SEARCH

SEARCH BY CITATION

Keywords:

  • meteorites;
  • meteors;
  • meteoroids

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

The aim of this paper is to demonstrate the capabilities of a new automated analysis scheme developed for meteor head echo observations by the Shigaraki middle and upper atmosphere (MU) radar in Japan (inline imageN, inline imageE). Our analysis procedure computes meteoroid range, velocity and deceleration as functions of time with unprecedented accuracy and precision. This is crucial for estimations of meteoroid mass and orbital parameters, as well as investigations into meteoroid–atmosphere interaction processes. We collected an extensive set of data (>500 h) between 2009 June and 2010 December. Here, we present initial results from data taken in 2009 October 19–21. More than 600 of about 10 000 head echoes recorded during 33 h were associated with the 1P/Halley dust of the Orionid meteor shower. These meteors constitute a very clear enhancement of meteor radiants centred around right ascension α=inline image and declination δ=inline image. Their estimated atmospheric entry velocity of 66.9 km s−1 is in good agreement with 1P/Halley dust ejected in the year 1266 bc, which, according to simulations, crossed Earth’s orbit at the time of our observation. The Orionid activity within the MU radar beam reached about 50 h−1 during radiant culmination. The flux of sporadic meteors in the MU radar data, coming primarily from the direction of the Earth’s apex, peaked at about 700 h−1 during the same observations.


1 INTRODUCTION

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

Meteor head echoes are radio waves scattered from the intense regions of plasma surrounding and comoving with meteoroids during atmospheric flight. Head echoes were first reported by Hey, Parsons & Stewart (1947) who made observations with a 150 kW VHF radar system during the Giacobinid (now called Draconid) meteor storm of 1946.

Evans (1966) used the Millstone Hill incoherent scatter radar system to conduct the first head echo measurements using the kind of radar system later termed a high-power large-aperture (HPLA) radar (Pellinen-Wannberg 2001). This type of radar system has a peak transmitter power of the order of 1 MW and array or dish antenna apertures in the range of 800–7 × 104 m2. The transmitted radiation and receiver sensitivity is focused into narrow beams of the order of 1° full width at half-maximum (FWHM) at the VHF and/or UHF operating frequencies. This high power density permits numerous head echo detections from faint meteors.

Since the 1990s, head echo observations have been conducted with most HPLA radar facilities around the world (Pellinen-Wannberg & Wannberg 1994; Mathews et al. 1997, 2008; Close et al. 2000; Sato, Nakamura & Nishimura 2000; Chau & Woodman 2004). Modelling efforts of meteoroid flight in the atmosphere and electromagnetic plasma simulations of head echoes have been conducted to improve our understanding of the mass range of detected particles (Close et al. 2004, 2007; Dyrud et al. 2008; Kero 2008; Kero et al. 2008b).

Head echoes of shower meteors are quite rare in modern HLPA radar data. The reason for this is that sporadic meteors outnumber shower meteors in the low-mass regime observable with these radar systems. Meteors belonging to a well-known shower (even one for which the visual meteor activity is very high) are less numerous than the sporadic background when the mass limit of the detector is well below the visual range. The small collecting area of an HPLA radar system further limits successful observation of shower meteors. Analysis performed on a limited data set may, therefore, contain no or only a few shower meteors due simply to low statistical probability (e.g. Brown, Hunt & Close 2001; Szasz et al. 2008). Some orbital studies, e.g. by Janches, Meisel & Mathews (2001), have been conducted using the radial component of the meteoroid velocity vector. This is the property observed by a monostatic radar. This property does not allow orbits to be calculated, except when the meteor radiant is located exactly in the pointing direction of the radar beam.

The observability of shower meteors was, however, confirmed during the very first HPLA head echo measurements reported by Evans (1966). Evans observed the Geminid, Quadrantid and Perseid meteor showers by pointing the Millstone Hill radar towards the shower radiants at times when the radiants were located at very low elevations above the local horizon. This maximized the cross-beam detection area.

Chau & Galindo (2008) have again drawn attention to the observability of shower meteor head echoes with HPLA radars. Using the interferometric 49.92-MHz radar of the Jicamarca Radio Observatory (JRO) they detected meteors belonging to the η Aquarid and Perseid meteor showers.

We conducted a systematic set of monthly 24-h meteor head echo observations with the 46.5-MHz middle and upper atmosphere (MU) radar near Shigaraki, Japan. The MU radar has interferometric and post-beam-steering capabilities and is further described in Section 2. Compared with other HPLA radar systems, the MU radar gain pattern has a relatively large FWHM of inline image, resulting in a larger observing volume and, thus, longer event durations, at the expense of a lower peak power density.

Previous meteor head echo observations with the MU radar have been reported by Sato et al. (2000) and Nishimura et al. (2001). We developed improved analysis algorithms that provide precise geocentric velocities and directions for the observed meteors. These are described in detail in Kero et al. (in preparation). The radial velocity of each 6-ms time segment is determined to within a few tens of metres per second using a pulse-to-pulse phase-correlation technique. The radiants are estimated to within a fraction of a degree for about 3000 of a total number of about 10 000 head echoes per 24-h observation. The high number of detections allows us to map the seasonal variation of the sporadic meteor influx, as well as other characteristics, such as geocentric and heliocentric velocity distribution.

The aim of this paper is to demonstrate the capabilities of the meteor head echo observation technique we have developed by presenting initial results from 33 h of data taken on 2009 October 19–21. This data set contains primarily sporadic meteor detections, but more than 600 of a total of about 10 000 well-defined head echoes can be associated with the 1P/Halley dust of the Orionid meteor shower. These meteors constitute a very clear enhancement of meteor radiants centred around right ascension α=inline image and declination δ=+inline image and have an estimated atmospheric entry velocity of 66.9 km s−1. This agrees with the 1P/Halley dust ejected in the year −1265 (1266 bc), which, according to simulations, crossed Earth’s orbit at that time (Sato & Watanabe, private communication).

2 MU EXPERIMENT

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

The Shigaraki 46.5-MHz MU radar in Japan is located at inline imageN and inline imageE. In total, 33 h of data were collected during the 2009 Orionid meteor shower, October 19 23:00 JST to October 21 18:00 JST.

The MU radar was running in the general head echo mode briefly described below. The measurement mode is not optimized for detection of particular meteor showers, but the radar experiment ran with an identical setup each month from 2009 April until 2009 December (except 2009 August) to facilitate comparisons with sporadic detections.

The interferometric capabilities of the MU radar, described by Hassenpflug et al. (2008), make it an excellent tool for meteor head echo observations. It is a circular phased array antenna with a diameter of 103 m, consisting of 475 crossed Yagi antennas with one transmitter/receiver module each (Fukao et al. 1985). In our experiment, the output from 25 subgroups of 19 Yagi antennas each was stored as the data of separate digital channels.

The beamwidth of the MU radar system is wider than for most other HPLA radar systems, with a one-way FWHM of inline image. This gives a comparatively large observing volume and, thus, longer event durations. During the experiments described in this paper, the radar beam was pointed in the direction of the zenith.

Head echoes are detected over the entire experiment height range of 73–127 km, limited by the experiment settings adapted to the maximum data rate of 20 GB h−1 (Kero et al., in preparation). We transmitted 13-bit Barker-coded pulses with a total pulse length of 156 μs and interpulse period of 3.12 ms to use the 5 per cent duty cycle. The received data were stored as 85 range values from each transmitted pulse, sampled at 6 μs intervals. The sampling corresponded to a range resolution of 6 × 10−6c0/2 ≃ 900 m, where c0 is the speed of light. The range of meteor targets was determined with a precision of of the order of 10 m, or to within about one-hundredth of a range gate, using a range interpolation technique described in detail by Kero et al. (in preparation).

3 OBSERVATION OF ORIONID SHOWER METEORS

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

Sato & Watanabe (2007) developed simulations of the comet 1P/Halley dust trails responsible for the 2006 Orionid outburst. Their simulations for 2009 suggest that on October 19–21, Earth passed a dust train from the year 1266 bc, producing meteoroids with a geocentric entry velocity of 66.9 km s−1 from a radiant located at right ascension α=inline image and declination δ=inline image (Sato & Watanabe, private communication).

Fig. 1 shows the hourly detection rate of all meteors (blue) and meteors classified as belonging to the Orionid shower (red) versus local time (JST). The elevation of the Orionid radiant is plotted as a green curve. Orionids were here selected as meteors with radiants within a radial distance of ±5° from the Orionid radiant and a geocentric velocity in the range of 67 ± 3 km s−1. The radiant and velocity properties of this subgroup are described in detail below.

image

Figure 1. The hourly detection rate of all meteors (blue) and meteors classified as belonging to the Orionid shower (red) as a function of time during MU radar observations from 2009 October 19 23:00 JST to 2009 October 21 08:00 JST (Japan Standard Time). The hourly rate is given in 30-min bins, derived from the number of detections per half-hour. The elevation of the Orionid radiant (α=inline image, δ=inline image) is plotted as a green curve, culminating at about 71° around 04:30 JST. The noise temperature (inline image, dashed line) peaks at around 03:30 and 18:00 JST, when the strong radio sources Taurus A and Cygnus A passes close to zenith.

Download figure to PowerPoint

Sporadic meteors are those that cannot be directly ascribed to a parent body. Sporadics are the most numerous among our observed particles, and the main contributors to the mass influx into the Earth atmosphere. In this paper, ‘sporadic detections’ refer to all detections, except meteors classified as Orionids. However, it should be noted that these detections also contain activity from minor streams, which we will examine in detail in future papers.

The radiant distribution of the detected meteors is provided in Sun-centred ecliptic coordinates in Fig. 2. The Sun is located at 0° Sun-centred ecliptic longitude, regardless of time of year. The direction of the apex is 270° Sun-centred ecliptic longitude. The enhancement in the Orionid radiant region is very clear, although a few detections are likely due to sporadic meteors. In addition to the Orionids, enhancements appear close to the radiant region of the Leonis Minorids (LMI) and the South Taurids (STA). More than 600 of the detections are likely to belong to the Orionid shower, out of a total number of about 10 000 head echoes in the 33 h of consecutive data, spanning over almost two full Orionid radiant culminations (upper transit of the radiant).

image

Figure 2. The radiant distribution of the ∼10 000 meteors detected during the 33-h experiment plotted in Sun-centred ecliptic coordinates. The meteor radiant with geocentric velocity (the Earth’s velocity not subtracted) is expressed in terms of the Sun-centred ecliptic longitude and ecliptic latitude. The position of the ecliptic longitude of the Sun at the moment of detection has thus been subtracted from each meteor radiant, positioning the Sun at 0° ecliptic longitude regardless of time of year. The direction of the apex is = 270° Sun-centred ecliptic longitude.

Download figure to PowerPoint

The plotted hourly rate (Fig. 1) is given at half-hour intervals. The number of detections each half-hour is counted and multiplied by a factor of 2. There are no unintentional gaps in the data set. The radar is, however, stopped and restarted every full hour to reboot the system and avoid memory buffer overflow, an occasional consequence of the data rate being close to the system limit. Due to this procedure, we have no data for the first 1–2 min of every full hour. This is taken into account when calculating the hourly rates.

Fig. 3 displays a histogram of the geocentric velocity distribution of all October detections, corrected for Earth focusing (zenith attraction) and Earth rotation using the routines described in Szasz (2008) and Szasz et al. (2008). Each bar represents the number of detections in a bin of 1 km s−1 width. The grey bars are the number of detections with Orionids excluded. The solid line is the sporadic distribution, weighted with the geocentric velocity v as v−3. The solid line represents the true distribution, assuming that a mass–velocity selection effect similar to that described by Close et al. (2007) and Szasz et al. (2008) is valid for our data. The black bars plotted over the grey bars are the meteors classified as Orionids. Their sum thus represents the velocity distribution of all detections.

image

Figure 3. Histogram of the sporadic meteor geocentric velocity (grey) and this distribution weighted with v−3 (solid line) to account for mass–velocity selection effects (Close et al. 2007; Szasz et al. 2008). The weighted distribution is normalized to cross the observed distribution at 58 km s−1, the maximum of the observed sporadic distribution. The black bars show the observed distribution including Orionids.

Download figure to PowerPoint

Fig. 4 shows the entry velocity of the meteors close to the Orionid radiant. The estimated entry velocity of the main bulk is in satisfactory agreement with the predicted 1P/Halley dust entry velocity of 66.9 km s−1. Some faster detections appear, as well as a tail of slower detections.

image

Figure 4. Distribution of detected initial velocities for meteors close to the Orionid radiant, corrected for zenith attraction and Earth rotation according to the procedures described by Szasz (2008).

Download figure to PowerPoint

We have not yet tried to model deceleration during the atmospheric passage of each meteoroid before detection. However, we plotted the velocity as derived for the outset of each event at the altitude observed by the radar. The tail of slower meteors is thus likely caused by some meteoroids having experienced a significant amount of deceleration before detection in the radar beam. In an attempt to remedy this, we developed and used a post-analysis beam-steering routine (Palmer et al. 1990), in which we increased the signal-to-noise ratio (S/N) far from beam centre and extended the duration of each event as far as possible (Kero et al., in preparation). In this routine, we used the trajectory determined without beam steering to re-analyse all events with the receiver beam redirected towards the anticipated position of the target. Inevitably, despite this precaution, some particles will experience considerable ablation (and deceleration) before entering regions of highenough transmitted power density due to the fixed transmitter beam radiation pattern, which was always directed towards the zenith.

In the worst case scenario, short-duration events give rise to additional errors of up to a few km s−1, especially when the radiant is close to the horizon and the directly measured radial velocity component is only a small fraction of the meteoroid velocity. This will be investigated further in future papers.

4 RADIANT DISTRIBUTION AND DETECTION RATES

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

Orionid activity among the well-defined MU radar head echoes reached about 50 h−1 during radiant culmination. The simultaneous flux of sporadic meteors, originating primarily from the direction of the Earth’s apex (Fig. 2), peaked at about 700 h−1.

Meteors belonging to meteor showers still retain orbital properties similar to the active or extinct comets (or possibly asteroids) from which they were released. Thus, shower meteors are highly interesting in their own right, even though they do not contribute as much mass to the total mass influx as the sporadic meteors. Shower meteors have been investigated with many different methods and instruments over long periods of time (Jenniskens 2006). Consequently, we can compare the MU radar shower meteor characteristics, such as velocities, radiant directions and flux, with results achieved using other methods.

Fig. 5 displays a scatter diagram in equatorial coordinates of the region close to the Orionid radiant. Fig. 6 displays two contour plots of the radiant density for a part of the same region. In this figure, the left-hand panel includes, and the right-hand panel excludes, compensation for the drift of the radiant during the observation. The shower radiant drifts as a function of time relative to the stars due to Earth moving through the comet dust trail, a phenomenon termed the radiant drift. We were guided by Lindblad & Porubcan (1999), but instead of reducing the radiant positions to a solar longitude of 208°, we reduced them to inline image to permit comparison with the 2009 simulations developed by Sato and Watanabe (private communication). Our radiant drift compensation is therefore

  • image(1)

where ⊙ is the solar longitude at the time of the observation for each meteor. The radiant motion in 24 hours (from one culmination to the next) is about 0°7, meaning that the total motion from start to end of measurement (33 h) was about 1°.

image

Figure 5. Scatter diagram of the meteor radiants in equatorial coordinates (declination δ versus right ascension α) of the region close to the Orionid radiant, with colour-coded velocity.

Download figure to PowerPoint

image

Figure 6. Contour plots of the meteor detections displayed in the scatter diagram of Fig. 5. The left-hand panel includes radiant drift; the right-hand panel excludes radiant drift. The radiant drift reported by Lindblad & Porubcan (1999) was used. The red curves show the FWHM of the radiant density (FWHMδ).

Download figure to PowerPoint

The red curves in Fig. 6 show the FWHM of the radiant density (FWHMδ). The compensated contour plot is improved in the sense that the contour distribution is more compact and circular than the original contour plot. When the radiant drift is taken into account, FWHMδinline image–2°. This result indicates that a significant part of the spread of the original contour plot can be ascribed to the radiant drift rather than to uncertainties in our measurement technique. Our radiant density distribution is as compact as the distribution of all precisely reduced Orionids (66 photographic and 19 video meteors) of the IAU Meteor Data Center presented by Lindblad & Porubcan (1999). This indicates that our radar method provides precision and accuracy comparable to the photographic reduction of much brighter meteors with longer detectable trajectories. The centre of both the compensated and uncompensated radiant distribution agree very well with the prediction made by Sato & Watanabe (private communication).

5 COLLECTION AREA

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

The detection rates of visual meteors are converted to a flux of meteors by taking into account the detection probability as a function of visual magnitude and distance from the centre of the observer’s field of view (e.g. Jenniskens 1994). As an example, six observers at the Skalnaté Pleso Observatory simultaneously monitored the same portion of the local sky. By examining the statistical distribution of 1351 observations, subsequently found to refer to 476 individual meteors, they calculated the probability that an observer would record a meteor as a function of its visual magnitude and angular distance from the centre of the team’s field of view (Kresáková 1966).

Detection probability can be integrated over an observer’s field of view to produce an equivalent collection area of the atmosphere, generally defined as a horizontal area at 100 km altitude (Koschack & Rendtel 1988). An observed count rate divided by the collection area provides an estimate of the meteor flux, i.e. the number of meteors in Earth’s atmosphere per unit of area and time.

This section is devoted to an outline of our MU radar collection area estimation, enabling meteor flux calculations from the MU Orionid detection rate and comparison with other methods. The result of the estimation is presented in Fig. 7, where the equivalent collection area of the MU radar is plotted versus radar cross-section (RCS). For the remainder of this paper, we adopt the antenna beam gain integration (red curve in Fig. 7) as the MU radar collection area. The estimated collection area varies with at least three orders of magnitude, from about 1 to 1000 km2 within the RCS range of the detected meteors.

image

Figure 7. Estimated meteor collection area as a function of meteor radar cross-section (RCS). The blue curve is calculated using a probability of detection approach and assuming a uniform flux of meteors as a function of distance from the bore axis. The red curve is based on an integration of the antenna radiation pattern.

Download figure to PowerPoint

5.1 Radar cross-section

RCS is the radar equivalent of visual magnitude and is given relative to a metallic sphere with an area of 1 m2. The higher the RCS, the more detectable an object is with a radar. As the RCS spans several orders of magnitude, it is expressed in units of decibel-relative-to-a-square-metre (dBsm), where 0 dBsm is equivalent to the return from a 1-m2 sphere (Close 2004) according to

  • image(2)

 We have evaluated the RCS of detected targets by rewriting the classical radar equation (e.g. Skolnik 1962) as

  • image(3)

where Pr is the received power, R is the target range, Gt is the transmitter antenna gain, Gr is the receiver antenna gain, θ is the azimuth (positive east of north), ϕ is the zenith distance of target, λ is the radar wavelength and Pt is the transmitted power. The received power is given by

  • image(4)

where Tmet is the equivalent signal temperature of the echo, kB= 1.3810−23 J K−1 is the Stefan–Boltzmann constant and bw≃ 1/6 μ s ≈ 167  kHz is the receiver bandwidth. The equivalent signal temperature, Tmet, is calculated by multiplying the S/N with the equivalent noise temperature, inline image, which is the sum of the system noise (∼3000 K) and the fluctuating cosmic background radio noise (∼6000–15 000 K every diurnal cycle). inline image is plotted as a dashed line in Fig. 1. The peak-to-peak variation is ≃3 dB and could in principle modulate the low RCS event rate. However, no such trends are visible in the data. The collection area estimation is based on the whole observation period and is therefore representative of the average inline image.

The peak transmitted power is optimally Pt= 1  MW, but is not continuously monitored and is therefore an estimated value. The attenuation of the signal as compared to a target at the boresight axis depends on the one-way directional gain of the MU radar antenna, GMU, displayed in Fig. 8. The peak value at boresight axis is ≃34 dB and the gain variations in the azimuthal direction are small inside the first side lobe, located about 15 km from the bore axis at a range of 100 km. For the purpose of the meteor head echo collection area estimation described here, we treated the gain pattern as azimuthally symmetric. We calculated an average gain at each radial distance. Furthermore, we used the target angular zenith distance at maximum RCS and calculated an equivalent radial distance at 100 km altitude.

image

Figure 8. The theoretical gain pattern of the MU radar antenna, expressed in terms of the one-way directional gain (GMU; the colour-coding and the vertical axis) as a function of radial displacement from the bore axis at a range of 100 km.

Download figure to PowerPoint

5.2 MU radar collection area estimation

As an initial MU radar collection area estimate, we used a probability-of-detection approach similar to that of visual meteors (Kresáková 1966). Such an approach could only be applied to the part of the meteor RCS distribution where the number of meteors is large enough (−35 to −5 dBsm or equivalently 3 cm2 to 0.3 m2). Using these results, we investigated how detection probability depends on a meteor’s maximum RCS and the antenna gain at its location. Finally, we integrated the antenna gain pattern to get an equivalent collection area for the whole range of detectable RCS. We assumed the same detection probability dependence on gain (and thus sensitivity of the radar system) for all RCS as for the RCS interval investigated in the initial estimation.

5.2.1 Probability-of-detection approach

The MU radar provides high enough meteor head echo detection rates to compare a normalized flux of various meteor RCS intervals in different concentric ring-shaped areas around the beam centre having successively lower antenna gain. A plot of the meteor flux (colour-coded) versus maximum RCS (vertical axis) and radial distance (horizontal axis) is provided in Fig. 9. The white curve shows an adaptation, a, of the antenna gain pattern, described in Section 5.2.2. The flux was obtained from dividing the number of meteors in each ring by its area. Due to the stochastic nature of the occurrence of meteors, the distribution of meteoroid trajectories in the atmosphere above the radar is random. Given that a large enough number of events is recorded, the flux of meteors in adjacent regions of the atmosphere is thus uniform. The relative probability of detecting a meteor in regions with different antenna gains can therefore be obtained by comparing the observed flux in each region.

image

Figure 9. Observed number of meteors (colour coded), normalized by beam area, versus RCS (vertical axis) and radial distance from beam centre (horizontal axis); and an adaptation, a, of the antenna gain pattern to the meteor RCS distribution according to equation (5.2.2) (white curve).

Download figure to PowerPoint

Fig. 9 shows the expected result that the RCS distribution above an antenna gain dependent threshold is essentially independent of radial distance. The distribution of small RCS meteors visible near beam centre (left) is progressively missing towards beam edge (right). This effect is equivalent to a larger MU collection area for larger RCS meteors.

The blue curve in Fig. 7 represents the collection area for RCS ranging from −35 to −5 dBsm. We integrated the flux of each RCS interval between −35 and −5 dBsm radially and normalized with the flux at the beam centre, assuming a detection probability there equal to unity. This assumption can if necessary be adjusted at a later stage by dividing the presented collection area, and multiplying the estimated fluxes, by a constant factor.

5.2.2 Antenna gain pattern integration

The white curve in Fig. 9 shows an adaptation, a, of the antenna gain pattern to the meteor RCS distribution (expressed in dBsm) according to

  • image(5)

where GMU is expressed in dB, as visualized in Fig. 8, and has a value of 34 dB at the beam centre. It is evident from Fig. 9 that the smallest RCS meteors that can successfully be observed with our present analysis technique and quality criteria are well described by a. At beam centre, meteors down to RCS ≃−45 dBsm ≃ 0.3 cm2 were detected. The detection limit outside the beam centre increased from −45 dBsm towards larger RCS as 1.5GMU. The reason for the 1.5 factor is likely an effect of the statistical probability of successfully detecting and determining the trajectory of a meteor, as a function of its location in the gain pattern.

The collection area estimated from the antenna beam pattern is presented as a red curve in Fig. 7. It was calculated from integrating the area at 100-km range where 1.5(34 −GMU) > RCS + 39 for −40 ≤ RCS ≤ 20 and dividing the result by a factor of 2 to fit the collection area calculated from the flux distribution (blue curve). That division by a factor of 2 was necessary to make the estimations agree is consistent with the detection probability not being unity inside a, but decreasing with radial distance.

The collection area for meteors with RCS < −30 dBsm is likely overestimated. The reason is that we expect the flux of small RCS meteors to be greater than the flux of large RCS meteors, due to the fact that RCS is proportional to meteoroid mass (similarly to visual magnitude) and small-mass meteoroids are more numerous than large-mass meteoroids (Hughes 1972). The relatively low beam centre flux for RCS < −30 dBsm displayed in Fig. 9 indicates therefore that the detection probability of these meteors is smaller than unity. In Section 7 we have defined the limit RCScut≃− 20  dBsm, above which observational selection does not appear to affect the RCS distribution.

6 ORIONID ACTIVITY AND ZHRMU

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

Experimental investigations by Close et al. (2002) and Kero et al. (2008c) and simulations by Dyrud et al. (2008) suggest that meteor head echo RCS depends very weakly on viewing geometry. This is confirmed by the MU Orionid activity, presented in Figs 10 and 11.

image

Figure 10. Number of detected Orionids per hour (red) and simulated activity (green) versus local time. The radiant elevation (el) is taken into account according to sin   (el)1.47 (Zvolankova 1983; Jenniskens 1994) and a constant flux of 50 detectable Orionids from a radiant in zenith is assumed. The vertical bars represent Poisson-distributed statistical errors of the meteor counts.

Download figure to PowerPoint

image

Figure 11. MU radar zenithal hourly rate (ZHR) of Orionids (red) and radiant elevation (green) versus local time. The vertical bars represent Poisson-distributed statistical errors, inline image, where N is the hourly meteor count rate displayed in Fig. 10. The horizontal line at ZHR = 50 is equivalent to the simulated activity curve in Fig. 10.

Download figure to PowerPoint

A flux of Orionids is already discernable when the radiant is 10° above the horizon, and thus the angle between the trajectory and the beam is close to perpendicular. The error bars represent Poisson-distributed statistical errors of the meteor counts. Each data point represents the number of Orionids during the 60 min centred around each full hour. The first data point, thus, shows the number of Orionids from 22:30–23:30 JST on October 19.

The Orionid detection rate increases as a function of the radiant elevation (el), as expected from the increasing flux density of meteors per unit volume in the atmosphere, and the ablation-profile dependence of the atmospheric density gradient along the trajectory. The green curve in Fig. 10 is a simulated activity profile, assuming a constant flux of Orionids and the observed rate depending on elevation as sin γ (el), where γ= 1.47. The parameter γ is called the radiant altitude exponent. For visual meteors, it is generally found to have a value in the range of 1 < γ < 2, depending on meteor shower, varying slightly between different investigations (Jenniskens 1994).

Zvolankova (1983) investigated 17 000 visual Perseids observed at the Skalnaté Pleto Observatory during nine years and found a best-fitting value of γ= 1.47 ± 0.11. As evident from Fig. 10, γ∼ 1.47 agrees well with the MU radar activity of Orionids. Jenniskens (1994) calculated empirical and theoretical values of the radiant altitude exponent for various meteor showers. His empirical value for the Orionids is γ= 1.68 ± 0.14, while the theoretical exponent is γthe= 1.47.

In Fig. 11, we report what is hereafter called the MU radar zenithal hourly rate, ZHRMU. It is computed by dividing the count rate of Orionids in Fig. 10 with sin 1.47 (el). If the radiant altitude exponent is correctly chosen and the Orionid flux is constant, the data in Fig. 11 should be consistent with the horizontal line at ZHRMU= 50 within the error bars. These represent Poisson-distributed statistical errors, inline image, where N is the hourly meteor count rate displayed in Fig. 10. The singular strongly deviating data point at 23:00 JST on October 19 is likely a statistical outlier. The four-hour interval from October 20 21:30 JST to October 21 01:30 JST, however, is consistently above the constant level and suggest an increased Orionid activity, by a factor of 2 ± 0.4.

The good agreement between the radiant altitude dependence of the observed head echo rate and visual meteors encourages the use of radar head echo measurements for activity monitoring and flux measurements of meteor showers.

7 RCS MAGNITUDE DISTRIBUTION INDEX

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

Fig. 12 presents a logarithmic plot of the total number of Orionids (red) and sporadics (blue) per one dBsm interval, detected during our 33-h experiment. The Orionid distribution contains a greater relative fraction of large-RCS meteors than the sporadic distribution does. One of the reasons for this is the higher mean velocity of the Orionids compared to the sporadics (see Fig. 3). The RCS of meteor head echo targets has been found to depend on meteoroid mass m and velocity v approximately as RCS ∝mv3 (e.g. Close et al. 2007; Szasz et al. 2008). Fig. 13 compares the Orionid (red) and the sporadic (blue) distribution, the latter being restricted to the ≈1900 sporadic meteors within the velocity range of 64–69 km s−1 to suppress effects from the RCS-velocity dependence on the comparison. We took the collection area dependence on RCS into account by calculating the ratio of the number of meteors and the MU radar collection area. RCScut≃− 20  dBsm in Fig. 13 represents the limit above which the distribution can be converted into a RCS magnitude distribution index, similar to the magnitude distribution index of visual observations (see e.g. fig. 2 of Hughes 1987). To the right of RCScut, the RCS distribution is a straight line when visualized in logarithmic form. Its slope is proportional to the RCS magnitude distribution index.

image

Figure 12. Number of detected Orionids (red) and sporadics (blue) per integer value of dBsm.

Download figure to PowerPoint

image

Figure 13. Differential Orionid (red) and sporadic (blue) RCS distributions, calculated by taking the ratio of the number of Orionids and sporadics (in the velocity range 64–69 km s−1 and the MU radar collection area). RCScut represents the limit above which the distribution can be converted into a RCS magnitude distribution index, similar to the magnitude distribution index of visual observations (see e.g. fig. 2 of Hughes 1987).

Download figure to PowerPoint

The slope and zero crossing of the least-squares-fitted straight lines shown in Fig. 13 are −0.0994(±0.0045) and −1.924(±0.0505) for the sporadics and −0.0753(±0.0025) and −1.863(±0.0287) for the Orionids, where the values in parentheses are standard errors of the fit. The ratio of the slopes of the sporadic and Orionid RCS distributions is therefore 1.32±0.11. The ratio of the slopes may reflect the ratio of the mass coefficients of the sporadic and the Orionid meteoroids, sspo s−1ORI. The meteoroid mass distribution, s, is usually expressed by the relation

  • image(6)

where n(m) is the number of meteors in the mass interval m to m+ d m (Hughes 1972). The differing slopes may also represent different physical characteristics of Orionid relative to sporadic meteoroids. The sporadics might hence show more varied fragmentation potential. Kero et al. (2008a) showed that fragmentation and quasi-continuous disintegration result in RCS pulsations. The implications of such features in shower and sporadic MU radar meteor detections will be investigated in a future publication.

8 CUMULATIVE METEOR FLUX

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

To compare the detection rates of meteor head echoes with meteor detections using other methods, we calculated the flux of detected particles per unit area and time. In Section 5, we estimated the collection area of the MU radar as a function of meteor RCS. In Section 6, we found the encouraging result that the measured rate of Orionids corresponds well to a rate of 50 h−1 (except for a short burst of up to 100 h−1) had the radiant been located in zenith. Because the collection area is determined as a horizontal surface, it is natural to determine the equivalent zenith flux of Orionids. This parameter corresponds to the flux of visual meteors estimated from visual ZHR.

The cumulative flux of Orionids (red) and all sporadic meteors (blue) is presented in Fig. 14. Each point on the curve represents the number of observed meteors with a RCS greater than the RCS of that point. The reason for the 8 dBsm upper limit on the limited cumulative distributions (grey and black) is evident from Fig. 12; namely the number of detected meteors is very low. These few large RCS meteors significantly affect the original sporadic cumulative distribution (blue). As pointed out by Hughes (1987), cumulative distribution plots are very sensitive to errors and/or uncertainties in the tail of the distributions, affecting all data points. The statistics are not good enough to investigate whether the deviation from a constant slope of the sporadic distribution is truly due to large RCS meteors having a different magnitude distribution index than small RCS meteors or not. Thus, we focus on the gradient of the distribution of sporadic meteors with a limited RCS range and the total cumulative sum. The gradient of the Orionid RCS distribution is constant, regardless of the upper limit of the RCS.

image

Figure 14. Cumulative Orionid (red) and sporadic (blue) meteor flux versus RCS, and corresponding distributions when the upper limit is 8 dBsm (grey and black). Plotted in green are the visual and radar fluxes of Orionid meteoroids of mass greater than m (kg) reported by Jones (1983).

Download figure to PowerPoint

The total cumulative sum of MU Orionids amounts to about 1 km−2 h−1. According to the International Meteor Organization (IMO), the visual ZHR of the 2009 Orionids was 30–40 during MU observations and may have peaked at about 45 on October 22–23, after our observations had ended. Their statistics are based on 3688 reported Orionids from 63 observers in 23 countries.

In Fig. 14 we also plotted the visual and the radar specular meteor trail fluxes of Orionids (green) reported by Jones (1983). The visual flux was estimated as 4 × 10−3 km−2 h−1 from a reported visual Orionid ZHR of 22 observed 1944–1950 (Stohl & Porubcan 1981). The visual flux during our measurements may have been a factor of 1.5–2 higher, given the reported ZHR of 30–40. The calculated flux based on visual observations corresponds to meteoroids of masses larger than about 10−7 kg, as indicated on the green abscissa in Fig. 14.

According to Jenniskens (1994), the magnitude and corresponding mass distribution indices found for various annual meteor showers in the +5 to −2 mag interval is constant for meteoroid masses down to 10−9 kg. One of the data sets this conclusion is derived from is the specular meteor radar observations reported by Jones (1983). That radar system was determined to have a limiting magnitude of +8.1. The rate of Orionid detections was quite low, with a maximum number of about 20 events d−1 stronger than +7.35 and about 33 stronger than +8.1. The total number of Orionids was 342, recorded during 29 days of observation, about a fortnight each year 1980–81 centred around the Orionid activity peak at ⊙∼ 209°.

The reason for this relatively small number of detections, given the low limiting magnitude and a large effective collection area of the order of thousands of  km2, is the initial trail radius attenuation effect (Greenhow & Hall 1960). It is particularly severe for high-speed meteors, as these ablate at high altitudes. In the case of the Orionids, Jones estimates that only 2.83 per cent of the total number was detected. This grave attenuation introduces large uncertainties when converting the observed number of detections to meteor influx. Additionally, determining the collecting area of a specular meteor trail radar system is not a trivial task. The methods used are generally based on the theory described by Kaiser (1960). Cervera & Elford (2004) developed a meteor radar response function including non-uniform meteor ionization profiles.

Jones (1983) estimated the flux of +8.1 mag Orionids (when assuming that 2.83 per cent of all meteors within the collection area were detected) to be about 0.1 km−2 h−1. Orionids of +8.1 limiting magnitude correspond to meteoroid masses larger than 10−8 kg.

Given that the MU radar flux of Orionids is about an order of magnitude higher, and comparing the two abscissae in Fig. 14, a crude estimate of the smallest Orionids detected with MU in the present data set is 10−9 kg. It should be noted that the head echo RCS of individual meteors may be differently affected by meteoroid density, fragmentation, differential ablation, etc., than the luminosity and specular trail echo RCS. The comparison of cumulative distributions in Fig. 14 is valid when head echo RCS, at least when regarded as a statistical distribution, can be used as a proxy for meteoroid mass, similar to visual and specular meteor trail radar magnitude.

9 CONCLUSIONS

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

We have developed and used a new automated analysis scheme for meteor head echo observations with the MU radar. In this paper, we reviewed 33 h of data from 2009 October 19 to 21, for which about 600 of 10 000 meteor head echoes were associated with the 1P/Halley dust of the Orionid meteor shower. These meteors constituted a very clear enhancement of meteor radiants centred around right ascension α=inline image, declination δ=+15°4, and had an atmospheric entry velocity of ∼67 km s−1. The Orionid detection rate reached a maximum of 50 ± 7 h−1 close to radiant culmination. The rate of sporadic meteors, coming primarily from the direction of the Earth’s apex, peaked at about 700 ± 26 h−1.

In this work, we showed that the Orionid meteoroid stream activity could be accurately tracked with the MU radar when the radiant is at least 10° above the local horizon. The Orionid detection rate was consistent with the radiant altitude exponent γ= 1.47 ± 0.11 derived for visual meteors by Zvolankova (1983). The resulting ZHRMU∼ 50 h−1 of Orionids was fairly uniform during the measurements, except for October 20 22:00 JST to October 21 02:00 JST (October 20 13–17 ut), when it was enhanced by a factor of 2 ± 0.4. The agreement with visual observations may be fortuitous, because the true time variation of the rate of faint Orionids during the measurement is unknown. We will investigate the radiant altitude exponent for other showers in forthcoming papers.

We estimated the collection area as a function of meteor RCS for the MU radar and converted the meteor count rates into differential and cumulative fluxes. The estimated collection area varies by at least three orders of magnitude, from about 1 to 1000 km2 within the RCS span of the detected meteors. The zenithal equivalent cumulative flux of Orionids is ∼1 km−2 h−1, while the sporadic flux peaked at ∼30 km−2 h−1.

The ratio of the RCS magnitude distribution of 64–69 km s−1 sporadics and the Orionids is 1.32 ± 0.11. This may reflect the Orionid to sporadic mass coefficient ratio, or that they have different physical characteristics.

HPLA radar meteor head echo observations complement specular meteor trail radar orbital surveys (e.g. Galligan & Baggaley 2004; Brown et al. 2010), as the observational biases are different, and a wealth of information can be determined from each particular head echo event (e.g. Kero et al. 2008b).

ACKNOWLEDGMENTS

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES

This study is supported by JSPS grants-in-aid 20-08730, 20-08731 and 21340141. JK and CS are financed by JSPS postdoctoral fellowships for foreign researchers, P08730 and P08731. The MU radar belongs to and is operated by the Research Institute of Sustainable Humanosphere, Kyoto University, Kyoto, Japan.

We acknowledge Mikiya Sato and Junichi Watanabe for fruitful discussions of their 2009 Orionid simulations.

REFERENCES

  1. Top of page
  2. ABSTRACT
  3. 1 INTRODUCTION
  4. 2 MU EXPERIMENT
  5. 3 OBSERVATION OF ORIONID SHOWER METEORS
  6. 4 RADIANT DISTRIBUTION AND DETECTION RATES
  7. 5 COLLECTION AREA
  8. 6 ORIONID ACTIVITY AND ZHRMU
  9. 7 RCS MAGNITUDE DISTRIBUTION INDEX
  10. 8 CUMULATIVE METEOR FLUX
  11. 9 CONCLUSIONS
  12. ACKNOWLEDGMENTS
  13. REFERENCES