Accepted for publication in Meteoritics & Planetary Sciences, September 2, 1999

First results from video spectroscopy of 1998 Leonid meteors

J. Borovicka, R. Stork, and J. Bocek

Astronomical Institute of the Academy of Sciences of the Czech Republic, 251 65 Ondrejov Observatory, The Czech Republic

Abstract. We report spectroscopic observations of meteors made from the FISTA aircraft on November 17, 1998 as a part of the Leonid Multi-Instrument Aircraft Campaign. Low resolution spectra of 119 meteors of apparent visual magnitudes from +3 to –4, corresponding to meteoroid masses from 10–6 to 10–3 kg, were obtained. After analyzing a representative sample of the spectra and comparing them to the spectra of Perseids from the Ondrejov archive, the following conclusions were reached: Leonid meteoroids are very loose and disintegrate easily in the atmosphere. This leads to much faster evaporation of volatile sodium than of other elements, an effect which is not observed in Perseids. Relative bulk abundances of Mg, Fe, Ca, and Na in Leonids are nearly CI-chondritic within the uncertainty of the method (factor of three). Smaller meteoroids tend to be poorer in sodium, which is true also for Perseids. Most meteoric vapor emissions could be reasonably well explained with the temperature of 4500 K. High temperature meteoric emissions (Ca+, Mg+) are present only in bright meteors. Leonid spectra are very rich in atmospheric emissions of O, N, and N2 – even at high altitudes and in faint meteors. These emissions are therefore not connected with meteor shock wave. Thermal continuum is also present in the spectra. Organic material was not revealed.

1. Introduction

Meteor spectroscopy enables the study of chemical composition and structure of interplanetary bodies (meteoroids) entering the atmosphere of the Earth, the process of the meteoroid-atmosphere interaction and the deposition of meteoric atoms in the atmosphere. Although the precision of chemical analysis made this way is much lower than laboratory studies of meteorites and interplanetary dust particles (and the analysis is restricted to elements observable in the given spectral range), meteor spectroscopy has the advantage that it is usable for all types of asteroidal and cometary material coming to the Earth. Moreover, the origin of a studied meteor is often apparent from meteor trajectory and heliocentric orbit. This is always true for meteor showers with a known parent body.

Up to now, compositional analyses were restricted to bright photographic meteor spectra, typically of magnitude of –10 corresponding to meteoroid mass of 50 g or more (see Ceplecha et al. 1999 for a review). A number of effects complicating the compositional analysis was identified. Two spectral components with temperatures of about 4500 and 10 000 K were identified in spectra of bright meteors (Borovicka 1993, 1994a). The high temperature component is much stronger in fast meteors. Refractory elements (Ca, Al, Ti) were often found to be evaporated incompletely during the atmospheric entry and thus depleted in the radiating gas. Nevertheless, the depletion does not occur in fast meteors (Borovicka and Betlem 1997) and in slow meteors at very low altitudes (Borovicka and Spurny 1996). The chemical composition of more than 90% meteoroids was found to be nearly chondritic (Borovicka 1994b, 1999).

The predicted rich occurrence of Leonid meteors in 1998 represented a clear opportunity to learn more on meteoroids ejected from comet Tempel-Tuttle. A video spectroscopy experiment was therefore included in the 1998 Leonid Multi-Instrument Aircraft Campaign (Jenniskens and Buttow 1999). We have obtained a number of spectra of Leonid meteors from +2 to –6 absolute stellar magnitude. Such meteors are produced by meteoroids of the masses of the order of 10–3 to 1 g, i.e. millimeter to centimeter size (for assumed density of 1 g cm–3). By a statistical way, we have been able to obtain for the first time some quantitative information on the composition of such small meteoroids. A number of spectra of meteors with identical velocity enabled us to study also the dependence of meteor spectrum on meteoroid mass.


Fig. 1: Relative spectral sensitivities of the video camera with Dedal image intensifier (bold curve) and the SET vidicon TV camera. Both curves have been normalized to unity at 550 nm.


2. Observation and data calibration

The spectroscopy of Leonid meteors presented here was carried out from the FISTA aircraft flying at altitudes between 11.3 and 13 km around Okinawa island (see the article by Jenniskens and Buttow in this issue for details). Our instrumentation consisted of a commercial video camcorder Panasonic NV-S88E, a second generation Russian image intensifier Dedal 41 equipped with photographic lens Arsat 1.4/50 mm, and a spectral grating Milton Roy with 600 grooves per mm blazed to 470 nm. The grating was mounted in front of the lens. This set was placed on a tripod just behind a single-layer circular glass window of the aircraft (diameter of 30 cm, BK7 glass transparent down to 300 nm). The axis of the window pointed only 12.5 degrees above horizon but the camera was adjusted to look higher (30–40° ). The grating was rotated by hand after each aircraft maneuver to adjust the direction of dispersion to be nearly perpendicular to the direction of flight of Leonid meteors.

The video signal was recorded in the S-VHS PAL system giving 25 full frames per second. The record was searched later by human inspection and all recorded spectra were digitized with a Miro DC-30 framegraber on a PC computer. After digitization, each frame was transformed into a 768´ 576 pixel / 8 bit monochrome image. Individual spectra have been stored as sequences in the AVI/DIB (Audio Video Interleave/Device Independent Bitmap) format.

The system provided a circular field of view with diameter of 25° , somewhat cut at the top and bottom. High order spectra of bright meteors located up to 50° from the center of the field of view in the direction of dispersion could be detected. On the other hand, only zero order images were recorded for meteors appearing near the center of the field of view. The resulting dispersion was 1.15 nm/pixel in the first order and the field of view therefore represents 750 nm across (for the first order).

An effective spectral sensitivity curve was constructed by measuring a number of spectra of bright stars taken from the aircraft. The curve is shown in Fig. 1 by a bold line. The maximum sensitivity lies near 550 nm. The sensitivity curve of the vidicon camera used at the Ondrejov Observatory (Borovicka and Bocek 1995) is also shown. That camera was used to take the spectra from the Ondrejov archive mentioned below. Both cameras are sensitive in the 380–880 nm range. The vidicon camera has better performance in the blue part, the image intensified camera is more sensitive in the infrared part of the spectrum.

The observations were done between 15–21 UT on November 17, 1998. During this period, 119 meteor spectra were obtained, at least 98 of them being Leonids. 28 more spectra were obtained on other November nights but no observations were made on November 16, when the Leonid component rich in bright meteors peaked (Arlt 1998). Using the stellar spectra calibration, the faintest recorded spectra were calculated to belong to meteors of apparent visual magnitude of +3.

Fig. 2: Example of one video frame containing the Leonid spectrum SZ 82. The meteor was moving from upper right to lower left. The short lines following that direction represent individual spectral lines (in fact, monochromatic images of the meteor). The length of the lines is equal to the distance traveled by the meteor during the 0.04 s exposure time. The bright quasi-continuous emission belongs to the N2 molecule. Wavelengths increase from the left to the right. The bright streak at upper right is the spectrum of the star a Aurigae.

For this study we selected 20 Leonid, 2 Taurid, and 4 sporadic spectra. The selection preferred spectra covering large wavelength intervals at whole or at majority of the meteor trajectory. Brightness was not the criterion. In addition, 7 Perseid spectra obtained with the vidicon camera at the Ondrejov Observatory in 1997–1998 were used for comparison. 7 Orionid spectra were also selected from the Ondrejov archive but they are relatively faint and were used only for some purposes.

The list of Leonid meteors studied in this paper is given in Table 1. The time of appearance on Nov 17 (± 2 s), the spectral orders in which the meteor was recorded, the slope of the trajectory (i.e. elevation of the radiant), and the approximate elevation of the meteor above horizon are given. The last two columns contain the apparent maximum visual (more precisely V-band) magnitude derived from the spectrum and the approximate absolute (100-km distance) magnitude. The absolute magnitude was computed using meteor distance estimated from meteor elevation. Actual meteor distances and heights are not known at the moment. In future it will be possible to compute trajectories of three meteors which were recorded simultaneously by the HD-TV cameras onboard the Electra aircraft (H. Yano, private communication). The meteors from the list were also captured by non-spectral video cameras onboard FISTA . The estimates of their brightness made to ± 1 mag by P. Jenniskens (private communication) agree within ± 1.5 mag with the apparent magnitudes from Table 1 for most meteors. Three meteors disagree but this may be due to incorrect identificaction.

As an example, one frame of the spectrum SZ 82 is shown in Fig. 2. This spectrum of a medium brightness meteor covers the whole wavelength range. The raw image is shown. All raw images were subject to dark frame subtraction and flat-fielding using a flat image taken from the aircraft during twilight. The spectra were then extracted from the images using a self-written software for slot-scanning along a selected path. To

account for the radiation of the sky and background stars, the same path was scanned also on frames taken shortly before the meteor appearance and both scans were subtracted.

Table 1: Leonid meteors analyzed in this study

No     time     order     slope     elev.     mag     Mag
UT abs
SZ 21 16:13:12     0,+1   14°     45°     –1.3     –1.9
SZ 22 16:26:37     –2     16°     11°     –2.3     –5.9
SZ 34 17:14:51     +1     27°     30°     +2.2     +0.9
SZ 37 17:19:54     0,+1   28°     29°     +0.8     –0.7
SZ 43 17:36:10     –1,–2  32°     10°     +0.2     –3.5
SZ 48 17:46:10     +1     34°     57°     +0.4     +0.2
SZ 66 18:54:18     +1     49°     42°     +1.0     +0.3
SZ 70 18:58:25     0,+1   50°     36°     +1.5     +0.5
SZ 72 19:12:41     0,+1   53°     35°     –0.4     –1.4
SZ 77 19:19:29     +1     55°     60°     –0.9     –1.1
SZ 82 19:29:15     +1     57°     51°     +0.7     +0.3
SZ 91 19:50:51     –1,–2  62°     18°     +0.6     –1.9
SZ 94 19:56:49     0,+1   63°     35°     –3.6     –4.6
SZ 99 20:07:36     0,+1   66°     26°     –2.8     –4.5
SZ 102 20:09:40    0,+1   66°     50°     +1.6     +1.2
SZ 103 20:10:48    –1,–2  66°     17°     –2.7     –5.3
SZ 109 20:20:23    –1,–2  68°     31°     +0.9     –0.4
SZ 113 20:29:48    +1     71°     44°      0.0     –0.7
SZ 116 20:31:39    +1     71°     32°     –1.5     –2.7
SZ 120 20:37:46    +1     72°     27°     +2.4     +0.8

The wavelength scale was determined using known lines and the zero order image if present. The observed intensities at different wavelengths were corrected according to the spectral sensitivity curve (Fig. 1). In case of spectra captured in other than the blazed first order (order +1), the relative sensitivity in the respective order to the order +1 as a function of wavelength was taken into account. We used spectra of an emission line lamp to determine relative sensitivities in different spectral orders. The sensitivity in order –1 (non-blazed) proved to be only 5% of the order +1 at 400 nm, but 50% at 850 nm. For order –2 the value lies between 10–15% in the whole wavelength range, while the sensitivity in order +2 drops from 15% at 400 nm to 5% at 500 nm. Bright features in bright spectra had also to be corrected for system non-linear response since very strong signals saturated the image intensifier.

After applying all the above corrections, the intensities were converted to absolute units (W m-2 nm-1) by multiplying by a factor determined from spectra of bright stars of known magnitudes. For the purpose of comparing meteors at various distances from the observer, the intensities were converted to the luminosity at the source (W ster-1 nm-1) using the meteor distance estimated from its elevation above the horizon.


3. General appearance of the spectra, spectral components

A typical medium brightness Leonid spectrum SZ 82 is shown in Fig. 3. The upper part represents the spectrum as observed, in the lower part the spectrum is shown after calibration for spectral sensitivity. The brightest lines lie in the infrared region and belong to atomic oxygen and nitrogen. Lines of O, and N together with the first positive bands of the nitrogen molecule N2, which are also quite pronounced, represent most of the energy radiated between 380–880 nm. These emissions come from the heated atmosphere. Meteor vapors are represented by the lines of Mg, Na, Fe and Ca. The brightest meteoric feature is probably formed by a group of Mg and Fe lines near 380 nm but these lie at the edge of sensitivity and cannot be studied well in our spectra. Well observable is the (unresolved) Mg I triplet at 517 nm and Na I dublet at 593 nm. Meteoric emissions represent about 1/6 of the energy, atmospheric emissions form 2/3, and the remaining 1/6 is supplied by thermal continuum.


Fig. 3: The spectrum SZ 82 (summed from all frames) with the identification of main features. The upper panel shows the spectrum as it was observed. The lower panel shows the spectrum after calibration for spectral sensitivity of the camera (see Fig. 1). The dashed line represents thermal continuum.

Table 2: Elements and multiplets (numbers according to Moore 1945) seen in Leonid spectra.
element     major             faint
            multiplets        multiplets
O         1,4,9,10,11,12;1F     13,14,15,16,17,18,21
N         1,2,3                 8,21,24
N2        1st positive
Mg        2                     1,9,11
Mg+       4
Na        1
Fe        41                    1,2,15,42,43,318
Ca        2
Ca+       1


In Table 2, the elements unambiguously detected in Leonid spectra are listed. For each element, numbers of detected multiplets (Moore 1945) are given. Individual lines within multiplets are not listed because they are usually unresolved. The lines of ionized Ca and Mg (at 395 nm and 448 nm, respectively, not present in Fig. 3) belong to the high temperature component, other meteoric emission belong to the low temperature component. Some additional elements, e.g. Cr, probably contribute to some features in the spectra but remain unresolved. The spectrum of a bright sporadic meteor SZ 117 (–4 mag) contains lines of ionized Si, ionized Fe and hydrogen Ha in the 2nd and 3rd order, all belonging to the high temperature component. No bright Leonid spectrum was conveniently placed to show these lines.

Most surprising in the Leonid spectra is the presence of many oxygen lines including some obscure high excitation multiplets. Nitrogen lines are not so numerous. A substantial fraction of the N remains probably in molecular form. A comparison of Leonids with other meteors of similar magnitude showed that Perseids and Taurids contain lower amount of atmospheric emissions, about 1/2 and 1/4, respectively. This can be explained by their lower velocity – the velocity is 71 km s-1 for Leonids, 60 km s-1 for Perseids and 33 km s-1 for Taurids. In Taurids, N and O lines are very faint and the atmospheric emissions are almost exclusively due to the N2 molecule.

The main meteoric component can be satisfactorily described by a temperature between 4000 and 5000 K. The value of 4500 K seems to be the best. The column density of Fe and Mg atoms was found to be the order of 1014 cm–3. To explain the observed absolute brightness of the Mg and Fe lines in a meteor of stellar magnitude zero (such as SZ 82), the cross-section of the radiating volume containing Mg and Fe must correspond to the diameter of the radiating volume of about 10 cm. This is one or two orders of magnitude more than the size of the meteoroid itself. However, meteoroids are probably disintegrated into a swarm of microscopic fragments at the time of maximum brightness and may occupy relatively large volume. Note also that the number of metal atoms involved in the radiation is so large that any contribution from metal atoms pre-existing in the atmosphere is quite insignificant.

The meteor observations done so far show that the temperature of the main meteoric component does not exhibit any strong dependence on meteoroid mass and velocity. The radiation probably originates in a region in the vicinity of, or just behind, the ablating body. Frequent collisions among atoms constitute conditions close to thermal equilibrium. The intercombination lines (Fe I multiplets 1 and 2 and Mg multiplet 1), which are often strongly enhanced above their equilibrium intensities, probably originate at the outskirts of this region and/or in the wake further behind the body where the atomic collisions are much less frequent.

The second, high temperature, component was tentatively identified with meteor shock wave and both meteoric and atmospheric emissions were assigned to this component by Borovicka (1994a). The current observation show that this picture must be partly revised. The atmospheric lines (N,O) behave differently from the meteoric lines of the high temperature component (Mg+, Ca+). Atmospheric lines are present even in faint meteors and also at the beginning of the trajectory. They are therefore not a part of the meteor shock wave which develops only in brighter meteors and at lower heights. The observations show that atmospheric lines are excited also in the free molecular regime by direct collisions between molecules and atoms. Since all atmospheric lines have relatively high excitation potential (> 10 eV), they are fainter in slow meteors where the energy of the collisions is smaller. In contrast to atmospheric lines, the high temperature meteoric emissions develop only in relatively bright meteors (see Sect. 5). Whether or not they are connected with the shock wave remains an open question.

The present observations also show that the role of a thermal continuum in meteor radiation is larger than previously thought. Thermal continuum is usually not detected in photographic spectra of not-very-bright meteors. Low resolution video spectroscopy is, however, more suitable for detection of continuous radiation and the present spectra show that continuum is present. The signal at some wavelengths, especially at 450–480 nm cannot be explained by atomic and molecular emissions only. We fitted the continuum with a 4500 K blackbody which resulted in no conflict with observation but the information on the continuum is limited and the real temperature may well be different. Part of the continuum can be produced by the heated surface of the meteoroid but this source is insufficient and the gas phase probably contributes as well. The observed continuum intensity in zero magnitude Leonids corresponds to radiation of a 0.1 cm2 black-body surface at 4500 K or of 20 cm2 at 2500 K. Thermal continuum is present in Perseids and Taurids too.

Leonids are also rich in meteor trails. The short duration trail is due to the well known green forbidden line of oxygen at 557.7 nm. This line reach its maximum intensity with some delay (» 0.1 sec) after the meteoroid passage at a given location and then decays gradually for 1–2 sec. No spectrum of a long duration meteor train (> 3 s) was captured.

A search for organic carbon bearing molecules CN and C2 was unsuccessful. Although the presence of these bands was suspected at the beginning of few meteors, the bands could not be confirmed on the following frames and were probably not real. The search was negative also in Perseids. Little organic material was found in meteor spectra up to now – relative strong CN bands were detected in only one faint Orionid meteor (Stork et al. 1999). Marginal detection of CN and C2 was reported by Ceplecha (1971).


Fig. 4: Temporal evolution of spectrum SZ 82. The spectra extracted from individual video frames are shown as individual curves from the top to the bottom. The consecutive spectra have been shifted vertically by 500 units for clarity. Three brightest lines are identified.


4. Temporal evolution of the spectra and the structure of meteoroids

All spectra were measured on each video frame, i.e. with time resolution of 0.04 seconds. The temporal evolution of spectrum SZ 82 is shown in Fig. 4. The most striking feature is a peculiar behavior of the Na line. Sodium is well visible at the beginning (t=0.04 s), together with O and N2. Magnesium is still missing at that time but later it brightens more rapidly than Na. After the maximum, Mg and most other lines fade gradually, but Na disappears abruptly between t=0.20 and 0.24 s, when other lines are still quite bright.


 Fig. 5: Intensity of the Na (thin curve) and Mg (bold curve) line in 12 Leonid meteor spectra as a function of time. Time zero corresponds to the maximum brightness of the magnesium line. Individual spectra were offset by two units vertically for clarity and are marked by their numbers (see Table 1).

To evaluate this effect, monochromatic light curves in Na and Mg lines were constructed for more Leonids and for other meteors. The Leonid curves are shown in Fig. 5. The effect is present in all Leonids of low and medium brightness. Typically, Na line intensity drops by an order of magnitude at the time when Mg is still bright. The intensities shown after the drop are in fact upper limits because the Na line becomes fainter than neighboring bands of N2 and cannot be measured precisely. Meteor SZ 34 is somewhat different and shows a Na decrease from the beginning. The length of this meteor is related to its rather horizontal trajectory.

In Leonids brighter than about magnitude –2 (represented by SZ 22, 99, 103, and 116 in Fig. 5), Na is well visible until the end of the trajectory and the abrupt drop does not occur. Nevertheless, also here the Na/Mg ratio decreases along the trajectory.

Figure 6 shows the Mg-Na light curves for 5 Perseids, 3 Orionids, 2 Taurids, 1 Geminid and four sporadic meteors. Meteors designed with a TVS number were taken with the vidicon camera. From the 15 meteors, only the sporadic TVS 264 shows the same behavior of Na as Leonids. A somewhat rapid decrease of Na is seen also in the Orionid TVS 125. No Perseid shows similar effect.


 Fig. 6: The same as Fig. 5 for non-Leonid meteors.


The Na drop typical for Leonids is therefore only rarely seen in other meteors. A literature search revealed another interesting case. Millman (1972) reported the Na line to be shifted up in one photographic spectrum of a Draconid meteor taken during the meteor storm in 1946. The brightness of the meteor was about –2 mag. The Draconid meteors are a prototype of the most fragile meteoroids of type IIIB as defined by Ceplecha and McCrosky (1976). Meteor storms contain meteoroids relatively recently released from the parent comet.

Sodium is the most volatile of the observed elements. Its absence in the terminal part of the trajectory suggests that it was already lost from the whole meteoroid volume. Alternative explanations such as changes in temperature and/or ionization have no support in other lines of the spectrum. The Na line vanishes very quickly, often within few pixels in one video frame. The meteoroids exhibiting the Na drop probably have different structure and can be probably best described in terms of the dustball model of Hawkes and Jones (1975). In that model, meteoroids are composed of grains held together by a low boiling point material. When the melting point of that volatile material is reached, grains are released. Hawkes and Jones (1975) assumed no light to be produced by the volatile 'glue'. In our interpretation, sodium is a part of the glue and the glue therefore contributes some light when it is being evaporated independently on grains (more quickly). Larger meteoroids disintegrate later and keep a part of the volatile material until the terminal part of the trajectory.




Fig. 7: Multiplet intensity ratios as a function of intensity of magnesium multiplet 2. All intensities have been integrated along meteor trajectories and summed for all lines of a multiplet. The multiplet numbers are given in parentheses in the axis labels. For N2, the intensity of the band near 659 nm was used. Multiplet 10 at 618 nm was used for O because the brightest multiplet 1 at 777 nm is overexposed in some spectra and lies far from other lines. Spectral sensitivity of the camera was taken into account when computing intensity ratios. Intensities are in arbitrary units and are displayed in logarithmic scale. The error of individual points is about ± 0.1 for Na and N2, ± 0.15 for O, and ± 0.25 for other ratios. An empirically found approximate conversion of Mg intensities to meteor maximum magnitudes is given on the upper axes. The dashed lines represent average values for Leonids.

The difference of Leonids from Perseids suggests either a different structure of parent comets or different processes involved in the ejection of meteoroids from the comets or different history of meteoroids in interplanetary space. The possible similarity of Leonids to Draconids may favor the last alternative – younger meteoroids being more similar to the dustball model.

McNeil et al. (1998) concluded that differential meteoroid ablation is responsible for the distribution of metal atoms in the upper terrestrial atmosphere and for the underabundance of atomic Ca relative to Na in the atmosphere. Although the effect found in Leonids can clearly be called differential ablation, we do not consider it to be responsible for the atmospheric metal distribution. Leonids form only a tiny part of incoming cosmic material and most of other meteors behave differently. The effect found by Borovicka (1993) and called incomplete evaporation is probably the main mechanism causing the differentiation of Na and Ca. Incomplete evaporation takes place in slower meteors which represent a majority of the incoming material.

5. Integrated line intensities and elemental abundances

Only bright lines can be measured with reasonable signal-to-noise ratio on individual frames of faint meteors. To obtain information on more lines, the spectra extracted from individual frames were summed and total line intensities integrated along meteor trajectories were measured.

The ratios of intensities of some representative lines to the magnesium line are shown in Fig. 7. The ratios are displayed as a function of the intensity of the Mg line (i.e. El/Mg versus Mg). The Mg line was chosen because it is a bright line near the middle of the visual passband and is measurable in most spectra. Its intensity is, of course, proportional to the total meteor brightness and to the meteoroid mass. In each plot, one point represents one meteor. Leonids, Perseids and other meteors (Orionids and Geminids are not included) are displayed with different symbols. The plot thus enables to study line ratios for as a function of meteoroid mass in different showers. Of course, not all lines were measurable in all spectra.

No difference is seen between Leonids and Perseids in the total Na/Mg, Ca/Mg and Fe/Mg ratios. This suggests that the difference in behavior of the Na line is due to the different meteoroid structure and not by different bulk composition. Interestingly, in both Leonids and Perseids, smaller meteoroids tend to have somewhat lower content of Na. Other meteors vary in Na content strongly, one sporadic meteor (SZ 9) was especially poor in sodium. For Ca and Fe the differences are not so large, the overall scatter is caused by the relative faintness of the Ca and Fe lines.

Only the Mg+ /Mg ratio shows strong dependence on meteoroid mass. The ionized Mg line is a part of the high temperature meteoric component. The importance of this spectral component therefore grows with meteoroid mass. The dependence on velocity, which is also known to exist (Borovicka 1994a), is not clearly seen from the plot.

The N2/Mg and O/Mg ratios do not depend on mass but are dependent on velocity, which is demonstrated by lower ratios in Perseids and Taurids than in Leonids. This means that for a given velocity the amount of heated atmosphere is directly proportional to the meteoroid mass and the proportionality factor increases with meteor velocity.

The average mutual ratios of Na, Fe, Mg, and Ca line intensities have been converted to relative abundances of these elements. Thermal equilibrium (TE) at a temperature of 4500 K and a column density of Mg I and Fe I atoms of the order of 1014 cm-2 were assumed for the main spectral component. This explained reasonably the intensities of most meteoric lines during the bright phase of the meteors. Besides the intercombination lines, only the lines of Fe I multiplet 318 and Mg I multiplets 9 and 11 are brighter than expected for these parameters. This means that either TE is not strictly valid or some unrecognized emissions mix with these relatively faint lines.


From the mean multiplet intensity ratios plotted in Fig. 7, we derived the following atomic ratios in the radiating gas of Leonid meteors: Fe I/Mg I = 0.4; Ca I/Mg I = 7´ 10–4; and Na I/Mg I  = 4–6´ 10–4 (lower for faint meteors, higher for bright meteors). Owing to all uncertainties of the procedure used, the estimated error of these values is by a factor of three. Since sodium and calcium are highly ionized at 4500 K, the Saha equation (see e.g. Mihalas 1978, chapter 5) had to be used to convert neutral atom ratios to elemental ratios. The ionization degree depends also on density. Fortunately, changes of density affect all elements in nearly the same way for a given temperature and the density estimate is therefore not so critical. We estimated the density of Fe atoms in the radiating gas to be 5´ 1012 cm–3 and free electron density to be 7´ 1012 cm–3.

Table 3: Elemental ratios by number in different sources.
logarithm   video             photographic C1-chondrites   comet
of ratio    Leonids           Perseids                     Halley
Fe/Mg     –0.5 ± 0.5         –0.3 ± 0.4     –0.08         –0.3 ± 0.1
Ca/Mg     –1.7 ± 0.5         –1.6 ± 0.4     –1.24         –1.2 ± 0.1
Na/Mg     –1.5 to –1.3 ± 0.4 –1.3 ± 0.3     –1.27         –1.0 ± 0.2



The resulting elemental abundances are given in Table 3 together with the values derived for photographic Perseids (Borovicka and Betlem 1997), carbonaceous chondrites (Anders and Grevesse 1989) and comet Halley dust (Jessberger et al. 1988). Within the errors, the values for Leonids are not distinct from other values. Slightly higher abundance of Mg is suggested but this is probably insignificant. Thanks to their large velocity, neither Leonids nor Perseids are affected by an incomplete calcium evaporation.

Information on more elements might be obtained by high dispersion photographic spectroscopy, which, however, is possible only for bright meteors. The main aim of the video campaign was to reveal possible individual differences among faint meteors. The Leonids analyzed so far form a relatively homogeneous sample. Only the spectrum SZ 34 shows a different Na light curve. Faint meteors generally tend to be poorer in Na by about 30%. This is true for both Leonids and Perseids. The analysis of the available Orionid spectra suggest that the depletion of Na is even more severe in faint Orionids. A tentative explanation of this effect may be that volatile Na is lost from a thin surface layer of meteoroids when they are orbiting Sun. That process may be reflected in the formation of sodium tails of comets (Cremonese et al. 1997). Such a loss of Na would be, of course, more important for small meteoroids.


6. Conclusions

The present analysis confirms that video spectroscopy is a useful method of studying faint meteors inaccessible for other methods. The main findings concerning the Leonids are the following:

  1. The Mg, Fe, Ca, Na ratios are generally consistent with normal cosmic composition of Leonids. Small meteoroids, however, are poorer in Na by about 30%. This is true also for Perseids and even to larger extent in Orionids.
  2. Leonids show preferential ablation of Na on the upper part of the trajectory. This points out their fragile structure. In this respect Leonids are distinct from Perseids and most of other meteors, they may be, however, similar to Draconids.
  3. The search for organic molecules (CN, C2) in Leonid spectra was unsuccessful. At least some Orionids seem to be richer in carbon.
Conclusions which are probably general for all meteors have been also reached, namely the importance of atmospheric radiation in even faint fast meteors, the non-connection of atmospheric radiation with meteor shock wave, the dependence of the strength of the high temperature meteoric spectral component on meteoroid mass, and the importance of thermal continuum in meteor spectra.

We plan to analyze more spectra of Leonids as well as other meteors from Ondrejov archives by the procedures developed for this paper. Further observational campaign is planned for 1999 Leonids and regular observations at the Ondrejov Observatory will also continue.

Acknowledgements. This work is based on the data obtained during the NASA's 1998 Leonid Multi-Instrument Aircraft Campaign. We thank Peter Jenniskens for inviting us to the participation on this campaign and for his support. On the Czech side the work was supported by project no. 205/99/0146 from the Grant Agency of the Czech Republic.



Anders E. and Grevesse N. (1989) Abundances of the elements: Meteoritic and solar. Geochim. Cosmochim. Acta 53, 197–214

Arlt R. (1998) Bulletin 13 of the International Leonid Watch: The 1998 Leonid meteor shower. WGN, Journal of IMO 26, 239–248

Borovicka J. (1993) A fireball spectrum analysis. Astron. Astrophys. 279, 627–645

Borovicka J. (1994a) Two components in meteor spectra. Planet. Space Sci. 42, 145–150

Borovicka J. (1994b) Meteor spectra – possible link between meteorite classes and asteroid families. In Seventy-Five Years of Hirayama Asteroid Families (eds. Y. Kozai, R.P. Binzel, and T. Hirayama), Astron. Soc. Pacific Conf. Series 63, 186–191

Borovicka J. (1999) Meteoroid properties from meteor spectroscopy. In Meteoroids 1998 (ed. V. Porubcan), Slovak Acad. Sci., Bratislava, Slovakia (in press)

Borovicka J. and Betlem H. (1997) Spectral analysis of two Perseid meteors. Planet. Space Sci. 45, 563–575

Borovicka J. and Bocek J. (1995) Television spectra of meteors. Earth, Moon, and Planets 71, 237–244

Borovicka J. and Spurny P. (1996) Radiation study of two very bright terrestrial bolides and an application to the comet S-L 9 collision with Jupiter. Icarus 121, 484–510

Ceplecha Z. (1971) Spectral data on terminal flare and wake of double-station meteor No. 38421 (Ondrejov, April 21, 1963). Bull. Astron. Inst. Czech. 22, 219–304

Ceplecha Z. and McCrosky R.E. (1976) Fireball end heights: a diagnostic for the structure of meteoric material. J. Geophys. Res. 81, 6257–6275

Ceplecha Z., Borovicka J., Elford W.G., ReVelle D.O., Hawkes R.L., Porubcan V., and Simek M. (1998) Meteor phenomena and bodies. Space Science Reviews 84, 327–471

Cremonese G., Boehnhardt H., Crovisier J., Rauer H., Fitzsimmons A., Fulle M., Licandro J., Pollacco D., Tozzi G.P., and West R.M. (1997) Neutral sodium from Comet Hale-Bopp: A third type of tail. Astrophys. J. 490, L199–L202

Hawkes R.L. and Jones J. (1975) A quantitative model for the ablation of dustball meteors. Mon. Not. R. astr. Soc. 173, 339–356

Jenniskens P. and Buttow S. (1999) The 1998 Leonid Multi-Instrument Aircraft Campaign. Meteorit. Planet. Sci., this volume

Jessberger E.K., Christoforidis A. and Kissel J. (1988) Aspects of the major element composition of Halley's dust. Nature 332, 691–695

McNeil W.J., Lai S.T. and Murad E. (1998) Differential ablation of cosmic dust and implications for the relative abundances of atmospheric metals. J. Geophys. Res. 103, D3, 10899–10911

Mihalas D. (1978) Stellar Atmospheres. Freeman & Co., San Francisco, USA

Millman P.M. (1972) Giacobinid meteor spectra. J. Royal Astron. Soc. Canada 66, 201–211

Moore C.E. (1945) A multiplet table of astrophysical interest. Contr. Princeton Univ. Obs. 20

Stork R., Borovicka J., Bocek J. and Solc M. (1999) TV spectra of shower meteors. In Meteoroids 1998 (ed. V. Porubcan), Slovak Acad. Sci., Bratislava, Slovakia (in press)