Implications of Meteor Observations by the MU Radar

Implications of Meteor Observations by the MU Radar Qihou H. Zhou Arecibo Observatory, National Astronomy and Ionosphere Center, Arecibo, Puerto Rico John D. Mathews Communications and Space Sciences Laboratory, Pennsylvania State University, State College Takuji Nakamura Radio Atmospheric Science Center, Kyoto University, Kyoto, Japan 1

Abstract. We report high resolution meteor echo observations using the Kyoto University Middle and Upper (MU) Atmosphere 46.5 MHz Radar. When the MU radar was pointed perpendicular to the geomagnetic field lines (B), numerous long-lived range spread trail echoes were observed which were largely absent when the beam was pointed in the vertical and parallel-to-b directions. This shows that this type of trail echo is largely due to scattering structures aligned along B. Additionally, nearly all the head echoes displaying an along-the-beam velocity component were followed by range spread echoes in the perpendicular-to-b pointing geometry. This demonstrates that meteoric field aligned irregularity is present in essentially all meteors up to the detection limit of the MU radar. Practically all the spectra are limited within a bandwidth corresponding to a Doppler shift of 320 m/s, suggesting that the two stream instability is absent most of the time. Meteoric field aligned structures can be a potential error source for aeronomical applications if they are not appropriately considered. 1. Introduction Meteor observations using narrow beam radars in recent years have added to the fascination and complexity of meteoric phenomena. Results prior to 1997 using several incoherent scatter facilities were summarized by Zhou et al. [1998]. More recently, Raghava Reddi and Nair [1998] studied meteor trails using the India 53 MHz MST radar and interpreted the range spread echoes to be due to plasma striations parallel to the geomagnetic field. Pellinen- Wannberg et al. [1998] reported observations using the EISCAT incoherent scatter radars in conjunction with optical methods. Chang et al. [1999], using a narrow-beam meteor radar, reported observations of large apparent vertical velocities near geomagnetic equator, which they believe to be most likely due to plasma irregularity present in meteor trails. Janches et al. [2000a, b] have taken advantages of the narrow Arecibo 430 MHz incoherent scatter radar beam width to yield highly accurate deceleration measurements and meteoroid orbits. Close et al. [2000] reported polarization observations of meteor head-echoes using the Kwajalein Atoll ALTAIR VHF radar. The VHF echo characteristics presented by Zhou et al. [1998] using the Arecibo 300 m dish appear to have different echo characteristics from the traditionally wide beam VHF meteor radars. However, the Arecibo radar pointing is limited to within a 20 o zenith angle and it is possible that some of the peculiar characteristics may be pointing related. Here we report observations using the middle and upper (MU) atmosphere radar located at Shigaraki, Japan. The MU radar is similar in sensitivity to the Arecibo VHF radar but can be steered to 35 o in zenith angle without grating lobes and to 60 o if grating lobes in different azimuth angles are permitted. Our present study focuses on the role that geomagnetic field plays on range spread trail echoes. Previous uses of the MU radar to measure aeronomical parameters from meteor trails are summarized in Nakamura et al. [1997]. Recent researches on aeronomical applications using the decay characteristics of underdense echoes can be found in Cervera and Reid [2000] and references therein. 2. Observations The MU radar is located at Shigaraki, Japan (geographic 34.9 o N and 136.1 o E; geomagnetic dip: 51 o ). It has an aperture diameter of 103 m and operates at 46.5 MHz, yielding a 3-db full beam width of 3.6 o. The transmitter has a peak power of 1 MW. In 2

our experiments, we transmitted a square pulse of 2 µs duration every 3 ms using the full array. On the night of Sept. 5, 1998, observations was made with the beam pointed vertically. On the night of Sept. 7, 1998, the pointing (k) was alternatively in the directions parallel (//) and perpendicular ( ) to B about every half hour. For both nights, data were taken between 00:00 and 08:30 LT. Echoes in the altitude range of ~80-100 km were collected for every pulse at a resolution of 300 m. The echo characteristics were very different in the three pointing directions. Most of the echoes in the zenith pointing (ZP) observation were similar to the Arecibo VHF observation reported by Zhou et al. [1998]. A majority of the echoes were slant head echoes (displaying a range rate) and a few were either constant range echoes (CRE), or range spread trail echoes (RSTE), i.e., trail echoes appearing over more than two range bins simultaneously. The head echo rate observed by the MU radar was a few times higher than the Arecibo VHF result as expected because of the wider beam width and higher transmitter power of the MU radar. Eight RSTEs were observed between 07:20-08:30 LT while typically only one such echo was observed per hour prior to 07:00 LT. Since our observation stopped at 8:30 LT, we are not certain whether the background ionosphere makes a difference for the hourly rate of RSTEs. In the ZP geometry, RSTEs typically lasted several hundred milliseconds but one such echo lasted about 3 s. Most of the echoes in the k//b scenario were CREs and slant head echoes were often observed as well. A majority of the CREs were short-lived (lasting less than 100 ms) and did not exhibit an exponential decay in echo power while the long lasting CREs typically showed such a behavior. It is likely that the short-lived CREs are associated with head-echoes and the long-lived CREs are the trail echoes often observed by traditional meteor radars although this cannot be ascertained from our present observations. A few RSTEs were also observed in this pointing direction. In the k B pointing scenario, many trail echoes were longlasting RSTEs and there were also a significant number of CREs. In the dawn hours, over 100 RSTEs per hour were observed. Although RSTEs typically lasted several seconds, some of them lasted over several minutes. RSTEs occurred over all our sampling altitude range of 84 km to 100 km. For the entire duration of our observation, essentially all head echoes had an accompanying RSTE. There was often a few hundred ms delay between the headecho and the ensuing RSTE. In Figure 1, we show RSTE examples observed when the radar was pointed in the three pointing directions: k B, k//b and ZP, respectively. Figure 2 shows the power variation of two CREs. In the following, we discuss the most salient features of RSTEs. Detailed statistics of the echoes will be reported elsewhere. Figure 1 Figure 2 3. Implications Range spread trail echoes (RSTE), as shown in Fig 1, are variously called unusual or long enduring echoes [McKinley, 1961]. They were observed at least as early as 1950 s. This type of echo was believed to be overdense, i.e., the plasma frequency of the meteor trail exceeding the probing radar frequency. A head echo was often seen to precede the RSTE and there was typically a delay time (sometimes up to several seconds) between the head echo and the ensuing RSTE. Those unusual echoes have been explained via the so called glint theory and the blob theory [McKinley, 1961]. The former assumes that wind-shears distort the meteor trail such that parts of the trail become perpendicular to the radar beam. In the latter theory, fragmented meteors form numer- 3

ous ionization centers which then expand and part of the surface of the expanded plasma eventually becomes specular reflectors. Heritage et al. [1962] appear to be the first to report meteor echo-associated field aligned irregularities using a bistatic system. Haldoupis and Schlegel [1993], observing with a 50.5 MHz CW bistatic system, reported Doppler velocity at the ion-acoustic speed, which made them suggest the presence of Farley-Buneman, or two stream, instabilities in the meteor trails. Those bistatic observations, however, could not determine whether the echoes were spread in range and thus neither Heritage et al. nor Haldoupis and Schlegel associated their observations with previously observed unusual echoes. Chapin and Kudeki [1994], using the Jicamarca VHF radar, reported observations of RSTEs over the geomagnetic equator. Their echoes were in many aspects similar to those unusual echoes referred to in McKinley. However, they concluded that the electrojet is essential for the unusualness of their observations because of Jicamarca s equatorial location and the k B geometry. Subsequently, Raghava Reddi and Nair [1998] reported observations of RSTEs outside equatorial electrojet region with the radar pointing ~10 degrees away from the B direction. Although they could not observe in the k B or the k//b direction, they interpreted plasma striations along the field line to be responsible for the echo power. Recently, Oppenheim et al. [2000] have modeled the plasma instability process for equatorial region and found ExB drift velocity plays an essential role in all stages of meteor trail development. By showing that RSTEs are most easily observed in the k B direction and largely absent in other directions, our observations yield unequivocal evidence that RSTEs are associated with field aligned irregularities (FAI). More importantly, our observations show practically all the head echoes having a non-zero line-ofsight velocity were followed by a RSTE in the k B geometry. This implies that, excluding the relatively few meteors entering the atmosphere in the B direction, all other meteors, up to the detection limit of the MU radar, deposit trails that evolve into FAI structure. The effective scattering cross-section of the MU echoes is expected to be the same as that of the Arecibo VHF radar, at ~10-3 m 2, because of the similar sensitivity of the two radars [Zhou et al., 1998]. Even those meteors that appear to be classical specular echoes, i.e., having a near zero radial velocity, often do not exhibit the classical exponential decay characteristics, as shown in Fig. 2. RSTEs are essentially not seen in the k//b direction for (nonclassical) meteors with a radial velocity component. This demonstrates that the scattering mechanism requires the development of FAI. While the initial ionization trail produced by a meteoroid has some roughness to it, such a roughness by itself is generally not sufficient to produce observable echoes even when there is no diffusion. This is because if the uneven ablation were the main mechanism to produce the necessary irregularity, the time gap between the head-echo and the ensuing RSTE should not exist. Thus the initial ionization distribution along the meteor trail must be reorganized and it is reorganized in such a way as to enhance FAI, i.e., plasma instability has to occur. The smaller-scale irregularity structure within the FAI that produces radar echoes can be either at the half-wavelength scale (~3 m) or at a very large scale (e.g.,a few scattering centers within the 300 m range resolution). If the latter is more applicable, one should expect a rather large Doppler spread in the RSTEs since it takes about a couple of hundred ms to structure the electrons to be radar visible. Assuming an average distance of 40 m between the scattering centers, electrons need to move at about 200 m/s in order for the large-scale irregularity hypothesis to be viable. While the Farley-Buneman instability, which has been 4

suggested by Haldoupis and Schlegel [1993] as well as by Chapin and Kudeki [1994] for their observations, can produce electron-ion drift velocity exceeding 300 m/s, it does not appear to be the dominant mechanism in our observation, which we discuss in the following. In regard to the delay time between the headecho and the accompanying RSTE, we interpret it as the time needed for plasma instabilities to grow strong enough to be radar visible, instead of the time for windshear or diffusion to make the ionization favorable for specular reflection to occur as hypothesized in glint and blob theories. Because we took raw voltage samples in our observations, we can find the power spectra as well. Figure 3 shows the power spectra at different heights from a RSTE. We note that the spectral power is largely confined within a Doppler shift of 327 m/s, the ion-acoustic speed, assuming an ion/electron temperature of 200 K and ion mass of 31 amu. This is true for practically all the RSTE spectra. Although we typically used 256 data points (768 ms) to form a spectrum, we have also used 64 data points FFT to examine a large number of trails and did not find any significant amount of power spreading beyond 320 m/s Doppler shift. Thus we conclude that Farley-Buneman instability is not important on a time scale of ~200 ms unless the ion-acoustic speed in a meteor trail is much smaller than that in the ambient ionosphere. We further note that many RSTE spectra are highly skewed, a feature also observed by Chapin and Kudeki [1994]. In order to see a RSTE, each scattering region within the FAI must be largely confined within one Fresnel zone, i.e., within two radar wavefronts separated by a quarter-wavelength or about 1.4 km along B at 100 km altitude for the MU radar wavelength and geometry. If we assume that ambipolar diffusion dominates along B, it takes about 10 4 seconds for a meteor trail to diffuse into an entire Fresnel zone at 100 km assuming a diffusion coefficient of 10 m 2 /s. In the k B scenario, the lifetime of a RSTE likely depends more on recombination than on diffusion. However, as k is further off the B direction, diffusion along the field line becomes increasingly important in diminishing the echo power. Clearly, the process that results in the FAI structure is electrodynamic in origin. However, the motion along B may not be the standard ambipolar diffusion because of the existence of plasma instability. Nevertheless, the exponential decay of the long lifetime CRE powers in the //B geometry suggests that diffusion dominates at the largest scale although the diffusion rate can be affected by the FAI generating mechanism. One may ask why the FAI nature of a meteor trail has not been noticed from observations using conventional meteor radars if it is so prevalent in our observation. This is because: (1) the echo power from FAI is much smaller than the classical echo power, which is obtained when the meteor trail is perpendicular to the radar pointing; and (2) the existence of FAI does not affect the echo power significantly in the perpendicular-to-trail geometry. In the perpendicular-to-trail geometry, most of the electrons in the entire trail scatter coherently since most of them are in one Fresnel zone. A RSTE, on the other hand, derives its power from the density fluctuation in the FAI. RSTE power matches the classical specular echo power only when all the electrons in the FAI scatter coherently, which obviously seldom happens. Since conventional meteor radars typically have a much lower gain than the MU or the Arecibo radar, they cannot observe RSTEs very easily. Further, since FAI is largely confined within the trail, it does not affect the echo power very much in the perpendicular-to-trail geometry. Although FAI does not change the total scattering power in classical specular echoes significantly, it may still have an undes- Figure 3 5

ired effect on extractions of aeronomy parameters, i.e., the ambipolar diffusion coefficient and neutral winds. Ambipolar diffusion rate cannot be accurately determined if plasma instability exists. In our observation, many specular echoes do not decay exponentially which is likely indicative of FAI activity. For example, the echo shown by the thick line in Figure 2 may be dominated by FAI after 300 ms. One simply throws out this type of echoes for aeronomical applications. The more detrimental cases are those FAI echoes which happen to have a quasi exponential decay as shown by the thin line in Figure 2. Although the echo shown by the thick line in Figure 2 has a clear exponential decay for the portion between 240 to 300 ms, the diffusion coefficient calculated from the decay time is ~9 m 2 /s. This value is inconsistent with commonly accepted ambipolar rate of ~3 m 2 /s at 90.6 km [e.g., Cervera and Reid, 2000], the altitude of the echo if it was in the mainbeam. Further, despite that the decay constant (~80 ms) as shown by the thin line in Figure 2 is largely consistent with ambipolar diffusion at 91 km, the power fluctuation suggests the existence of another process in addition to diffusion. The often reported large spread of diffusion rates measured from meteor trails, such as those found in Cervera and Reid [2000], may partially be caused by FAI. In addition, the existence of FAI in meteor trails may also affect the measurements of neutral winds. Chang et al. [1999], in fact, have suggested that the large vertical motion observed near a geomagnetic equator site may be due to meteoric plasma irregularities. Our results further indicate that the problem may exist in other geomagnetic latitudes as well. Conceivably, it is possible that while the plasma instability process affects the wind interpretation of each individual trail, the velocity averaged over many trails largely represent the background neutral wind. In any event, although the effect of FAI on aeronomical applications needs further investigation, one should be aware that FAI can be a potential error source as a precaution. 4. Conclusion We have reported the most salient characteristics of the meteor echoes observed by the MU radar. When the radar was pointed B, essentially all the head echoes were followed by range spread trail echoes (RSTE), which were largely absent when the radar was pointed in other directions. This demonstrates that practically all meteors have field aligned structures in them. The fact that RSTEs are not observed immediately after a head-echo but with a delay time suggests that plasma instability must occur. From the RSTE spectra, we conclude that two stream instability is not important most of the time. In addition, we suggest that meteoric field aligned irregularity can be an error source for the determination of ambipolar diffusion and/or neutral winds if proper care is not taken. Although our present study is qualitative, it is evident that there are complicated electrodynamic processes occurring within meteor trails. Revelation of these processes, especially the field aligned irregularity generating mechanism, may prove fascinating and challenging in the coming years. Acknowledgments: The MU radar belongs to and is operated by Radio Science Center for Space and Atmosphere, Kyoto University with financial aid from the Monbusho. The Arecibo Observatory is part of the National Astronomy and Ionosphere Center, which is operated by Cornell University under a cooperative agreement with the National Science Foundation. One author (JDM) is supported by NSF grants AST 98-01590 and ATM 98-09998 to the Pennsylvania State University. 6

References Cervera, M. A. and I. M. Reid, Comparison of atmospheric parameters derived from meteor observations with CIRA, Radio Sci., 35, 833-843, 2000. Chang, J. L., S. K. Avery, and R. A. Vincent, New narrow-beam meteor radar results at Christmas Island: Implications for diurnal wind estimation, Radio Sci., 34, 179-197, 1999. Chapin, E., and E. Kudeki, Plasma wave excitation on meteor trails in the equatorial electrojet, Geophys. Res. Let., 21, 2433, 1994. Close, S., S. M. Hunt, M. J. Minardi, and F. M. McKeen, Analysis of Perseid meteor head echo data collected using the Advanced Research Projects Agency Long-Range Tracking and Instrumentation Radar (ALTAIR), Radio Sci., 35, 1233-1240, 2000. Heritage, J. L., W. J. Fay, and E. D. Bowen, Evidence that meteor trails produce fields aligned scatter signals at VHF, J. Geophys. Res., 67, 953-959, 1962. Haldoupis, C., and K. Schlegel, A 50 MHz radio Doppler experiment for midlatitude E-region coherent backscatter studies: system description and first results, Radio Sci., 28, 959-979, 1993. Janches, D., J. D. Mathews, D. D. Meisel, V. S. Getman, and Q. H. Zhou, Doppler studies of near-antapex UHF radar meteors, ICARUS, 143, 347-353, 2000a. Janches, D., J. D. Mathews, D. D. Meisel, and Q. H. Zhou, Micrometeor observation using the Arecibo 430 MHz Radar: I. Determination of the ballistic parameter from measured Doppler velocity and deceleration results, ICARUS, 145, 53-63, 2000b. McKinley, D. W. R., Meteor Science and Engineering, McGraw-Hill, New York, 1961. Nakamura, T., T. Tsuda, and M. Tsutsumi, Development of an external interferometer for meteor wind observation attached to the MU radar, Radio Sci., 32, 1203-1214, 1997. Oppenheim, M. M., A. F. vom Endt, and L. P. Dyrud, Electrodynamics of meteor trail evolution in the equatorial E-region ionosphere, Geophys. Res. Lett., 27, 3173-3176, 2000. Pellinen-Wannberg, A., A. Westman, G. Wannberg, and K. Kaila, Meteor fluxes and visual magnitudes from ESICAT radar event rates: a comparison with cross-section based magnitude estimates and optical data, Ann. Geophysicae, 116, 1475-1485, 1998. Raghava Reddi, C. and S. M. Nair, Meteor trails induced backscatter in MST radar echoes, Geophys. Res. Lett., 25, 473-476, 1998. Zhou, Q., P. Perillat, J. Y. N. Cho, and J. D. Mathews, Simultaneous meteor echo observations by large aperture VHF and UHF radars, Radio Sci., 33, 1641, 1998. J. D. Mathews, CSSL, The Pennsylvania State University, 323A EE East, University Park, PA16802-2707. (Email: JDMathews@psu.edu) T. Nakamura, Radio Atmospheric Science Center, Kyoto University, Uji, Kyoto 611, Japan. (nakamura@kurasc.kyoto-u.ac.jp) Q. H. Zhou, Arecibo Observatory, HC3 Box 53995, Arecibo, Puerto Rico 00612. (zhou@naic.edu) (Received October 18, 2000; Accepted December 12, 2000.) 7

Figure Captions: Figure 1. (a) Two range spread trail echoes (RSTE) observed when the radar was pointed perpendicular to B. There was no distinct head echo associated with the upper RSTE while the lower RSTE had an accompanying head-echo. RSTEs in B direction can be much more easily observed than in other directions, and also last much longer. (b) A RSTE observed when the radar was pointed vertically. In this pointing geometry, the angle between B and the radar pointing is 51 o. (c) A RSTE observed when the radar was pointed parallel to B. Figure 2. Temporal development of two constant range echoes observed at 00:53 LT (thin line) and 07:45 LT (bold line), Sept. 7, 1998 when the radar was pointed B. The thin line echo decayed quasi-exponentially and the power fluctuation suggests that the decay was not a purely diffusive process. Figure 3. Power spectra of a RSTE observed at 04:53 LT when the radar was pointed B. The lowest and highest spectra correspond to a range of 146.1 and 151.8 km respectively. The spectra, averaged over 7 s, have a resolution of 4.2 m/s. 8

Figure Captions: Figure 1. (a) Two range spread trail echoes (RSTE) observed when the radar was pointed perpendicular to B. There was no distinct head echo associated with the upper RSTE while the lower RSTE had an accompanying head-echo. RSTEs in B direction can be much more easily observed than in other directions, and also last much longer. (b) A RSTE observed when the radar was pointed vertically. In this pointing geometry, the angle between B and the radar pointing is 51 o. (c) A RSTE observed when the radar was pointed parallel to B. Figure 2. Temporal development of two constant range echoes observed at 00:53 LT (thin line) and 07:45 LT (bold line), Sept. 7, 1998 when the radar was pointed perpendicular to the B field. The bold line echo decayed exponentially from 240 ms to 300 ms only. The thin line echo decayed quasi-exponentially and the power fluctuation suggests that the decay was not a purely diffusive process. Figure 3. Power spectra of a RSTE observed when the radar was pointed perpendicular to the field line. The lowest and highest spectra correspond to a range of 146.1 and 151.8 km respectively. The spectra, averaged over 7 s, have a resolution of 4.2 m/s. 9

Top side Altitude (km) 95 90 85 (a) 0 2000 4000 6000 8000 "!#%$ &(' )*,+-./ 0.0 10.0 20.0 S/N (db) 102 0 20 40 S/N (db) 101 S/N (db) 0 20 Altitude (km) 100 98 96 100 99 94 98 (b) 0.0 200.0 400.0 600.0 800.0 Time Elapsed from Sept. 5, 00:37 LT (ms) (c) 0.0 100.0 200.0 300.0 Time Elapsed from Sept 7, 08:26 LT (ms) Zhou et al. Fig. 1

TOP side 8:94;<03 IJ(KMLONOP I#Q:NOP GH F 8:94;<0= 8:94;<021 8:94;<08 0 12020 34040 5040 67020 8"02020 >?A@ 9CB @ D E Zhou et al. Fig. 2

Top side SUT2VW R XZY [ [(\ ]%^ _a` b cudegfihjk T4VW Zhou et al. Fig 3