JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 110, A05311, doi: /2004ja010795, 2005

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 110,, doi: /2004ja010795, 2005 Artificial disturbances of the ionosphere over the Millstone Hill Incoherent Scatter Radar from dedicated burns of the space shuttle orbital maneuver subsystem engines Paul A. Bernhardt Plasma Physics Division, Naval Research Laboratory, Washington, DC, USA Philip J. Erickson, Frank D. Lind, and John C. Foster Atmospheric Sciences Group, MIT Haystack Observatory, Westford, Massachusetts, USA Bodo W. Reinisch University of Massachusetts-Lowell, Lowell, Massachusetts, USA Received 17 September 2004; revised 22 December 2004; accepted 19 January 2005; published 27 May [1] Two ionospheric modification experiments were carried out over the incoherent scatter radar (ISR) located at Millstone Hill, Massachusetts. These experiments are part of the Shuttle Ionospheric Modification with Pulsed Localized Exhaust (SIMPLEX) program at the Naval Research Laboratory. The experiments use 10-s burns of the dual orbital maneuver subsystem (OMS) engines to produce the injection of high-speed molecules in the ionosphere near 380 km altitude. Charge exchange between the high-speed exhaust molecules and the ambient oxygen ions yields molecular ion beams that disturb the natural state of the ionosphere. Radar scatter provides measurements of the ion velocity distributions and plasma turbulence that result from the ion beam interactions. Groundbased observations with the University of Massachusetts Digisonde record the ionospheric density depressions resulting from recombination of the molecular ions with electrons. Prompt signatures of nonequilibrium ion distributions in the OMS engine plume are seen in the data taken during the SIMPLEX III and IV experiments for the space shuttle flights STS-108 and STS-110, respectively. The SIMPLEX III observations are much weaker than those during SIMPLEX IV. These differences are primarily attributed to the changes in the viewing directions for the radar beam. During SIMPLEX IV, the radar is looking more downstream from the exhaust injection and the stimulation of plasma turbulence is seen with the ISR for over 30 s at distances up to 200 km from the burn altitude along the radar beam. Strong backscatter in the radar spectra is attributed to ion acoustic waves driven by the pickup ion beams. Both experiments provide large-scale cavities detected by the Digisonde for up to 20 min after the engine burn. These cavities are the result of ion-electron recombination of the pickup ions. Citation: Bernhardt, P. A., P. J. Erickson, F. D. Lind, J. C. Foster, and B. W. Reinisch (2005), Artificial disturbances of the ionosphere over the Millstone Hill Incoherent Scatter Radar from dedicated burns of the space shuttle orbital maneuver subsystem engines, J. Geophys. Res., 110,, doi: /2004ja Introduction [2] The Shuttle Ionospheric Modification with Pulsed Localized Exhaust (SIMPLEX) series of experiments has been conducted by the Naval Research Laboratory for the past 6 years. The experiments use the space shuttle orbital maneuver subsystem (OMS) engines to inject exhaust over ground radar sites. The radar detects the changes in the ionosphere with incoherent scatter. SIMPLEX I was conducted over the Jicamarca Radio Observatory in Peru during STS-86 [Bernhardt et al., 2001]. This experiment monitored Copyright 2005 by the American Geophysical Union /05/2004JA the recovery of the ionosphere by field-aligned transport of plasma. The SIMPLEX II experiment was conducted over the Arecibo Observatory in Puerto Rico during STS-93 [Bernhardt and Sulzer, 2004]. For SIMPLEX II, the ion beams from the exhaust produced nonthermal spectra with the incoherent scatter radar. The SIMPLEX III and IV experiments over the Millstone Hill incoherent scatter radar in Massachusetts are reported here. [3] The SIMPLEX experiments use the space shuttle engines to produce regions of large relative neutral plasma convection. The exhaust of the space shuttle provides a high-speed neutral gas that streams through the ambient plasma of the ionosphere. The exhaust molecules exchange charge with the ambient O + ions in the ionosphere to give 1of10

2 Figure 1. Comparison of ring distribution functions for (a) high-speed plasma convection in the auroral ionosphere and (b) high-speed neutral injection into a stationary ionosphere. ring-beam ion distributions. These distributions are studied with ground radars using incoherent scatter of radio waves from the electrons in the ionosphere. The electron radar scatter is affected by both the ion distributions and any plasma turbulence in the modified ionosphere. The effects of the ion beams are transient because the ion particle motion is damped by ion-neutral collisions, the ring distributions are damped by instabilities, and the molecular ions created by the charge exchange eventually recombine with the electrons. [4] All of the SIMPLEX experiments for the modification of the ionospheric plasma in the exhaust plume of the space shuttle are designed to mimic naturally occurring ionospheric disturbances. The SIMPLEX III and IV experiments focus on the artificial generation of ion-ring and beam velocity distributions. A ring of plasma particles in velocity space may be formed in the ionosphere through coupling of the plasma and neutral interactions. In the high-latitude auroral region, the ambient O + ions and electrons in the ionosphere can be accelerated to high velocities by external electric fields (Figure 1a). Charge exchange of the highspeed oxygen ions chemically reacting with a stationary neutral molecule or atom produces an ion velocity ring distribution. In the reference frame of the plasma, the ions are at rest and the background neutrals are streaming across the ambient magnetic field lines. The space shuttle exhaust directly produces this motion of the neutrals through the plasma to give the distribution function of Figure 1b. The primary difference between auroral plasma convection and space shuttle OMS burns is the reference frame for the relative ion and neutral motion. In either case, if a neutral atom or molecule charge exchanges with an ion, the pickup ion that is created responds to the magnetic field by gyrating around a magnetic field line. The gyroradius is the perpendicular velocity of the initial neutral divided by the ion gyrofrequency of the pickup ion. Any initial velocity parallel to the magnetic field becomes a parallel streaming velocity of the pickup ions. Thus the initial perpendicular and parallel velocities of the ring-beam velocity distribution are entirely determined by the relative velocities between the ion and neutral streams. [5] Knowledge of ring distributions is important to understand the physics of both space and laboratory plasmas. Magnetic mirror geometries found both in magnetic fusion devices and in the Earth s radiation belts have loss-cone-type distributions which may be regarded as ring distributions. Ion rings are produced (1) as the solar wind interacts with the bow shock of the Earth s magnetosphere, (2) in comets as neutral atoms become ionized, and (3) when a solar loop collides with the interplanetary magnetic fields. Instabilities are excited by ion ring distributions leading to regions of plasma turbulence. The instabilities can generate radiation and the turbulence can scatter the charged particles. Ions injected perpendicular to the external magnetic field B can move in relatively large gyro-orbits, whereas the electrons are tied to the field lines. The ratio of these orbits from the charge exchange process is the particle mass ratio (m i /m e ) 33,000. [6] A description of the SIMPLEX experiments over Millstone Hill in Westford, Massachusetts is given in the next section. This is followed with observations using both ground-based radar and ionosonde diagnostics. The analysis section gives an interpretation of the results in terms ion velocity distributions that excite plasma waves in the ionosphere. In the final section it is concluded that these plasma waves scatter radar signals to yield a unique spectral signature of the exhaust plume interactions. 2. Experiment Description [7] The Millstone Hill incoherent scatter radar (ISR) operates at 440 MHz with transmissions from a fully steerable 46-m diameter dish. The Millstone Hill radar is located in Westford, Massachusetts (42.62 N latitude, E longitude, and 146 m altitude). Incoherent backscatter from the ionosphere is observed using the ISR and fitted to produce measurements of electron density, lineof-sight plasma velocity, and electron and ion temperatures. The backscatter from the radar can also be used to detect plasma turbulence and nonthermal ion velocity distributions. The Millstone Hill ISR has been previously used to observe ionospheric perturbations produced by the space shuttle during the flight of Spacelab-2 [Mendillo et al., 1987; Foster et al., 2000]. [8] Millstone Hill also hosts the Digisonde Portable Sounder (DPS-4) from the University of Massachusetts- Lowell. The Digisonde [Reinisch, 1996] produces ionograms that show the electron density profiles and regions of ionospheric density disturbances. The calibration of the absolute electron densities from the radar profiles obtained using the peak layer densities obtained from the critical frequencies is determined by the Digisonde. Ionosonde traces of ionospheric holes from rocket exhaust have been recorded for over 40 years [Booker, 1961; Reinisch, 1973]. The formation of these holes by dissociative recombination of electrons and pickup ions has been known for almost 30 years [Mendillo et al., 1975; Bernhardt et al., 1975]. [9] The two SIMPLEX experiments near Millstone Hill were conducted during the orbits of STS-108 and STS-110 of the space shuttle. The primary objective of the space shuttle flights was to rendezvous with the International Space Station (ISS). For these missions, the orbit inclination is 51.6 degrees. With this inclination, Millstone Hill is favorably placed to get many overflights. Both burns were conducted at the ends of the shuttle flights 1 day before deorbit. The SIMPLEX III and IV experiments occurred during the daylight hours with posigrade (ram facing) burns. 2of10

3 Table 1. STS-108 (SIMPLEX III) and STS-110 (SIMPLEX IV) Ion Beam Injection Experiment: Ionospheric Parameters and Plume Conditions SIMPLEX III SIMPLEX IV Date 16 December April 2002 Burn start time, UTC 1851: :19 Electron density n e,cm Electron temperature T e,k Ion temperature, T i0,k Pickup ion temperature, T i1, K Burn altitude, Z 0, km Atm. neutral density at burn, n n,cm Neutral atm. scale height, H n, km Average atm. neutral mass, m n, kg Atm. neutral temperature, T n, K Burn termination latitude, N Burn termination longitude, E Two engine injection flow rate, Q 0, molecules/s Molecular exit velocity, V E, km/s Shuttle velocity, V S, km/s Molecular injection velocity, V m, km/s Magnetic field, B 0, T Ion gyrofrequency, W i, Radians/s Electron gyrofrequency, W e, Radians/s Injection angle with B, a 0, degrees Parallel injection velocity, V mk, km/s Perpendicular injection velocity, V m?, km/s Radar scatter wave number, k =2k 1, radians/m Radar angle with magnetic field, q, degrees A comparison of the parameters for the two burns is given in Table 1. [10] The orbital maneuver subsystem (OMS) engines were fired for 10 s releasing exhaust molecules per second. The mole fractions of exhaust material are N 2 (31.3%), H 2 O (27.4%), H 2 (24.1%), CO 2 (12.2%), and CO (5%). The most important species for ion beam formation is H 2 O because of its reactivity with O +. The exhaust nozzle produces a 30 degree cone about the thrust axis. The exit velocity of the exhaust along the axis of the nozzle is 3.07 km/s. The exhaust from the nozzle cools by expansion to 120 K. Any ion created by charge exchange has the initial low temperature and high speed of the exhaust molecule. [11] Table 1 shows that the two Millstone Hill experiments had nearly identical conditions. The SIMPLEX III burn during STS-108 took place on 16 December 2001 at 1851:37 UT (1351:37 local time). The SIMPLEX IV burn during STS-110 was initiated on 8 April 2002 at 1726:19 UT (1226:19 local time). The molecular exit velocity (3.07 km/s around a 30 degree cone) was added to the orbit velocity (7.68 km/s) of each orbiter to give a vector sum for the exhaust injection speed ranging from km/s at the axis to km/s on the edge of the cone. The exhaust cone angle including the orbiter velocity is 8.4 degrees. Both injections took place with the engine thrust axes at an angle of 83.5 degrees relative to the Earth s magnetic field B. Because of the exhaust cone, the injection angle relative to B ranged from 75 to 90 degrees. Most of the exhaust velocity was directed perpendicular to the magnetic field, but there was a significant component of the velocity along B. The perpendicular-to-b injection velocity spread was 10.1 to 10.7 km/s for both experiments. The velocities parallel to B had a spread of 0.2 to 2.6 km/s for both STS-108 and STS-110. [12] All burns on ascending orbits of the space shuttle with 51.6 degrees inclination will have these injection parameters over Millstone Hill. For a descending orbit over Millstone Hill, the spread in the perpendicular-to-b injection angles will be 97 to 108 degrees and the perpendicular velocity spread will be 9.5 to 10.4 km/s. The parallel velocity spread will range between 97 and 108 km/s. In all cases, the exhaust material is released nearly perpendicular to the magnetic field but there is a significant component of parallel velocity. This is important for excitation of electrostatic waves perpendicular and parallel to B. The altitudes of the two burns were 365 and 395 km, respectively. [13] The first major difference between the burn observations was the pointing of the radar. The Millstone Hill radar was pointed at the closest approach to the burn with an elevation of 85 and an azimuth of 329 east of north for the STS-108 (SIMPLEX III) burn. At this position the radar made an angle of 154 degrees with the Earth s magnetic field B. For the STS-110 (SIMPLEX IV) burn the radar beam was pointed directly east (azimuth = 90 ) with an elevation of 74.6 degrees. With this pointing angle, the radar beam makes an angle of degrees relative to B. From the aspect of radar beam angle relative to the magnetic field, both experiments primarily view the plasma within about 25 degrees of the magnetic field direction. This is because the radar is pointed at high elevation in a magnetic field with a 69 degree dip angle. [14] The ionospheric plasma densities were similar at the exhaust release point. For SIMPLEX III, the electron density at the burn altitude was cm 3. The electron density at the SIMPLEX IV burn altitude was cm 3. The background electron density proportionally affects the density of the pickup ions. At the experimental altitudes, the electron density is equal to the O + ion density. The pick up ions (H 2 O + ) are made by the charge exchange reaction H 2 O+O +! H 2 O + + O. Consequently, the highspeed ion beam formed during the SIMPLEX III experiment should be about 1.4 times denser than the beam produced during the SIMPLEX IV experiment. 3of10

4 Figure 2. Map of orbital maneuver system (OMS) burn for the Shuttle Ionospheric Modification with Pulsed Localized Exhaust (SIMPLEX) III experiment. The burn occurred on 16 December 2001 with the ignition at 1851:37 UT, and the termination at 1851:47 UT. The map shows a projection of the radar beam, burn track, and ground site locations relative to the STS 108 orbit track. [15] The second major difference between the two experiments was the neutral densities at the burn altitude. This is because the SIMPLEX III burn occurred 30 km below the altitude of the SIMPLEX IV burn. The NRLMSISE-00 model [Hedin, 1987; J. M. Picone, private communication, 2003] is used to estimate the neutral densities at the burn altitudes. At 365 km altitude, the background neutral density for the SIMPLEX III burn was about cm 3. At 395 km altitude, the background neutral density for the SIMPLEX III burn was about cm 3. The lower densities for SIMPLEX IV might lead to greater transit distances for the injected particles through reduced collisions. A reduction in collision frequency can also provide longer lifetimes for any electrostatic waves generated by the ion beam distributions. [16] For both experiments, the radar was fixed to an initial location for the first 180 s after the burn. The radar was subsequently scanned to yield the spatial extent of the ionospheric disturbance. The Digisonde was operated continuously to yield ionograms at 5 min intervals during both experiments. The space shuttle exhaust produces an ionospheric hole that is easily detected with the wide field of view provided by the ionosonde. The plasma in the modified region is lost by dissociative recombination with the ion-electron reaction H 2 O + +e! OH + H. Thus after each exhaust burn, a hole is left in the plasma. This has been observed on many occasions [Bernhardt et al., 2001]. What is novel about the Millstone Hill experiments is the remote detection of the transient ion beam distributions and their effects immediately after the space shuttle OMS burns. 3. SIMPLEX III Experimental Observations [17] The ground track of the SIMPLEX III experiment is illustrated in Figure 2. SIMPLEX III data were taken with the radar pointed at the closest approach for the orbit. The 10-s burn was terminated at that point (Figure 2). The range to the burn termination point for SIMPLEX III was 366 km. The radar views the injection nearly perpendicular to the orbit vector. [18] The SIMPLEX III data were obtained with the Millstone Hill radar using line-of-site backscatter. The backscatter data before and just after the burn are illustrated in Figure 3. Prior to the burn, a thermal ion line spectra are recorded (Figure 3a) and the F region had a peak density of cm 3 near 300 km altitude. The time of ignition (TIG) was 1851:37 UT. At the time of the burn, the radar Figure 3. Range-time-intensity (RTI) plot and sample spectra for the OMS burn over Millstone Hill on 16 December The temporal and spatial resolutions are 10 s and 24 km, respectively. (a) and (d) The double peaked ion line spectra are seen in most of the spectra. (b) The reflection from the orbiter is clearly seen in the backscatter data and the Doppler shifted line. (c) The only unusual feature in the ion line spectra is a central peak near a zero frequency shift. Following the burn, the backscatter profiles show a density reduction. The range of the display saturates the echoes from the space shuttle which is 200 times the backscatter level of the ionosphere. The contouring introduces some smoothing of the data. 4of10

5 overwhelming evidence of ion beam effects from the burns. They are significant, however, in light of the results of the SIMPLEX IV experiment given in the next section. Figure 4. Digisonde record of the ionospheric hole produced by the SIMPLEX III burn over Millstone Hill on 16 December shows a large backscatter profile from the orbiter and a Doppler shifted echo (Figure 3b). The radar echoes from the space shuttle are consistent with the line-of-sight velocity of the orbiter. [19] The burn occurred at between 1851:37 and 1851:47 UT in the red area of Figure 3. The radar profile modifications resulting from the STS-108 OMS burn are shown in the right half of Figure 3. At 1852:50 near a range of 451 km, a small central peak is observed in the ion line spectra (Figure 3c). The ionospheric hole extends for about 40 km above and below the burn range of 366 km. The radar data show a density depression that is centered at the altitude of the space shuttle above the layer peak. By 1957:00 (not shown), the ionospheric hole has disappeared from the radar data, as the horizontal plasma drift transports the localized density hole away from the radar beam path. [20] The existence of the hole past 1957:00 is demonstrated by using the ground-based Digisonde near Millstone Hill. The density hole refills by flow of plasma along the magnetic field lines. This eventually causes a localized reduction in the density at the layer peak and below. When the hole penetrates though the layer peak, oblique echoes from the ground-based ionosonde were observed. The ionogram records from the Millstone Hill Digisonde (Figure 4) shows an oblique echo of the ionospheric hole from 1900 through 1920 UT, 8 to 28 min after the burn. After this time, the density hole is convected away from the field-of-view of the Digisonde. [21] Examination of individual spectra from the Millstone Hill radar showed very little distortions from the usual ion line shapes. Under normal conditions, the incoherent scatter spectra show thermal ion line peaks located above and below the radar frequency. These peaks are caused by Landau-damped ion acoustic waves at offsets equal to the local ion acoustic frequency. When the electron temperature is greater than or equal to the ion temperature, the incoherent scatter spectra show two peaks displaced above and below the radar frequency by the ion acoustic frequency (Figures 3a and 3d). Individual spectra for the SIMPLEX III data were examined, and only three spectra at the time illustrated by Figure 3c showed any unusual features. These spectra are characterized by a central line in addition to the double peaks of the thermal ion line. The distorted spectra are found about 100 km above the burn 1 min after the burn was terminated. By themselves, these data do not provide 4. SIMPLEX IV Experimental Observations [22] During SIMPLEX IV the radar was pointed directly to the east at an elevation of 74.6 degrees at the termination point of the burn (Figure 5). The radar range to that point was 410 km. With this radar geometry, much of the exhaust expands along the radar path. [23] The results from the SIMPLEX IV experiment during STS-110 on 18 April 2002 are more complex than the results for the SIMPLEX III experiment. Figure 6 illustrates the range corrected backscatter with 2-s time resolution and 24 km spatial resolution. There are three unusual features in the data. First, a strong echo from the space shuttle occurs just after the engine burn. Second, a strong echo is detected from the International Space Station (ISS). The STS-110 orbiter had recently undocked from the ISS and they were still in the same orbit. Finally, strong clumps of backscatter enhancements were observed for 1-min duration over distances 100 km below and 200 km above the burn altitude. [24] The radar data from the SIMPLEX IV (STS-110) experiment has been analyzed to provide spectra and backscatter intensity. The radar backscatter profiles from the time of the burn to 30 s after the burn show the same features (electron density profile and reflections from the space shuttle), as does Figure 3 for SIMPLEX III. The period of time after the burn is significantly different. Figure 6 displays range corrected backscatter with the data analyzed with 2-s samples. The ignition time for the 10-s burn was 1726:19 UT. Starting at 60 s after the burn, clumps of anomalous scatter were observed in the ISR backscatter. The clumps are the result of turbulence driven by the ion ring-beam velocity distributions. The scattered power spectra from the turbulence (Figures 6c and 6d) can be as much as two times that of the thermal ion line (Figure 6a) Figure 5. Map of OMS burn for the SIMPLEX IV experiment. The burn took place on 18 April 2002 with the ignition at 1726:18.95 UT and the termination at 1726:28.95 UT. The map shows a projection of the radar beam looking into the modified volume. 5of10

6 Figure 6. Radar spectra and total backscatter power during the SIMPLEX IV burn over Millstone Hill. The data were obtained on 18 April 2002 with a 10-s burn (in red) starting at 1726:19 UT. The temporal and spatial resolutions are 2 s and 24 km, respectively. The spectral data show (a) undisturbed ion lines, (b) echoes from the space shuttle and International Space Station, and (c) and (d) scatter from irregularities. The range of the display saturates the echoes from the space shuttle and the ISS which are 1400 and 430 times the backscatter level of the ionosphere, respectively. The contouring introduces some smoothing of the localized turbulent patches. and one-third of the echo power from the space shuttle (Figure 6b). [25] Following time 1728:12 (TIG s), a depression was seen in the backscatter data of Figure 6. This depression is the result of ion-electron recombination that yields reductions in electron density and is also responsible for the oblique echoes for the Digisonde records that last for 15 min (Figure 7). To better study the evolution of the hole, the Digisonde range was extended to beyond the usual limit of 1000 km. The chemical and dynamical processes for the formation and recovery on an ionospheric hole are well known [Bernhardt et al., 2001]. The new discovery of enhanced scatter in the form of localized clumps, however, is subject for further investigation. [26] To aid analysis of the enhanced scatter, the individual spectra from the ISR given in Figure 8 were obtained with 10-s integrations of the total power in a frequency spectrum centered at the 440 MHz radar frequency. As explained previously for the SIMPLEX III data, the double-peaked spectrum (Figure 8a) is typical of the thermal ion line in the unmodified ionosphere. Analysis of this type of thermal spectra yields electron density, line-of-sight velocity, and both electron and ion temperatures. The single shifted peak (Figure 8b) is the echo from the space shuttle. [27] Multipeaked spectra observed between 1227:22 and 1728:02 are very unusual, and they provide the first radar evidence of ion beam driven turbulence from a space shuttle OMS burn. The multipeaked spectra are primarily asymmetric (Figure 8c) with some indication of harmonic structure. The first and second downshifted harmonics and the first upshifted harmonic of the ion acoustic frequency are labeled 3f IA, 2f IA, and 2f IA, respectively. The modified spectra have a central peak that is not found in the thermal ion line spectra. Some of the modified spectra have Figure 7. Digisonde record of the ionospheric hole produced by the SIMPLEX IV burn over Millstone Hill on 18 April The O-mode and X-mode traces are red and green, respectively. The oblique O-mode echoes from the ionospheric hole created by the engine burn are outlined in black. Figure 8. Sample spectra for radar scatter at 440 MHz. The spectra are identified as (a) thermal ion line, (b) Doppler shifts from space shuttle, (c) strong scatter from electrostatic waves, and (d) thermal ion line with additional scatter from ion ring velocity distribution. 6of10

7 line are due to scattering from electron density fluctuations produced by electrostatic waves. 5. Sources of Spectral Features [29] The ion beams created by the high-speed exhaust produce a number of additional lines in the radar scatter spectra. The features in the spectra are the result of electromagnetic wave scatter from the electron turbulence in the plasma. The resulting scattered wave frequency (w) and wave number (k) are given by the matching conditions for wave-wave interactions [Nicholson, 1983] w S ¼ w R þ w 1 ; k S ¼ k R þ k 1 ; ð1þ Figure 9. Comparison of thermal ion-line with and multipeaked ion-line before and after the SIMPLEX IV burn over Millstone Hill. The observations were made at 436 km altitude where the undisturbed plasma has T e = 2690 K and T i = 1480 K. several peaks that are downshifted by as much as 25 khz. As the ionosphere recovers, the harmonic components disappear and only an ion line with a central spectral peak remains (Figure 8d). Eventually, the central peak vanishes and the plasma relaxes to a thermal state with ion line spectra similar that of Figure 8a. Physical explanations of these features are given in the next section. [28] An overlay of the thermal ion line and multipeaked modified ion line spectra is presented in Figure 9 with a tentative assignment for the line features. The thermal ion line and the modified ion line both show symmetric peaks for frequency shifts of ±4.8 khz, as indicated by the dash-dot line in the figure. This frequency shift is near the ion acoustic frequency for the plasma. The downshifted peak in the modified ion line is found at 9.6 khz, twice the frequency of the thermal ion line peaks. The small upshifted peak in the modified ion line is found near 11 khz. Finally, the modified ion line has an additional peak at a zero frequency shift from the 440 MHz radar frequency. It seems likely that some of the additional peaks in the modified ion where subscripts S, R, and 1 designate the scattered, radar, and plasma wave quantities. For backscatter, k S = k R, and the Bragg condition is found from (1) as k 1 =2k S. The frequency shift component represented by w 1 in (1) for the fixed value of wave number k 1 is determined by examining the fluctuations in the plasma. Two processes for creation of these frequencies are considered. First, the scatter from the electrons is modified by the nonthermal distribution of ion velocities. Second, the scatter is modified by electrostatic waves from instabilities driven by the ion velocity distribution. [30] A general treatment of incoherent scatter without instabilities from ring-beam distributions has been described by Bernhardt et al. [1998]. The theory assumes that the plasma is in quasi-equilibrium where the relaxation times for the ion-velocity distribution function are much longer than the correlation times of the density fluctuations [Swartz and Farley, 1979]. Using this theory, the parameters for the ring-beam distribution were adjusted to produce scatter at a central peak as shown in the observations. The thermal and ion-beam distributions for this calculation are graphed in Figure 10. The initial injection velocities have been reduced to 3.2 km/s perpendicular to B and 0.4 km/s parallel to B to represent the effects of collisions. Ion neutral collisions have also increased the temperature of the pickup ions from the initial value of 120 K to 300 K. [31] The computed ion line spectra from the thermal and ion-beam velocity distributions are shown in Figure 11. The dashed spectrum is the typical thermal ion line. The solid Figure 10. Ambient and plume ion distributions. The velocities are normalized by the ion thermal speed, v i = km/s. Figure 11. Radar spectrum simulation using a thermal (dashed) and a ring-beam (solid) velocity distribution. 7of10

8 line spectrum has a central peak that is Doppler shifted by the 0.4 km/s motion parallel to the magnetic field. The combination of the central peaked spectra and the thermal ion line spectra is similar to ISR data shown in Figures 2c and 8d. The bimodal velocity distributions can produce many of the three-line spectra obtained at Millstone Hill during the SIMPELX IV experiment. For additional spectral lines, generation of waves by ion-beam plasma instabilities must be considered. [32] Several instability mechanisms can be considered. The pickup ions in the exhaust form a ring in velocity space which is a source of free energy for several unstable plasma modes. The nature of the instabilities depends on a number of plasma parameters. [33] The dynamic plasma beta (b0) is defined as the dynamic pressure (n i m i v i 2 ) divided by the magnetic pressure (B 2 /2 m 0 ). For the SIMPLEX experiments, b0 is about and is much less than unity. For low-beta plasma, the magnetic field is not distorted by the ion beam injections. Hydromagnetic waves in a plasma move along magnetic field lines at the Alfvén speed (B/m 0 m i n i ) 1/2. For the SIMPLEX experiments the Alfvén speed is approximately 230 km/s. This speed is much larger than the ion injection speed (v i = m/s), and hydromagnetic effects on the instabilities are not important. [34] The ambient ions travel with a thermal speed (k T i / m i ) 1/2 of about 0.87 km/s which is much lower than the injection speed km/s. When the neutrals charge exchange with the background ions they will be supersonic. Initially, the pickup ions will move in orbits with a fixed ion frequency (W i =eb/m i ) and ion gyroradius (v i /W i ). The ion gyrofrequency for the SIMPLEX experiments is about 235 Radians/s. This is much less than the corresponding electron cyclotron frequency (W e =eb/m e )of Radians/s. The plasma gyroradii are about 45 m for the ions and m for the electrons. For this reason, the electrons are often considered to be magnetized relative to the ions. Finally, the electron thermal velocity (k T e /m e ) 1/2 is about 200 km/s, much larger than the neutral or ion speed. The electron velocity distribution is not expected be directly modified by the exhaust injection. [35] The gyromotion of the ions will continue until their orbits are disturbed by electrostatic wave interactions or by collisions. The ion-neutral collision cross section (Q D )is approximately m 2. For supersonic ion particles the mean free path (l D ) of ions in the neutral atmosphere is (N n Q D ) 1. For the SIMPLEX experiments the mean free path for the pickup ions is 2.6 km at 365 km altitude and 4.2 km at 395 km altitude. This ion travel distance increases with altitude. Even with a substantial parallel velocity component, the ions will not move more than a few kilometers from the point of charge exchange. The pickup ions, however, will undergo many gyrations before they are damped by collisions with the background neutrals. The average time between collisions (l D /v i ) is 0.24 and 0.38 s, respectively, for SIMPLEX III and IV. The average number of gyrations per collision (W i /2p)(l D /v i )is about nine orbits/collision at 365 km altitude, increasing to 14 orbits/collision at 395 km altitude. With each collision, each ion will loose a fraction of its perpendicular speed. Consequently, the gyromotion of each ion should last a few seconds after creation. During these few seconds, these gyrating ions comprise the ion ring velocity distribution that can excite electrostatic instabilities. [36] Electrostatic instabilities may be excited by the ion beam injections. To investigate possible waves excited by the space shuttle pickup ions, the wave modes in the plasma are determined. These modes satisfy the electrostatic dispersion equation given by Nicholson [1983] k 2 x W2 i 1 C2 S k2 w 2 þ k2 x W iw e k 2 w 2 k 2 z W2 e k 2 ¼ 0 w 2 k 2 z W2 i where C 2 S ¼ p ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðt e þ T i Þ=m i ; W e;i ¼ q e;ib ; m e;i k x ¼ ksinq; k z ¼ k Cos q Using the specific numerical values for the parameters from Table 1, the specific wave modes for the SIMPLEX III and IV experiments are calculated. [37] Assuming that W i 2 C S 2 k 2 W e 2, the three wave modes in the plasma are found to be (1) ion-cyclotron waves with the dispersion w 2 1 ¼ k2 z W2 i k 2 ; (2) ion acoustic waves with an electrostatic ion cyclotron correction and the dispersion w 2 2 ¼ C2 S k2 þ k2 x W2 i k 2 ; and (3) electron cyclotron/lower-hybrid waves with the dispersion w 2 3 ¼ k2 z k 2 W2 e k2 x k 2 W ew i : Solving the dispersion equation (2) for the three wave modes yields the curves in Figure 12, where all the experimental parameters in Table 1 are used except that the radar angle with the magnetic field is varied. The only waves that yield frequencies near the observed shifts of a few kilohertz are lower-hybrid waves and ion acoustic waves. [38] The lower-hybrid instability is the most common mode driven by an ion ring velocity distribution in a lowb, sub-alfvénic plasma. The lower-hybrid frequency (w lh ) for conditions of the SIMPLEX experiments is given by (W i W e ) 1/2. The calculated lower-hybrid frequency at the 365 km injection altitude for the SIMPLEX III experiment is 6950 Hz. The lower-hybrid frequency at a height 395 km for SIMPLEX IV is 6800 Hz. These waves are excited with propagation angles near 90 degrees to the magnetic field and they may be detected by scattering radar signals from the disturbed plasma. Unfortunately, as indicated in Figure 9, the Millstone Hill radar is limited to angles near 155 degrees for the scatter measurements. Under these conditions it is unlikely that the lower-hybrid waves will be seen even if they are produced with large amplitudes. [39] The streaming of an ion beam through an electron gas can destabilize ion acoustic waves. Figure 12 illustrates that the frequency of the ion acoustic waves is nearly ð2þ 8of10

9 Figure 12. Electrostatic wave modes for the SIMPLEX experiments with radar scatter wave number k = Radians/s. For oblique observations, ion-acoustic waves are the most likely to be responsible for the observed radar scatter. constant except for perpendicular scatter. With the SIMPLEX experiments, the ion cyclotron effects on the ion acoustic waves are negligible. The ion parallel velocities from the nozzle cone of the OMS engines are spread in the range of 0.27 to 2.60 km/s. For excitation of ion acoustic waves, this parallel speed must be greater than the ion sound speed (C s ). The ion acoustic frequency (w ia ) for the radar scatter wave number (k s ) is given by C s k s, where C s = (k T e /M i ) 1/2 is the ion sound speed. For the SIMPLEX III and IV experiments the ion sound speed is near 1100 m/s. The exhaust produces ions which stream along the magnetic field lines at over twice this speed. Landau damping of ion acoustic waves limits their growth. The electron temperatures were more than 20 times the pickup ion temperatures. With this temperature ratio, Landau damping of the ion acoustic waves is not important. [40] The formulas for the ion acoustic instability given by Goldston and Rutherford [1995] include both ion beam excitation and Landau damping. With these formulas, the instability growth rate is computed as shown in Figure 13 for the range of parallel ion speeds. The instability grows for all ions with parallel speeds greater than 1.4 km/s. With a positive growth rate on the order of 50 s 1 and a mean collision time of 0.3 s, the ion acoustic instability should not be significantly damped. The ion acoustic instability produces density fluctuations which scatter radar signals. The ion acoustic frequency for the SIMPLEX III and IV radar geometry is 3260 Hz. Figure 6 shows spectral lines that are harmonics of 4800 Hz. These harmonics may be the result of a strongly driven ion acoustic instability. [41] Two features of the radar spectra remain as mysteries. First, the enhanced scatter power is nearly the same as the power from the thermal ion line. The intensity of the coherent scatter from beam driven instabilities is usually much larger than that of incoherent power. The weak power in the anomalous lines is probably due to the nonoptimal orientation of the radar relative to the magnetic field. Table 1 shows that the ions are injected with an angle of 83.4 degrees relative to B. The radar was viewing the plasma at an angle of 23 degrees. The excitation of plasma turbulence is expected to be directed along the beam injection angle, so as the ions spiral around the magnetic field lines, the strongest wave should be launched at an angle of 83.4 degrees. Future radar scatter observations should be attempted with radar viewing angles matching the ion beam injection angles. [42] Second, the formation of harmonics is difficult to explain. The radar k vector is fixed and the corresponding plasma wave frequency is uniquely determined from the plasma dispersion relations (2). Consequently, only the fundamental frequency shifts are expected with three-wave interactions given by (1). For harmonic generation in the ion line spectra, the radar wave must couple to two electrostatic waves to yield induced scatter. In this four-wave interaction, the frequency/wave number conservations equations are given as w S ¼ w R þ w 1 þ w 2 ; k S ¼ k R þ k 1 þ k 2 : A solution to these equations for the ion acoustic instability with backscatter geometry is w 1;2 ¼ k 1;2 CS ; k S ¼ k R ; k 1 þ k 2 ¼ 2k S : The corresponding solution for the scattered wave has the second harmonic of the ion acoustic frequency ð3þ ð4þ w S ¼ w R þ 4jk S jc S ¼ w R þ 2w 1 : ð5þ Thus using four or more waves, the harmonics can be generated. Further investigation of the ion-beam interac- Figure 13. Ion acoustic instability excitation by streaming ions. The spread in exhaust from the OMS engine nozzle yields a range in parallel ion speeds from 0 to 2.6 km/s. 9of10

10 tions is required to justify the conditions in (4), which will yield frequency shifts at exact multiples of the ion acoustic frequency. 6. Conclusions [43] Unique experiments involving the space shuttle and the Millstone Hill Incoherent Scatter Radar (ISR) were conducted during SIMPLEX III and SIMPLEX IV. Dedicated ion beams were created with large beams using the space shuttle OMS exhaust. The space shuttle burn over Millstone Hill produced unusual incoherent scatter spectra which have been interpreted as incoherent scatter from electrons in the presence of non-maxwellian ion velocity distributions and turbulence that results from these distributions. The spectra were the result of exhaust injection speeds of over Mach 10 which yielded high-speed molecular (H2O+, etc.) ion beams. Theoretic analysis with a ring-beam velocity distribution produced a central peak of ISR spectra that was similar to the observations. The ion beam drives instabilities that create lower-hybrid and ion acoustic waves. [44] Future experiments should be conducted with changes in observation geometry and time of day. From the geometry of the SIMPLEX III and IV observations, only the effects of the ion acoustic waves were detected by the radar scatter. The new observations of ion beams exciting ion acoustic waves were recorded for almost 60 s with ranges greater than 150 km from the burn. Much stronger ion acoustic waves should be detected matching the radar view angle with the exhaust injection angle relative to B. The ion acoustic waves were detected during the daytime where the electron temperature was 20 times larger than the 120 K pickup ion temperature. The ion acoustic waves should be Landau damped for experiments that occur at night when the electron temperature is lower and the ratio of electron to pickup ion temperature falls below about ten [Goldston and Rutherford, 1995]. This explains the lack of detection for ion acoustic turbulence during the SMIPLEX II experiments which occurred at night over the Arecibo Incoherent Scatter Radar [Bernhardt and Sulzer, 2004]. Only lower-hybrid waves, which are not affected by Landau damping, are expected to be excited by the pickup ion beams at night. The effects of the lower-hybrid waves may be observed with a radar beam oriented perpendicular to the ambient magnetic field. Unfortunately, for the high latitudes of the Millstone Hill radar, the geometry of the magnetic field does not permit perpendicular backscatter measurements with ground radars. Future experiments using lower-latitude radars should be employed to detect the lower-hybrid waves. Finally, further modeling and more radar observations at Millstone Hill are required to describe the harmonic features seen in the UHF radar spectra. In situ observations with satellite plasma wave receivers should be employed to detect the full spectra of turbulence excited by the space shuttle OMS burns. Opportunities for these measurements may occur with the launch of the Air Force Communication/Navigation Outage Forecasting System (C/NOFS) satellite [de la Beaujardière et al., 2004] and the Canadian Enhanced Polar Outflow Probe (epop) Mission on the Cascade, Smallsat and Ionospheric Polar Explorer (CASSIOPE) satellite (A. Yau, private communication., 2004). [45] Acknowledgments. This research was sponsored at the Naval Research Laboratory by the Office of Naval Research. The space shuttle operations are scheduled by the Air Force Space Test Program with the assistance of Capt Y.M. Fedee and D.T. Walker at the Johnson Space Center. The Millstone Hill radar is operated by the Massachusetts Institute of Technology (MIT) under an NSF cooperative agreement. B.W Reinisch acknowledges support by Air Force grant F C [46] Arthur Richmond thanks Michael Sulzer and Wesley Swartz for their assistance in evaluating this paper. References Bernhardt, P. A., and M. J. Sulzer (2004), Incoherent scatter measurements of ion beam disturbances produced by space shuttle exhaust injections into the ionosphere, J. Geophys. Res., 109, A02303, doi: / 2002JA Bernhardt, P. A., C. G. Park, and P. M. Banks (1975), Depletion of the F2 region ionosphere and the protonosphere by the release of molecular hydrogen, Geophys. Res. Lett., 2, 341. Bernhardt, P. A., J. D. Huba, W. E. Swartz, and M. C. Kelley (1998), Incoherent scatter from space shuttle and rocket engine plumes in the ionosphere, J. Geophys. Res., 103, Bernhardt, P. A., J. D. Huba, E. Kudeki, R. F. Woodman, L. Condori, and F. Villanueva (2001), The lifetime of a depression in the plasma density over Jicamarca produced by space shuttle exhaust in the ionosphere, Radio Sci., 36, Booker, H. G. (1961), A local reduction of F-Region ionization due to missile transit, J. Geophys. Res., 66, de la Beaujardière, O., et al. (2004), C/NOFS: A mission to forecast scintillations, J. Atmos. Sol. Terr. Phys., 66, Foster, J. C., J. M. Holt, and L. J. Lanzerotti (2000), Mid-latitude ionospheric perturbation associated with the Spacelab-2 plasma depletion experiment at Millstone Hill, Ann. Geophys., 18, Goldston, R. J., and P. H. Rutherford (1995), Introduction to Plasma Physics, Inst. of Phys. Publ., Bristol, Penn. Hedin, A. E. (1987), MSIS-86 thermospheric model, J. Geophys. Res., 92, Mendillo, M., G. S. Hawkins, and J. A. Klobuchar (1975), A sudden vanishing of the ionosphere caused by the launch of Skylab, J. Geophys. Res., 81, Mendillo, M., J. Baumgardner, D. P. Allen, J. Foster, J. Hold, G. R. A. Ellis, A. Klekociuk, and G. Reber (1987), Spacelab-2 plasma depletion experiments for ionospheric and radio astronomical studies, Science, 238, Nicholson, D. R. (1983), Introduction to Plasma Theory, John Wiley, Hoboken, N. J. Reinisch, B. W. (1973), Burnt-out rocket punches hole into ionosphere, in Space Research XIII, edited by M. J. Rycroft and S. K. Runcorn, pp , Akademie, Berlin. Reinisch, B. W. (1996), Modern ionosondes, in Modern Ionosphere Science, edited by H. Kohl, R. Ruester, and K. Schlegel, pp , Eur. Geophys. Soc., London. Swartz, W. E., and D. T. Farley (1979), A theory of incoherent scattering of radio waves by plasma: 5. The use of the Nyquist Theorem in general quasi-equilibrium situations, J. Geophys. Res., 84, P. A. Bernhardt, Plasma Physics Division, Naval Research Laboratory, 4555 Overlook Avenue, SW, Washington, DC , USA. (bern@ppd.nrl.navy.mil) P. J. Erickson, J. C. Foster, and F. D. Lind, Atmospheric Sciences Group, MIT Haystack Observatory, Westford, MA , USA. B. W. Reinisch, University of Massachusetts-Lowell, Lowell, MA 01854, USA. 10 of 10