Mid-Latitude Ionospheric Perturbation Associated with the Spacelab-2 Plasma Depletion Experiment at Millstone Hill

J. C. Foster (1), J. M. Holt (1), L. J. Lanzerotti (2) (return to John Foster's Homepage)

1 Haystack Observatory, Massachusetts Institute of Technology
2 Bell Laboratories, Lucent Technologies

Abstract

1. Introduction

Both daytime and nighttime Shuttle exhaust firings and ground-based measurements were successfully performed in the 1985 campaign using the Millstone Hill incoherent scatter radar, optical imaging, and radio-propagation diagnostics. The nighttime experiment on July 30, 1985 used a 47-sec OMS burn directly to the south of the Millstone Hill Observatory (42.6° N, 288.5° E, 55° invariant latitude). Some 830 kg of exhaust products were released near the peak altitude of the ionospheric F layer along a 360-km path of the Shuttle during that interval. Airglow emissions and the formation of an ionospheric hole were observed by the ground-based diagnostic instruments. A description of the Spacelab-2 plasma depletion experiments over Millstone Hill and the characteristics of the resultant ionospheric holes and optical emissions has been reported by Mendillo et al. [1987].

The July 30, 1985 OMS-burn experiment was conducted several hours after sunset, under nighttime conditions, in order to avoid simultaneous re-ionization of the ionosphere. It was planned to produce the largest and most long-lived ionospheric hole that might be produced by the Shuttle [Mendillo et al. 1987]. This experiment produced a prompt ionization depletion which spread along the orbital track of the burn and expanded in latitude and altitude to create a hole ~2° wide, centered at the 41°N latitude location of the exhaust deployment. The airglow and the active depletion of the local density subsided over a 20-min interval, but the residual ionospheric hole persisted for > 5 hours, until sunrise. In darkness, without the replenishing effects of the solar ionization source, an ionization depletion will persist as a "fossil" after its causative processes have ceased to operate, as will the effects of density gradients or plasma waves associated with its formation. Natural processes, such as transient disturbance-enhanced electric fields (e.g. Schunk et al [1976]), can disturb the normal F-layer chemistry resulting in enhanced recombination and the formation of a deep localized density trough at mid-latitudes. Evans et al. [1983] have found such troughs to persist through the night, corotating with the earth at the longitude at which they were formed. The mid-latitude ionospheric hole generated in the nighttime Spacelab-2 experiment followed such a pattern and persisted through the night. Ionospheric parameters oscillated with ~40-min and~ 80-min periods through the interval following the initial ionospheric depletion. It cannot be determined uniquely whether the ionospheric oscillations were related to the active experiment or were the result of a propagating ionospheric disturbance (TID) launched by enhanced auroral activity which began nearly simultaneously with the Shuttle-burn experiment. In this brief report, we discuss the temporal evolution of the perturbed ionosphere observed by the Millstone Hill radar on the night of July 30, 1985 and its relationship to the auroral disturbance.

2. Observations

2.1. Ionospheric Perturbation

A quasi-periodic oscillation of the bottomside iso-density contours, with a 40-minute period, is seen during the 3-hour interval following the burn. Associated variations in TEC were < 10%. The oscillation of the bottomside density is highlighted in Figure 2 which presents the temporal variation of the density at two altitudes, 300 km and 260 km. Vertical lines beginning at the time of the Spacelab-2 OMS engine burn (03:22 UT) and repeated with 40-min period highlight the quasi-periodic variation of the density. Plasma Doppler velocity was measured in the radar experiment and this varied quasi-periodically in synchronism with the density variations, but with twice the period (~80 min), such that a significant velocity variation was associated with every other density oscillation. Doppler velocities at this position are principally in the vertical direction, and this component varied in the range 40 m/s such that the velocity was downward as the iso-density contours decreased in altitude.

The lower panel of Figure 2 repeats the iso-density contours of Figure 1 with overscribed fiducial lines to delineate the apparent downward phase progression of the density enhancements. The 40-min separation of the parallel lines indicates a 0.15° s-1 phase variation for the density oscillation at a given altitude. The density peak (0° phase) is observed to descend 200 km in altitude in 1460 sec. Combining these numbers yields a phase variation of -1.12° km-1 with altitude, indicating a 320-km vertical wavelength.

The radar antenna was held in a fixed position during the interval around the Shuttle burn at 03:22 UT and elevation scans resumed after 03:36 UT, providing information on the latitude spread and temporal evolution of the ionospheric perturbation. Figure 3 presents plasma density data at fixed altitude (260 km) in the bottomside F layer. This shows the initial latitude spread of the hole to be ~2° and the subsequent oscillation of the density over this range of latitude to have an approximate 40-min period.

2.2. Perturbation of Local Magnetic Field

2.3. Magnetospheric Disturbance

The Kp index on July 30, 1985 (not shown) was basically steady in the range 2 to 3 for the first 9 hours of the day, indicating that a major global disturbance did not occur. Solar wind data from the IMP-8 spacecraft (GSM coordinates [-25, -17, 22] Re), shown in Panel B of Figure 4, indicate that the north-south component of the interplanetary magnetic field IMF Bz turned negative near 02:00 UT, followed by an increase in solar-wind number density and a brief (30-min) rotation of the magnetic field, near 03:00 UT. Beginning at ~03:20 UT, Bz turned southward and remained strongly negative (-8 nT) through the interval 03:30 - 06:00 UT. Southward-directed IMF results in energization of the magnetosphere and most often leads to substorm disturbance, with the mean time between the Bz southward turning and the onset of the substorm expansion phase observed to be ~80 min [Foster et al., 1971].

Satellite observations are used to characterize the extent and magnitude of the ensuing auroral disturbance. The zero-energy electron boundary determined from DMSP satellite measurements [Newell et al., 1996] is shown in Figure 4C. This boundary serves to locate the equatorward extent of the auroral convection electric field, and often the plasmapause, since the earthward extent of zero-energy particles is determined solely by ExB drifts [Newell et al., 1996]. This boundary moves equatorward between about 03:30 UT and 07:00 UT, with temporal characteristics similar to the Z-component magnetic deflection seen in Figure 4A. This suggests that the electrojet currents poleward of the Greenhill magnetometer site intensified, and perhaps moved equatorward, during this interval.

The hemispheric auroral precipitation energy input derived from NOAA-8 overflights of auroral latitudes [Foster et al., 1986; Fuller-Rowell and Evans, 1987] is presented in Figure 4D, together with the midnight-sector averaged electron precipitation flux determined in the central-auroral region from the DMSP data (following Newell et al. [1996]). Both these measures of auroral particle loss indicate that strong precipitation occurred between 05UT - 06UT (near magnetic midnight on the U.S. east coast), considerably after the Shuttle OMS burn and the onset of the ionospheric and magnetometer oscillations. These monitors do not provide evidence of an enhancement in auroral precipitation near the time of onset of the ionospheric variations (~0330 UT) shown in Figure 1.

3. Discussion

3.1. Persistent Effects of Artificial Ionospheric Perturbation

Rosenberg et al. [1981] reported observations of the excitation of Alfvén waves in the earth's magnetosphere following sudden changes in the ionospheric conductivity. In that study, the sudden increase in D and E region conductivity produced by solar-flare x-ray and ultraviolet fluxes was observed to produce magnetic field variations in the ~0.01 Hz band. Those authors postulated that the sudden conductivity change would lead to an impulsive change in the horizontal current flow and, thence, to a transverse magnetic perturbation sufficient to excite resonance oscillations of the field line. For field lines near L = 4, the eigenmodes of the fundamental oscillations are in the 45-150 sec (Pc4) and 15-45 sec (Pc3) frequency bands for field lines inside or outside the plasmapause, respectively. Rosenberg et al. [1981] speculated that the excitation of hydromagnetic field-line resonance might also be expected to accompany a rapid decrease of ionospheric conductivity such as could be produced by chemical releases in the ionosphere. The Greenhill magnetometer observations on July 30, 1985 were made to determine what magnetic field effects, if any, could be attributed to the active ionospheric experiment. Our observations of the magnetic field perturbation along the magnetic field line through the Shuttle burn indicated that the filtered variations in the band 45 - 150 s (Pc4) were strongest, with distinct enhancements evident between ~0326 - 0405 UT and again ~0420 - 0500 UT. Although these observations are in keeping with the Rosenberg et al. prediction, they are more easily explained as being related to the substorm onset.

3.2. Magnetic Storm/Substorm Effects

The magnetic variations following the Shuttle burn, reported above, occurred in approximate time coincidence with the southward turning of the IMF, and the likelihood that a geomagnetic substorm was initiated thereafter. Indeed, the magnetic pulsation frequencies shown in Figure 5 are consistent with those associated with Pi2 pulsations, which are known to be in approximate time coincidence with substorm onsets (e.g. Rostoker [1968]). If the initial magnetic variations observed in the 45 s - 120 s band are a burst of Pi2 pulsations associated with substorm onset, then these data indicate that the substorm began at ~03:26 UT, some 4 minutes after the firing of the Shuttle thrusters and the perturbation of the mid-latitude, pre-midnight F-region conductivity. Hence, given the near coincidence of the Shuttle burn with the southward turning of the IMF, it would be very difficult to argue convincingly for the Shuttle initiation of the magnetic variations through the mechanism suggested by Rosenberg et al. [1981].

3.3. Magnetosphere-Ionosphere Interaction

The stimulation of hydromagnetic waves in the Pc4 band is consistent with the theoretically-predicted range of (80 s - 100 s) for a L ª 3 field line located inside the plasmapause [Orr and Matthew, 1971]. Although D and E-region conductivities were not measured during the experiment, and are not believed to have been affected directly by the Shuttle-exhaust recombination, the conductivity of the nighttime mid-latitude ionosphere is dominated by the density in the lower F region, and this was reduced by a factor of ~4 during the formation of the ionospheric hole. This rapid depletion of the lower F-region density, and the proportional conductivity decrease, could have provided an impulsive perturbation of the mid-latitude current, as discussed by Rosenberg et al. [1981]. The lower F-region perturbation occurred in time coincidence with the onset of the magnetic variations and could therefore be responsible for exciting the waves. In this scenario, the effects of the localized ionospheric perturbation produced by the Shuttle burn could be coupled into the overlying magnetosphere by varying the conductance at the ionospheric foot of the field line. Such ionosphere-magnetosphere coupling could then bring into play the wide range of magnetospheric processes which could amplify or spread the effects of the initial perturbation through the coupled magnetosphere-ionosphere system. The following studies indicate that localized ionospheric perturbations do couple into the overlying magnetosphere, where their effects can be spread more rapidly than through normal ionospheric propagation paths. Coupling of ionospheric perturbations into the magnetosphere has been suggested to explain the prompt arrival of ionospheric perturbations over large distances following discrete natural and artificial disturbances. Taran et al. [1985] report an event in which a fast-propagating ionospheric disturbance was observed with the Kharkov incoherent scatter radar as well as at distributed ionosonde and Doppler radar sites across the Soviet Union in response to a large surface explosion (~300 tons TNT). Propagation speed for the direct arrival of the ionospheric disturbances at various sites was determined to be in the range 2 km s-1 - 10 km s-1. Sochatsky et al. [unpublished manuscript, 1990] have used powerful military radars to investigate ionospheric disturbances accompanying missile launches. They report the occurrence of quick ionospheric disturbances observed at distances of 1000 km to 10,000 km which they attribute to hydromagnetic waves propagating along magnetospheric paths and resulting in apparent ionospheric propagation speeds of between 1 km s-1 and 10 km s-1. A variety of radio techniques have been used by Chernogor et al. [1993] to investigate ionospheric characteristics and their response to discrete perturbations. They report a category of events which exhibit apparent ionospheric propagation speeds of 10 km s-1 - 30 km s-1. Such propagation speeds vastly exceed the limits of ionospheric acoustic propagation and these reports suggest that there are mechanisms which couple a localized ionospheric perturbation into the magnetosphere, where its effects can propagate at the Alfvén speed to produce observable ionospheric signatures on field lines far distant from the causative event. A corollary to these studies of fast trans-ionospheric propagation is that the causative ionospheric perturbations can penetrate the magnetosphere, where their effects might modify magnetospheric processes. The extent to which the ionospheric changes produced by the July 31, 1985 Shuttle burn coupled into the magnetosphere is unknown, but the synchronism of ionospheric and magnetospheric effects during the event suggests that further investigations into the workings of magnetosphere-ionosphere coupling and feedback are needed.

4. Conclusions

Acknowledgments

References

Figure Captions

Figure 2. Temporal variation of the density at two altitudes, 300 km and 260 km, observed with the radar is shown in the upper panel. Vertical lines beginning at the time of the Spacelab-2 OMS engine burn (03:22 UT) and repeated with 40-min period highlight the quasi-periodic variation of the density at the lower altitude. Velocity data for 41° latitude and 300 km altitude are shown in the middle panel. In the lower panel, iso-density contours with overscribed fiducial lines delineate the downward phase progression of the density enhancements. Density peaks were observed to descend 200 km in altitude in 1460 sec and to repeat with 40-min period.

Figure 3. Radar elevation scan data in the bottomside F layer show the initial latitude spread of the hole to cover ~2° of latitude and the subsequent oscillation of the density across the field of view of the experiment with ~40-min period. Iso-density contours at fixed altitude (260 km) are shown. Figure 4. (A) Magnetic field variations observed by the Bell Labs magnetometer located on the field line intersecting the Spacelab-2 OMS burn. A pronounced D-component oscillation with 80-min period began at 03:31 UT. A negative Z perturbation spanned the interval 03:30 - 07:00 UT.

(B) Interplanetary magnetic field and plasma parameters observed by IMP-8 indicate a southward turning of IMF Bz and increase in solar wind plasma density in the interval following 03:00 UT on July 30, 1985. Such conditions could lead to enhanced magnetospheric activity.

(C) DMSP zero-energy electron boundary [Newell et al., 1996] moved equatorward between 03:30 UT and 07:00 UT, indicative of an equatorward expansion of the extent of auroral electric fields.

(D) NOAA-8 precipitating power estimates and DMSP central-plasmasheet flux indicate an increase in auroral-latitude precipitation between 05:00 UT and 06:00 UT.

Figure 5. Magnetic field signals filtered in the 45 s - 150 s band show bursts of pulsation activity following the Shuttle-burn experiment at 03:22 UT. The onset of pulsation activity occurred at ~03:26 UT.