In situ observations during an HF heating experiment at Arecibo: Evidence for Z-mode and electron cyclotron harmonic effects

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. A10, 1382, doi: /2003ja009922, 2003 In situ observations during an HF heating experiment at Arecibo: Evidence for Z-mode and electron cyclotron harmonic effects L. J. Gelinas, 1 M. C. Kelley, 2 M. P. Sulzer, 3 E. Mishin, 4 and M. J. Starks 5 Received 5 March 2003; revised 18 July 2003; accepted 24 July 2003; published 29 October [1] On 11 March 1998 the Langmuir Turbulence sounding rocket was launched through the Arecibo heater beam during an experiment to measure electric fields and plasma densities in the heater interaction region. In spite of a serious degradation of the Arecibo heater, the rocket data has provided evidence of Z mode waves and field aligned striations above the O mode reflection height. These observations give credence to the theory that Z mode wave interactions with field-aligned striations may be, in part, responsible for some of the features of the reflected HF spectrum observed in heating experiments. We also find strong lowfrequency electric fields where the local plasma frequency matches an electron cyclotron harmonic. This suggests that electron Bernstein modes must be generated in the interaction process and can propagate to regions where they are severely damped. INDEX TERMS: 2411 Ionosphere: Electric fields (2712); 2439 Ionosphere: Ionospheric irregularities; 2483 Ionosphere: Wave/particle interactions; 2487 Ionosphere: Wave propagation (6934); 2403 Ionosphere: Active experiments; KEYWORDS: ionospheric heating, Z mode, striations, electron Bernstein modes Citation: Gelinas, L. J., M. C. Kelley, M. P. Sulzer, E. Mishin, and M. J. Starks, In situ observations during an HF heating experiment at Arecibo: Evidence for Z-mode and electron cyclotron harmonic effects, J. Geophys. Res., 108(A10), 1382, doi: /2003ja009922, Introduction [2] When a high-power radio wave is reflected from the ionosphere a number of fascinating phenomena occur. Ground-based radar studies of ionospheric heating show a rich spectrum of reflected waves, both upshifted and downshifted from the transmitted frequency. Over the years several sounding rockets have been flown through highpower radio wave beams [Rose et al., 1985; Kelley et al., 1995; Arce and Kelley, 1994; Peria et al., 1999; Peria and Kelley, 2001]. In each case the goal was to diagnose how the high power waves interacted with the ionosphere near the reflection altitude. At high power levels the wave does not simply reflect from the critical level where the wave frequency equals the plasma frequency but instead nonlinearly interacts with the plasma, creating a variety of other waves and plasma irregularities. [3] Parametric three and four-wave processes describing the interaction of the heater wave and the plasma near reflection have been reviewed extensively in the literature 1 Laboratory for Plasma Studies, Cornell University, Ithaca, New York, USA. 2 Department of Electrical and Computer Engineering, Cornell University, Ithaca, New York, USA. 3 Arecibo Observatory, Arecibo, Puerto Rico. 4 Boston College, Institute for Scientific Research Chestnut Hill, Massachusetts, USA. 5 Space Weather Center of Excellence, Air Force Research Laboratory, Hanscom Air Force Base, Massachusetts, USA. Copyright 2003 by the American Geophysical Union /03/2003JA [e.g., Vas kov and Gurevich, 1977; Huang and Kuo, 1994; Istomin and Leyser, 1997; Gurevich et al., 1995; Gurevich, 1999; Frolov et al., 2000a, 2000b; Leyser, 2001; Robinson, 1989] (see also the collection of papers in Radio Science, 9(11), 1974). In addition, field-aligned striations (FAS) are generated during O-mode heating in all but the very early times in cold start experiments at Arecibo [Sulzer and Fejer, 1994]. The FAS must be generated by some interaction of the O mode heater wave in the region near the upper hybrid layer, H UH [Gurevich, 1978]. At this height the heater wave frequency matches the local upper hybrid frequency. Interaction between the O mode waves and FAS may occur throughout the region between the upper hybrid layer, H UH and the reflection height, H O. At high latitudes, due to the magnetic dip angle, interaction near the upper hybrid layer is more likely; at lower latitudes, conversion to Langmuir waves near the reflection height is favored [Leyser, 2001]. [4] In addition to processes involving direct electrostatic conversion of the heater O mode wave, there is some evidence that the O mode wave may first be converted to the Z mode, which can then propagate above the O mode reflection height. These Z mode waves can then be scattered on the FAS, producing the HF outshifted plasma line (HFOL), which has been observed above the O mode reflection height [Mishin et al., 1997; Mishin et al., 2001]. Conversion of O mode radiation to Z mode is possible via a narrow HF transmission window or due to scattering from FAS, allowing for parametric processes at altitudes above the O mode reflection height. The HFOL has been observed several kilometers above O mode reflection and, in the case SIA 7-1

2 SIA 7-2 GELINAS ET AL.: HF HEATING EXPERIMENT that the F peak maximum is below the cutoff frequency for the Z mode, just below the topside O mode reflection region. Current theory requires the presence of both Z mode waves and FAS above O mode reflection to explain these outshifted lines. [5] With the exception of the few in situ rocket probes, referenced above, and a single satellite fly-through [Farley et al., 1983], these interactions have mainly been studied remotely by using probing radio waves and airglow. These ground-based techniques are very powerful and much has been learned by their use; however, in situ rocket measurements provide a very unique viewpoint and can help distinguish between the several alternative theories which have been developed by examination of remote sensing data. [6] In a previous sounding rocket experiment, the 1992 Soliton rocket flew through the Arecibo heater beam and observed several bunches of density depletions, interpreted to be the field-aligned striations predicted by theory [Kelley et al., 1995; Franz et al., 1999]. In the 1992 experiment, the heater had been on for several tens of minutes, and the ionosphere was in a steady state. Three bunches of 10 m scale density depletions were detected, one just above and two below the O mode reflection height. The magnitude of the density depletions above O mode reflection were largest, estimated to be on the order of 10%, with less intense depletions at lower altitudes. The measured depletions were also found to be associated with temperature enhancements [Peria et al., 1999]. The formation mechanism for the striations was favorably compared to the theory of Gurevich et al. [1995] by Franz et al. [1999]. The latter authors went on to show that short wavelength waves seemed to be present inside the filaments, possibly drift waves. [7] The experiment reported here was performed at Arecibo in March It was planned to be somewhat different in nature from the 1992 rocket, with the goal of studying ionospheric processes earlier in the heating cycle, on the order of a few tens of seconds. The experiment, unfortunately, encountered problems, described in the next section. Problems notwithstanding, the data has been unexpectedly rich. Here we present and discuss electric field data throughout the flight, with the hope that these data will provide motivation for new investigations. The focus of the data presentation emphasizes regions likely to play a role in observed RF heating phenomena. In particular, the discussion of the data will focus on interactions of the heater wave near the reflection height, existence of Z mode waves, waveforms associated with FAS, and unexpected electric field structures possibly related to electron cyclotron harmonics Experiment Description [8] The Langmuir Turbulence sounding rocket was launched from Tortugerro, Puerto Rico on 11 March 1998 and flew through the Arecibo 5.1 MHz heater beam, which had been on at a 10% duty cycle for several tens of minutes prior to launch. About 40 s before the rocket reached the predicted O mode reflection height, the heater was turned to full power. Unfortunately, just after the switch to high power, a set of phase shifters designed to tip the beam into the rocket s path failed catastrophically and caught on fire. From the data shown below it is clear that the radiation continued, but at an unknown power and in a linear polarization mode, rather than the desired O-mode. The resulting signal seems to consist of roughly equal parts O and X mode, a unique situation for such experiments. We mention here and will reiterate later for emphasis that the O mode power near reflection was severely reduced by the partial diversion of the heater power to the X mode, by chaotic modulation of the transmitted power as a result of the heater fire, and possibly by anomalous absorption. Heater wave power measured in situ by the rocket fluctuated by more than 50% over a 1-s time scale (note upgoing heater wave power in Figure 4). The indication is that the O mode power is likely very near the threshold predicted to excite parametric instabilities so that this experiment not only explores the early stages of ionospheric heating, but also the development of irregularities, waves, and instabilities at low heating power levels. [9] The Arecibo ISR provided incoherent scatter power profiles up to the time the heater went to full power, yielding electron density profiles for both the launch decision and for later data analysis. After the heater was brought to full power the Arecibo ISR was used to measure the intensity, spectral character, and height distribution of the upshifted plasma line. No electron density profiles are available during the heating experiment. [10] The payload instrumentation was somewhat similar to the to 1992 Soliton rocket, including measurements of electric fields, plasma density, and superthermal electrons inside the heater beam. Electric field measurements were made by two sets of sensors: a single pair of 10 cm diameter carbon-coated spheres deployed on 1.5 m booms, measuring electric fields from DC to 30 khz, and a set of six 3 cm spheres deployed on orthogonal sets of 0.15 m booms, for electric field measurements both perpendicular and parallel to B up to 60 khz. Electron and ion Langmuir probes measured the plasma density, although the Langmuir probe saturated just above the reflection height. A single search coil magnetic field detector was also flown, tuned to the 5.1 MHz heater frequency, with both power and the heterodyned signal transmitted to ground. The superthermal electron detector was swamped by photoelectrons and provided no information on the heating physics. The payload was oriented with its long axis parallel to B during the flight. 2. Data Presentation 2.1. Flight Overview [11] We begin by presenting a description of the F region plasma density profile as measured by the Arecibo ISR just prior to the launch. In Figure 1 the plasma density is shown plotted both versus altitude and also as converted to flight time of the rocket, using postflight trajectory data. The heater was switched to full power as the rocket approached 170 km, at a flight time of 101 s. The rocket reached the O mode reflection height (239 km, as indicated by the dashed lines in Figure 1) 142 s after launch. For consistency, in the remainder of this paper we will plot rocket data versus flight time, rather than absolute altitude. [12] The right-hand panel of Figure 1 shows low-frequency electric field (20 to Hz) data, labeled VLF in the

3 GELINAS ET AL.: HF HEATING EXPERIMENT SIA 7-3 Figure 1. (a) Plasma density measured by the Arecibo ISR plotted versus altitude and interpolated rocket flight time. O-mode reflection height for the 5.1 MHz heater is indicated by the dashed lines. (b) Flight summary plot of low-frequency electric field data showing fields and structure well above the O mode reflection height. plots, taken by the long-boom electric field sensors during the flight. The electric field detector signals extend more than 10 kilometers above and below the O mode reflection height. There is also a solitary structure just above the O mode reflection height, which appears only as a line in Figure 1 due to the compressed altitude scale. The presence of these structures above and below the reflection height so early in the heating cycle and at low power levels is surprising. Also surprising is the apparent absence of any spatial structures at either the upper hybrid layer, located approximately 1 km below the O mode reflection height, or at the O mode reflection height itself. Discussion of these spatial structures and higher-frequency waves will be the focus of the following sections. Note that in this flight, as in the 1992 experiment, no electric field signatures remotely resembling these were seen during the downleg portion of the flight, which occurred well outside the heated region. Indeed, in the several dozen ionospheric flights made by our group, these waveforms are unique to the heater fly-through experiments. In this context it is important to note that in plasma structures supported by heating, electric field detectors are more sensitive to electron temperature variations, if present, than the corresponding electric potential variations [Peria et al., 1999] Arecibo ISR Data [13] As stated earlier, the Arecibo 430 MHz ISR was used to monitor the upshifted plasma line during the heating experiment. The plasma line power is a result of scattering from heater-enhanced Langmuir waves, with the frequencyupshifted plasma line corresponding to downgoing Langmuir waves, and is an indication of heater power near reflection. A range-time-intensity plot of the plasma line power versus time, beginning 120 s after launch, is shown in Figure 2. The rocket passed through the O mode reflection height (H O ) at a flight time of 142 s, corresponding to 22 s on this plot. At the time of rocket passage, the plasma line power is very weak, indicating that full development of heater-enhanced Langmuir waves had not yet occurred. The rocket remained within the heater beam and below the F peak until approximately 165 s after launch, corresponding to 45 s in Figure 2. By this time the plasma line power has increased significantly, though still quite a bit weaker than that observed during previous experiments (M. P. Sulzer, Figure 2. Arecibo IS upshifted plasma line range-time-intensity plot. Range resolution is 150 m, time resolution is 1 min.

4 SIA 7-4 GELINAS ET AL.: HF HEATING EXPERIMENT Figure 3. Arecibo IS plasma line spectra for 10 altitude ranges versus frequency (khz from MHz), centered near the O mode reflection height. Spectra were taken at 165 s after launch, as the rocket passed out of the heated region. personal communication, 2001). Note that the O mode reflection height measured by the rocket corresponds to an altitude of 239 km or a range of km; thus the plasma line power seems to be roughly centered at O mode reflection. [14] Plots of the plasma line frequency upshifted from MHz for a set of 10 altitudes centered near the O mode reflection height are shown in Figure 3. These upshifted plasma line spectra shown here were taken at the time of the highest plasma line power, just as the rocket was exiting the heated region (at 45 s in Figure 2). The shortest range plotted in Figure 3 corresponds to an altitude of 238 km and the longest to an altitude of km, corresponding to rocket flight times of s and s, respectively. We give the rocket flight times here for reference only, since the rocket had actually passed out of this height region at the time these profiles were taken. The peak in the plasma line intensity occurs just near the O mode reflection height (range km), and the spectra show upshifted and downshifted peaks near 2 khz, with a broader upshifted peak near 10 khz. The 2 khz bands are probably associated with parametric generation of acoustic waves. However, the broader 10 khz peak does not appear to be associated with a cascade of acoustic wave frequencies and could be an indication that lower hybrid waves were generated near reflection; the lower hybrid frequency is approximately 6 khz. The absence of higher-order multiples of the acoustic frequency in the plasma line spectrum is also further indication that the O mode transmission was fairly weak Measured Heater Wave Power [15] In the previous section it was shown that the plasma line power was fairly weak and not fully developed at the time the rocket was passing through H O ; therefore the in situ measurements may give some idea what is happening in the early stages of ionospheric heating. The heterodyned magnetometer data, shown as a frequency spectrogram in Figure 4, gives an indication of the in situ heater wave power during the flight. The coiled loop magnetic field sensor was tuned to the heater frequency of 5.1 MHz, and the resulting signal was heterodyned to a baseline offset from zero frequency by 1790 Hz. If the rocket were stationary, both the upgoing and reflected heater waves would be observed at the same frequency, 1790 Hz. In the moving rocket frame, the two signals are Doppler shifted from 5.1 MHz (corresponding to 1790 Hz): the frequency shift for a 60 m wave at the rocket velocity of 1500 m/s is 25 Hz, exactly that observed early in the rocket flight, as shown in Figure 4. The upgoing wave is downshifted in frequency, and the downgoing wave is upshifted. [16] As the wave nears the reflection height, it slows due to the increased index of refraction of the medium and eventually is reflected. This process appears to occur twice in Figure 4: once at a flight time of approximately 137 s (231 km) and once at approximately 142 s (239 km). The latter is the O mode reflection height, according to the plasma density measured by the Arecibo ISR (see Figure 1). The former corresponds to reflection of the X mode, with a reflection height near 233 km, again according to the Arecibo IS plasma density profile. For reference, the X mode wave is reflected at an altitude where X =1 Y, where: X ¼ w2 pe w 2 Y ¼ e w w is the transmitted wave frequency and w pe and e are the local plasma and electron cyclotron frequencies, respectively. O mode reflection occurs where X = 1. Figure 4 shows that the signal is therefore a measure of the relative ð1þ ð2þ

5 GELINAS ET AL.: HF HEATING EXPERIMENT SIA 7-5 Figure 4. Frequency sonogram of heterodyned magnetic field data, showing heater wave power received at rocket, 1790 Hz corresponding to 5.1 MHz waves. Upgoing waves are downshifted in frequency and downgoing waves are upshifted in the reference frame of the rocket. heater power in each of the two modes, proof of the mixed polarization of the transmitted wave. X mode radiation is reflected near 137 s, and the remaining O-mode component of the heater wave clearly can be seen to continue up to the O-mode reflection height. An estimate of the O mode electric field strength near reflection is problematic since the signal strength is very near the lower threshold of the detectors. The threshold electric field for development of density depleted striations is 100 mv/m [Gurevich et al., 1995], so it is to be expected that only weak, early stage, or slowly developing striations would be observed in this experiment. [17] There is also evidence of some weak HF wave power above H O, near 143 s, and a still weaker signal near 144 s, both located very close to the center frequency of 1790 Hz, indicating longer wavelength (slow) waves. This 5.1 MHz wave power above the O mode reflection height likely corresponds to Z mode (or slow X mode) waves, the result of O mode wave conversion. Note that in Figure 4 this wave power is upshifted in frequency, corresponding to downgoing waves, which must have originated above the O mode reflection altitude. This O! Z conversion process is most efficient for HF wave transmission in a narrow window near the Spitze angle but may occur over a larger area if the incident O mode waves scatter are scattered from FAS [cf., Mishin et al., 2001]. Figure 5 shows ground plane projection of the rocket trajectory and timeline relative to Arecibo. The Spitze angle at Arecibo is approximately 17 degrees, and the radio window horizontal dimension, where O mode is readily converted to Z mode, is on the order of 4 km. If the observed electromagnetic waves above O mode reflection are indeed Z mode waves, it seems unlikely that transmission through the narrow radio window is their source, and scattering from FAS, if they are present, is a better candidate for Z mode production. Z mode waves would reflect where X =1+Y (at an altitude of 249 km or flight time of 147 s) and would be reabsorbed just below the O mode reflection height [e.g., Mjolhus, 1990]. In the following section we present electric field data from the regions near O mode, X mode, and Z mode reflection indicating FAS did exist and discuss further evidence that Z mode waves are present Electric Field/Electron Temperature Structures [18] We now examine in more detail the electric field detector signals introduced earlier, discussing the altitudes at which they are located and the spatial scale sizes of the structures. Figure 6 shows the frequency spectrogram of the electric field measured by DC-coupled electric field probes deployed on 1.5 m booms (3.0 m tip-to-tip). These data were obtained using the same sensors as those plotted earlier in Figure 1 and were oriented perpendicular to B. Most of the wave power here is below 100 Hz. If we make the hypothesis that these are spatial features, they correspond to structures larger than 15 m. As in the 1992 experiments, however, significant power extends to over 1 khz (less than 1 m scale length). Figure 5. Ground plane projection of rocket trajectory, with the location of the Arecibo HF facility indicated. Numbers along trajectory correspond to rocket flight times; the rocket passed O mode reflection at 142 s.

6 SIA 7-6 GELINAS ET AL.: HF HEATING EXPERIMENT Figure 6. Frequency sonogram of DC-coupled electric field data. [19] We begin by identifying regions of possible striationlike electric field signatures. According to Gurevich et al. [1995], the striations are due to growth of a thermal instability, resulting in striations that are tens of kilometers long along the field lines and with perpendicular scale sizes not larger than meters. Self-focusing of the heater wave tends to eventually create bunches of striations, as observed in the 1992 rocket data [Kelley et al., 1995]. Here, it is expected that the striations may not be strongly bunched, due to the weak heating and early-stage evolution of the striations. There are several instances of possible striations scattered throughout the data; two examples, one above and one below O mode reflection, are shown in Figure 7. These electric field structures are almost identical to those observed by the 1992 rocket [Peria et al., 1999]. We take this as evidence that FAS develop and grow along B very early in the heating cycle, which is required if Z mode waves are to be present above O mode reflection, as stated above. Langmuir probe data is shown for the event below reflection. Density variations are barely detectable at this altitude, but several depletions on the order of 2% can be seen, for example at s and s. Following Peria et al. [1999], we can estimate the temperature gradient to be a few K/m, which yields a total temperature change of a few tens of K associated with these features. These relatively reproducible electric field structures we refer to as striations. [20] Not all of the electric field structures shown in Figure 6 are similar in waveform to the well-developed striations detected in the 1992 rocket flight. Figure 8 shows electric field data from three regions, both below and above O mode reflection, which exhibit less regular electric field structures than those associated with striations, as well as a more pronounced high-frequency/short scale component. Two of these events have density information and are plotted below the electric field data. Well-defined density depletions are seen, particularly at s and s. Unlike the striations observed on the 1992 rocket and data presented here in Figure 7, the structures shown in Figure 8 show evidence of high-frequency/small-scale electric fields. The long-boom electric field sensor pair, with a separation distance of 3.0 m, is not well-suited for investigating these smaller-scale structures. Instead, we examine these chaotic electric field signatures with the short-boom electric field sensors, with a 0.3 m sensor separation. [21] Electric field data taken by the short-boom sensor pair (parallel to the long-boom pair discussed above) is shown in Figure 9. The narrow band of electric field signals descending from 50 Hz are associated with what we Figure 7. Regions approximately 8 km below O mode reflection (top four panels) and 8 km above reflection (bottom panel) showing 10 m scale electric field/electron temperature structures characteristic of striations. Density data are shown only for the lower altitude case (top panels) since the Langmuir probe saturated just above reflection.

7 GELINAS ET AL.: HF HEATING EXPERIMENT SIA 7-7 source zones the waveforms are smoother and appear to look more like striations, though rather weak ones. Surprisingly, and as discussed next, the strongest electric field structures appear to be located where the local plasma frequency equals a multiple of the electron cyclotron frequency, both above and below the O mode reflection height. The higher-frequency components of the spectrum are consistent with spatial dimensions on the order of a few meters, approximately an ion gyroradius. Since the rocket velocity is comparable to the ion sound speed, ion acoustic waves in the few meter wavelength range are also a candidate for explaining these fluctuations. [23] We now consider the spatial (altitude) locations of the electric field structures discussed above. For reference in identifying the structures, we tentatively identify the altitude regions with various 5.1 MHz wave cutoffs and resonances, shown plotted versus rocket flight time in Figure 10. Local plasma frequency at the rocket location is calculated using the Arecibo ISR electron density profile. The intense structure at 138 s seems to be associated with the fifth electron cyclotron harmonic, a possible result of interaction with or damping of an electron Bernstein mode generated in the RF interaction process. Damping is expected to be particularly strong in the region of the so-called double plasma resonance, that is, w pe n e. The other two areas of intense electric field signals and possible heating, at 133 s and 155 s, appear very close to the fourth and sixth electron cyclotron harmonics, respectively. As expected, the effect of the former is much weaker than at the fifth harmonic, which is close to the heater frequency. The sixth harmonic region is well above the O and Z modes reflection heights. Figure 8. Waveforms for electric field structures and density data near 133 s, 138 s, and 155 s, apparently quite different from striations. Density data are shown for the first two cases, below O mode reflection. The Langmuir probe went off-scale just above the reflection height. interpret to be artificial periodic irregularities (APIs) generated by the heater. Note that 50 Hz in the moving rocket frame corresponds to 30 m spatial scale, half the heater wavelength. We postpone analysis of these structures for future work. [22] The short boom pairs are also more sensitive to changes in the floating potential ( 50kT e ) variations than to electric fields [Peria et al., 1999], so Figure 9 may represent structure in the electron temperature due electron heating. In looking at Figure 9, there appear to be three specific regions of strong signals, at 133 s, 138 s, and 155 s, the same regions as the waveforms shown above in Figure 8. In general, the waveforms in regions with strong signals and higher-frequency electric field components (above 100 Hz), appear to be in a strong interaction region, possibly associated with heating. Outside these 2.5. Solitary Structure [24] Looking at Figure 6, it is curious that there is very little electric field signal between the heating structure near the fifth cyclotron harmonic (138 s) and the Z mode reflection height (147 s) except for a solitary structure. This structure was seen at a flight time of s, almost 2 km above the O mode reflection height, and in the region where there is evidence of Z mode waves, as discussed above. Parallel and perpendicular electric fields measured by the short boom pairs, as well as the long boom pair are shown in Figure 11, along with the Langmuir probe data. [25] The scale size of the solitary structure is approximately 30 m, half the heater wavelength. There are also higher-frequency components inside the solitary structure, which correspond to a few meters if they are spatial features. At an ion temperature of 0.03 ev early in the heating cycle, the ion gyroradius for O + ions is 2 m; for NO + and O 2 + the gyroradius is 2.8 m. The plasma composition at this altitude is approximately 60% O +. These highfrequency structures are detected more clearly with the the short-boom pairs than with the longer booms. This indicates that the features are spatial and have scales which are on the order of or less than the long boom probe separation. The waveforms from the short-boom pairs appear to be consistent with a m structure with ion-gyroradius fine-scale structure. These few meter scale sizes also appeared in the electron cyclotron harmonic related structures described above and may indicate that heating is also taking place inside this solitary structure. This structure may be a developing striation connected to a classic upper hybrid

8 SIA 7-8 GELINAS ET AL.: HF HEATING EXPERIMENT Figure 9. Frequency sonogram of electric field data taken by short boom pair, which is less sensitive than the long-boom pair of Figure 6. Data is high-pass filtered at 27 Hz. interaction region below it by magnetic field lines or a ponderomotive-type structure associated with the Z-mode waves observed above O mode reflection. 3. Discussion [26] A detailed discussion of these data in the context of ionospheric heating is beyond the scope of this paper, the goal of which is to provide motivation and direction for future work on electron harmonic heating and SEE theory. In this section we summarize the main features (as we see them) of this data set and mention a few possible links between these data and current theories of ionospheric heating. [27] The conventional approach to interpret plasma waves generated by the injection of intense ordinary HF radio waves is based on the paradigm that there are two chief generation regions, that is, layers near upper hybrid and the reflection heights, H UH and H O [e.g., Stubbe and Hagfors, 1997]. A parametric decay of the HF wave into ion-acoustic (S) and Langmuir (L) waves with weak turbulence cascade spectra is the dominating feature at the matching height, while the continuum and free mode of strong turbulence are characteristic of the critical layer (H O ). Interactions where the heater frequency equals the upper hybrid frequency and possibly where the plasma frequency matches the electron cyclotron harmonics, dominate at h H UH, also leading to striations. [28] Unfortunately, the electromagnetic radiation in these regions appears to have been too weak for the instruments to observe any residual or secondary interactions, at least at the location where the rocket penetrated the generation regions. The upper hybrid layer and O mode reflection height are separated by only 2 km, and changes in heater 5.1 MHz magnetic field strength in Figure 4 are not easily resolved on this spatial scale. The 5.1 MHz electric field power was also below the detector thresholds, so secondary interactions recorded by the electric field long-boom and short-boom pairs would be the best indicator of activity in these areas. Activity in both the upper hybrid layer and the O mode reflection layer is extremely weak in both sets of electric field detectors, though the electric field power in the upper hybrid layer does appear to be slightly suppressed. This suppression may be consistent with the hypothesis that Langmuir turbulence in the O mode reflection layer is the dominant heating mechanism at low latitudes [Leyser et al., 2000], but the data is far from conclusive on this point. [29] However, this experiment does point to the existence of Z mode waves above O mode reflection, electric field structures consistent with FAS and regions of electron heating near multiples of the electron cyclotron frequency. One Figure 10. Locations of 5.1 MHz wave cutoffs and resonances plotted with respect to the local plasma frequency, as determined by the Arecibo ISR, fitted to the time-altitude profile of the rocket. Labels O, X, Z, U, 4, 5, and 6 show the location of the O, X, Z mode reflection heights, the upper hybrid layer, and the fourth, fifth, and sixth electron cyclotron harmonics, respectively. Earth magnetic field strength near 240 km is taken to be 35,000 nt.

9 GELINAS ET AL.: HF HEATING EXPERIMENT SIA 7-9 Figure 11. Perpendicular (top two panels) and parallel (third panel) electric fields of solitary structure measured by the short boom pairs. Perpendicular electric fields measured by DC-coupled long boom pair are plotted in fourth panel; the partially clipped wave is an instrumental effect. The fifth panel shows the density depletion associated with the solitary structure. phenomenon, the outshifted plasma line (HFOL), observed during ground-based experiments, may be more directly related to Z mode waves, as it occurs above the O mode reflection height [Isham et al., 1996]. Mishin et al. [1997] and Kuo et al. [1998] have shown that Z mode may decay into electron Bernstein (EB) and lower hybrid (LH) waves, the latter tending to collapse, with subsequent acceleration of electrons. The accelerated electrons enhance Langmuir waves, the source of the HFOL. The O! EB + LH decay is also thought to contribute to the SEE when the heater frequency is near a harmonic of the electron gyrofrequency [Istomin and Leyser, 1995; Tripathi and Liu, 1993]. Thus one may assume the HFOL and at least some part of the SEE to be the outcomes of the process of generation of Langmuir waves by electrons accelerated in the course of the O/Z! EB + LH decay and which propagate below and above the critical layer. Intense plasma turbulence may also be generated by the Z! L transformation [e.g., Mjolhus and Fla, 1984], independent of how far the heater frequency is away from a gyroharmonic. [30] In this section we discuss some of the mechanics of the O to Z conversion and further decay to electron Bernstein waves, lower hybrid waves, or Langmuir waves. The experiment data indicates that Z mode waves, FAS, and EB waves may be present, so further discussion of these processes seems warranted Conversion Into the Z Mode [31] The classical process of conversion of the O wave into Z mode takes place pat ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi the so-called Spitze angle, q c = q ± = ± arcsin (sin c Y=1 þ Y) [e.g., Ginzburg, 1970; Mjolhus, 1990]. Theoretically, at q = q c ( 17 for Arecibo), the conversion efficiency reaches 100%. The angular width of the conversion window is of the order of 1 [Mjolhus, 1990]. Evidently, in this specific angular range the Airy pattern does not occur at H O, as the incident energy flux is converted into the outgoing Z mode. Note that ± signs correspond, respectively, to the southward and northward injections in the Northern Hemisphere. [32] For the southward process, after reflection at X Z =1+Y, the Z mode reaches the height of transformation slightly below H O. The horizontal displacement to south of the Z! L transformation region is expected to be as large as 25 km [cf., Mjolhus and Fla, 1984]. On the other hand, for the northward injection, Z! L transformation occurs almost at once after the O! Z conversion [Mjolhus, 1990]. Thus in contrast to the southward process, the two regions practically coincide. As the rocket trajectory was to the north of Arecibo, the classical O! Z conversion does not seem to contribute to perturbations above H O, particularly near H Z (around 148 s). [33] On the other hand, Z mode waves can be generated between H UH and H O due to leakage of trapped UH waves [Gurevich, 1999] and due to resonant scattering of the heating (O mode) wave on small-scale field-aligned density irregularities (FAIs) [Mityakov et al., 1975] O þ dn q! Z This process is believed to be a plausible cause of the socalled anomalous absorption [e.g., Vas kov and Karashtin, 1980] of the SEE with frequencies above w 0 by khz [Frolov, 1990] and of topside echoes observed in a wide angular range [Mishin et al., 2001]. The necessary FAIs can be generated due to either the resonant instability of the heating wave [Vas kov and Gurevich, 1977] or a second-order four wave interaction [Huang and Kuo, 1994]. [34] Ray paths of the generated Z mode waves are computed using the Power Tracing Code, operated by the Space Weather Center of Excellence in the Air Force Research Laboratory s Space Vehicles Directorate. The code solves the Haselgrove [1963] form of the canonical ray tracing equations in a three-dimensional magnetized cold plasma using a stiff ordinary differential equations solver. Details of the code are given by Starks [2002], and we note that only ray paths are used in this work rather than ð3þ

10 SIA 7-10 GELINAS ET AL.: HF HEATING EXPERIMENT and is converted into an electrostatic wave just below H O [e.g., Mjolhus, 1990]. Near the upper hybrid layer, H UH,the electric field becomes parallel to k, and the wave can couple to the electron Bernstein mode (EB). The electrostatic EB waves propagate, but are easily absorbed at the nearby fifth electron cyclotron resonance and likely also at the fourth and sixth harmonics, resulting in electron heating. Our detection of enhanced electric fields at these locations seems to argue strongly that electron Bernstein modes were generated. This type of heating process has been observed in laboratory experiments, there referred to as O-X-B conversion processes, and result in electron cyclotron heating [Laqua et al., 1997]. The conversion process from Z mode to EB waves generates decay waves via parametric instabilities in the upper hybrid layer, with the resulting waves upshifted and downshifted by multiples of the lower hybrid frequency, along with lower hybrid waves, such that Figure 12. Results of ray tracing of Z mode injected in the magnetic meridian plane from the origin (0, H O ) with different angles off vertical a = bk z. 1,2,... designate rays with initial angles off vertical that are sequentially offset by 5 from south to north (10 corresponds to almost vertical injection). full power flux estimates. The density profile was chosen to be the Chapman layer fitted to the actual profile (Figure 4). [35] Figure 12 shows ray trajectories of Z mode waves injected in the magnetic meridian plane from the origin (0, H O ) with different angles off vertical a = bk z. As expected, after reflection all rays end up 600m below H O. Apparently, the region occupied by Z mode waves is rather wide with a significant portion to the north of the heater. This may explain the observations near H Z. Furthermore, we have calculated the frequency shift for the distribution of Z mode waves below H O. Depending on the injection angle, it turns out to be within the range of 0 to 70 Hz, which does not disagree with Figure O/Z! EB + LH [36] In the case of a Z mode conversion process, the Z mode wave is reflected from its cutoff altitude downwards w 0 ¼ w m w lh k 0 ¼ k m k lh where m = 1, 2, etc. [37] There is some evidence for the existence of waves near the lower hybrid frequency in the rocket data, as shown in Figure 13. The rocket passed through the upper hybrid layer at approximately 141 s and the O mode reflection height at 142 s. The lower hybrid frequency at this altitude is close to 6 khz. There appears to be some wave power near the upper hybrid layer that could be associated with multiples of the lower hybrid frequency, but the data is not conclusive on this point. [38] Another possible interpretation of the signals at 5f ce (at 230 km) and 6f ce (260 km) is in terms of the so called ballistic transformation [e.g., Vodyanitskij et al., 1974]. This process has been suggested as an explanation of the topside enhancements over the midlatitude heater facilities [Vas kov et al., 1995]. For thermal electrons, this seems to be unlikely. Indeed, the memory of the plasma oscillations at the bottomside carried by resonant electrons to the ð4þ ð5þ Figure 13. High-frequency electric field spectrum measured by long booms. Lower hybrid frequency is approximately 5 khz. Upper hybrid layer is near 141 s, and the O mode reflection is near 142 s.

11 GELINAS ET AL.: HF HEATING EXPERIMENT SIA 7-11 topside must be lost when the electrons cover a distance corresponding to a large number of mean free paths. However, the Langmuir Turbulence rocket was launched in twilight when photoelectrons are still present near the reflection height. As the mean free path of a few 100 ev electrons is of order of a few 10 km, these electrons can preserve the phase coherence required. Therefore fast photoelectrons modified by the fifth harmonic (v k 5f ce /k k ) can propagate upwards after the wave itself is ceased. When they reach the region of the double plasma resonance again, i.e. f pe =6f ce, the wave is regenerated Z! L Transformation [39] As a result of the transformation into the electrostatic mode, the amplitude of Langmuir wave, E l, obtained from the usual WKB argument of the energy flux conservation, is (after Mjolhus and Flaa [1984]) E l ab 2 1=6 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4pS Z =c Here a 2Y 5/2 sin 2 c 10 2, b 3 T e /mc , T e 1500 K is the electron temperature, and S Z is the Poynting flux of the Z mode. Note that the wave vector points vertically downward and so does the field E l. [40] Given the efficiency of the O! Z conversion 1%, the energy flux density above H O can be estimated as S Z 0.03P/R 2 mw/m 2 (P is the effective radiative power in MW, R is the distance to the layer in kilometers). pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi The free space p field of the injected wave is E O = 4pS O =c 4.7 ffiffiffi P /R 0.2 V/m for P 100 and R = 250. From equation (6) the value of E l is expected to be by a factor of 10 larger than E O, which is comparable to to the swelling factor in the Airy pattern of the heating wave. Therefore all well-known manifestations of the strong turbulent regime of Langmuir turbulence in the reflection layer of the heating wave [e.g., DuBois et al., 1990] may be expected in the Z! L transformation region as well. There is also an additional possibility of creating wave perturbations due to the asymmetric acceleration of suprathermal electrons driven by the electrostatic wave structure [e.g., Maggs and Morales, 1990]. 4. Summary [41] In summary: [42] 1. The experiment conditions were mixed O and X mode heating, resulting in a severely weakened O mode wave at the reflection height. The in situ measurements were made early in the heating cycle. The result is a study of early-onset, at-or-just-above-threshold ionospheric heating. [43] 2. Evidence for structures like striations were found above and below reflection altitude, even though heating had occurred for only a few tens of seconds. [44] 3. Strong electric field/electron temperature structures at multiples of the electron cyclotron harmonics were observed, with particularly intense waves near the fifth harmonic. The structures could be the result of plasma heating, possibly due to damping of Bernstein modes, or a nonlinear interaction of Bernstein modes leading to low frequency electrostatic waves. Unlike waveforms detected far from resonant altitudes, these harmonic related features had colocated small-scale (meter) oscillations. ð6þ [45] 4. The 5.1 MHz heterodyned magnetic field data indicate that the O mode wave undergoes at least partial conversion to Z mode waves, which can propagate above reflection. [46] 5. A solitary structure was observed 2 km above the O mode reflection height, consistent with a Z mode source, with fine-structure on the order of the ion gyroradius, possibly due to plasma heating inside the structure. [47] 6. Ray tracing results show that oblique Z mode waves generated in the region between upper hybrid and O mode reflection heights can propagate to the north of the heater and reach the rocket above O mode reflection. [48] Acknowledgments. The work at Cornell was supported by ONR under grant N EVM was supported in part by the HF Active Auroral Research Program (HAARP) under AFRL contract F C-0012 with Boston College. [49] Arthur Richmond thanks A. V. Gurevich and Terence Ronald Robinson for their assistance in evaluating this paper. References Arce, T. L., and M. C. Kelley, Density depletions at the scale of the ion gyro radius in the Arecibo heater volume, Radiophys. Quantum Electron., 37, , DuBois, D. F., H. A. Rose, and D. Russell, Excitation of strong Langmuir turbulence in plasmas near critical density: Application to HF heating of the ionosphere, J. Geophys. Res., 95, 21,221 21,272, Farley, D. T., C. LaHoz, and B. G. Fejer, Studies of the self-focusing instability at Arecibo, J. Geophys. Res., 88, , Franz, T. L., M. C. Kelley, and A. V. Gurevich, Radar backscattering from artificial field-aligned irregularities, Radio Sci., 34, , Frolov, V. L., A new artificial ionospheric radio emission component, Geomagn. Aeron., 30, , Frolov, V. L., et al., Study of large-scale irregularities generated in the ionospheric F-region by high-power HF waves, Radiophys. Quantum Electron., 43, , 2000a. Frolov, V. L., E. N. Ermakova, L. M. Kagan, G. P. Komrakov, E. N. Sergeev, and P. Stubbe, Features of the broad upshifted structure in stimulated electromagnetic emission spectra, J. Geophys. Res., 105, 20,919 21,933, 2000b. Ginzburg, V., The Propagation of Electromagnetic Waves in Plasmas, 2nd ed., Pergamon, New York, Gurevich, A. V., Nonlinear Phenomena in the Ionosphere, Springer-Verlag, New York, Gurevich, A. V., Modern problems of ionospheric modification, Radiophys. Quantum Electron., 42, , Gurevich, A. V., K. P. Zybin, and A. V. Lukyanov, Stationary striations developed in the ionospheric modification, Phys. Rev. Lett., 75, , Haselgrove, J., The Hamilton ray path equations, J. Atmos. Terr. Phys., 25, , Huang, J., and S. P. Kuo, A theoretical model for the broad upshifted maximum in the stimulated electromagnetic emission spectrum, J. Geophys. Res., 99, 19,569 19,576, Isham,B.,C.LaHoz,H.Kohl,T.Hagfors,T.B.Leyser,andM.T. Rietveld, Recent EISCAT heating results using chirped ISR, J. Atmos. Terr. Phys., 58, , Istomin, Y. N., and T. B. Leyser, Small-scale magnetic field-aligned density irregularities excited by a powerful electromagnetic wave, Phys. Plasmas, 4, , Kelley, M. C., T. L. Arce, J. Salowey, M. Sulzer, W. T. Armstrong, M. Carter, and L. Duncan, Density depletions at the 10-m scale induced by the Arecibo heater, J. Geophys. Res., 100, 17,367 17,376, Kuo, S. P., E. Koretzky, and M. C. Lee, Parametric excitation of lower hybrid waves by Z-mode waves near electron cyclotron harmonics at Tromso, J. Geophys. Res., 103, 23,373 23,379, Laqua, H. P., V. Erckmann, H. J. Hartfuss, and H. Laqua, Resonant and nonresonant electron cyclotron heating at densities above the plasma cutoff by O-X-B mode conversion at the W7-AS stellarator, Phys. Rev. Lett., 78, , Leyser, T. B., Stimulated electromagnetic emissions by high-frequency electromagnetic pumping of the ionospheric plasma, Space Sci. Rev., 98, , Maggs, J. E., and G. J. Morales, Nonlinear dynamics of electrons accelerated by resonant fields in nonuniform plasmas, Phys. Fluids B, 2, , 1990.

12 SIA 7-12 GELINAS ET AL.: HF HEATING EXPERIMENT Mishin, E., T. Hagfors, and W. Kofman, On origin of outshifted plasma lines during HF modification experiments, J. Geophys. Res., 102, 27,265 27,269, Mishin, E., T. Hagfors, and B. Isham, A generation mechanism for topside enhanced incoherent backscatter during high frequency modification experiments in Tromsoe, Geophys. Res. Lett., 28, , Mityakov, N. A., V. O. Rapoport, and V. Y. Trakhtengerts, Scattering of an ordinary wave near the reflection point by small-scale irregularities, Radiophys. Quantum Electron., 18, , Mjolhus, E., On linear conversion in a magnetized plasma, Radio Sci., 25, , Mjolhus, E., and T. Fla, Direct access to plasma resonance in ionospheric radio experiments, J. Geophys. Res., 89, , Peria, W. J., and M. C. Kelley, Convection electric field observations near the Arecibo HF heater beam, J. Geophys. Res., 106, 18,517 18,524, Peria, W. J., M. C. Kelley, and T. Franz, Double-probe measurements in field-aligned irregularities produced by intense electromagnetic radiation, J. Geophys. Res., 104, , Robinson, T. R., The heating of the high latitude ionosphere by high power radio waves, Phys. Rep., 179, , Rose, G., B. Grandal, E. Neske, W. Ott, K. Spenner, J. Holtet, K. Maseide, and J. Troim, Experimental results from the HERO project: In situ measurements of ionospheric modification using sounding rockets, J. Geophys. Res., 90, , Starks, M. J., Effects of HF heater-produced ionospheric depletions on the ducting of VLF transmissions: A ray-tracing study, J. Geophys. Res., 108(A11), 1336, doi: /2001ja009197, Stubbe, P., and T. Hagfors, The Earth s ionosphere: A wall-less plasma laboratory, Surv. Geophys., 18, , Sulzer, M. P., and J. A. Fejer, Radar spectral observations of HF-induced ionospheric Langmuir turbulence with improved range and time resolution, J. Geophys. Res., 99, 15,035 15,050, Tripathi, V. K., and C. S. Liu, O mode decay and upshifted electromagnetic emissions near cyclotron harmonics in the ionosphere, J. Geophys. Res., 98, , Vas kov, V. V., and A. V. Gurevich, Resonance instability of small-scale plasma perturbations, Sov. Phys. JETP, Engl. Transl., 73, , Vas kov, V. V., and A. N. Karashtin, Resonance absorption of radio waves at frequencies close to the gyrofrequency of electrons, Geomagn. Aeron., Engl. Transl., 20, , Vas kov, V. V., N. I. Budko, O. V. Kapustina, Yu. M. Mikhailov, N. A. Ryabova, G. P. Komrakov, A. N. Maresov, and G. L. Gdalevich, Appearance of VLF and ELF-noises in topside ionosphere under the action of high power radio wave from data of satellite Intercosmos-24, Adv. Space Res., 15, 49 56, Vodyanitskij, A. A., N. S. Erokhin, V. V. Lisitchenko, S. S. Moiseev, and V. N. Oraevskij, Transillumination of the wave barriers in a plasma as a result of kinetic effects, Nucl. Fusion, 14, , L. J. Gelinas, Laboratory for Plasma Studies, Cornell University, Ithaca, NY 14853, USA. (lynett@ece.cornell.edu) M. C. Kelley, Department of Electrical and Computer Engineering, Cornell University, Ithaca, NY 14853, USA. (mikek@ece.cornell.edu) E. Mishin, Boston College, Institute for Scientific Research, 402 St. Clements Hall, Chestnut Hill, MA 02467, USA. (evgenii.mishin@ hanscom.af.mil) M. J. Starks, Air Force Research Laboratory, Space Weather Center of Excellence, AFRL/VSBXI, Hanscom AFB, MA, USA. (michael.starks@ hanscom.af.mil) M. P. Sulzer, Arecibo Observatory, HC 3 Box 53995, Arecibo, PR 00612, Puerto Rico. (msulzer@naic.edu)