Sporadic meteor sources as observed by the Jicamarca high-power large-aperture VHF radar

Icarus 188 (2007) 162–174 www.elsevier.com/locate/icarus

Sporadic meteor sources as observed by the Jicamarca high-power large-aperture VHF radar Jorge L. Chau ∗ , Ronald F. Woodman, Freddy Galindo Radio Observatorio de Jicamarca, Instituto Geofísico del Perú, Apartado 13-0207, Lima 13, Peru Received 27 May 2006; revised 13 October 2006 Available online 15 December 2006

Abstract We present, for the first time, the main sources of sporadic meteors as inferred from meteor-head echoes obtained by a high-power large-aperture radar (HPLAR). Such results have been obtained at the Jicamarca HPLAR (11.95◦ S, 76.87◦ W, 1◦ dip angle). Observations are based on close to 170,000 meteors detected in less than 90 h spread over 14 days, between November 2001 and February 2006. Meteors with solar orbits are observed to come from basically six previously known sources, i.e., North and South Apex, Helion, Anti-Helion, and North and South Toroidal, representing ∼91% of the observations. The other ∼9% represents meteors with observed velocities greater than the Sun’s escape velocity at 1 AU, most of them of extra-solar origin. Results are given before and after removing the Earth’s velocity and the sources are modeled with two-dimensional Gaussian distributions. In general, our results are in very good agreement with previously known sources reported by Jones and Brown [Jones, J., Brown, P.G., 1993. Mon. Not. R. Astron. Soc. 265, 524–532] using mainly specular meteor radar (SMR) data gathered over many years and different sites. However, we find slightly different locations and widths, that could be explained on the basis of different sensitivities of the two techniques and/or corrections needed to our results. For example, we find that the North and South Apex sources are well defined and composed each of them of two collocated Gaussian distributions, one almost isotropic with ∼10◦ width and the other very narrow in ecliptic longitude and wide in ecliptic latitude. This is the first time these narrow-width sources are reported. A careful quantitative analysis is needed to be able to compare the strengths of meteor sources as observed with different techniques. We also present speed and initial altitude distributions for selected sources. Using a simple angular sensitivity function of the combined Earth–atmosphere–radar instrument, and an altitude selection criteria, the resulting meteor sources are in better qualitative agreement with the results obtained with SMRs. © 2006 Elsevier Inc. All rights reserved. Keywords: Meteors; Ionospheres; Radar observations

1. Introduction Since the early 1940’s specular meteor radars (referred to as SMR hereafter) have been the main source of meteor observations entering the Earth’s atmosphere (e.g., Jones and Brown, 1993; Galligan and Baggaley, 2005; and references therein). These systems use wavelengths between 3 and 15 m, on few occasions systems with larger wavelengths have been used (e.g., Steel and Elford, 1987). The combined results from these systems gathered over many years of observations have allowed the identification of the currently well accepted six sources of sporadic meteors, i.e., North Apex (NA), South Apex (SA), Helion, * Corresponding author. Fax: +51 (1) 317 2312.

E-mail address: [email protected] (J.L. Chau). 0019-1035/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.icarus.2006.11.006

Anti-Helion (AH), North Toroidal (NT), and South Toroidal (ST) (e.g., Jones and Brown, 1993; Taylor and Elford, 1998; Galligan and Baggaley, 2005). In this work we present for the first time the sources of sporadic meteors as inferred from the meteor-head echoes detected by the Jicamarca High-Power Large-Aperture Radar (HPLAR). This type of radar (e.g., ALTAIR, Arecibo, EISCAT, Jicamarca, Millstone Hill, MU, and Sondrestorm) has been used for meteor-head studies since mid 1990’s originally motivated by predictions of several returns of the Leonids meteor showers (e.g., Janches et al., 2000b; Close et al., 2002; Sato et al., 2000; Chau and Woodman, 2004). The radar frequencies from these radars start around 50 MHz (at Jicamarca and MU) and go as high as 1.29 GHz (at Sondrestorm). These radars are mainly sensitive to the plasma surrounding the meteoroid as it enters

Sporadic meteor sources observed at Jicamarca

the Earth’s atmosphere (between 70 and 140 km). Faster geocentric meteoroids will ionize at higher altitude and will have more chance to be observed by the HPLARs (e.g., Janches and ReVelle, 2005). More details on the scattering mechanism of meteor-head echoes are given by Close et al. (2004) and Mathews (2004). We first present the experimental setup and meteor parameters either directly measured or derived. The expected statistical errors for some derived parameters are presented in Appendix A. In Section 3 we show the meteor sources as seen by the Jicamarca radar and the speed and initial altitude distributions of selected meteors sources, including the not well defined Antapex (or Apex prograde) source. Finally results are interpreted and discussed, special emphasis is given to the comparison with results obtained by SMRs. 2. Experimental setup and meteor parameters Meteor-head observations have been performed using the large Jicamarca array (∼300 m × ∼300 m) for transmission and at least three quarter sections for reception (∼75 m × ∼75 m). The same linear polarization (north–east) has been used in both modes. The antennas were phased to point on-axis (1.46◦ from zenith towards the south–west). Complex voltages (raw data) from at least three quarters were recorded. Note that we only need the information from three non-collinear antennas in order to locate the meteors inside the transmitting beam. Since November 2001, two modes of operations have been used. In Table 1 we summarize the main parameters used in the two different modes. The main differences between the modes are: (a) range resolution, (b) unambiguous pulse-to-pulse Doppler velocities, (c) duty cycle, (d) use of analog or digital receivers, and (e) automatic or manual detection. Table 2 summarizes the observational campaigns used for the results presented in this paper, indicating the observational times, the mode used, and the percentage of meteors from the close to 170,000 meteor heads detected. Although not a direct subject of this work, one of the biggest difficulties for Jicamarca to be used for meteor-head studies, is the usual occurrence of strong echoes from equatorial electrojet (EEJ) irregularities between 90 and 120 km of altitude (e.g., Farley, 1985). Fortunately, these EEJ echoes are usually weak around sunrise times and when zonal electric field are close to zero. A close look at Table 1 Experimental parameters for meteor-head radar observations over Jicamarca Parameter

Mode 0

Mode 1

Interpulse period (km) Code Baud width (km) Altitude coverage (km) Sampling (µs) Nyquist Doppler (km/s) Peak transmitter power Duty cycle (%) Number of receivers Receiver type Processing

200 Barker-13 0.750 80–130 5 ±1.125 2 MW 4.875 4 Analog Automated

60 Barker-13 0.150 80–120 1 ±3.750 2 MW 3.25 3 Digital Manual

163

Table 2 shows that most of our initial observations were concentrated around sunrise periods (i.e., between 0500 and 0800 LT) and observed with Mode 0. As indicated by Chau and Woodman (2004), the automatic processing scheme (Mode 0) was very conservative in the sense that only clear meteor-head echoes were selected, i.e., no meteors were selected when EEJ echoes were present. In order to try to minimize the effects of EEJ echoes, in 2005 we introduced a new mode (Mode 1). The higher range resolution of Mode 1 (150 m compared to 750 m) allowed better discrimination of the meteor-head echoes from other irregularity echoes like EEJ and non-specular meteor trails (e.g., Chapin and Kudeki, 1994). In addition, we implemented a labor intensive but more aggressive manual processing scheme. The manual processing allows us to detect meteor-head echoes occurring: (a) simultaneously but at different altitudes, (b) those below and above EEJ region, and (c) in few but strong-echo cases, those inside the EEJ region. In Table 2 we can see that the percentage of detected echoes with Mode 1 almost doubles those detected with Mode 0 during comparable campaigns. Jicamarca’s interferometry capability allows precise angular measurements of meteoroids along its ionized trajectory. Combined with the classical monostatic radar measurements, we are able to provide, among others, the following meteor parameters: three dimensional geocentric velocities, elevation and azimuth angles of the meteor trajectories, initial altitudes of ionization, decelerations, etc. More details on the measured and estimated parameters, and experimental setup can be found in Chau and Woodman (2004). The measurement procedure of the HPLARs can be summarized as being detectors of meteoroids ionizing inside a small volume around 100 km. The radar scattering is produced from the plasma surrounding the meteoroid as it enters the Earth’s atmosphere/ionosphere. The small volume consists of a narrow truncated cone (or cylinder in the case of Arecibo), defined by the antenna pattern and a properly selected sampled altitude range. For Jicamarca, such truncated cone has ∼2◦ full-width (including the first sidelobe) and a height of 40–50 km around 100–105 km altitude. As we will see in Section 5, we should Table 2 Meteor campaigns at Jicamarca Date

Time (LT)

Mode

%

18-Nov-2001 20-Nov-2001 6-Dec-2001 19-Nov-2002 25-Feb-2003 3-Jun-2003 28-Aug-2003 19-Feb-2004 8-Sep-2004 9-Sep-2004 15-Oct-2004 23-Nov-2005 27-Feb-2006 28-Feb-2006

0.0:8.0 0.1:8.0 5.7:8.0 0.7:8.0 4.2:8.7 3.4:8.6 3.0:8.5 2.5:7.9 2.3:7.3 0.2:6.5 5.9:8.3 −2.8:8.1 15.7:24.0 0.0:17.8

0 0 0 0 0 0 0 0 0 0 0 1 1 1

6.0 8.8 5.1 8.4 6.8 8.8 12.7 4.2 8.4 5.8 1.0 14.0 1.0 9.3

Note. A total of 172,916 meteor heads were observed during ∼90 h of observations spread in 14 days between November 2001 and February 2006.

164

J.L. Chau et al. / Icarus 188 (2007) 162–174

consider as part of the instrument the Earth and the atmosphere, e.g., meteors coming “below” the selected range of the illuminated volume are blocked by the Earth, those coming at low elevation angles have less probability to ionize inside the truncated cone. HPLARs take advantage of the Earth rotation and translation to sample different radiant regions of the celestial sphere, when they occur at favorable locations (sufficiently above the horizon) in the combined Earth–atmosphere–radar instrument. 3. Meteor sources Before determining the radiant, i.e., where meteors are coming from (derived from the meteor vector velocities), we have performed Earth zenith attraction and Earth rotation corrections, to get the undisturbed meteor velocity with respect to the Earth (instantaneous) inertial frame of reference (also called geocentric velocities in the literature). This correction is mainly important for meteors with slow geocentric velocities (the Earth’s escape velocity is ∼11.2 km/s). Note that we have not attempted any other corrections at this point like deceleration by the atmosphere, attraction from giant planets, radiation pressure, etc. In the case of deceleration due to the atmosphere, we have not done such correction since the meteoroid masses are required, a parameter that requires careful analysis and will be left for future work. However, we have noticed that the deceleration is very small or zero at the initial stage of the ionized meteoroid trajectory, particularly for those ionizing at higher altitudes. Nonetheless, the implications due to possible underestimation of the absolute velocities will be discussed below along with the discussion of the expected statistical errors. Then, the corrected meteors are characterized with respect to Jicamarca coordinates (e.g., Chau and Woodman, 2004; Figs. 7 and 8). In order to compare our results with previous works, in the following section we show the results from the perspective of an observer looking towards the Earth’s Apex. The coordinates are expressed in terms of an ecliptic rotating frame of reference where the Sun is always at zero longitude, L = 0◦ , the Apex at L = −90◦ and both at zero latitude (β = 0◦ ), i.e., in the ecliptic plane. In performing the statistics we are assuming that they are independent of the seasonal and time-of-day position of the observer (Earth–atmosphere–radar instrument) when represented in this set of coordinates. Results will be shown before/after the Earth velocity (VEarth km/s) has been removed, i.e., to show them in the Earth’s/Sun’s frame of reference. Using the position and velocity vector of the meteors with respect to the Sun (at ∼1 AU, the exact distance depends on the time of the year), we have solved a well known orbital problem and determined the meteor’s heliocentric orbital parameters (e.g., eccentricity, semimajor axis length, and inclination), for each of the meteors (like shooting the meteors backwards). The inclination of orbits have been used to classify the meteor as prograde for those with inclinations between 0◦ and 90◦ , and retrograde those with inclinations between 90◦ and 180◦ . Similarly heliocentric speeds (i.e., velocities referenced to the Sun) have been compared to the Sun’s escape velocity at ∼1 AU (Vesc ∼ 42 km/s), to separate Solar System (Vesc ) meteors. Here we are using the definition

suggested by Meisel et al. (2002), where all meteors detected with velocities greater than the Sun’s escape velocity are called extra-solar. From this population not all of them are of interstellar origin, perturbations due to solar radiation pressure, solar wind magnetic field deflections and gravitational influence from giant planets (e.g., Jupiter, Saturn, Neptune) as well as expected statistical errors and other velocity corrections, should be carefully considered. The results presented below are obtained from the meteor parameters either directly measured or derived. But what are the implications of underestimation or overestimation of these parameters? In Appendix A we present the expected statistical errors for the elevation angle and the absolute velocity. Since errors in elevation and azimuth angles are small, the effects on the reported sources before VEarth is removed are negligible. These results are independent of the velocity magnitude. On the other hand, results obtained after VEarth is removed will be affected as follows. For overestimated velocities (positive errors), meteor sources will be narrower and closer to the Apex, while for underestimated velocities (negative errors) meteor sources will be wider and farther away from the Apex. Since statistical errors go in both directions, the overall effect will be to make the meteor sources wider. For 10% error in absolute velocity, the meteor source will be less than 5% wider. Similarly if the velocities are underestimated, e.g., due to not including the deceleration of the atmosphere, the resulting sources will be slightly wider and few degrees off the true sources (moving away from the Apex). As indicated in Appendix A, for most meteors, velocity errors are less than 4%. 3.1. Solar System meteor sources In Fig. 1 we show the radiant distributions of meteor heads with solar orbits in the Earth’s frame of reference. These meteors constitute ∼91% of the observed meteors from which 72% present retrograde directions and 28% have prograde directions. We have overlaid six modeled radiant sources with their characteristic ellipses. Each of the sources are characterized by a Gaussian distribution as follows:   (L − L¯ i )2 (β − β¯i )2 − , Si (L, β) = Ai exp − (1) 2 2 2σλi 2σβi where (Li , βi ) are the centers, (σλi , σβi ) the widths, and Ai the relative amplitude for source i. It is clear that source in the Apex part of the sky have the most meteors per degree square. These Apex sources are observed to be composed of a NA and SA source in agreement with previous SMRs observations (e.g., Jones and Brown, 1993; Galligan and Baggaley, 2005). However, our observations indicate that each of the Apex sources are in turn composed of two collocated Gaussian distributions, one almost circular with less amplitude and another one very narrow in ecliptic longitude and wide in ecliptic latitude with larger amplitude. Below we will show that most of these Apex meteors have retrograde directions. From Fig. 1 it appears that only Apex sources are detected, however by showing only the prograde meteors in Fig. 2, we

Sporadic meteor sources observed at Jicamarca

165

Fig. 1. The radiant distribution of all solar meteor heads detected by the Jicamarca HPLAR in the Earth’s frame of reference. The solar (Helion) direction is placed at an ecliptic longitude of 0◦ , and the Earth’s Apex at −90◦ . The color values represent number of meteors per square degree. Six modeled radiant sources are overlaid with the characteristic ellipse appropriate to each source. Note that both Apex sources are composed of two collocated distributions. See text for details.

Fig. 2. Similar to Fig. 1 but only for prograde meteors. Note that the maximum value is two orders of magnitude smaller than that from Fig. 1. In this case only the Helion, Anti-Helion, North, and South Toroidal modeled sources are overlaid.

are now able to see four other sources: Helion, AH, NT, and ST. The Helion and AH sources are symmetric in longitude with respect to the Apex, while the NT and ST sources are symmetric in latitude, in agreement with previous SMRs results. Amplitude differences of these symmetric sources which are expected

to be comparable, are mainly due to the angular sensitivity function of our instrument (see Section 5). Note that so far we have not been able to detect a clear prograde source around the Apex (i.e., meteors coming from the Antapex). Therefore we have not attempted to model such

166

J.L. Chau et al. / Icarus 188 (2007) 162–174

Fig. 3. Similar to Fig. 1 but in the Sun’s frame of reference, i.e., after removing the Earth’s velocity. To show both the retrograde and prograde sources, we have rotated the plot a few degrees. Again, we overlaid the modeled radiant sources with their corresponding characteristics ellipse. Note that the North Toroidal source is not visible with the current color scale.

source. Few meteors from this region have been also observed by the Arecibo HPLAR (e.g., Janches et al., 2000a). Now in Fig. 3 we show the distribution of all meteors but in the Sun’s frame of reference, i.e., after removing the VEarth . In this new frame of reference, we have also modeled the radiant sources with Gaussian distributions, but as expected, with different parameters. This time we are able to see five of the six detected sources, i.e., NA, SA, Helion, AH, and ST. The NT source is there but with less amplitude, not visible with the current color scale. Again the Antapex or Apex-prograde source in this frame of reference, is not well characterized, although there are meteors around that region. The summary of the characteristics of all six radiant sources in both Earth’s and Sun’s frame of reference are given in Table 3. We are including the direction, the number of meteors detected within the characteristic ellipse (in percent), the locations and widths, and the mean values of eccentricity. In addition, although not well identified we are including some characteristics of the Antapex (or Apex-prograde) region. Note that for the Apex sources we provide the widths of the Gaussian distributions with larger amplitude, i.e., the one with narrow width in longitude. In both Apex sources, i.e., NA and SA, the almost circular distributions have the same location as the distributions in the table but with (σλ = 12◦ , σβ = 9◦ ), (σλ = 28◦ , σβ = 18◦ ) for the Earth’s and Sun’s frame of reference, respectively. In both frames of reference, the amplitude of this almost circular distribution are ∼30% the amplitude of the highly elliptical source shape distributions. From the percentage of meteors, we can see that SA population is the highest followed by the NA. The number of detected

meteors from the AH and ST sources are slightly higher than those observed in the Helion region. Antapex and NT meteors constitute less than 0.6%. Comparing the radiant sources results in the Earth’s frame of reference with results obtained by SMRs, e.g., Table 3 in Jones and Brown (1993), and Table 3 in Galligan and Baggaley (2005), we find that: • Excellent agreement is obtained in the longitude location of the Helion, AH, NA, and SA sources, in particular with the corrected distributions of Galligan and Baggaley (2005). • The latitude location is in reasonable agreement for the Helion, AH, ST, and NT sources. • There is poor agreement in general with the latitude location of the Apex sources (NA, SA). Our values (±13◦ ) are between the corrected values given by Galligan and Baggaley (2005) (−10◦ for the SA) and the NA (25◦ ) and SA (−15◦ ) values given by Jones and Brown (1993). Note that the Galligan and Baggaley (2005) results are not accurate for the NA and NT sources due to a poor coverage of northern latitudes. • There is also poor agreement in the widths of the NA and SA populations. We clearly identify highly elliptical sources while SMR results report almost circular ones. • Comparisons of the widths for the other sources require more Jicamarca observations for optimal times, particularly for the Helion, AH, and NT sources, to yield representative results.

Sporadic meteor sources observed at Jicamarca

167

Table 3 Main orbital and location parameters for identified meteor sources Source North Apex South Apex Helion Anti-Helion North Toroidal South Toroidal Antapex

Direction Retrograde Retrograde Prograde Prograde Prograde Prograde Prograde

Eccentricity 0.15–0.2 0.15–0.2 0.8–0.9 0.8–0.9 0.15–0.2 0.15–0.2 0.2–0.3

% 17.5 20.9 0.9 1.0 0.3 1.3 0.6

Earth’s reference ¯ β) ¯ (L,

(σλ , σβ )

Sun’s reference ¯ β) ¯ (L,

(σλ , σβ )

(−89.5, 13.0) (−89.0, −13.0) (−20.0, 0.0) (−160.0, 0.0) (−90.0, 55.0) (−90.0, −55.0) (−90.0, 0.0)

(3,9) (3,9) (12,10) (12,10) (15,10) (15,10) −

(−89.5, 22.0) (−88.0, −22.0) (42.0, 0.0) (138.0, 0.0) (90.0, 50.0) (90.0, −50.0) (90.0, 0.0)

(6,14) (6,14) (20,15) (20,15) (20,17) (20,17) −

¯ β) ¯ and widths (σλ , σβ ) are given in degrees for a modified ecliptic longitude (L) and ecliptic latitude (β), Note. The units of the mean angular location (L, respectively.

Fig. 4. Similar to Fig. 3 but only for extra-solar meteors, i.e., with heliocentric velocities greater than the Sun’s escape velocity. Note that there are no clear clusters of meteors. There are no meteors observed around the Antapex region and there are more meteors observed around the Apex region.

3.2. Extra-solar meteors The distribution of extra-solar meteors is presented in Fig. 4. These meteors represent less than 9% of the total observations. The results are given in the Sun’s frame of reference. As expected no clear clustering is found in this meteor population, they appear to come from almost all directions. However, due to (a) the time-integrated angular sensitivity function of the instrument, (b) the high probability for Jicamarca to detect meteors with high geocentric speeds, and/or (c) the probability that some of these meteors are of solar origin (e.g., Meisel et al., 2002), there are regions with no coverage (e.g., Antapex) or region with more occurrence (around Apex). See Section 5 for details. 4. Speed and initial altitude distributions Now that we have identified different meteor populations, we could proceed to present statistics for any meteor parame-

ter or pair of parameters. However in this work we only present the speed and initial altitude distributions of selected sources. In Fig. 5 we show the geocentric (in black) and heliocentric (in red) speed distributions for the following meteor sources: (a) All retrograde Apex sources, (b) narrow SA, (c) Antapex (Apex-prograde), (d) AH, (e) ST, and (f) extra-solar. Although not shown here, NA and SA distributions are similar. The same occur for the Helion and AH distribution. The results of the NT meteors are not conclusive due to a poor representation (only 0.3% of all the detected meteors). The escape velocity is indicated with a vertical dashed line. We have been tempted to compare our geocentric velocity results with previous works (e.g., Janches et al., 2003; Hunt et al., 2004), but we think those comparisons need to be done carefully. Geocentric velocities are highly dependent on the radar location and the observational time and season. In the case of HPLARs at low latitudes, the distribution of geocentric velocities of meteors observed around noon will be mainly

168

J.L. Chau et al. / Icarus 188 (2007) 162–174

Fig. 5. Distributions of geocentric (black) and heliocentric (red) speed for the selected sporadic meteor sources: (a) all Apex retrograde, (b) South Apex, (c) Antapex (Apex prograde), (d) Anti-Helion, (e) South Toroidal, and (f) extra-solar. The Sun’s escape velocity is shown with a vertical dashed line. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Table 4 Speed (in km/s) and initial altitude (in km) characteristics of selected meteor sources: Mean values, and peak and widths from fitting a Gaussian function around the peaks Source

Geocentric speeds

Heliocentric speeds

Initial altitude

Mean

Peak

Sigma

Mean

Peak

Sigma

Mean

Peak

Sigma

Apex retrograde South Apex Antapex Anti-Helion South Toroidal Extra-solar

57.8 57.8 08.1 25.4 26.0 72.2

57.3 57.4 06.63 24.4 25.9 71.1

3.50 2.65 2.93 5.29 5.26 3.88

29.5 29.1 24.6 31.1 29.6 53.6

28.9 28.8 26.6 33.7 29.6 –

3.32 2.45 4.01 4.15 3.61 –

109.0 108.9 100.2 98.1 98.5 107.9

109.9 109.5 95.7 97.26 98.9 110.2

5.29 5.27 6.45 4.35 5.12 7.04

Note. Speed values are given for both geocentric (Earth velocity removed) and heliocentric distributions.

composed of slow-moving meteors, while those from meteors around sunrise will be mainly composed of high-velocities due to the Apex meteors. Similarly, for HPLARs located at polar regions there will be seasons where the Apex will be below the horizon and the high-velocity meteors will be less or not detected compare to other seasons (e.g., Janches et al., 2006; Fig. 12 for Sondrestorm expected results). However, it is clear that we can see both the high geocentric velocity population due to the Apex sources, and the slow geocentric velocity populations due to the prograde sources, the Antapex meteors being the slowest. In the case of the extra-solar meteors, their geocentric velocities are centered around 72 km/s. We have fitted a Gaussian function to the velocity distributions (when possible) to determine their peak and width. In Table 4 we summarize the mean, peak and width values for each of the distributions. From the analysis of the heliocentric speeds of Table 4, we find that: • The fastest meteors come from the Helion and AH sources (∼33.7 km/s) and the slowest from the Apex (29 km/s) and Antapex (26.6 km/s) sources. Meteors coming from

• •

•

•

the NT and ST have peak velocities in between Helion/AH and Apex sources (∼29.5 km/s). The NA and SA sources, as described by the highly elliptical regions, present a narrower speed distribution than when all Apex meteors are included. Antapex meteors show a wider speed distribution. Most of these meteors have been observed with low elevation angles, as shown in Appendix A, they are expected to have larger statistical errors, explaining the wider speed distribution. The AH and ST meteors show wider speed distributions than the Apex meteors. Similarly, most AH and ST meteors are observed with low elevation angles, therefore with larger statistical errors than the Apex meteors observed usually at high elevation angles. The majority of extra-solar meteors present heliocentric velocities close to the escape velocity. Given that statistical errors could be significant for this classification, particularly those with low elevation angles, some of this population are not of extra-solar origin. Similarly, given that the velocities could be underestimated for not including the atmospheric

Sporadic meteor sources observed at Jicamarca

169

Fig. 6. Distributions of initial altitude for the same sources shown in Fig. 5.

deceleration, some potential extra-solar meteors might not be included in this distribution. In Fig. 6 we show the distribution of the altitude where meteors were initially detected for the same meteor sources used in Fig. 5. The mean, peak and width of these distributions are also shown in Table 4 (again peaks and widths were obtained from fitting a Gaussian function). We can see that Apex distributions peak around 110 km with a half width of ∼5 km. The AH and ST sources peak at a lower altitude (around 98 km) with comparable widths. The Antapex meteors show a peak at the lowest altitude (95.7 km). It is well known that as the meteor geocentric velocity decreases, they tend to ionize at lower altitudes (e.g., Janches and ReVelle, 2005; Close et al., 2004), explaining most of these results. Janches et al. (2003) showed a similar result in their Fig. 9, i.e., meteors with slower geocentric velocities were observed at lower altitudes. The extra-solar meteors peak at similar altitudes as the Apex meteors, but with a wider height range. 5. Discussion As shown in previous works (e.g., Janches et al., 2003; Close et al., 2002; Chau and Woodman, 2004), the HPLARs provide meteor measurements with very high accuracy in velocity, initial altitude, and deceleration, among other parameters. Using interferometry, Jicamarca’s results show that the location (inferred from the velocities) of the meteors on a modified (non-inertial) ecliptic coordinate system are also excellent, being able to observe pretty much all the known sporadic meteors sources with known and new features, based on only few days of observations from a single site. In this section we will limit our discussion to what we observe, i.e., to what the instrument is sensitive to, and compare them with results reported by other instruments. It is important to be aware that HPLARs as well as any other instrument might

have a sensitivity region (e.g., in mass vs velocity space) different than other instruments. Given that most of what we currently know about meteors come from specular meteor radars (SMRs), it is convenient to compare our results with SMR results, at least qualitatively. SMRs are sensitive to the irregularities inside the meteor trail and get strong echoes when pointed perpendicular to these trails. Measurements from these type of systems started in the early 1940’s. Since then many theoretical and empirical corrections have been applied to their measurements in order to improve, among other meteor parameters: meteor sources, geocentric velocity distributions, meteor fluxes, mass estimates, etc. Most of these corrections are associated with instrumental effects (e.g., pulse repetition frequency, transmitting and receiving antenna patterns), to atmospheric and ionospheric effects (e.g., ionization efficiency, Faraday rotation, initial train radius), and also to space effects (e.g., orbital geometry upon detection). Most of these effects are described in detail by Galligan and Baggaley (2004) and Jones and Campbell-Brown (2005) and references therein. In the case of the HPLARs, this type of radars has been used relatively recently for meteor-head observations (e.g., PellinenWannberg et al., 1998; Sato et al., 2000; Close et al., 2002; Janches et al., 2003; Chau and Woodman, 2004). They are sensitive to the sharp discontinuity in plasma density at the meteor head, thus to the meteoroid position and dynamics. They are also sensitive to irregularities in their trail, but these echoes can be easily distinguished due to their slower Doppler velocities. Most of these alternative echoes are referred to as non-specular trail echoes and are being studied separately. Contrary to the SMR observations, very little work has been done to improve the direct HPLAR measurements, except for the work of Hunt et al. (2004) who tried to correct the measured geocentric velocity distribution taking into account assumed sensitivity biases. It is important to mention that some SMRs are also able to measure few meteor-head echoes or so called down-the-beam

170

J.L. Chau et al. / Icarus 188 (2007) 162–174

Fig. 7. Time-integrated angular sensitivity function for the Earth–atmosphere–radar instrument at Jicamarca in the Earth’s frame of reference. The simple function indicated in Eq. (2) has been used for the observational periods indicated in Table 2.

echoes (e.g., Elford, 2001). Similarly, HPLARs are also able to detect few specular echoes. In both cases, those unusual echoes are not considered in their analysis, for example at Jicamarca specular echoes are removed from the meteor-head statistics. Instead of correcting the direct measurements, Janches and Chau (2005) and Janches et al. (2006), more recently, have added atmospheric and ionospheric effects into a Monte Carlo model with expected meteor sources and fluxes. These modelling efforts aim to understand the meteor populations as a function of radiant location, time of day and season, by comparing the outputs of the model to the diurnal and seasonal characteristics of meteor heads detected by HPLARs. Janches et al. (2006) found excellent agreement with Arecibo diurnal and seasonal meteor observations introducing not only the Apex source as originally done by Janches and Chau (2005) but also the Helion and AH sources. They also found very good agreement with diurnal curves from Jicamarca and Sondrerstorm observations. In order to arrive at these results, they introduced an empirical atmospheric filtering effect that prevented meteors from low elevation angles (below 20◦ ) being detected by the HPLRs. Such effect is probably produced by a combination of factors related to the interaction of the meteor with the air molecules such electron production and/or ablation at higher altitudes. In this work we invoke mainly two effects that will help us understand our observations and possibly would allow us to do qualitative comparisons with SMR results. The first effect is a simple atmospheric filtering or what we have called an instrumental angular sensitivity function. The second effect is the ionization efficiency which favors meteors with high geocentric speeds. Note that other effects or improved versions of what we are doing in this paper, might bring the comparisons closer, par-

ticularly quantitative ones (e.g., strengths of meteors sources), but that will be left for a future work. The simple angular sensitivity function represents the effects of the atmosphere/ionosphere and Earth around our instrument. Due to more efficient ablation at higher altitudes, meteors coming from low elevation angles are expected to have less probability of detection since they would have ablated at higher altitudes. Meteors coming below the horizon but above the tangent to solid Earth, are not expected to be detected by our truncated narrow cone around 100 km due to a longer path of denser atmosphere. Similarly, the Earth will block meteors coming from below the illuminated volume (below the tangent to solid Earth). Moreover, we expect this function to be smooth (without discontinuities) and independent of azimuth direction. The simple function we are using is 1 − cos f (θ ) (2) 2 for 0◦  θ < 180◦ and 0 otherwise, where θ is the elevation angle of the meteor trajectory as it enters the Earth’s atmosphere, and f (θ ) is function of θ . In this work we have used f (θ ) = 2θ . Our function in a way is similar to the atmospheric filter used by Janches et al. (2006) but much simpler, since it only depends on the elevation angle of the meteor trajectory. For example, we do not include other meteor parameters like initial altitude or mass. Fig. 7 shows the time-integrated angular sensitivity function in the Earth’s frame of reference. It can be interpreted as the total count of meteor observations that we would have observed if the meteors had a uniform distribution in latitude and longitude before being “filtered” by the Earth–atmosphere, considering only the times of the day and the days listed in A(θ ) =

Sporadic meteor sources observed at Jicamarca

171

Fig. 8. Similar to Fig. 1 but only for meteors detected below 100 km. In addition the distribution has been corrected (divided) by the time-integrated angular sensitivity function shown in Fig. 7.

Table 2. The count has been normalized with respect to the maximum. A quick analysis of the time-integrated function, indicate that: • Our observations are heavily weighted towards the southern latitudes around the Apex longitude. • The SA has more weight than the NA source. • The AH source is also weighted more than the Helion source since more observations have been performed around midnight than around noon. • The ST source has been observed more than the NT source, explained by positive radiant latitudes being further away from Jicamarca’s zenith during most of the observational periods. • Except for few regions of the sky (in white), from Jicamarca we have been able to cover most of the celestial sphere. We are aware that this is a simple procedure, but it help us to understand important features of our observations. A more sophisticated formalism including physical arguments is being currently conducted with Jicamarca and Arecibo data (D. Janches, personal communication, 2006). Dividing the meteor distributions in Fig. 1 by the timeintegrated function of Fig. 7, the resulting distribution shows that the NA and SA Apex have the same amplitudes, as expected. Similarly the AH and Helion sources have similar amplitudes, as well as the ST and NT sources. However, the Apex source amplitude (or number of meteors) are still significant larger (15–20 times) than that from the other sources. Uncorrected SMR observations also show Apex sources with larger amplitudes than the other sources but by a factor of 2 (e.g.,

Galligan and Baggaley, 2005; Table 2). We can explain this apparent disagreement as follows. It is well known that typical SMRs (with wavelengths around 3–15 m) are not sensitive to the meteors with high geocentric velocities, since these meteors ionize at higher altitudes and their initial train radii are larger, causing an attenuation due to destructive interference (e.g., Steel and Elford, 1987; Galligan and Baggaley, 2004; Jones and Campbell-Brown, 2005). Therefore SMRs are more sensitive to the slower populations. The altitude distributions of our slow velocity populations are in very good agreement with the altitude distribution of typical SMRs meteors (e.g., Elford, 2001; Galligan and Baggaley, 2004). Moreover if we do a geocentric speed distribution of meteors below 100 km, the result is bimodal (i.e., slow and fast populations), and very similar to the one shown by Galligan and Baggaley (2004, Fig. 27). Taking into account the time-integrated angular sensitivity function and the altitude selection of SMRs, in Fig. 8 we show the meteor distribution ionizing at altitudes below 100 km, corrected (divided) by the time-integrated angular sensitivity function. The resulting distributions show comparable amplitudes for all sources, and in the case of the SA, NA, ST, and NT sources, they present their mean latitude center at higher latitudes. These preliminary qualitative results with simple corrections are encouraging, since they have better agreement with the corrected results obtained with SMRs. The ionization sensitivity of HPLARs is discussed in detailed in Janches and ReVelle (2005) and Close et al. (2004). Faster geocentric meteors will ionize at higher altitudes and will have more chance to be observed by the HPLRs, but not by the SMRs due to the difference in the scattering mechanisms they are sensitive to. Given that the Earth velocity is

172

J.L. Chau et al. / Icarus 188 (2007) 162–174

comparable to the heliocentric velocities of all the sporadic meteors, then it is expected that meteors coming from the Apex with suitable geometries (e.g., medium to high elevation angles) will have more probability of being detected. This effect explains why HPLRs observe many more meteors coming from the Apex than SMRs, particularly at Arecibo, ALTAIR and Jicamarca radars which are located at low geographic latitudes, and therefore with convenient elevation angles. The sensitivity of Jicamarca to the higher altitude echoes, allow the observations of the highly elliptical Apex sources on top a more circular source, so far not reported by any other instrument. Therefore, HPLARs in addition to observing some of the meteor population observe by SMRs, also observe the faster meteors ablating at higher altitudes. Similar conclusions were postulated by Janches et al. (2003) based on monostatic HPLAR observations at Arecibo. Another strong evidence of the ionization effects in HPLRs, that allow them to see meteors not detected by the SMRs, is the distribution of extra-solar meteors shown in Fig. 4. The distribution of extra-solar echoes are expected to be isotropic, particularly those visiting the Solar System for the first time (interstellar), since they do not know where the ecliptic plane or Earth’s Apex are. Our results indicate that the observed extra solar meteors are observed with greater probability coming from the Apex. But this can be equally explained by the fact that meteors coming from the Apex directions will have higher geocentric velocities and therefore they have higher probability to be detected. A deeper analysis of the extra-solar meteors, for example of those that are of interstellar origin after considering the expected statistical errors, will be a left as a future work. The Jicamarca results could help improve the understanding of the theoretical and empirical models used for HPLARs meteor observations. For example, we plan to continue the work initiated by Janches et al. (2006) and do similar Monte Carlo simulations with more sources (they used a single Apex source and the AH and Helion sources) and study the model response to each of the sources one at the time. In addition, it will be important to find empirically and theoretically, the meteor distributions as a function of absolute velocity and mass in order to determine the region in this parameter space, to which Jicamarca is sufficiently sensitive. Although we expect that HPLARs are sensitive to the meteor-head echoes in front of the meteor trails that SMRs are sensitive to, a more quantitative work, including meteor mass estimations, is needed, for example to decide if there is overlap in their sensitive regions, or if these techniques complement to each other.

and North Toroidal. These sources have been previously identified by combining information from different specular meteor radars (SMRs) and other instruments (e.g., satellites, optical instruments). Based on our observations we have modeled these sources with two dimensional Gaussian distributions with different longitudinal and latitudinal widths. The comparison of the meteor sources in modified ecliptic coordinates, speed and initial altitude distributions for different sources indicate that Jicamarca and therefore the other HPLARs (e.g., Arecibo, ALTAIR) are sensitive to at least some meteors usually detected by SMRs. In addition, HPLARs are sensitive to the highest velocity meteors mainly occurring at altitudes above 100 km, complementing SMRs. For example, this is the first time Apex sources (north and south) with very narrow width in ecliptic longitude and wide width in ecliptic latitude are reported. These two sources are superimposed on top of a weaker and more isotropic sources that are in better agreement with SMR results. Discrepancies between symmetrical sources, i.e., North vs South Apex, Helion vs Anti-Helion, and North vs South Toroidal, are shown to result from the time-integrated angular (elevation) sensitivity function of the Earth–atmosphere–radar instrument. Although we have shown, qualitatively, that Jicamarca observations are in very good agreement with SMR observations (location and shape of sources), a careful quantitative analysis, for example, to compare the actual strengths of meteor sources, will be left for future work.

6. Concluding remarks

V=

Using the Jicamarca (11.95◦ S, 76.87◦ W, 1◦ dip angle) high-power large-aperture radar (HPLAR), we have detected 170,000 meteor heads in less than 90 h of observations spread in 14 days, between November 2001 and February 2006. From these observations 91% are of from the Solar System and the remaining appear to be extra-solar. The solar meteors are mainly clustered in six main sources, i.e., South and North Apex, Helion, Anti-Helion, and South

where Vr , α, R0 , and R are the radial velocity (m/s), angular coverage (rad), initial range (m), and range coverage (m), respectively. The expected errors in R0 and R (∂R0 and ∂R) depend on the radar precision and are constant (75 or 350 m, depending on the mode, for Jicamarca meteor experiments). On the other hand, since Vr and α are obtained from phase of time and space complex correlation functions, respectively, their statisti-

Acknowledgments We thank P. Brown, D. Holdsworth, D. Janches, and M. Oppenheim for their comments, suggestions, and encouragement during different stages of this work. The Jicamarca Radio Observatory is a facility of the Instituto Geofísico del Perú and is operated with support from the NSF Cooperative Agreement ATM-0432565 through Cornell University. Appendix A. Expected statistical errors on derived meteor parameters Chau and Woodman (2004) show that the elevation angle and the absolute velocity of meteor-head echoes are derived quantities from directly measured parameters, as follows: θ = arctan Vr , sin θ

R , R0 α

(A.1) (A.2)

Sporadic meteor sources observed at Jicamarca

173

typical altitudinal ranges. Similarly meteors coming with low elevation angles are not expected to be seen with small angular coverage, since they will be confused with specular meteors occurring at only one range. In general, errors for most meteors are below 4%. At low elevation angles, errors in absolute velocity could be greater than 20%. References

Fig. 9. Expected statistical errors in absolute velocity meteor-head measurements (in percent) as function of elevation angle. Different colors represent values for different SNRs, while different symbols represent different radial velocities. Small discontinuities represent the use of different angular coverage values, depending on the elevation angles.

cal errors (∂VR and ∂α), are proportional to 1 (1 + 1/SNR)2 − ρ 2 , 2M ρ2

(A.3)

where SNR, M, and ρ are the signal to noise ratio, number of independent samples, and coherence, respectively (Woodman and Hagfors, 1969). Propagating the expected errors, we found that the expected statistical errors of the elevation angle and absolute velocity are approximately given by  2  2  2  sin θ cos θ   cos2 θ  sin θ cos θ   ∂R0  +  ∂θ ≈  ∂R  +  ∂α  , R0 R0 α α   2   cos θ 2  1 ∂V    ≈  ∂Vr  +  ∂θ  . |V | Vr sin θ 

(A.4) (A.5)

For typical meteor-head observations at Jicamarca the errors in elevation angles are between 2◦ and 5◦ . From (A.5) even small errors in θ produce huge percentage errors in velocity at low elevation angles. Fig. 9 shows the expected errors for typical observational parameters. Values for different SNRs are represented with different colors, while errors for different radial velocities are represented by symbols. Note that at high elevation angles, percentage velocity errors are greater for smaller radial velocities. In addition, we have used reasonable angular coverage values depending on the elevation angle. For example, at high elevation angles, we do not expect large angular coverage since that would mean meteors ionizing outside

Chapin, E., Kudeki, E., 1994. Radar interferometric imaging studies of longduration meteor echoes observed at Jicamarca. J. Geophys. Res. 99, 8937– 8949. Chau, J.L., Woodman, R.F., 2004. Observations of meteor-head echoes using the Jicamarca 50 MHz radar in interferometer mode. Atmos. Chem. Phys. 4, 511–521. Close, S., Hunt, S.M., McKeen, F.M., Minardi, M.J., 2002. Characterization of Leonid meteor head echo data collected using the VHF–UHF advanced research projects agency long-range tracking and instrumentation radar (ALTAIR). Radio Sci. 37, doi:10.1029/2000RS002602. Close, S., Oppenheim, M., Hunt, S., Coster, A., 2004. A technique for calculating meteor plasma density and meteoroid mass from radar head echo scattering. Icarus 168, 43–52. Elford, W.G., 2001. Novel applications of MST radars in meteor studies. J. Atmos. Sol. Terr. Phys. 63, 143–153. Farley, D.T., 1985. Theory of equatorial electrojet plasma waves: New developments and current status. J. Atmos. Sol. Terr. Phys. 47 (8), 729–744. Galligan, D.P., Baggaley, W.J., 2004. The orbital distribution of radar-detected meteoroids of the Solar System dust cloud. Mon. Not. R. Astron. Soc. 353, 422–446. Galligan, D.P., Baggaley, W.J., 2005. The radiant distribution of AMOR radar meteors. Mon. Not. R. Astron. Soc. 359, 551–560. Hunt, S., Oppenheim, M.M., Close, S., Brown, P., McKeen, F., Minardi, M., 2004. Determination of the meteoroid velocity distribution at the Earth using high-gain radar. Icarus 168, 34–42. Janches, D., Chau, J.L., 2005. Observed diurnal and seasonal behavior of the micrometeor flux using the Arecibo and Jicamarca radars. J. Atmos. Sol. Terr. Phys. 67, 1196–1210. Janches, D., ReVelle, D., 2005. The initial altitude of the micrometeor phenomenon: Comparison between Arecibo radar observations and theory. J. Geophys. Res. 110, doi:10.1029/2005JA011022. A08307. Janches, D., Mathews, J.D., Meisel, D.D., Getman, V.S., Zhou, Q.-H., 2000a. Doppler studies of near-Antapex UHF radar micrometeors. Icarus 143, 347–353. Janches, D., Mathews, J.D., Meisel, D.D., Zhou, Q.H., 2000b. Micrometeor observations using the Arecibo 430 MHz radar. I. Determination of ballistic parameter from measured Doppler velocity and deceleration results. Icarus 145, 53–63. Janches, D., Nolan, M.C., Meisel, D.D., Mathews, J.D., Zhou, Q.H., Moser, D.E., 2003. On the geocentric micrometeor velocity distribution. J. Geophys. Res. 108, doi:10.1029/2002JA009789. Janches, D., Heinselman, C.J., Chau, J.L., Chandran, A., Woodman, R.F., 2006. Modeling the global micrometeor input function in the upper atmosphere observed by high power and large aperture radars. J. Geophys. Res. 111, doi:10.1029/2006JA011628. Jones, J., Brown, P.G., 1993. Sporadic meteor radiant distribution: Orbital survey results. Mon. Not. R. Astron. Soc. 265, 524–532. Jones, J., Campbell-Brown, M., 2005. The initial train radius of sporadic meteors. Mon. Not. R. Astron. Soc. 359, 1131–1136. Mathews, J.D., 2004. Radio science issues surrounding HF/VHF/UHF radar meteor studies. J. Atmos. Sol. Terr. Phys. 66, 285–299. Meisel, D.D., Janches, D., Mathews, J.D., 2002. Extra solar micrometeors radiating from the vicinity of the local interstellar bubble. Astrophys. J. 567, 323–341. Pellinen-Wannberg, A., Westman, A., Wannberg, G., Kaila, K., 1998. Meteor fluxes and visual magnitudes from EISCAT radar event rates: A compari-

174

J.L. Chau et al. / Icarus 188 (2007) 162–174

son with cross-section based magnitudes estimates and optical data. Ann. Geophys. 116, 1475–1485. Sato, T., Nakamura, T., Nishimura, K., 2000. Orbit determination of meteors using the MU radar. IEICE Trans. Commun. E83-B, 1990–1995. Steel, D.I., Elford, W.G., 1987. The height distribution of radio meteors: Observations at 2 MHz. J. Atmos. Sol. Terr. Phys. 49, 243–258.

Taylor, A.D., Elford, W.G., 1998. Meteoroid orbital element distributions at 1 AU deduced from the Harvard radio meteor project. Earth Planets Space 50, 569–575. Woodman, R.F., Hagfors, T., 1969. Methods for the measurement of vertical ionospheric motions near the magnetic equator by incoherent scattering. J. Geophys. Res. 74, 1205–1212.