 1 AND 2.
position of the satellite is quoted at 30min intervals above each plot in the GME coordinate system (R, Cartesions: see text). The plots are arranged from top to bottom in order of decreasing angular distance from noon. All the data were obtained Earthward of the magnetopause.
1969 FEB 9
2CJAN 1969 UT
pulsations in outer magnetosphere from HEOS-1 data
A PLOTSIMILARTOFIG. 3 FORTHETHREEORBITSEGMENTSNEARDAWN. The arrows indicatethe crossing of the magnetopause.
by the digitization of the three components in 0.5 y increments. The trace is interrupted at data gaps in excess of one missing vector.
points owing to transmission errors. This effect occurs at several places in each of these events.
1. Waveform and period
Close inspection of Fig. 5 reveals that as the field magnitude increases the vector tips Earthward, while Fig. 6 suggests that the small transverse perturbation is approximately circularly polarized. These characteristics were investigated by constructing hodograms for the three pairs of field components measured during a one hour period of each event. A more exhaustive statistical analysis (e.g. Paulson et al., 1965) was not considered appropriate owing to the infrequent measurement rate relative to the periodicities and the several missing data points in each sequence. Figure 7 displays the hodograms in the XZ plane for the four events. The view is from the east and the line represents the locus of the tip of the field vector whose origin is displaced by the indicated mean field magnitude along the -Z axis. It should be noted that these hodograms have different scale factors. Clearly, in all four cases the angular deviation is approx. *5” about the mean direction in the sense of a tip towards the Earth (-X) for increasing field magnitude. The corresponding hodograms for
The periods vary both from event to event and
through the individual events, the latter possibly being the result of the changing radial position. Estimating the number of zero crossings per hour, we obtain periods of -5OOsec for 12 October, 450-SOOsec for 22 October, -360 set for 4 November and -700 set for 9 February. Some of the apparent waveform irregularity in these events is the result of the rather low rate of measurement. This is illustrated in Fig. 6 which is a 15 min “snapshot,” at the 1.5 set sampling rate, obtained via the instruments’ core memory near the start of the 4 November data sequence. This display is similar to Fig. 5 except that the ambient field has not been subtracted from the Z data and the time scale is in minutes of elapsed time starting at 23: 29+30 set UT. Since the smaller ripples are entirely attributable to the effects of digitization, the smoothness of the Z component fluctuation is striking. The apparent ragged nature of the same two cycles in Fig. 5 is caused by the loss of two data
OCT 12 1969
FIG. 5. THE
IN GAMMA AND
The data are expressed in a right-handed coordinate system with Z aligned with the slowly varying mean-field direction, Y eastward and X outward from Earth in the plane defined by the mean field vector and the radius vector from Earth (see text). The satellite position (R, Cartesians) is quoted in GME coordinates at the top and the field vectors were sampled at 48 set intervals. the YZ plane, looking towards Earth, are displayed in a similar format in Fig 8 with the origin of the total field vector again displaced along -Z. No systematic deviation from the mean field direction is evident in the 4-5 November and 22 October hodograms while only a slight deviation is apparent in the other two events, indicating the mainly compressional nature of these events. The hodograms of the transverse (XY) component fluctuations are displayed in Fig. 9, the view being in the direction of the mean field vector. Here, the right-hand polarization of the 9 February
event near dawn is unambiguous. In contrast, the three events near dusk are less easy to classify although they are left-handed for part of the time. To clarify the situation we have constructed cumulative area plots (Dungey and Southwood, 1970) for these events. The cumulative area CA is obtained by forming the running sum of the vector product of consecutive pairs of transverse perturbation vectors, i.e. EA increases positively for righthand polarization relative to the mean field vector. These plots, for the 1 hour intervals represented in Fig. 9, are displayed in Fig. 10. Again, the right
Giant PC5 pulsations in outer magnetosphere
from HEOS-1 data
handed polarization of the 9 February event is contrasted with the mainly left-handed polarization of the events near dusk. It is interesting that there is an apparent reversal in the sense of polarization during the 22 October and 4 November events as the satellite proceeds Earthward. 3. Equatorial symmetry At the geocentric distance of these events, HEOS-1 was located south of the magnetospheric equator only during particularly favourable orienta-
FINER TIME RESOLUTION MEASUREMENTS OF THE FIELDCOMPONENTFLUCllJATIONSINGAMMARBcORDEDBY THE EXPERIMENT'S CORE MEMORY NEAR THE END OFTHE DATAGAPINTHE 4/5 NOVEMBEREVENT (FIG. 3).
measurements and are expressed
are sampled at 1.4 second intervals in the mean-field-aligned coordinate system (see text).
HODOGRAMS A RADIALLY
OF THE EVENTS OF FIG. 7 AS SEEN
The display is similar to that of Fig. 7 with Z in the locally-defined magnetic meridian and Y easterly. The mean-field magnitudes indicate the displacement of the origin of the mean field vector along -Z and “S,” “F” indicate the start and finish, respectively. Note the different scale factors.
FIG. 7. HODOGRAMS OF THE DISTURBANCE CORDEDDURINGTHEFOURLARGEPULSATION The
mean field magnitude is indicated in each plot and the origin of the mean field vector is displaced by this amount along the -Z axis. Each plot is a view of the perturbation in the local magnetic meridian plane as seen from an Easterly direction (see text). The starting and end points are indicated by the “S” and “E” Note the different scale factors.
tions of the geomagnetic dipole and, unfortunately, these orientations generally precluded good telemetry reception at the northern hemisphere ground stations. Thus, very few southern hemisphere orbit segments in the dusk sector were obtained and only one showed any pulsation activity. This rather poorly developed event occurred between 05 :00 and 06:OOUT on 14 November. As indicated in Fig. 1, the event was situated rather near the normal magnetopause location but on this occasion the magnetopause was located -3 R, outward from its
was close to the nomi-
as shown in Fig. 2 but the sign of the radial field component confirmed that the satellite was, in fact, in the southern hemisphere. The field component fluctuations in MFA coordinates and the satellite position every 30 min in GME coordinates are displayed in Fig. 11. The satellite encountered the magnetopause just prior to the start of this data sequence and crossed the equator at 05:OO UT, near the beginning of the pulsation event. The hodograms for this event are x 6.69 'I 13.85 2 2.79
5.15 4.54 4.01 3.38 12.97 12.54 12.16 11.61 0.60 -0.02 -0.53 -1.01
FIG. 9. HODOGRAMSOF THE TRANSVERSECOMPONENT OF THE EVENTSOF FIG. 7 LOOKING IN THE DIRECTIONOF THE MEANFIELD. The coordinates are arranged with Y easterly and X outward from Earth. Note the different scale factors.
UT NOV 14
FIG. 11. FIELD
COMPONENT IWKTLIATIONS MEASURED DURING A POORLY DEVELOPED HEMISPNEREEVENTDESCRIBEDINTHETEXT.
IN GAMMA SOUTHERN
data are expressed as in Fig. 5 with the field components in the mean-field-aligned coordinate system and the satellite position in GME coordinates. The
FIG. 10. PLOTS OF THE CUMULATIVEAREA SWEPTOUT BY THE PERTURBATION
VECI-OR PLOTTED IN THE HODOGRAMS OFFIG. 9.
The function CA increases positively for right-hand polarization (see text). Each plot is evaluated for the indicated l-hour time span with measurements at the indicated points. Note the polarization reversal displayed by the 22 October and 4/5 November events as the satellite proceeds Earthward.
shown together in Fig. 12. The phase relationship between the X and 2: fluctuations on this occasion is clearly opposite to that seen in the events described previously. This phase corresponds to a deflection of the field vector away from Earth for increasing field strength as would happen if the change in the configuration of the field line for increasing field strength were the same in both hemispheres. Comparing this event with the four previous events, it is interesting that the two events nearest the equator (12 October and 14 November) display the smallest fluctuation in the Y component. The properties of these events are summarised in Table 1, where (B), LIB,,, AB, and (7’) represent,
Giant PCS pulsations in outer magnetosphere from HBOS-1 data
F10.12. A SECTION
COMPOSITEDISPLAYOFTHEHODOGRAMSFORA OF THE 14 NOVEMBER EVENT DISPLAYED
FIG. 11. The three hodograms represent the three views displayed in Figs. 7, 8 and 9 for the larger events, with the same notation. Note the different scale factors.
respectively, the mean field magnitude, the amplitudes of the components of the perturbation field parallel and perpendicular to the mean field direction and the mean period. The polarization is relative to the direction of the mean-field vector. It will be noticed that ABll is 2-4 times AB, and that the perturbation amplitude is a significant fraction of (B) in all cases. 6.
Although we have observed only four giant pulsation events near dusk this represents a rather large probability of occurrence. The orbit geometry and telemetry coverage were particularly favourable in this region in 1969 and all but two of the ten inbound passes in the 16:00-20:00 LT zone provided good, near-equatorial coverage in to -7.5 R, geocentric distance, where the magnetometer satu-
* Where polarization reversal occurs the polarization quoted is that occurring in the region furthest from Earth.
rated. Events were observed on half of these occasions. Only two of the corresponding passes in 1970 had sufficient telemetry coverage in the 812 R, range. One of these (Fig. 3) showed no pulsation whereas the other (not shown) displayed pulsations comparable to the events described here but with a somewhat less sinusoidal waveform. Outbound orbit segments near dusk also were inspected but, owing to the shape of the orbit, these occurred three months earlier when the orientation of the Earth’s spin axis was less favourable. These mid-summer trajectories were at even higher latitude than the 13 February segment in Fig. 2 and no events were observed. This, probably, could be attributed to the latitude of observation which was too large for the satellite to reach the 74” field line before the instrument saturated (cJ Figure 2). The 1970 outbound observations suffered from a similar constraint. Thus, pulsations were observed in the 16: 00-20:00 LT zone on half of the occasions when the satellite was passing through the 8-14 R, geocentric distance range within 3 R, of the equator. Owing to the different season of observation, the orbital geometry was more favourable for observations near dawn. Eighteen outbound and fourteen inbound passes in the 04:00-08:OOLT zone were inspected and the outbound pass on 9 February 1969 was the only one to show any sign of pulsation activity. It is clear, therefore, that the event occurrence rate near dawn is much lower, with events occurring at 8-12 R, distance probably no more than a few percent of the time, but it is worth noting that this single dawn event was the most spectacular of all the events seen and was first noticed by ESRO engineers viewing real-time telemetry data. No giant PC5 activity was observed in the 08:00-16:00 LT zone, although 41 outbound and 39 inbound orbit segments during 1969 and 1970 were inspected. In this region the magnetosphere is more compressed and the HEOS satellite spends rather little time crossing the field lines from a high magnetic latitude. Nevertheless, it seems that pulsation activity of this type is low in the bulk of the Sunward hemisphere, as found by Heppner et al. (1970). The mainly-compressional nature and nearsinusoidal waveform of the giant pulsation in the outer magnetosphere is clearly revealed in Fig. 5 and the character of the hodograms (Figs. 7-9 and 12) suggests that these events are very closely similar to the storm-time PC5 events recorded at ATS 1 (Barfield et al., 1970, 1972) and at smaller radial distances (Brown et al., 1968, Sonnerup et al.,
main field. The geographic locations of these intercepts are listed in Table 2. Unfortunately, they are remote from land in most cases, especially in the southern hemisphere. Rapid-run magnetograms from Leirvogur and Great Whale River were inspected and indeed show some pulsation activity with a similar period during parts of the 12 and 22 October and 4 November events but, owing to the resolution of the records and the much larger fluctuations produced by local ionospheric currents, no detailed correlation could be made. This should not have been the result of ionospheric attenuation since the few degrees range of latitude and longitude implied by Figs. 1 and 2 for the wavefront exceeds the 100 km limit set by Hughes and Southwood (1975a, Fig. 6) for significant attenuation to occur. Nevertheless, these events might have been observed near the ground as magnetic perturbations in suitably filtered magnetograms or, for example, as long period fluctuations in either ionospheric absorption (Brown, 1975) or ionospheric electric field (D’Angelo et al., 1975, Mozer, 1971). Data obtained by OGO-5 and ATS-1 during these events were inspected but these satellites were displaced relatively far from HEOS in radial distance or longitude and no similar fluctuations could be identified. Several properties of these events are in agreement with current theoretical ideas involving a Kelvin-Helmholtz source at the magnetopause which excites a field-line resonance deeper in the magnetosphere (Chen and Hasegawa, 1974; Southwood, 1974; Type A events of Hasegawa and
1969, Lanzerotti et al., 1969, 1974, 1975a; Lanzerotti and Hasegawa, 1975b). In contrast, the events reported here show little correlation with geomagnetic activity. All of the pulsation events occurred during intervals when D,, was more positive than --15 y and the only transits to occur during more negative values of D,, showed no pulsation activity. Considerable substorm activity occurred in the six hours preceding the 12 October event and, to a lesser extent, in the interval preceding the 4 November event, but these events were actually observed during intervals of relatively low AE and similar AE enhancements preceded periods when no pulsation was observed (e.g. 9 and 26 October). Bearing in mind that the AE index is not an infallible indicator of substorm activity (Akasofu, 1974), some substorm association is possible but so far we have found no evidence to support an association for these events. Pate1 and Cahill (1964), Pate1 (1965, 1966) and later, in more detail, Lanzerotti and Tartaglia (1972) and Lanzerotti et al. (1974) demonstrated good correlation between pulsation events observed simultaneously in the magnetosphere and at ground stations near the “foot” of the field line passing through the satellite’s location. We have traced the field lines from the satellite at the beginning and end of each event to identify the northern and southern hemisphere ground intercepts. This was performed with a field model of the MeadFairfield type (previously described in connection with Fig. 2) but with an additional modification to use IGRF6.5 (IAGA 1969) to represent the Earth’s
Reg I on
Zemlya Feb 9
w Ott 22
w NO" 5
-72.0 1 193.9
Giant PC5 pulsations in outer magnetosphere Chen, 1974). The limit of large longitudinal wave number assumed in calculating the resonance behaviour is well justified by the longitude distribution implied by Fig. 1. Since all of these events were of comparable amplitude, occurred relatively frequently near dusk but were never seen (even with reduced amplitude) outside the 1600-2000 local time zone, they must appear as a rather sharply defined “wave packet” in longitude. Two of the dusk events (21/22 October, 4/5 November, Fig. 10) show the expected polarization reversal as the satellite moves Earthward through the field-line resonance condition. These two reversals occurred very close to the 72” field line (the error in field line location with the given field model is - *OS” in this region). If the data during the other two events were recorded on the magnetopause side of the resonance shell, the dawn-dusk polarization difference and the sense of polarization reversal are exactly that reported for ground station data by Samson et al. (1971) and discussed by Chen and Hasegawa (1974), Southwood (1974) and Hasegawa and Chen (1974, Figs. 2, 3). Owing to the rather large noise level of the infrequently sampled data points, little information can be obtained from the hodograms of Fig. 9 concerning the orientation of the polarization ellipse at resonance. Interestingly, however, the orientation of the polarization ellipse for the 9 February (dawn) event is that quoted by Hasegawa and Chen (1974, Fig. 1) for their values q < 0, m < 0 (on making due allowance for our differently directed axes). This implies, as assumed above, that the data were obtained on the magnetopause side of the resonance shell and, as noted by Southwood (1975), since these measurements were well above the ionosphere and the Hughes rotation (Hughes, 1974) need not be considered, this orientation implies that the wave phase-velocity vector is directed towards the geomagnetic tail, with the energy flux directed Earthward. So far, we have been unable to process sufficient data from the region between the resonance shell and the magnetopause to establish with reasonable significance that the “noise,” referred to previously as generally being present in that region, has the polarization appropriate to a Kelvin-Helmholtz source. However, we note that this type of polarization pattern was reported by Dungey and Southwood (1970) for several measurement sequences in that region. Finally, since the resonance shell should be the location of maximum disturbance amplitude (Southwood, 1974), it is gratifying to find that our latitude of 72” coincides with the latitude at which, statistically, Rostoker et
from HEOS-1 data
al. (1972, Fig. 1) find the region of maximum disturbance amplitude in ground station data, for wave periods -5OOsec near dusk. The data in Figs. 3 and 4 clearly show that when pulsations are observed, the measured field magnitude is somewhat less than the reference field, particularly Earthward of the event location, while no pulsation is observed if the measured field is comparable to, or exceeds, the reference field magnitude. This suggests that a particular plasma condition is required for adequate resonance in this region and since the Earthward limits of these events are located several R, outward from the mean location of the plasmapause (Chappell et al., 1971), it is quite possible that they may be associated with the detached “plasma islands” reported out to -9 R, geocentric radius near dusk (Chappell et aI., 1971; Maynard and Chen, 1975). Lacking magnetospheric plasma measurements from HEOS-1 it is impossible to resolve this question. However, recent data from HEOS-2 at high latitudes suggests a further possibility. Paschmann et al. (1975) report many observations of a thick magnetosheath-like plasma layer Earthward of the magnetopause in the Sunward hemisphere. This layer is in some respects similar to the “plasma mantle” observed by the same satellite in the highlatitude tail (Rosenbauer et al., 1975). If, like the “plasma mantle” (Hones et al., 1972; Akasofu et al., 1973), this plasma extends downwards to the equator in the Sunward hemisphere, it would occupy just the region we have found to contain an enhanced level of low-frequency “noise” in the HEOS-1 data (Section 3). It is tempting to speculate that the radial gradient in plasma parameters at the inner edge of this region near dusk (i.e. close to the trapping boundary) is responsible for the form of pulsation reported here; certainly, there is an abundance of low-frequency “noise” to provide an energy input. 7. SUMMARY
The pulsations described here provide a graphic illustration of the large-scale coherent oscillation of a localized section of the outer magnetosphere. Their analysis has led to the following conclusions: 1. The giant PC5 pulsation is the most spectacular phenomenon in the outer magnetosphere Sunward of the tail region. 2. These pulsations occur relatively frequently (i.e. at least 25% of the time) 8-12 R, from Earth in the 1600-2000 LT zone but are exceedingly rare Sunward of this zone and are relatively rare at dawn.
P. c. HEDc3EUXK
3. The pulsations are predominantly compressional but have a small transverse component. 4. The polarization characteristics of the transverse component and the longitudinal extent of the phenomena are in excellent agreement with the assumptions of present micropulsation theory. 5. Polarization reversal (i.e. field-line resonance) appears to occur on the field line which reaches Earth at the latitude corresponding to the statistically-defined position of maximum disturbance amplitude in ground station data, as required by theory. 6. The pulsations are associated with regions of plasma pressure enhancement which cause a depression in the magnitude of the ambient magnetospheric field. 7. The deformation of the field line configuration with changing perturbation field strength is the same in the northern and southern hemispheres but with single-point measurements in the magnetosphere and lacking supporting ground measurements, no north-south phase relationships could be determined. It would be most interesting to define the magnetospheric distribution of these events and the plasma conditions which lead to their generation in more detail. This will require more comprehensive measurements from several satellites located simultaneously in the dusk sector at 8-12 R, geocentric distance. It is hoped that the clear character and relatively high occurrence rate demonstrated here will stimulate a search of other data to this end. Acknowledgements-l am indebted to Professor J. W. Dungey and Dr. D. J. Southwood for many stimulating discussions and to Mrs. Anne Evans for performing all of the programming and data management activities associated with this analysis. I would like to thank the World Data Centres at Slough, U.K. and Boulder, U.S.A. for making available geomagnetic data and Joe Allen of the latter for answering in detail many questions conceming the characteristics of the AE index. My thanks are also extended to Drs. C. T. Russell and R. L. McPherron for permitting me the opportunity of inspecting plots of OGO 5 and ATS 1 data during these events and to Dr. L. J. Lanxerotti for suggesting several improvements to the manuscript. Professor H. Elliot was Principal Investigator for the HEOS-1 magnetometer experiment which was funded by the British Science Research Council.
Akasofu, S.-l., Hones, E. W. Jr., Bame, S. J., Asbridge, J. R. and Lui, A. T. Y. (1973). J. geophys. Res. 78,7257. Akasofu, S.-l. (1974). Space Sci. Rev. 16, 617. Barfield, J. N. and Coleman, P. J., Jr. (1970). J. geophys. Res. 75, 1943.
Barfield, J. N., Lanxerotti, L. J., Maclennan, C. G., Paulikas, G. A. and Schulz, M. (1971). J. geophys. Res. 76, 5252. Barfield, J. N. and McPherron, R. L. (1972). J. geophys. Rex 77, 4720. Brown, R: R. (1975). J. geophys. Res. 80, 1023. Campbell. W. H. (1968). In Phvsics of Geomagnetic Phenomena (Eds. ‘S. Matsushita and W. H. Cam$ell), p. 821. Academic Press, NY. Chappell, C. R., Harris, K. K. and Sharp, G. W. (1971). J. geophys. Rex 76, 7632. Chen, L. and Hasegawa, A. (1974). J. geophys. Res. 7!J, 1024. Coleman, P. J. Jr., Sonett, C. P., Judge, D. L. and Smith, E. J. (1960). J. neophys. Res. 65, 1856. Cummings, W. D, Mason, F. and Coleman, P. J., Jr. (1972). J. eeoohvs. Res. 72. 748. Cummings, W. D.,‘Contee, 6, Lyons, D. and Wiley, Ill. W. (1975). Preprint, Grambling State University, Los Angeles 71245, CA. D’Angelo, N., lversen, I. B. and Mohl Madsen, M. (1975). J. geophys. Res. 80, 1353. Davis, T. N. and Sugiura, M. (1966). J. geophys. Res. 71, 785. Dungey, J. W. and Southwood, D. J. (1970). Space Sci. Rev. 10,672. Dwarkin, M. L., Zmuda, A. J. and Radford, W. E. (1971). J. aeophvs. Res. 76, 3668. F&field, D. H. *(1971). J. geephys. Res. 76, 6700. Fairfield. D. H. and Ness. N. F. (1972). J. _ geophys. _ _ Res. 75, 7432. Gupta, J. C. (1973). J. amws. rerr. Phys. 35, 2217. GUDU. J. C. 11974). Radio Sci. 9, 757. Gupta; J. C. (1975) Planet. Space Sci. 23, 733. Hasegawa, A. (1969). Phys. Fluids 12,2642. Hasegawa, A. and Chen, L. (1974). Space Sci. Rw. 16, 347. Hedgecock, P. C. (1975a). Space Sci. Instrum 1,61. Hedgecock, P. C. (1975b). Space Sci. Instrum. 1,83. Hedgecock, P. C., Thomas, B. T., Cornwall, A. M. and Davis. C. W. (1976). Imperial College Preprint. To hesubmitted to l&et. Space Sci. _ Hedaecock. P. C. and Thomas. B. T. (1975). Geophvs. _ _ J. R:astr kc. 41, 391. Heppner, J. P., Ledley, B. G., Skillman, T. L. and Sugiura, M. (1970). Ann Geophys. 26, 709. Hones, E. W., Jr., Asbridge, J. R., Bame, S. J., Montgomerv. M. D., Singer, S. F. and Akasofu, S.-l. (1972). J. geb;phys. Res. 77, 5503. Hushes. W. J. (1974). Planet. Soace Sci. 22, 1157. Hu&es: W. J. and Southwood, D. J. (1975):Submitted to J. geophys. Res. IAGA Comm. 2, W.G. 4 (1969). J. geophys. Res. 74, 4407. Jacobs, J. A. (1970). Geomagnetic Pufsarions. Springer, NY. Jacobs, J. A., Kato, Y., Matsushita, S. and Troitskaya, V. A. (1964). J. geophys. Res. 69, 180. Judge, J. L. and Coleman, P. J., Jr. (1962). J. geophys. Res. 67, 5071. Kaneda, E., Kokubun, S., Oguti, T. and Nagata, T. (1964). Rep. lonosph. Space Res. Japan 18,165. Kato, Y. and Utsumi, T. (1964). Rep. Ionosph. Space Res. Japan 18,214. Kitamura, T. (1963). Rep. lonosph. Space Res. Japan 17, 67.
Giant PCS pulsations in outer magnetosphere Kivelson, M. G. (1975). Preprint 1493-86, Inst. Geophys. UCLA. Submitted to J. a0nos. te?r. phys. Kokubun, S., Kivelson, M. G., McPherron, R. L. and Russell. C. T. (1975). EOS Trans. A.G.U. 56,423. Lanzero&, L. J., Hasegawa, A. and MacLenn&, C. G. (1969). .I. eeovhvs. Res. 74. 5565. La’&ero& Ly J: &d Tartaglia, N. A, (1972). J. geophys. Res. 77, 1934. Lanzerotti, L. J., Fukunishi, H., Lin, C. C. and Cahill, L. J. (1974). J. geophys. Res. 79, 2420. Lanzerotti, L. J., MacLennan, C. G., Fukunishi, H., Walker, J. K. and Williams, D. J. (1975a). J. geophys. Res. 80, 1014. Lanzerotti, L. J. and Hasegawa, A. (1975b). J. geophys. res. 80, 1019. Maynard, N. C. and Chen A. J. (1975). J. geophys. Res. 80,1009.
Mead, G. D. and Fairfield, D. H. (1975). J. geophys. Res. 80,523.
McPherron, R. L., Russell, C. T. an! Coleman, P. J., Jr. (1972). Space Sci. Reu. 13, 411. Mozer, F. S. (1971). J. geophys. Res. 76, 3651. Nagata, T., Kokubun, S. and Iijima, T. (1963). J. geophys. Res. 68, 4621. Oberts, P. and Raspopov, 0. M. (1968). Geomagn. Aeron. 8, 534. 01’ A. I. (1963). Geomagn. Aeron. 1, 90. Orr D. (1973). 1. atmos.-ten. Phys. 35, 1. Paschrnann. G.. Haerendel. G.. Sckouke, N., Rosenbauer, H. and &&ecock, P. d. (i975). ‘M& Pianck, Institu; Preprint MPI-PAI/Extraterr. 114. Submitted to J. geophys. Res. Pate], V. L. (1965). Planet. Space Sci. 13, 485. Patel, V. L. (1966a) Space Res. 6, 758. Patel, V. L. (1966b). Earth Planet Sci. Mt. 1, 282.
from HEOS-1 data
Pate], V. L. and Cahill, L. J. (1964). Phys. Reu. IAt. 12, 213. Paulson, K. V., Egeland, A. and Eleman, F. (1965). J. amm. terr. Phys. 27, 943. Raspopov, 0. M., Afanas’yeva Kiselev, B. V. and Loginov, G. A. (1972). Geomagn. Aeron. 12, 128. Rolf, B. (1931). Ten. Magn. 36, 9. Rosenbauer, H., Grunwaldt, H., Montgomery, M. D., Paschmann, G. and Sckopke, N. (1975). J. geophys. Res. 80, 2723. Rostoker, G., Samson, J. C. and Higuchi, Y. (1972). J. geophys. Res. 77, 4700. Saito, T. (1964). J. Geomugn Geoelecr. 16, 115. Saito, T. (1969). Space Sci. Rev. 10, 319. Samson, J. C., Jacobs, J. A. and Rostoker, G. (1971). J. geophys. Res. 76, 3675. Southwood, D. J. (1973). Planet. Space Sci. 21, 53. Southwood, D. J. (1974). Space Sci. Reo. 16, 413. Southwood. D. J. (1975a). Submitted to J. eeovhvs. Res. Southwood; D. J. (1975b). Geophys. JR: a&.’ Sot. 41, 425. Sonett, C. P. (1963). J. geophys. Res. 68, 6371. Sonnerup, B. U. O., Cahill, L. J. and Davis, L. R. (1969). J. geophys. Res. 74, 2276. Sucksdo& E. (1939). Ten. Magn. 44, 157. Sugiura, M. (1964). Ann. IGY 35, 9. Sugiura, M. and Poros, D. J. (1971). GSFC Preprint X645-71-278 July 1971. Thomas B. T. and Hedgecock P. C., 1976, Imperial College Preprint. To be submitted to Planet. Space Sci. Troitskaya, V. A. (1967). Micropulsations and the state of the magnetosphere, in Solar Terrestrial Physics (Eds. J. W. King and W. S. Newman). Academic Press, New York.