ASTROBIOLOGYVolume 4 Number 1 2004copy Mary Ann Liebert Inc

Research Paper

Meteor Wake in High Frame-Rate ImagesmdashImplicationsfor the Chemistry of Ablated Organic Compounds

PETER JENNISKENS1 and HANS C STENBAEK-NIELSEN2

ABSTRACT

Extraterrestrial organic matter may have been chemically altered into forms more ameanablefor prebiotic chemistry in the wake of a meteor after ablation We measured the rate of cool-ing of the plasma in the meteor wake from the intensity decay just behind a meteoroid byfreezing its motion in high frame-rate 1000 framess video images with an intensified cam-era that has a short phosphor decay time Though the resulting cooling rate was found to belower than theoretically predicted our calculations indicated that there would have been in-sufficient collisions to break apart large organic compounds before most reactive radicals andelectrons were lost from the air plasma Organic molecules delivered from space to the earlyEarth via meteors might therefore have survived in a chemically altered form In addition wediscovered that relatively small meteoroids generated far-ultraviolet emission that is absorbedin the immediate environment of the meteoroid which may chemically alter the atmosphereover a much larger region than previously recognized Key Words Prebiotic chemistrymdashOri-gin of lifemdashMeteorsmdashMeteor wakemdashTemperature Astrobiology 4 95ndash108

95

INTRODUCTION

INFALLING EXTRATERRESTRIAL MATTER has longbeen argued to be an important source of pre-

biotic molecules for the origin of life (eg Pon-namperuna 1981 Thomas et al 1997) Oroacute andLazcano (1997) have discussed the long history ofthe field Much effort has been made to identifyprebiotic molecules in meteorites and interplan-etary dust particles which are available for studyin the laboratory In a new chapter to the storywe argue that the more important source of or-ganic molecules is the meteoric matter ablated inthe Earthrsquos atmosphere (Jenniskens et al 2000a

Jenniskens 2001) Ablation in the Earthrsquos atmos-phere is the fate of the bulk of infalling organicladen extraterrestrial matter (Ceplecha 1992Love and Brownlee 1993) Because atmosphericchemistry in the meteor and ambient environ-ment can chemically change those ablation prod-ucts into forms more amenable for prebioticchemistry this can be a very significant processof relevance to astrobiology

In an earlier paper (Jenniskens et al 2004a) weargued that much of the organic matter may leavethe meteoroid in the form of relatively large compounds This paper addresses the questionwhether ablated organic compounds can survive

1Center for the Study of Life in the Universe SETI Institute Mountain View California2University of Alaska Fairbanks Alaska

the chemistry with reactive species in the airplasma in a chemically altered but still reducedform It has always been assumed that the organicmatter is consumed by oxygen atoms to form CO2and water However the concentration of reac-tive species is time and temperature dependentHence an important parameter for understand-ing the fate of exogenous organic matter is aknowledge of the cooling rate of the plasma es-pecially in the wake of relatively small mete-oroids of size 150 mm at the peak of the massinflux curve (Ceplecha 1992 Love and Brownlee1993)

Our current understanding of how meteors de-posit cometary and asteroidal material in the at-mosphere and how its chemical composition ischanged during ablation and subsequent chem-istry in the air plasma in the wake of the mete-oroid is primarily based on spatially unresolvedoptical and radar observations (Oumlpik 1958 Bron-shten 1983 Ceplecha et al 1998) Immediatelybehind the meteoroid an air plasma is created byair molecules that are accelerated by collidingwith the meteoroid and its ablation vapor cloudand decelerated by subsequent collisions with theambient air environment Order of magnitude es-timates on the rate of elastic and inelastic colli-sions within the air plasma are obtained frommeasurements of the trail width (Oumlpik 1958Hawkes and Jones 1978) In contrast the loss oftranslational energy in the wake of meteoroidshas recently been calculated using a DirectMonte-Carlo Simulation model that described therarefied flow around a 1 g Leonid (Boyd 2000Popova et al 2000 Jenniskens et al 2000a) Thatmodel predicts a time-dependent temperature(T t212) from about 6300 K at 10 m behind themeteoroid to 3400 K at a distance of 40 m Thenew model shows that the ablation vapor cloud(due to molecules and atoms sputtered off sur-rounding and traveling with the meteoroid) de-termines to a large extent the size of the wake bydramatically increasing the collision cross sectionof the meteoroid

While the process of excitation in suprathermalcollisions implies a lack of thermodynamic equi-librium the typical air plasma vibrational andelectronic excitation temperatures and chemicalequilibrium temperatures measured for such me-teors all suggest T 4300 K a value surprisinglyindependent of meteoroid mass and speed (Jen-niskens et al 2004b)

This paper deals with the temperature decay

following that excitation The wake of actual me-teors is complicated by unknown infrared and ul-traviolet radiative cooling mechanisms nonequi-librium chemistry in the meteor wake (Park andMenees 1978) fragmentation of the meteoroid(Hawkes and Jones 1978 Murray et al 1999 vonZahn 2001) and by fragments ejected at highspeed from a spinning grain (Le Blanc et al 2000Taylor et al 2000)

Until now the only useful measurement of thetemperature decay in the meteor wake was thatby BorovicIuml ka and Jenniskens (2000) who discov-ered an afterglow of cooling meteoric plasma inthe wake of a bright Leonid fireball with a cool-ing time much slower than predicted by theMonte-Carlo Simulation model In at least onepart of the meteoroid trajectory the afterglowwas associated with secondary ablation fromsolid debris detected by way of spectroscopic ev-idence

Measurements of meteor afterglow in faintermeteors are hampered by motion blurring andoverexposure In conventional imaging the rapidmotion of the meteoroid blurs all morphologicaldetail Instantaneous photography using rapidlyrotating shutters has detected meteor wake out to150 m behind the meteoroid (Halliday et al1980) but it is not very sensitive and is limited tometeors so bright that they do not characterizethe typical rarefied flow conditions of small me-teoroids Moreover instantaneous photographydoes not provide the high temporal resolutionthat can be achieved with todayrsquos high frame-rateimagers

Here we report high frame-rate imaging of arelatively faint 23 magnitude Leonid meteor froma ground location during the 2001 Leonid Multi-Instrument Aircraft Campaign (MAC) (Jenniskensand Russell 2003) The most striking aspect of theimages is the unexpected development of a bow-shock-like structure and a spherical luminosity be-fore the shock which is discussed elsewhere (Sten-baek-Nielsen and Jenniskens 2003) This paperdiscusses the wake in the meteor images

METHODS

The observations were made at the Universityof Alaskarsquos Poker Flat Research Range 30 milesnortheast of Fairbanks AK (6512N 14746W and039 km altitude) with an intensified charge cou-pled device (CCD) camera operating at 1000

JENNISKENS AND STENBAEK-NIELSEN96

framess The intensifier used was sensitive tolight in the wavelength range 500ndash900 nm withpeak sensitivity at 700 nm where strong atomicline emissions of oxygen and nitrogen and theFirst Positive band emission of N2 are present inmeteor spectra The intensifier phosphor had abrief decay time constant of 08 ms ideally suitedto study the natural afterglow in rapidly movingtargets

The CCD images were 256 3 256 pixels with256 gray levels (8 bits) and 64 3 64deg field of viewTo facilitate the high frame rate each image quad-rant was read out in parallel through separateelectronics Because of small unavoidable differ-ences between the four-quadrant electronicssome differences in intensity were expected to bevisible in the images

At 1000 framess gray level 255 was reachedat a surface brightness of 3 Mega-Rayleigh (at 700nm) which is about 25 of the CCD well depthTo further prevent blooming of the images thegain of the intensifier was set for saturation to oc-cur only at a higher brightness level Images werecontinuously entered into the high-speed imager(HSI) which was equipped with a 4000-frame cir-cular digital buffer (ie the buffer can contain 4 sof data) Upon recognizing an event the operatorwould then intervene and save a selected se-quence of the buffer to the computer disk storage

The HSI was bore-sighted with a conventionalwide-field low-light-level TV system which pro-vided a more classical video image of the meteorand a larger star field for orientation purposesThis system also consisted of an intensified CCDbut with a 50 ms phosphor decay time and thedata were recorded on videotape with standardNTSC resolution (5994 interlaced fieldss and avertical resolution of 525 horizontal lines) GlobalPositioning System time was encoded on eachfield The field of view was 21 3 16deg and the in-tensifier responded to wavelengths of 400ndash800 nm

RESULTS

On the night of November 18 universal time(UT) 2001 the weather was mostly clouded Af-ter 10 UT the eastern sky cleared Near 1020 UTa persistent meteor train was recorded on thewide-field imager for almost 20 min At 104259UT a bright Leonid meteor passed through thefield of view of both imagers This was close tothe peak of the Leonid storm at 1040 UT which

was caused by Earthrsquos crossing of the 1767 dusttrail of comet 55PTempel-Tuttle (Kondratrsquoevaand Reznikov 1985 Jenniskens 2003)

The wide-field imager

The wide-field imager captured most of themeteor (Fig 1) which moved from the bottom tothe top across the central part of the 21 3 16degwide field over 29 frames (a little less than 1 s)The spatial resolution of the wide field camerawas fairly similar to that of the HSI which cov-ers the central 64 3 64deg of the field Howeverthe frame rate was lower and the meteor movedsignificantly across each frame during the expo-sure Also the bright emissions caused the de-tector to saturate and ldquobloomrdquo These effects com-bined to make it impossible to resolve any of thestructures within the head of the meteor A hazeor thin clouds was present in the wide-field im-ages For ground-based observations at maxi-mum instrument gain and with a clear transpar-ent dark sky we would expect to see stars up tomagnitude approximately 185 The limiting stel-lar magnitude in the images was slightly less than17 A persistent train would have been muchbrighter but it was not detected

The meteor in the narrow-field HSI

Figure 2 shows representative frames acquiredfrom a 400-ms HSI sequence of this event Eachimage revealed an 094 3 094deg section of the orig-inal 64 3 64deg image centered on the meteor Themeteor entered the field of view at frame 60within the sequence The meteor was initially apoint source saturated at the center and slightlybroader than the unsaturated star images withonly a faint trace of wake The onset of subse-quent features was gradual The selected framesin Fig 2 are those in which a new feature is wellpronounced Around frame 170 two unresolvedlines developed at the tail of the meteor whichgrew around frame 235 into a distinct spatialstructure reminiscent of a shock front Initiallyonly a diffuse triangular-shaped wake was visi-ble between the two fronts (frame 200) Aroundframe 235 a wake developed inside the shockfront that persisted for a relatively long time Atthe same time the meteor brightness increaseddramatically The emission on the outside of theldquoshockrdquo structure gradually brightened filling analmost perfect circle of light with a piece cut out(to which we will refer as the ldquobite-outrdquo) At the

METEOR WAKE IN HIGH FRAME-RATE IMAGES 97

same time the ldquoshockrdquo opened up to about 45degThe structure was fully developed around frame330 This is the first time such spatial structurehas been reported The HSI recording ended atframe 463

The spatial structure was not the result of op-tical reflections in the camera We noted that theimages were already distinctly asymmetric whenthe meteor passed very close to the optical axisof the camera (frame 300) where any instrumen-tal distortion of light would be expected to besymmetric We did however find evidence ofscattered light Centered on the meteoroid posi-tion there appeared to be an underlying diffuseglow that was much wider than any spatial struc-ture observed in the images This diffuse glowmay have been scattered light by thin clouds inthe field of view

The centers of the images were saturated How-ever the blooming of the brightest part remainedmodest throughout the exposure period and wasmuch less than the observed spatial structureThis implies that the intensity of the meteor at itspeak was at most a factor of a few brighter thanthe saturation level

The wake

Following the images with a small delay wasa wake of emission that persisted over the dura-tion of the exposure The characteristic delayidentified this wake as being caused by the for-bidden green line emission of O I at 557 nm (Hal-liday 1958b Baggaley 1976 1977) That O I wakewas clearly seen as well in the TV imager (Fig 2)where it remained visible for about 2 s

A second wake phenomenon became visible at lower altitudes which was identified (as dis-cussed below) as the meteor afterglow resultingfrom collisional excitation of metal atoms and airplasma compounds (Halliday 1958a BorovicIuml kaand Jenniskens 2000) This feature was first seenaround frame 300 and is a tail of emission grow-ing from the saturated part of the meteor imageIf this tail was due to phosphor decay from over-exposure of the meteor then there would havebeen an increase in blooming which was not ob-served Indeed the tail faded slower than ex-pected for an overexposed phosphor

We conclude that during the development ofthe tail the meteor did not brighten enough to

JENNISKENS AND STENBAEK-NIELSEN98

FIG 1 Conventional intensified TV images of the event The field of view is 21 3 16deg and the HSI (Fig 2) cov-ers the central portion of the field The first image shows the meteor just before it entered the HSI field of view Thetwo next are within the HSI field of view and the last just after the meteor exited the HSI field of view

overexpose the phosphor The meteor did notleave what is called a persistent train a chemilu-minescence from the catalytic recombination ofoxygen atoms and ozone molecules which wouldbe expected if the Leonid meteor had beenbrighter than magnitude 24 (Jenniskens et al2000b) Such persistent trains have peak emissionin the center of the systemrsquos response curve andare characterized by an apparent brightness ofapproximately 14 magnitude at this spatial res-olution (Jenniskens et al 2000c)

ANALYSIS

Geometry of the observations

Both data sets had a sufficient number of starsdistributed across the images to permit good spa-tial calibration Using computer routines devel-oped for analysis of image data in connectionwith auroral research (Stenbaek-Nielsen et al1984) stars in the Smithsonian Astronomical Ob-servatory star catalog were overlaid and fitted tothe stars present in the images The fitting pro-gram is interactive and can accommodate variousdisplay formats as well as nonlinearity associatedwith the optical system used in the imager The

resulting star fit provided directional informationfor all pixels within the individual images

The center location of the meteor in each im-age was measured for all images in the HSI andthe wide-field video data sequences A computerprogram was written to calculate the location ofthe meteor in each image within the two imagesequences using a fixed meteoroid velocity vec-tor and an initial starting point derived from themeteor location in one of the early HSI imagesThe motion across the field depends upon the me-teoroid angular velocity the assumed range tothe selected initial point and to a lesser degreethe radiant

The meteoroidrsquos velocity vector given by theLeonid shower radiant position measured at thesame time in Arizona by multistation photogra-phy in a Leonid MAC-related effort was 1541 602deg right ascension and 214 6 01deg declination(H Betlem personal communication) The speedwas 716 6 04 kms which is sufficiently high sothat changes due to the Earthrsquos gravitational fieldcould be neglected At the time of the event104259 UT the radiant was at 865deg azimuth (eastof north) and 223deg elevation The camera orien-tation was 761deg azimuth and 539deg elevation re-sulting in an angle between camera orientation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 99

FIG 2 Sections of the original highframe-rate images Each is a 094 3 094degsection from an original 64 3 64deg im-age which shows the development ofthe meteor morphology The imageswere recorded at 1000 framess Theframe number within the sequence isshown in the upper right corner The cal-culated position of the meteoroid (seetext) is shown by a dot

and the velocity vector varying from 28deg to 33degacross the field of view Thus structures alongthe trajectory in the images were foreshortenedby roughly a factor of 2 The pixel size in the high-speed images is 0025 3 0025 At a range of 135km (altitude of 110 km) the spatial resolution was57 mpixel This corresponds to the distance trav-eled by a meteor during the time its image on thephosphor screen fades by a factor e With the me-teor trajectory at an angle of 30deg to the line ofsight the 1-pixel resolution along the trajectorywas 114 m In 1 ms the Leonid moved 72 m Thusthe meteor moved less than 1 pixel betweenframes and the phosphor intensity decayed by86 from one pixel to the next Consequentlythere was little ldquosmearingrdquo of spatial structuresdue to the motion of the Leonid meteor

The calculated positions were fitted to the ob-served positions by adjusting the assumed rangeto the initial point and the Leonid velocity vec-tor A very good fit was obtained for a right as-cension of 1540deg and a declination of 1214degwhich is in agreement with the position measuredphotographically A small systematic deviation of1 pixel in the y direction later in flight was dueto the difficulty of choosing the center of the im-age once the ldquoshockrdquo forms Indeed the differ-ences between calculated and observed positionsin both data sets (Fig 3) were similar to the ac-curacy by which the center could be determinedand hence the analysis showed the meteoroid ve-locity was constant during the time covered bythe optical observations This finding agrees withSpurny et al (2000) who reported that the decel-eration of similar Leonid meteors was below mea-surement accuracy in photographic data

The positions for selected frames of the highframe-rate imager are given in Table 1 The me-teor entered the wide field of view camera at analtitude of 1229 km It brightened very rapidlyin the next two frames The meteoroid entered theHSI field of view at an altitude of 1156 km (frame65) The bow shock and wake started to developat about 1104 km The final HSI image with theshock fully developed was at 1044 km when themeteor brightness started to level off A largerfraction of the path was recorded in the wide-field TV imager which showed that the meteorhad a broad maximum in luminosity centered atan altitude of 1006 km after which the meteordecreased in intensity It left the TV camera fieldof view at an altitude of 95 km when it was stillof magnitude 11 The position analysis is rela-

tively insensitive to the assumed meteor velocitysince varying the assumed range to the initialpoint can compensate for any error in velocitywithin a reasonable range For example if the ve-locity decreased by 1 kms the altitude of thepath would decrease by 2 km

Relative brightness

The meteorrsquos light curve (Fig 4) was derivedfrom the integrated intensity of both the HSI andTV images Both data sets were processed differ-ently in response to the amount of blooming withoverall good agreement to within 05 magnitudesThe HSI data in Fig 4 show the integrated in-tensity over an area of 21 3 21 pixels centered onthe position of the meteoroid This covered the

JENNISKENS AND STENBAEK-NIELSEN100

FIG 3 Differences between observed and calculatedpositions of the meteor image centerThe meteor crossedthe field from the bottom to the top and hence dy is es-sentially the difference along the track while dx is thecross track difference The abscissa is the altitude derivedfor each frame while the ordinate is pixels The observedcenter for the high frame-rate images was estimated to anaccuracy of 02 pixels while the TV record has an accu-racy of about 1 pixel

meteor head but not much of the tail It did notinclude intensity lost in the central saturated pix-els but added intensity from the spatially ex-tended component and the diffuse scattered lightThe TV data were integrated over a 65 3 71 pixelrectangle that covered the full size of the bloomedarea Saturation effects were corrected for bytreating the measured intensity as an opticaldepth In the video there is a substantial back-ground signal which was subtracted by com-puting the average signal in two similar-size rec-tangles located on either side of the meteor To

increase its temporal resolution we measured atthe beginning of data collection the intensity vari-ation across the meteor images in each 130 s ex-posure while the blooming was still modest (solidline in Fig 4) Each data set was calibrated to theV magnitude of the stars in the field of view overthe range V 5 131 to 161 magnitude For sim-ilar cameras it was found that blooming offsetthe effects of saturation over a much wider mag-nitude range the relationship between bloomingand saturation being close to that expected if theelectron production is linear with incident lightbut the electrons distribute over more pixels V 525 log SIpixel (Jenniskens 1999) Indeed the ex-pected intensity in the central pixels of the HSIwas not much above the measured level thus ex-plaining the lack of blooming The slightly higherbrightness may reflect the fact that the HSI im-ager is sensitive to wavelengths longer than theV band which overestimates the brightness in thecalibration procedure However the limitednumber of calibration stars introduces a similarsystematic uncertainty

At the peak of its brightness the meteor hadan absolute magnitude (ie as seen from a dis-tance of 100 km) of 227 6 04 magnitude In-deed a meteor of this brightness would not havea bright persistent train Other Leonid meteors of227 magnitude have their peak brightness at97 6 4 km in agreement with the value of 1006km found here (Jenniskens et al 1998 Betlem etal 2000) On video records with limiting magni-tude approximately 16 such Leonids are usuallydetected first at altitudes of 143 km but theirbrightness does not increase to the photographiclimiting magnitude of about 11 until they decend

METEOR WAKE IN HIGH FRAME-RATE IMAGES 101

TABLE 1 OBSERVED AND CALCULATED POSITIONS OF THE METEOR IN THE HSI DATA

Calculated Observed

Frame x y x y Range Latitude Longitude Altitude

65 1552 14 1552 14 1493 65281 2145492 1156100 1488 188 1489 188 1471 65280 2145541 1146150 1396 446 1396 446 1440 65279 2145610 1132200 1300 714 1299 715 1409 65279 2145679 1118250 1198 996 1198 996 1378 65278 2145749 1104300 1090 1294 1092 1290 1348 65277 2145818 1090350 980 1610 981 1598 1317 65276 2145887 1076400 862 1932 865 1920 1287 65275 2145957 1062450 740 2270 743 2257 1258 65274 2146026 1048463 708 2358 711 2347 1250 65274 2146044 1044

xy are pixel location in the images with (00) in lower left The range is in km from the camera location while latitude and longitude are in geographic degrees N and E Altitude is in km

FIG 4 Light curve of the meteor and its OI wake emis-sion derived from the integrated images () and fromtracings across individual images (solid line) Dashedlines are classical light curves for a single solid body withpeak intensity at different altitudes The wake follows themeteor light curve between 115 and 97 km but is 28 mag-nitudes fainter as demonstrated by the dotted line whichis the meteor light curve shifted by 128 magnitudes

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

the chemistry with reactive species in the airplasma in a chemically altered but still reducedform It has always been assumed that the organicmatter is consumed by oxygen atoms to form CO2and water However the concentration of reac-tive species is time and temperature dependentHence an important parameter for understand-ing the fate of exogenous organic matter is aknowledge of the cooling rate of the plasma es-pecially in the wake of relatively small mete-oroids of size 150 mm at the peak of the massinflux curve (Ceplecha 1992 Love and Brownlee1993)

Our current understanding of how meteors de-posit cometary and asteroidal material in the at-mosphere and how its chemical composition ischanged during ablation and subsequent chem-istry in the air plasma in the wake of the mete-oroid is primarily based on spatially unresolvedoptical and radar observations (Oumlpik 1958 Bron-shten 1983 Ceplecha et al 1998) Immediatelybehind the meteoroid an air plasma is created byair molecules that are accelerated by collidingwith the meteoroid and its ablation vapor cloudand decelerated by subsequent collisions with theambient air environment Order of magnitude es-timates on the rate of elastic and inelastic colli-sions within the air plasma are obtained frommeasurements of the trail width (Oumlpik 1958Hawkes and Jones 1978) In contrast the loss oftranslational energy in the wake of meteoroidshas recently been calculated using a DirectMonte-Carlo Simulation model that described therarefied flow around a 1 g Leonid (Boyd 2000Popova et al 2000 Jenniskens et al 2000a) Thatmodel predicts a time-dependent temperature(T t212) from about 6300 K at 10 m behind themeteoroid to 3400 K at a distance of 40 m Thenew model shows that the ablation vapor cloud(due to molecules and atoms sputtered off sur-rounding and traveling with the meteoroid) de-termines to a large extent the size of the wake bydramatically increasing the collision cross sectionof the meteoroid

While the process of excitation in suprathermalcollisions implies a lack of thermodynamic equi-librium the typical air plasma vibrational andelectronic excitation temperatures and chemicalequilibrium temperatures measured for such me-teors all suggest T 4300 K a value surprisinglyindependent of meteoroid mass and speed (Jen-niskens et al 2004b)

This paper deals with the temperature decay

following that excitation The wake of actual me-teors is complicated by unknown infrared and ul-traviolet radiative cooling mechanisms nonequi-librium chemistry in the meteor wake (Park andMenees 1978) fragmentation of the meteoroid(Hawkes and Jones 1978 Murray et al 1999 vonZahn 2001) and by fragments ejected at highspeed from a spinning grain (Le Blanc et al 2000Taylor et al 2000)

Until now the only useful measurement of thetemperature decay in the meteor wake was thatby BorovicIuml ka and Jenniskens (2000) who discov-ered an afterglow of cooling meteoric plasma inthe wake of a bright Leonid fireball with a cool-ing time much slower than predicted by theMonte-Carlo Simulation model In at least onepart of the meteoroid trajectory the afterglowwas associated with secondary ablation fromsolid debris detected by way of spectroscopic ev-idence

Measurements of meteor afterglow in faintermeteors are hampered by motion blurring andoverexposure In conventional imaging the rapidmotion of the meteoroid blurs all morphologicaldetail Instantaneous photography using rapidlyrotating shutters has detected meteor wake out to150 m behind the meteoroid (Halliday et al1980) but it is not very sensitive and is limited tometeors so bright that they do not characterizethe typical rarefied flow conditions of small me-teoroids Moreover instantaneous photographydoes not provide the high temporal resolutionthat can be achieved with todayrsquos high frame-rateimagers

Here we report high frame-rate imaging of arelatively faint 23 magnitude Leonid meteor froma ground location during the 2001 Leonid Multi-Instrument Aircraft Campaign (MAC) (Jenniskensand Russell 2003) The most striking aspect of theimages is the unexpected development of a bow-shock-like structure and a spherical luminosity be-fore the shock which is discussed elsewhere (Sten-baek-Nielsen and Jenniskens 2003) This paperdiscusses the wake in the meteor images

METHODS

The observations were made at the Universityof Alaskarsquos Poker Flat Research Range 30 milesnortheast of Fairbanks AK (6512N 14746W and039 km altitude) with an intensified charge cou-pled device (CCD) camera operating at 1000

JENNISKENS AND STENBAEK-NIELSEN96

framess The intensifier used was sensitive tolight in the wavelength range 500ndash900 nm withpeak sensitivity at 700 nm where strong atomicline emissions of oxygen and nitrogen and theFirst Positive band emission of N2 are present inmeteor spectra The intensifier phosphor had abrief decay time constant of 08 ms ideally suitedto study the natural afterglow in rapidly movingtargets

The CCD images were 256 3 256 pixels with256 gray levels (8 bits) and 64 3 64deg field of viewTo facilitate the high frame rate each image quad-rant was read out in parallel through separateelectronics Because of small unavoidable differ-ences between the four-quadrant electronicssome differences in intensity were expected to bevisible in the images

At 1000 framess gray level 255 was reachedat a surface brightness of 3 Mega-Rayleigh (at 700nm) which is about 25 of the CCD well depthTo further prevent blooming of the images thegain of the intensifier was set for saturation to oc-cur only at a higher brightness level Images werecontinuously entered into the high-speed imager(HSI) which was equipped with a 4000-frame cir-cular digital buffer (ie the buffer can contain 4 sof data) Upon recognizing an event the operatorwould then intervene and save a selected se-quence of the buffer to the computer disk storage

The HSI was bore-sighted with a conventionalwide-field low-light-level TV system which pro-vided a more classical video image of the meteorand a larger star field for orientation purposesThis system also consisted of an intensified CCDbut with a 50 ms phosphor decay time and thedata were recorded on videotape with standardNTSC resolution (5994 interlaced fieldss and avertical resolution of 525 horizontal lines) GlobalPositioning System time was encoded on eachfield The field of view was 21 3 16deg and the in-tensifier responded to wavelengths of 400ndash800 nm

RESULTS

On the night of November 18 universal time(UT) 2001 the weather was mostly clouded Af-ter 10 UT the eastern sky cleared Near 1020 UTa persistent meteor train was recorded on thewide-field imager for almost 20 min At 104259UT a bright Leonid meteor passed through thefield of view of both imagers This was close tothe peak of the Leonid storm at 1040 UT which

was caused by Earthrsquos crossing of the 1767 dusttrail of comet 55PTempel-Tuttle (Kondratrsquoevaand Reznikov 1985 Jenniskens 2003)

The wide-field imager

The wide-field imager captured most of themeteor (Fig 1) which moved from the bottom tothe top across the central part of the 21 3 16degwide field over 29 frames (a little less than 1 s)The spatial resolution of the wide field camerawas fairly similar to that of the HSI which cov-ers the central 64 3 64deg of the field Howeverthe frame rate was lower and the meteor movedsignificantly across each frame during the expo-sure Also the bright emissions caused the de-tector to saturate and ldquobloomrdquo These effects com-bined to make it impossible to resolve any of thestructures within the head of the meteor A hazeor thin clouds was present in the wide-field im-ages For ground-based observations at maxi-mum instrument gain and with a clear transpar-ent dark sky we would expect to see stars up tomagnitude approximately 185 The limiting stel-lar magnitude in the images was slightly less than17 A persistent train would have been muchbrighter but it was not detected

The meteor in the narrow-field HSI

Figure 2 shows representative frames acquiredfrom a 400-ms HSI sequence of this event Eachimage revealed an 094 3 094deg section of the orig-inal 64 3 64deg image centered on the meteor Themeteor entered the field of view at frame 60within the sequence The meteor was initially apoint source saturated at the center and slightlybroader than the unsaturated star images withonly a faint trace of wake The onset of subse-quent features was gradual The selected framesin Fig 2 are those in which a new feature is wellpronounced Around frame 170 two unresolvedlines developed at the tail of the meteor whichgrew around frame 235 into a distinct spatialstructure reminiscent of a shock front Initiallyonly a diffuse triangular-shaped wake was visi-ble between the two fronts (frame 200) Aroundframe 235 a wake developed inside the shockfront that persisted for a relatively long time Atthe same time the meteor brightness increaseddramatically The emission on the outside of theldquoshockrdquo structure gradually brightened filling analmost perfect circle of light with a piece cut out(to which we will refer as the ldquobite-outrdquo) At the

METEOR WAKE IN HIGH FRAME-RATE IMAGES 97

same time the ldquoshockrdquo opened up to about 45degThe structure was fully developed around frame330 This is the first time such spatial structurehas been reported The HSI recording ended atframe 463

The spatial structure was not the result of op-tical reflections in the camera We noted that theimages were already distinctly asymmetric whenthe meteor passed very close to the optical axisof the camera (frame 300) where any instrumen-tal distortion of light would be expected to besymmetric We did however find evidence ofscattered light Centered on the meteoroid posi-tion there appeared to be an underlying diffuseglow that was much wider than any spatial struc-ture observed in the images This diffuse glowmay have been scattered light by thin clouds inthe field of view

The centers of the images were saturated How-ever the blooming of the brightest part remainedmodest throughout the exposure period and wasmuch less than the observed spatial structureThis implies that the intensity of the meteor at itspeak was at most a factor of a few brighter thanthe saturation level

The wake

Following the images with a small delay wasa wake of emission that persisted over the dura-tion of the exposure The characteristic delayidentified this wake as being caused by the for-bidden green line emission of O I at 557 nm (Hal-liday 1958b Baggaley 1976 1977) That O I wakewas clearly seen as well in the TV imager (Fig 2)where it remained visible for about 2 s

A second wake phenomenon became visible at lower altitudes which was identified (as dis-cussed below) as the meteor afterglow resultingfrom collisional excitation of metal atoms and airplasma compounds (Halliday 1958a BorovicIuml kaand Jenniskens 2000) This feature was first seenaround frame 300 and is a tail of emission grow-ing from the saturated part of the meteor imageIf this tail was due to phosphor decay from over-exposure of the meteor then there would havebeen an increase in blooming which was not ob-served Indeed the tail faded slower than ex-pected for an overexposed phosphor

We conclude that during the development ofthe tail the meteor did not brighten enough to

JENNISKENS AND STENBAEK-NIELSEN98

FIG 1 Conventional intensified TV images of the event The field of view is 21 3 16deg and the HSI (Fig 2) cov-ers the central portion of the field The first image shows the meteor just before it entered the HSI field of view Thetwo next are within the HSI field of view and the last just after the meteor exited the HSI field of view

overexpose the phosphor The meteor did notleave what is called a persistent train a chemilu-minescence from the catalytic recombination ofoxygen atoms and ozone molecules which wouldbe expected if the Leonid meteor had beenbrighter than magnitude 24 (Jenniskens et al2000b) Such persistent trains have peak emissionin the center of the systemrsquos response curve andare characterized by an apparent brightness ofapproximately 14 magnitude at this spatial res-olution (Jenniskens et al 2000c)

ANALYSIS

Geometry of the observations

Both data sets had a sufficient number of starsdistributed across the images to permit good spa-tial calibration Using computer routines devel-oped for analysis of image data in connectionwith auroral research (Stenbaek-Nielsen et al1984) stars in the Smithsonian Astronomical Ob-servatory star catalog were overlaid and fitted tothe stars present in the images The fitting pro-gram is interactive and can accommodate variousdisplay formats as well as nonlinearity associatedwith the optical system used in the imager The

resulting star fit provided directional informationfor all pixels within the individual images

The center location of the meteor in each im-age was measured for all images in the HSI andthe wide-field video data sequences A computerprogram was written to calculate the location ofthe meteor in each image within the two imagesequences using a fixed meteoroid velocity vec-tor and an initial starting point derived from themeteor location in one of the early HSI imagesThe motion across the field depends upon the me-teoroid angular velocity the assumed range tothe selected initial point and to a lesser degreethe radiant

The meteoroidrsquos velocity vector given by theLeonid shower radiant position measured at thesame time in Arizona by multistation photogra-phy in a Leonid MAC-related effort was 1541 602deg right ascension and 214 6 01deg declination(H Betlem personal communication) The speedwas 716 6 04 kms which is sufficiently high sothat changes due to the Earthrsquos gravitational fieldcould be neglected At the time of the event104259 UT the radiant was at 865deg azimuth (eastof north) and 223deg elevation The camera orien-tation was 761deg azimuth and 539deg elevation re-sulting in an angle between camera orientation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 99

FIG 2 Sections of the original highframe-rate images Each is a 094 3 094degsection from an original 64 3 64deg im-age which shows the development ofthe meteor morphology The imageswere recorded at 1000 framess Theframe number within the sequence isshown in the upper right corner The cal-culated position of the meteoroid (seetext) is shown by a dot

and the velocity vector varying from 28deg to 33degacross the field of view Thus structures alongthe trajectory in the images were foreshortenedby roughly a factor of 2 The pixel size in the high-speed images is 0025 3 0025 At a range of 135km (altitude of 110 km) the spatial resolution was57 mpixel This corresponds to the distance trav-eled by a meteor during the time its image on thephosphor screen fades by a factor e With the me-teor trajectory at an angle of 30deg to the line ofsight the 1-pixel resolution along the trajectorywas 114 m In 1 ms the Leonid moved 72 m Thusthe meteor moved less than 1 pixel betweenframes and the phosphor intensity decayed by86 from one pixel to the next Consequentlythere was little ldquosmearingrdquo of spatial structuresdue to the motion of the Leonid meteor

The calculated positions were fitted to the ob-served positions by adjusting the assumed rangeto the initial point and the Leonid velocity vec-tor A very good fit was obtained for a right as-cension of 1540deg and a declination of 1214degwhich is in agreement with the position measuredphotographically A small systematic deviation of1 pixel in the y direction later in flight was dueto the difficulty of choosing the center of the im-age once the ldquoshockrdquo forms Indeed the differ-ences between calculated and observed positionsin both data sets (Fig 3) were similar to the ac-curacy by which the center could be determinedand hence the analysis showed the meteoroid ve-locity was constant during the time covered bythe optical observations This finding agrees withSpurny et al (2000) who reported that the decel-eration of similar Leonid meteors was below mea-surement accuracy in photographic data

The positions for selected frames of the highframe-rate imager are given in Table 1 The me-teor entered the wide field of view camera at analtitude of 1229 km It brightened very rapidlyin the next two frames The meteoroid entered theHSI field of view at an altitude of 1156 km (frame65) The bow shock and wake started to developat about 1104 km The final HSI image with theshock fully developed was at 1044 km when themeteor brightness started to level off A largerfraction of the path was recorded in the wide-field TV imager which showed that the meteorhad a broad maximum in luminosity centered atan altitude of 1006 km after which the meteordecreased in intensity It left the TV camera fieldof view at an altitude of 95 km when it was stillof magnitude 11 The position analysis is rela-

tively insensitive to the assumed meteor velocitysince varying the assumed range to the initialpoint can compensate for any error in velocitywithin a reasonable range For example if the ve-locity decreased by 1 kms the altitude of thepath would decrease by 2 km

Relative brightness

The meteorrsquos light curve (Fig 4) was derivedfrom the integrated intensity of both the HSI andTV images Both data sets were processed differ-ently in response to the amount of blooming withoverall good agreement to within 05 magnitudesThe HSI data in Fig 4 show the integrated in-tensity over an area of 21 3 21 pixels centered onthe position of the meteoroid This covered the

JENNISKENS AND STENBAEK-NIELSEN100

FIG 3 Differences between observed and calculatedpositions of the meteor image centerThe meteor crossedthe field from the bottom to the top and hence dy is es-sentially the difference along the track while dx is thecross track difference The abscissa is the altitude derivedfor each frame while the ordinate is pixels The observedcenter for the high frame-rate images was estimated to anaccuracy of 02 pixels while the TV record has an accu-racy of about 1 pixel

meteor head but not much of the tail It did notinclude intensity lost in the central saturated pix-els but added intensity from the spatially ex-tended component and the diffuse scattered lightThe TV data were integrated over a 65 3 71 pixelrectangle that covered the full size of the bloomedarea Saturation effects were corrected for bytreating the measured intensity as an opticaldepth In the video there is a substantial back-ground signal which was subtracted by com-puting the average signal in two similar-size rec-tangles located on either side of the meteor To

increase its temporal resolution we measured atthe beginning of data collection the intensity vari-ation across the meteor images in each 130 s ex-posure while the blooming was still modest (solidline in Fig 4) Each data set was calibrated to theV magnitude of the stars in the field of view overthe range V 5 131 to 161 magnitude For sim-ilar cameras it was found that blooming offsetthe effects of saturation over a much wider mag-nitude range the relationship between bloomingand saturation being close to that expected if theelectron production is linear with incident lightbut the electrons distribute over more pixels V 525 log SIpixel (Jenniskens 1999) Indeed the ex-pected intensity in the central pixels of the HSIwas not much above the measured level thus ex-plaining the lack of blooming The slightly higherbrightness may reflect the fact that the HSI im-ager is sensitive to wavelengths longer than theV band which overestimates the brightness in thecalibration procedure However the limitednumber of calibration stars introduces a similarsystematic uncertainty

At the peak of its brightness the meteor hadan absolute magnitude (ie as seen from a dis-tance of 100 km) of 227 6 04 magnitude In-deed a meteor of this brightness would not havea bright persistent train Other Leonid meteors of227 magnitude have their peak brightness at97 6 4 km in agreement with the value of 1006km found here (Jenniskens et al 1998 Betlem etal 2000) On video records with limiting magni-tude approximately 16 such Leonids are usuallydetected first at altitudes of 143 km but theirbrightness does not increase to the photographiclimiting magnitude of about 11 until they decend

METEOR WAKE IN HIGH FRAME-RATE IMAGES 101

TABLE 1 OBSERVED AND CALCULATED POSITIONS OF THE METEOR IN THE HSI DATA

Calculated Observed

Frame x y x y Range Latitude Longitude Altitude

65 1552 14 1552 14 1493 65281 2145492 1156100 1488 188 1489 188 1471 65280 2145541 1146150 1396 446 1396 446 1440 65279 2145610 1132200 1300 714 1299 715 1409 65279 2145679 1118250 1198 996 1198 996 1378 65278 2145749 1104300 1090 1294 1092 1290 1348 65277 2145818 1090350 980 1610 981 1598 1317 65276 2145887 1076400 862 1932 865 1920 1287 65275 2145957 1062450 740 2270 743 2257 1258 65274 2146026 1048463 708 2358 711 2347 1250 65274 2146044 1044

xy are pixel location in the images with (00) in lower left The range is in km from the camera location while latitude and longitude are in geographic degrees N and E Altitude is in km

FIG 4 Light curve of the meteor and its OI wake emis-sion derived from the integrated images () and fromtracings across individual images (solid line) Dashedlines are classical light curves for a single solid body withpeak intensity at different altitudes The wake follows themeteor light curve between 115 and 97 km but is 28 mag-nitudes fainter as demonstrated by the dotted line whichis the meteor light curve shifted by 128 magnitudes

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

framess The intensifier used was sensitive tolight in the wavelength range 500ndash900 nm withpeak sensitivity at 700 nm where strong atomicline emissions of oxygen and nitrogen and theFirst Positive band emission of N2 are present inmeteor spectra The intensifier phosphor had abrief decay time constant of 08 ms ideally suitedto study the natural afterglow in rapidly movingtargets

The CCD images were 256 3 256 pixels with256 gray levels (8 bits) and 64 3 64deg field of viewTo facilitate the high frame rate each image quad-rant was read out in parallel through separateelectronics Because of small unavoidable differ-ences between the four-quadrant electronicssome differences in intensity were expected to bevisible in the images

At 1000 framess gray level 255 was reachedat a surface brightness of 3 Mega-Rayleigh (at 700nm) which is about 25 of the CCD well depthTo further prevent blooming of the images thegain of the intensifier was set for saturation to oc-cur only at a higher brightness level Images werecontinuously entered into the high-speed imager(HSI) which was equipped with a 4000-frame cir-cular digital buffer (ie the buffer can contain 4 sof data) Upon recognizing an event the operatorwould then intervene and save a selected se-quence of the buffer to the computer disk storage

The HSI was bore-sighted with a conventionalwide-field low-light-level TV system which pro-vided a more classical video image of the meteorand a larger star field for orientation purposesThis system also consisted of an intensified CCDbut with a 50 ms phosphor decay time and thedata were recorded on videotape with standardNTSC resolution (5994 interlaced fieldss and avertical resolution of 525 horizontal lines) GlobalPositioning System time was encoded on eachfield The field of view was 21 3 16deg and the in-tensifier responded to wavelengths of 400ndash800 nm

RESULTS

On the night of November 18 universal time(UT) 2001 the weather was mostly clouded Af-ter 10 UT the eastern sky cleared Near 1020 UTa persistent meteor train was recorded on thewide-field imager for almost 20 min At 104259UT a bright Leonid meteor passed through thefield of view of both imagers This was close tothe peak of the Leonid storm at 1040 UT which

was caused by Earthrsquos crossing of the 1767 dusttrail of comet 55PTempel-Tuttle (Kondratrsquoevaand Reznikov 1985 Jenniskens 2003)

The wide-field imager

The wide-field imager captured most of themeteor (Fig 1) which moved from the bottom tothe top across the central part of the 21 3 16degwide field over 29 frames (a little less than 1 s)The spatial resolution of the wide field camerawas fairly similar to that of the HSI which cov-ers the central 64 3 64deg of the field Howeverthe frame rate was lower and the meteor movedsignificantly across each frame during the expo-sure Also the bright emissions caused the de-tector to saturate and ldquobloomrdquo These effects com-bined to make it impossible to resolve any of thestructures within the head of the meteor A hazeor thin clouds was present in the wide-field im-ages For ground-based observations at maxi-mum instrument gain and with a clear transpar-ent dark sky we would expect to see stars up tomagnitude approximately 185 The limiting stel-lar magnitude in the images was slightly less than17 A persistent train would have been muchbrighter but it was not detected

The meteor in the narrow-field HSI

Figure 2 shows representative frames acquiredfrom a 400-ms HSI sequence of this event Eachimage revealed an 094 3 094deg section of the orig-inal 64 3 64deg image centered on the meteor Themeteor entered the field of view at frame 60within the sequence The meteor was initially apoint source saturated at the center and slightlybroader than the unsaturated star images withonly a faint trace of wake The onset of subse-quent features was gradual The selected framesin Fig 2 are those in which a new feature is wellpronounced Around frame 170 two unresolvedlines developed at the tail of the meteor whichgrew around frame 235 into a distinct spatialstructure reminiscent of a shock front Initiallyonly a diffuse triangular-shaped wake was visi-ble between the two fronts (frame 200) Aroundframe 235 a wake developed inside the shockfront that persisted for a relatively long time Atthe same time the meteor brightness increaseddramatically The emission on the outside of theldquoshockrdquo structure gradually brightened filling analmost perfect circle of light with a piece cut out(to which we will refer as the ldquobite-outrdquo) At the

METEOR WAKE IN HIGH FRAME-RATE IMAGES 97

same time the ldquoshockrdquo opened up to about 45degThe structure was fully developed around frame330 This is the first time such spatial structurehas been reported The HSI recording ended atframe 463

The spatial structure was not the result of op-tical reflections in the camera We noted that theimages were already distinctly asymmetric whenthe meteor passed very close to the optical axisof the camera (frame 300) where any instrumen-tal distortion of light would be expected to besymmetric We did however find evidence ofscattered light Centered on the meteoroid posi-tion there appeared to be an underlying diffuseglow that was much wider than any spatial struc-ture observed in the images This diffuse glowmay have been scattered light by thin clouds inthe field of view

The centers of the images were saturated How-ever the blooming of the brightest part remainedmodest throughout the exposure period and wasmuch less than the observed spatial structureThis implies that the intensity of the meteor at itspeak was at most a factor of a few brighter thanthe saturation level

The wake

Following the images with a small delay wasa wake of emission that persisted over the dura-tion of the exposure The characteristic delayidentified this wake as being caused by the for-bidden green line emission of O I at 557 nm (Hal-liday 1958b Baggaley 1976 1977) That O I wakewas clearly seen as well in the TV imager (Fig 2)where it remained visible for about 2 s

A second wake phenomenon became visible at lower altitudes which was identified (as dis-cussed below) as the meteor afterglow resultingfrom collisional excitation of metal atoms and airplasma compounds (Halliday 1958a BorovicIuml kaand Jenniskens 2000) This feature was first seenaround frame 300 and is a tail of emission grow-ing from the saturated part of the meteor imageIf this tail was due to phosphor decay from over-exposure of the meteor then there would havebeen an increase in blooming which was not ob-served Indeed the tail faded slower than ex-pected for an overexposed phosphor

We conclude that during the development ofthe tail the meteor did not brighten enough to

JENNISKENS AND STENBAEK-NIELSEN98

FIG 1 Conventional intensified TV images of the event The field of view is 21 3 16deg and the HSI (Fig 2) cov-ers the central portion of the field The first image shows the meteor just before it entered the HSI field of view Thetwo next are within the HSI field of view and the last just after the meteor exited the HSI field of view

overexpose the phosphor The meteor did notleave what is called a persistent train a chemilu-minescence from the catalytic recombination ofoxygen atoms and ozone molecules which wouldbe expected if the Leonid meteor had beenbrighter than magnitude 24 (Jenniskens et al2000b) Such persistent trains have peak emissionin the center of the systemrsquos response curve andare characterized by an apparent brightness ofapproximately 14 magnitude at this spatial res-olution (Jenniskens et al 2000c)

ANALYSIS

Geometry of the observations

Both data sets had a sufficient number of starsdistributed across the images to permit good spa-tial calibration Using computer routines devel-oped for analysis of image data in connectionwith auroral research (Stenbaek-Nielsen et al1984) stars in the Smithsonian Astronomical Ob-servatory star catalog were overlaid and fitted tothe stars present in the images The fitting pro-gram is interactive and can accommodate variousdisplay formats as well as nonlinearity associatedwith the optical system used in the imager The

resulting star fit provided directional informationfor all pixels within the individual images

The center location of the meteor in each im-age was measured for all images in the HSI andthe wide-field video data sequences A computerprogram was written to calculate the location ofthe meteor in each image within the two imagesequences using a fixed meteoroid velocity vec-tor and an initial starting point derived from themeteor location in one of the early HSI imagesThe motion across the field depends upon the me-teoroid angular velocity the assumed range tothe selected initial point and to a lesser degreethe radiant

The meteoroidrsquos velocity vector given by theLeonid shower radiant position measured at thesame time in Arizona by multistation photogra-phy in a Leonid MAC-related effort was 1541 602deg right ascension and 214 6 01deg declination(H Betlem personal communication) The speedwas 716 6 04 kms which is sufficiently high sothat changes due to the Earthrsquos gravitational fieldcould be neglected At the time of the event104259 UT the radiant was at 865deg azimuth (eastof north) and 223deg elevation The camera orien-tation was 761deg azimuth and 539deg elevation re-sulting in an angle between camera orientation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 99

FIG 2 Sections of the original highframe-rate images Each is a 094 3 094degsection from an original 64 3 64deg im-age which shows the development ofthe meteor morphology The imageswere recorded at 1000 framess Theframe number within the sequence isshown in the upper right corner The cal-culated position of the meteoroid (seetext) is shown by a dot

and the velocity vector varying from 28deg to 33degacross the field of view Thus structures alongthe trajectory in the images were foreshortenedby roughly a factor of 2 The pixel size in the high-speed images is 0025 3 0025 At a range of 135km (altitude of 110 km) the spatial resolution was57 mpixel This corresponds to the distance trav-eled by a meteor during the time its image on thephosphor screen fades by a factor e With the me-teor trajectory at an angle of 30deg to the line ofsight the 1-pixel resolution along the trajectorywas 114 m In 1 ms the Leonid moved 72 m Thusthe meteor moved less than 1 pixel betweenframes and the phosphor intensity decayed by86 from one pixel to the next Consequentlythere was little ldquosmearingrdquo of spatial structuresdue to the motion of the Leonid meteor

The calculated positions were fitted to the ob-served positions by adjusting the assumed rangeto the initial point and the Leonid velocity vec-tor A very good fit was obtained for a right as-cension of 1540deg and a declination of 1214degwhich is in agreement with the position measuredphotographically A small systematic deviation of1 pixel in the y direction later in flight was dueto the difficulty of choosing the center of the im-age once the ldquoshockrdquo forms Indeed the differ-ences between calculated and observed positionsin both data sets (Fig 3) were similar to the ac-curacy by which the center could be determinedand hence the analysis showed the meteoroid ve-locity was constant during the time covered bythe optical observations This finding agrees withSpurny et al (2000) who reported that the decel-eration of similar Leonid meteors was below mea-surement accuracy in photographic data

The positions for selected frames of the highframe-rate imager are given in Table 1 The me-teor entered the wide field of view camera at analtitude of 1229 km It brightened very rapidlyin the next two frames The meteoroid entered theHSI field of view at an altitude of 1156 km (frame65) The bow shock and wake started to developat about 1104 km The final HSI image with theshock fully developed was at 1044 km when themeteor brightness started to level off A largerfraction of the path was recorded in the wide-field TV imager which showed that the meteorhad a broad maximum in luminosity centered atan altitude of 1006 km after which the meteordecreased in intensity It left the TV camera fieldof view at an altitude of 95 km when it was stillof magnitude 11 The position analysis is rela-

tively insensitive to the assumed meteor velocitysince varying the assumed range to the initialpoint can compensate for any error in velocitywithin a reasonable range For example if the ve-locity decreased by 1 kms the altitude of thepath would decrease by 2 km

Relative brightness

The meteorrsquos light curve (Fig 4) was derivedfrom the integrated intensity of both the HSI andTV images Both data sets were processed differ-ently in response to the amount of blooming withoverall good agreement to within 05 magnitudesThe HSI data in Fig 4 show the integrated in-tensity over an area of 21 3 21 pixels centered onthe position of the meteoroid This covered the

JENNISKENS AND STENBAEK-NIELSEN100

FIG 3 Differences between observed and calculatedpositions of the meteor image centerThe meteor crossedthe field from the bottom to the top and hence dy is es-sentially the difference along the track while dx is thecross track difference The abscissa is the altitude derivedfor each frame while the ordinate is pixels The observedcenter for the high frame-rate images was estimated to anaccuracy of 02 pixels while the TV record has an accu-racy of about 1 pixel

meteor head but not much of the tail It did notinclude intensity lost in the central saturated pix-els but added intensity from the spatially ex-tended component and the diffuse scattered lightThe TV data were integrated over a 65 3 71 pixelrectangle that covered the full size of the bloomedarea Saturation effects were corrected for bytreating the measured intensity as an opticaldepth In the video there is a substantial back-ground signal which was subtracted by com-puting the average signal in two similar-size rec-tangles located on either side of the meteor To

increase its temporal resolution we measured atthe beginning of data collection the intensity vari-ation across the meteor images in each 130 s ex-posure while the blooming was still modest (solidline in Fig 4) Each data set was calibrated to theV magnitude of the stars in the field of view overthe range V 5 131 to 161 magnitude For sim-ilar cameras it was found that blooming offsetthe effects of saturation over a much wider mag-nitude range the relationship between bloomingand saturation being close to that expected if theelectron production is linear with incident lightbut the electrons distribute over more pixels V 525 log SIpixel (Jenniskens 1999) Indeed the ex-pected intensity in the central pixels of the HSIwas not much above the measured level thus ex-plaining the lack of blooming The slightly higherbrightness may reflect the fact that the HSI im-ager is sensitive to wavelengths longer than theV band which overestimates the brightness in thecalibration procedure However the limitednumber of calibration stars introduces a similarsystematic uncertainty

At the peak of its brightness the meteor hadan absolute magnitude (ie as seen from a dis-tance of 100 km) of 227 6 04 magnitude In-deed a meteor of this brightness would not havea bright persistent train Other Leonid meteors of227 magnitude have their peak brightness at97 6 4 km in agreement with the value of 1006km found here (Jenniskens et al 1998 Betlem etal 2000) On video records with limiting magni-tude approximately 16 such Leonids are usuallydetected first at altitudes of 143 km but theirbrightness does not increase to the photographiclimiting magnitude of about 11 until they decend

METEOR WAKE IN HIGH FRAME-RATE IMAGES 101

TABLE 1 OBSERVED AND CALCULATED POSITIONS OF THE METEOR IN THE HSI DATA

Calculated Observed

Frame x y x y Range Latitude Longitude Altitude

65 1552 14 1552 14 1493 65281 2145492 1156100 1488 188 1489 188 1471 65280 2145541 1146150 1396 446 1396 446 1440 65279 2145610 1132200 1300 714 1299 715 1409 65279 2145679 1118250 1198 996 1198 996 1378 65278 2145749 1104300 1090 1294 1092 1290 1348 65277 2145818 1090350 980 1610 981 1598 1317 65276 2145887 1076400 862 1932 865 1920 1287 65275 2145957 1062450 740 2270 743 2257 1258 65274 2146026 1048463 708 2358 711 2347 1250 65274 2146044 1044

xy are pixel location in the images with (00) in lower left The range is in km from the camera location while latitude and longitude are in geographic degrees N and E Altitude is in km

FIG 4 Light curve of the meteor and its OI wake emis-sion derived from the integrated images () and fromtracings across individual images (solid line) Dashedlines are classical light curves for a single solid body withpeak intensity at different altitudes The wake follows themeteor light curve between 115 and 97 km but is 28 mag-nitudes fainter as demonstrated by the dotted line whichis the meteor light curve shifted by 128 magnitudes

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

same time the ldquoshockrdquo opened up to about 45degThe structure was fully developed around frame330 This is the first time such spatial structurehas been reported The HSI recording ended atframe 463

The spatial structure was not the result of op-tical reflections in the camera We noted that theimages were already distinctly asymmetric whenthe meteor passed very close to the optical axisof the camera (frame 300) where any instrumen-tal distortion of light would be expected to besymmetric We did however find evidence ofscattered light Centered on the meteoroid posi-tion there appeared to be an underlying diffuseglow that was much wider than any spatial struc-ture observed in the images This diffuse glowmay have been scattered light by thin clouds inthe field of view

The centers of the images were saturated How-ever the blooming of the brightest part remainedmodest throughout the exposure period and wasmuch less than the observed spatial structureThis implies that the intensity of the meteor at itspeak was at most a factor of a few brighter thanthe saturation level

The wake

Following the images with a small delay wasa wake of emission that persisted over the dura-tion of the exposure The characteristic delayidentified this wake as being caused by the for-bidden green line emission of O I at 557 nm (Hal-liday 1958b Baggaley 1976 1977) That O I wakewas clearly seen as well in the TV imager (Fig 2)where it remained visible for about 2 s

A second wake phenomenon became visible at lower altitudes which was identified (as dis-cussed below) as the meteor afterglow resultingfrom collisional excitation of metal atoms and airplasma compounds (Halliday 1958a BorovicIuml kaand Jenniskens 2000) This feature was first seenaround frame 300 and is a tail of emission grow-ing from the saturated part of the meteor imageIf this tail was due to phosphor decay from over-exposure of the meteor then there would havebeen an increase in blooming which was not ob-served Indeed the tail faded slower than ex-pected for an overexposed phosphor

We conclude that during the development ofthe tail the meteor did not brighten enough to

JENNISKENS AND STENBAEK-NIELSEN98

FIG 1 Conventional intensified TV images of the event The field of view is 21 3 16deg and the HSI (Fig 2) cov-ers the central portion of the field The first image shows the meteor just before it entered the HSI field of view Thetwo next are within the HSI field of view and the last just after the meteor exited the HSI field of view

overexpose the phosphor The meteor did notleave what is called a persistent train a chemilu-minescence from the catalytic recombination ofoxygen atoms and ozone molecules which wouldbe expected if the Leonid meteor had beenbrighter than magnitude 24 (Jenniskens et al2000b) Such persistent trains have peak emissionin the center of the systemrsquos response curve andare characterized by an apparent brightness ofapproximately 14 magnitude at this spatial res-olution (Jenniskens et al 2000c)

ANALYSIS

Geometry of the observations

Both data sets had a sufficient number of starsdistributed across the images to permit good spa-tial calibration Using computer routines devel-oped for analysis of image data in connectionwith auroral research (Stenbaek-Nielsen et al1984) stars in the Smithsonian Astronomical Ob-servatory star catalog were overlaid and fitted tothe stars present in the images The fitting pro-gram is interactive and can accommodate variousdisplay formats as well as nonlinearity associatedwith the optical system used in the imager The

resulting star fit provided directional informationfor all pixels within the individual images

The center location of the meteor in each im-age was measured for all images in the HSI andthe wide-field video data sequences A computerprogram was written to calculate the location ofthe meteor in each image within the two imagesequences using a fixed meteoroid velocity vec-tor and an initial starting point derived from themeteor location in one of the early HSI imagesThe motion across the field depends upon the me-teoroid angular velocity the assumed range tothe selected initial point and to a lesser degreethe radiant

The meteoroidrsquos velocity vector given by theLeonid shower radiant position measured at thesame time in Arizona by multistation photogra-phy in a Leonid MAC-related effort was 1541 602deg right ascension and 214 6 01deg declination(H Betlem personal communication) The speedwas 716 6 04 kms which is sufficiently high sothat changes due to the Earthrsquos gravitational fieldcould be neglected At the time of the event104259 UT the radiant was at 865deg azimuth (eastof north) and 223deg elevation The camera orien-tation was 761deg azimuth and 539deg elevation re-sulting in an angle between camera orientation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 99

FIG 2 Sections of the original highframe-rate images Each is a 094 3 094degsection from an original 64 3 64deg im-age which shows the development ofthe meteor morphology The imageswere recorded at 1000 framess Theframe number within the sequence isshown in the upper right corner The cal-culated position of the meteoroid (seetext) is shown by a dot

and the velocity vector varying from 28deg to 33degacross the field of view Thus structures alongthe trajectory in the images were foreshortenedby roughly a factor of 2 The pixel size in the high-speed images is 0025 3 0025 At a range of 135km (altitude of 110 km) the spatial resolution was57 mpixel This corresponds to the distance trav-eled by a meteor during the time its image on thephosphor screen fades by a factor e With the me-teor trajectory at an angle of 30deg to the line ofsight the 1-pixel resolution along the trajectorywas 114 m In 1 ms the Leonid moved 72 m Thusthe meteor moved less than 1 pixel betweenframes and the phosphor intensity decayed by86 from one pixel to the next Consequentlythere was little ldquosmearingrdquo of spatial structuresdue to the motion of the Leonid meteor

The calculated positions were fitted to the ob-served positions by adjusting the assumed rangeto the initial point and the Leonid velocity vec-tor A very good fit was obtained for a right as-cension of 1540deg and a declination of 1214degwhich is in agreement with the position measuredphotographically A small systematic deviation of1 pixel in the y direction later in flight was dueto the difficulty of choosing the center of the im-age once the ldquoshockrdquo forms Indeed the differ-ences between calculated and observed positionsin both data sets (Fig 3) were similar to the ac-curacy by which the center could be determinedand hence the analysis showed the meteoroid ve-locity was constant during the time covered bythe optical observations This finding agrees withSpurny et al (2000) who reported that the decel-eration of similar Leonid meteors was below mea-surement accuracy in photographic data

The positions for selected frames of the highframe-rate imager are given in Table 1 The me-teor entered the wide field of view camera at analtitude of 1229 km It brightened very rapidlyin the next two frames The meteoroid entered theHSI field of view at an altitude of 1156 km (frame65) The bow shock and wake started to developat about 1104 km The final HSI image with theshock fully developed was at 1044 km when themeteor brightness started to level off A largerfraction of the path was recorded in the wide-field TV imager which showed that the meteorhad a broad maximum in luminosity centered atan altitude of 1006 km after which the meteordecreased in intensity It left the TV camera fieldof view at an altitude of 95 km when it was stillof magnitude 11 The position analysis is rela-

tively insensitive to the assumed meteor velocitysince varying the assumed range to the initialpoint can compensate for any error in velocitywithin a reasonable range For example if the ve-locity decreased by 1 kms the altitude of thepath would decrease by 2 km

Relative brightness

The meteorrsquos light curve (Fig 4) was derivedfrom the integrated intensity of both the HSI andTV images Both data sets were processed differ-ently in response to the amount of blooming withoverall good agreement to within 05 magnitudesThe HSI data in Fig 4 show the integrated in-tensity over an area of 21 3 21 pixels centered onthe position of the meteoroid This covered the

JENNISKENS AND STENBAEK-NIELSEN100

FIG 3 Differences between observed and calculatedpositions of the meteor image centerThe meteor crossedthe field from the bottom to the top and hence dy is es-sentially the difference along the track while dx is thecross track difference The abscissa is the altitude derivedfor each frame while the ordinate is pixels The observedcenter for the high frame-rate images was estimated to anaccuracy of 02 pixels while the TV record has an accu-racy of about 1 pixel

meteor head but not much of the tail It did notinclude intensity lost in the central saturated pix-els but added intensity from the spatially ex-tended component and the diffuse scattered lightThe TV data were integrated over a 65 3 71 pixelrectangle that covered the full size of the bloomedarea Saturation effects were corrected for bytreating the measured intensity as an opticaldepth In the video there is a substantial back-ground signal which was subtracted by com-puting the average signal in two similar-size rec-tangles located on either side of the meteor To

increase its temporal resolution we measured atthe beginning of data collection the intensity vari-ation across the meteor images in each 130 s ex-posure while the blooming was still modest (solidline in Fig 4) Each data set was calibrated to theV magnitude of the stars in the field of view overthe range V 5 131 to 161 magnitude For sim-ilar cameras it was found that blooming offsetthe effects of saturation over a much wider mag-nitude range the relationship between bloomingand saturation being close to that expected if theelectron production is linear with incident lightbut the electrons distribute over more pixels V 525 log SIpixel (Jenniskens 1999) Indeed the ex-pected intensity in the central pixels of the HSIwas not much above the measured level thus ex-plaining the lack of blooming The slightly higherbrightness may reflect the fact that the HSI im-ager is sensitive to wavelengths longer than theV band which overestimates the brightness in thecalibration procedure However the limitednumber of calibration stars introduces a similarsystematic uncertainty

At the peak of its brightness the meteor hadan absolute magnitude (ie as seen from a dis-tance of 100 km) of 227 6 04 magnitude In-deed a meteor of this brightness would not havea bright persistent train Other Leonid meteors of227 magnitude have their peak brightness at97 6 4 km in agreement with the value of 1006km found here (Jenniskens et al 1998 Betlem etal 2000) On video records with limiting magni-tude approximately 16 such Leonids are usuallydetected first at altitudes of 143 km but theirbrightness does not increase to the photographiclimiting magnitude of about 11 until they decend

METEOR WAKE IN HIGH FRAME-RATE IMAGES 101

TABLE 1 OBSERVED AND CALCULATED POSITIONS OF THE METEOR IN THE HSI DATA

Calculated Observed

Frame x y x y Range Latitude Longitude Altitude

65 1552 14 1552 14 1493 65281 2145492 1156100 1488 188 1489 188 1471 65280 2145541 1146150 1396 446 1396 446 1440 65279 2145610 1132200 1300 714 1299 715 1409 65279 2145679 1118250 1198 996 1198 996 1378 65278 2145749 1104300 1090 1294 1092 1290 1348 65277 2145818 1090350 980 1610 981 1598 1317 65276 2145887 1076400 862 1932 865 1920 1287 65275 2145957 1062450 740 2270 743 2257 1258 65274 2146026 1048463 708 2358 711 2347 1250 65274 2146044 1044

xy are pixel location in the images with (00) in lower left The range is in km from the camera location while latitude and longitude are in geographic degrees N and E Altitude is in km

FIG 4 Light curve of the meteor and its OI wake emis-sion derived from the integrated images () and fromtracings across individual images (solid line) Dashedlines are classical light curves for a single solid body withpeak intensity at different altitudes The wake follows themeteor light curve between 115 and 97 km but is 28 mag-nitudes fainter as demonstrated by the dotted line whichis the meteor light curve shifted by 128 magnitudes

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

overexpose the phosphor The meteor did notleave what is called a persistent train a chemilu-minescence from the catalytic recombination ofoxygen atoms and ozone molecules which wouldbe expected if the Leonid meteor had beenbrighter than magnitude 24 (Jenniskens et al2000b) Such persistent trains have peak emissionin the center of the systemrsquos response curve andare characterized by an apparent brightness ofapproximately 14 magnitude at this spatial res-olution (Jenniskens et al 2000c)

ANALYSIS

Geometry of the observations

Both data sets had a sufficient number of starsdistributed across the images to permit good spa-tial calibration Using computer routines devel-oped for analysis of image data in connectionwith auroral research (Stenbaek-Nielsen et al1984) stars in the Smithsonian Astronomical Ob-servatory star catalog were overlaid and fitted tothe stars present in the images The fitting pro-gram is interactive and can accommodate variousdisplay formats as well as nonlinearity associatedwith the optical system used in the imager The

resulting star fit provided directional informationfor all pixels within the individual images

The center location of the meteor in each im-age was measured for all images in the HSI andthe wide-field video data sequences A computerprogram was written to calculate the location ofthe meteor in each image within the two imagesequences using a fixed meteoroid velocity vec-tor and an initial starting point derived from themeteor location in one of the early HSI imagesThe motion across the field depends upon the me-teoroid angular velocity the assumed range tothe selected initial point and to a lesser degreethe radiant

The meteoroidrsquos velocity vector given by theLeonid shower radiant position measured at thesame time in Arizona by multistation photogra-phy in a Leonid MAC-related effort was 1541 602deg right ascension and 214 6 01deg declination(H Betlem personal communication) The speedwas 716 6 04 kms which is sufficiently high sothat changes due to the Earthrsquos gravitational fieldcould be neglected At the time of the event104259 UT the radiant was at 865deg azimuth (eastof north) and 223deg elevation The camera orien-tation was 761deg azimuth and 539deg elevation re-sulting in an angle between camera orientation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 99

FIG 2 Sections of the original highframe-rate images Each is a 094 3 094degsection from an original 64 3 64deg im-age which shows the development ofthe meteor morphology The imageswere recorded at 1000 framess Theframe number within the sequence isshown in the upper right corner The cal-culated position of the meteoroid (seetext) is shown by a dot

and the velocity vector varying from 28deg to 33degacross the field of view Thus structures alongthe trajectory in the images were foreshortenedby roughly a factor of 2 The pixel size in the high-speed images is 0025 3 0025 At a range of 135km (altitude of 110 km) the spatial resolution was57 mpixel This corresponds to the distance trav-eled by a meteor during the time its image on thephosphor screen fades by a factor e With the me-teor trajectory at an angle of 30deg to the line ofsight the 1-pixel resolution along the trajectorywas 114 m In 1 ms the Leonid moved 72 m Thusthe meteor moved less than 1 pixel betweenframes and the phosphor intensity decayed by86 from one pixel to the next Consequentlythere was little ldquosmearingrdquo of spatial structuresdue to the motion of the Leonid meteor

The calculated positions were fitted to the ob-served positions by adjusting the assumed rangeto the initial point and the Leonid velocity vec-tor A very good fit was obtained for a right as-cension of 1540deg and a declination of 1214degwhich is in agreement with the position measuredphotographically A small systematic deviation of1 pixel in the y direction later in flight was dueto the difficulty of choosing the center of the im-age once the ldquoshockrdquo forms Indeed the differ-ences between calculated and observed positionsin both data sets (Fig 3) were similar to the ac-curacy by which the center could be determinedand hence the analysis showed the meteoroid ve-locity was constant during the time covered bythe optical observations This finding agrees withSpurny et al (2000) who reported that the decel-eration of similar Leonid meteors was below mea-surement accuracy in photographic data

The positions for selected frames of the highframe-rate imager are given in Table 1 The me-teor entered the wide field of view camera at analtitude of 1229 km It brightened very rapidlyin the next two frames The meteoroid entered theHSI field of view at an altitude of 1156 km (frame65) The bow shock and wake started to developat about 1104 km The final HSI image with theshock fully developed was at 1044 km when themeteor brightness started to level off A largerfraction of the path was recorded in the wide-field TV imager which showed that the meteorhad a broad maximum in luminosity centered atan altitude of 1006 km after which the meteordecreased in intensity It left the TV camera fieldof view at an altitude of 95 km when it was stillof magnitude 11 The position analysis is rela-

tively insensitive to the assumed meteor velocitysince varying the assumed range to the initialpoint can compensate for any error in velocitywithin a reasonable range For example if the ve-locity decreased by 1 kms the altitude of thepath would decrease by 2 km

Relative brightness

The meteorrsquos light curve (Fig 4) was derivedfrom the integrated intensity of both the HSI andTV images Both data sets were processed differ-ently in response to the amount of blooming withoverall good agreement to within 05 magnitudesThe HSI data in Fig 4 show the integrated in-tensity over an area of 21 3 21 pixels centered onthe position of the meteoroid This covered the

JENNISKENS AND STENBAEK-NIELSEN100

FIG 3 Differences between observed and calculatedpositions of the meteor image centerThe meteor crossedthe field from the bottom to the top and hence dy is es-sentially the difference along the track while dx is thecross track difference The abscissa is the altitude derivedfor each frame while the ordinate is pixels The observedcenter for the high frame-rate images was estimated to anaccuracy of 02 pixels while the TV record has an accu-racy of about 1 pixel

meteor head but not much of the tail It did notinclude intensity lost in the central saturated pix-els but added intensity from the spatially ex-tended component and the diffuse scattered lightThe TV data were integrated over a 65 3 71 pixelrectangle that covered the full size of the bloomedarea Saturation effects were corrected for bytreating the measured intensity as an opticaldepth In the video there is a substantial back-ground signal which was subtracted by com-puting the average signal in two similar-size rec-tangles located on either side of the meteor To

increase its temporal resolution we measured atthe beginning of data collection the intensity vari-ation across the meteor images in each 130 s ex-posure while the blooming was still modest (solidline in Fig 4) Each data set was calibrated to theV magnitude of the stars in the field of view overthe range V 5 131 to 161 magnitude For sim-ilar cameras it was found that blooming offsetthe effects of saturation over a much wider mag-nitude range the relationship between bloomingand saturation being close to that expected if theelectron production is linear with incident lightbut the electrons distribute over more pixels V 525 log SIpixel (Jenniskens 1999) Indeed the ex-pected intensity in the central pixels of the HSIwas not much above the measured level thus ex-plaining the lack of blooming The slightly higherbrightness may reflect the fact that the HSI im-ager is sensitive to wavelengths longer than theV band which overestimates the brightness in thecalibration procedure However the limitednumber of calibration stars introduces a similarsystematic uncertainty

At the peak of its brightness the meteor hadan absolute magnitude (ie as seen from a dis-tance of 100 km) of 227 6 04 magnitude In-deed a meteor of this brightness would not havea bright persistent train Other Leonid meteors of227 magnitude have their peak brightness at97 6 4 km in agreement with the value of 1006km found here (Jenniskens et al 1998 Betlem etal 2000) On video records with limiting magni-tude approximately 16 such Leonids are usuallydetected first at altitudes of 143 km but theirbrightness does not increase to the photographiclimiting magnitude of about 11 until they decend

METEOR WAKE IN HIGH FRAME-RATE IMAGES 101

TABLE 1 OBSERVED AND CALCULATED POSITIONS OF THE METEOR IN THE HSI DATA

Calculated Observed

Frame x y x y Range Latitude Longitude Altitude

65 1552 14 1552 14 1493 65281 2145492 1156100 1488 188 1489 188 1471 65280 2145541 1146150 1396 446 1396 446 1440 65279 2145610 1132200 1300 714 1299 715 1409 65279 2145679 1118250 1198 996 1198 996 1378 65278 2145749 1104300 1090 1294 1092 1290 1348 65277 2145818 1090350 980 1610 981 1598 1317 65276 2145887 1076400 862 1932 865 1920 1287 65275 2145957 1062450 740 2270 743 2257 1258 65274 2146026 1048463 708 2358 711 2347 1250 65274 2146044 1044

xy are pixel location in the images with (00) in lower left The range is in km from the camera location while latitude and longitude are in geographic degrees N and E Altitude is in km

FIG 4 Light curve of the meteor and its OI wake emis-sion derived from the integrated images () and fromtracings across individual images (solid line) Dashedlines are classical light curves for a single solid body withpeak intensity at different altitudes The wake follows themeteor light curve between 115 and 97 km but is 28 mag-nitudes fainter as demonstrated by the dotted line whichis the meteor light curve shifted by 128 magnitudes

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

and the velocity vector varying from 28deg to 33degacross the field of view Thus structures alongthe trajectory in the images were foreshortenedby roughly a factor of 2 The pixel size in the high-speed images is 0025 3 0025 At a range of 135km (altitude of 110 km) the spatial resolution was57 mpixel This corresponds to the distance trav-eled by a meteor during the time its image on thephosphor screen fades by a factor e With the me-teor trajectory at an angle of 30deg to the line ofsight the 1-pixel resolution along the trajectorywas 114 m In 1 ms the Leonid moved 72 m Thusthe meteor moved less than 1 pixel betweenframes and the phosphor intensity decayed by86 from one pixel to the next Consequentlythere was little ldquosmearingrdquo of spatial structuresdue to the motion of the Leonid meteor

The calculated positions were fitted to the ob-served positions by adjusting the assumed rangeto the initial point and the Leonid velocity vec-tor A very good fit was obtained for a right as-cension of 1540deg and a declination of 1214degwhich is in agreement with the position measuredphotographically A small systematic deviation of1 pixel in the y direction later in flight was dueto the difficulty of choosing the center of the im-age once the ldquoshockrdquo forms Indeed the differ-ences between calculated and observed positionsin both data sets (Fig 3) were similar to the ac-curacy by which the center could be determinedand hence the analysis showed the meteoroid ve-locity was constant during the time covered bythe optical observations This finding agrees withSpurny et al (2000) who reported that the decel-eration of similar Leonid meteors was below mea-surement accuracy in photographic data

The positions for selected frames of the highframe-rate imager are given in Table 1 The me-teor entered the wide field of view camera at analtitude of 1229 km It brightened very rapidlyin the next two frames The meteoroid entered theHSI field of view at an altitude of 1156 km (frame65) The bow shock and wake started to developat about 1104 km The final HSI image with theshock fully developed was at 1044 km when themeteor brightness started to level off A largerfraction of the path was recorded in the wide-field TV imager which showed that the meteorhad a broad maximum in luminosity centered atan altitude of 1006 km after which the meteordecreased in intensity It left the TV camera fieldof view at an altitude of 95 km when it was stillof magnitude 11 The position analysis is rela-

tively insensitive to the assumed meteor velocitysince varying the assumed range to the initialpoint can compensate for any error in velocitywithin a reasonable range For example if the ve-locity decreased by 1 kms the altitude of thepath would decrease by 2 km

Relative brightness

The meteorrsquos light curve (Fig 4) was derivedfrom the integrated intensity of both the HSI andTV images Both data sets were processed differ-ently in response to the amount of blooming withoverall good agreement to within 05 magnitudesThe HSI data in Fig 4 show the integrated in-tensity over an area of 21 3 21 pixels centered onthe position of the meteoroid This covered the

JENNISKENS AND STENBAEK-NIELSEN100

FIG 3 Differences between observed and calculatedpositions of the meteor image centerThe meteor crossedthe field from the bottom to the top and hence dy is es-sentially the difference along the track while dx is thecross track difference The abscissa is the altitude derivedfor each frame while the ordinate is pixels The observedcenter for the high frame-rate images was estimated to anaccuracy of 02 pixels while the TV record has an accu-racy of about 1 pixel

meteor head but not much of the tail It did notinclude intensity lost in the central saturated pix-els but added intensity from the spatially ex-tended component and the diffuse scattered lightThe TV data were integrated over a 65 3 71 pixelrectangle that covered the full size of the bloomedarea Saturation effects were corrected for bytreating the measured intensity as an opticaldepth In the video there is a substantial back-ground signal which was subtracted by com-puting the average signal in two similar-size rec-tangles located on either side of the meteor To

increase its temporal resolution we measured atthe beginning of data collection the intensity vari-ation across the meteor images in each 130 s ex-posure while the blooming was still modest (solidline in Fig 4) Each data set was calibrated to theV magnitude of the stars in the field of view overthe range V 5 131 to 161 magnitude For sim-ilar cameras it was found that blooming offsetthe effects of saturation over a much wider mag-nitude range the relationship between bloomingand saturation being close to that expected if theelectron production is linear with incident lightbut the electrons distribute over more pixels V 525 log SIpixel (Jenniskens 1999) Indeed the ex-pected intensity in the central pixels of the HSIwas not much above the measured level thus ex-plaining the lack of blooming The slightly higherbrightness may reflect the fact that the HSI im-ager is sensitive to wavelengths longer than theV band which overestimates the brightness in thecalibration procedure However the limitednumber of calibration stars introduces a similarsystematic uncertainty

At the peak of its brightness the meteor hadan absolute magnitude (ie as seen from a dis-tance of 100 km) of 227 6 04 magnitude In-deed a meteor of this brightness would not havea bright persistent train Other Leonid meteors of227 magnitude have their peak brightness at97 6 4 km in agreement with the value of 1006km found here (Jenniskens et al 1998 Betlem etal 2000) On video records with limiting magni-tude approximately 16 such Leonids are usuallydetected first at altitudes of 143 km but theirbrightness does not increase to the photographiclimiting magnitude of about 11 until they decend

METEOR WAKE IN HIGH FRAME-RATE IMAGES 101

TABLE 1 OBSERVED AND CALCULATED POSITIONS OF THE METEOR IN THE HSI DATA

Calculated Observed

Frame x y x y Range Latitude Longitude Altitude

65 1552 14 1552 14 1493 65281 2145492 1156100 1488 188 1489 188 1471 65280 2145541 1146150 1396 446 1396 446 1440 65279 2145610 1132200 1300 714 1299 715 1409 65279 2145679 1118250 1198 996 1198 996 1378 65278 2145749 1104300 1090 1294 1092 1290 1348 65277 2145818 1090350 980 1610 981 1598 1317 65276 2145887 1076400 862 1932 865 1920 1287 65275 2145957 1062450 740 2270 743 2257 1258 65274 2146026 1048463 708 2358 711 2347 1250 65274 2146044 1044

xy are pixel location in the images with (00) in lower left The range is in km from the camera location while latitude and longitude are in geographic degrees N and E Altitude is in km

FIG 4 Light curve of the meteor and its OI wake emis-sion derived from the integrated images () and fromtracings across individual images (solid line) Dashedlines are classical light curves for a single solid body withpeak intensity at different altitudes The wake follows themeteor light curve between 115 and 97 km but is 28 mag-nitudes fainter as demonstrated by the dotted line whichis the meteor light curve shifted by 128 magnitudes

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

meteor head but not much of the tail It did notinclude intensity lost in the central saturated pix-els but added intensity from the spatially ex-tended component and the diffuse scattered lightThe TV data were integrated over a 65 3 71 pixelrectangle that covered the full size of the bloomedarea Saturation effects were corrected for bytreating the measured intensity as an opticaldepth In the video there is a substantial back-ground signal which was subtracted by com-puting the average signal in two similar-size rec-tangles located on either side of the meteor To

increase its temporal resolution we measured atthe beginning of data collection the intensity vari-ation across the meteor images in each 130 s ex-posure while the blooming was still modest (solidline in Fig 4) Each data set was calibrated to theV magnitude of the stars in the field of view overthe range V 5 131 to 161 magnitude For sim-ilar cameras it was found that blooming offsetthe effects of saturation over a much wider mag-nitude range the relationship between bloomingand saturation being close to that expected if theelectron production is linear with incident lightbut the electrons distribute over more pixels V 525 log SIpixel (Jenniskens 1999) Indeed the ex-pected intensity in the central pixels of the HSIwas not much above the measured level thus ex-plaining the lack of blooming The slightly higherbrightness may reflect the fact that the HSI im-ager is sensitive to wavelengths longer than theV band which overestimates the brightness in thecalibration procedure However the limitednumber of calibration stars introduces a similarsystematic uncertainty

At the peak of its brightness the meteor hadan absolute magnitude (ie as seen from a dis-tance of 100 km) of 227 6 04 magnitude In-deed a meteor of this brightness would not havea bright persistent train Other Leonid meteors of227 magnitude have their peak brightness at97 6 4 km in agreement with the value of 1006km found here (Jenniskens et al 1998 Betlem etal 2000) On video records with limiting magni-tude approximately 16 such Leonids are usuallydetected first at altitudes of 143 km but theirbrightness does not increase to the photographiclimiting magnitude of about 11 until they decend

METEOR WAKE IN HIGH FRAME-RATE IMAGES 101

TABLE 1 OBSERVED AND CALCULATED POSITIONS OF THE METEOR IN THE HSI DATA

Calculated Observed

Frame x y x y Range Latitude Longitude Altitude

65 1552 14 1552 14 1493 65281 2145492 1156100 1488 188 1489 188 1471 65280 2145541 1146150 1396 446 1396 446 1440 65279 2145610 1132200 1300 714 1299 715 1409 65279 2145679 1118250 1198 996 1198 996 1378 65278 2145749 1104300 1090 1294 1092 1290 1348 65277 2145818 1090350 980 1610 981 1598 1317 65276 2145887 1076400 862 1932 865 1920 1287 65275 2145957 1062450 740 2270 743 2257 1258 65274 2146026 1048463 708 2358 711 2347 1250 65274 2146044 1044

xy are pixel location in the images with (00) in lower left The range is in km from the camera location while latitude and longitude are in geographic degrees N and E Altitude is in km

FIG 4 Light curve of the meteor and its OI wake emis-sion derived from the integrated images () and fromtracings across individual images (solid line) Dashedlines are classical light curves for a single solid body withpeak intensity at different altitudes The wake follows themeteor light curve between 115 and 97 km but is 28 mag-nitudes fainter as demonstrated by the dotted line whichis the meteor light curve shifted by 128 magnitudes

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

to 118 6 4 km which coincides with the rapidbrightening of the Leonid reported here Bothvideo and photographic records showed that theend point of the meteoroid trail was at 91 6 3 kmaltitude in good agreement with the altitude atwhich our Leonid meteor reached the photo-graphic limiting magnitude of 11 Hence a dif-ferent wide-field camera would not have detectedthe meteor much beyond the frame of the currentcamera field

Light intensity decay in the meteor wake

For each of the final frames 460ndash463 we sub-tracted the dark offset and the scattered lightcomponent The latter was found from a perpen-dicular scan over the calculated position of themeteoroid That scan is composed of a Lorentz-ian (full width at half-maximum 5 004 pixel155intensity units) and Gaussian profile (s 5 003pixel125 intensity units) We assumed that thelatter represented the spherical intensity halowith the bite-out from what appears to be a shockand the former due to scattered light We fittedthe Lorentzian component to the front part ofeach trace and found the peak to coincide to6150 m with the calculated position of the meteoroid After subtraction of this Lorentzian-shaped scattered light contamination the resultwas divided by the brightness of the meteor whenit was at that position (Fig 4) to obtain the decayof light intensity over time

The ratio of the final divided by the initial lightintensity is plotted in Fig 5 bottom trace (001ndash033s) on a log-log scale The graph shows two regimesof light decay the first being a continuous decayfrom 001 to 009 s This decay does not have a sin-gle 1e time scale but can be described by two 1edecay times of 65 ms (001ndash004 s) and 25 ms(004ndash009 s) Recall that the decay time of the in-tensifier phosphor is 08 ms a significantly smallervalue After 01 s there was a gradual increase ofintensity with time rather than a decrease

The cause of the two regimes of intensity de-cay could be identified from low resolution spec-tra of similar Leonids obtained during the 2001Leonid MAC mission (Fig 6) Two componentsare recognized in these spectra (1) a delayedgreen-line (557 nm) emission of O I and (2) awake of emission similar to that of the main me-teor that persisted for 1ndash2 frames or 006 s Theemission of Na I with a low upper energy levelof 210 eV persisted longer than the transitions

from higher energy states of Mg I (511 eV) theFirst Positive band of N2 (72 eV) O I (107 eV)and N I (118 eV)

This pattern is the same as the one found byBorovicIuml ka and Jenniskens (2000) who discovereda meteor afterglow in a 213 magnitude Leonidfireball ascribed to secondary ablation The newresults are the first confirmation that meteor af-terglow is a phenomenon present in relativelyfaint meteors

BorovicIuml ka and Jenniskens (2000) found that thedecay of line intensities depends on the excitationpotential rather than on the transition probabil-ity The intensity decay is therefore due primar-ily to the decrease of temperature rather thandensity We can use this property to calculate thetemperature variation from 001 to 009 s afterpassage of the meteoroid

BorovicIuml ka and Jenniskens (2000) found alsothat the exponential decay rate (B) defined asI(t) exp(2B t) of light intensity (I) of a giventransition depends linearly (factor D) on the ex-citation potential for most lines (E)

B 5 Bo 1 D E (1)

In our case the observed intensity decay is asum from the different components However at

JENNISKENS AND STENBAEK-NIELSEN102

FIG 5 Decay of light intensity and temperature be-hind the meteorData for t 00006 s are from the modelby Boyd (2000) Observations reported in this paper areshown as a dark band The solid line above is the inferredtemperature decay of the meteor vapor (marked ldquo23rdquo)Results from BorovicIuml ka and Jenniskens (2000) for a 213magnitude fireball afterglow are also shown The dashedline marked by a question is an extrapolated result forsmall meteoroids that is perhaps most relevant to the de-livery of organics in the origin of life

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

least two contributions from transitions withquite different decay rates are needed to explainthe observed decay of intensity The two mea-sured 1e decay rates correspond to B 5 153 andB 5 40 s21 respectively The spectral responsecurve of the HSI covered the range 400ndash800 nmwhich included Mg I Na I N2 and O I Hencethe shallow slope of the decay rate plot is mostlikely the response from sodium with E 5 21 eVThe initial steep decline is most likely due to theFirst Positive band of N2 but may have a contri-bution from O I (774 nm) and the much fainterMg I line If due to N2 then D 5 221 s21 eV21

and Bo 5 265 s21 and if due to O I then D 5131 s21 eV21 Bo 5 1124 s21

We adopt Bo 5 0 s21 (no production or de-struction of emitting compounds) and D 5 19 s21

eV21 The calculated time dependence of thetemperature of the meteoric vapor follows fromBorovicIuml ka and Jenniskens (2000)

5040T(t) 5 5040T(t 5 0) 2 D 3 t (2)

The result for T(t 5 0) 5 4500 K is shown in Fig5 and is compared with that calculated by

BorovicIuml ka and Jenniskens (2000) who had D 515 s21 eV21 a factor of 10 smaller This earlierresult pertains to a 213 magnitude fireball at 84km altitude and predicts correctly the tempera-ture of 50 K which was measured several min-utes later in the persistent train of a similar brightfireball by LIDAR (LIght Detection And Ranging)resonance scattering of Doppler-broadened sodium(Chu et al 2000) We conclude that D is either asteep function of the peak brightness of the me-teor or the rate of mass loss at the position of themeasured afterglow

DISCUSSION

Slowing down the loss of translational energy

The inferred rate of cooling is much slowerthan predicted in the model by Boyd (2000)shown on the left side of Fig 5 Boyd (2000) cal-culated the translational temperature for a rar-efied flow field around a 1-cm meteoroid at 95km altitude (close to the 04-cm meteoroid stud-ied here at an altitude of 105 km) The calcula-tions were performed for a grid that extended 40m behind the meteoroid representing times t 00006 s The calculated temperature decays ac-cording to T(t) t212 perhaps falling off morerapidly in exponential form at the end of the grid

The disagreement between the model and ob-servations is also apparent when one considersthe expected intensity from the model plasmaThe expected intensity of a local thermodynamicequilibrium (LTE) plasma is proportional to I exp(2EkT) where E is the excitation level of theupper state of an excited compound For the rel-evant E 72 eV the predicted line intensity isthe dashed line marked ldquometeorrdquo in Fig 6 Theemission is expected to peak sharply a few mean-free paths (1 m) from the position of the mete-oroid at a time t 2 3 1025 s where the (non-LTE) translation temperature is at its peak of4 3 106 K representative of a mean root meansquare molecule velocity of 05 times the impactspeed However that would have created a morepoint-like meteor with a strong blooming at themeteoroid position The dashed line on Fig 6shows the probable emission distribution with-out effects of blooming subtracted from the HSIimages We note that this more gradual intensitydecay favors the cooler LTE part of the tempera-ture curve and can perhaps help explain the low

METEOR WAKE IN HIGH FRAME-RATE IMAGES 103

FIG 6 Slit-less low-resolution spectrum of a 23 mag-nitude Leonid at 084531 UT November 18 2001 Thereis delayed emission of the O I green line and an after-glow of Mg Na O I N I and N2 The meteor moved fromleft to right against a backdrop of stars in the constella-tion of Orion The emission lines in the first order spec-trum are elongated because of the meteor motion duringthe 130 s exposure

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

T 4300 K in meteor spectra (Jenniskens et al2000a)

The role of meteor debris

Having established that something slowsdown the loss of translational energy or kinetictemperature of the meteor plasma the questionarises as to what is responsible for this effect Mc-Crosky (1955) first suggested that the luminosityof the wake is caused by the ablation of minuteparticles which become detached from the me-teoroid and lag behind it However Oumlpik (1958)criticized this idea because very small particlesvaporize before falling far enough behind to pro-vide an explanation for the observed length of thewake in bright fireballs (Bronshten 1983) AlsoHalliday (1958a) pointed out that changes in theluminosity of the wake are intimately linked tochanges in the luminosity of the meteor with nonoticeable lag more than 001 s This demands arapid deceleration of the responsible agent Andfinally the H and K lines of Ca1 were detectedin the wake of a bright fireball (with a peculiarV-shape splitting) which would demand colli-sions powerful enough to ionize calcium (61 eV)which are lacking in the thermal vapor of evap-orating fragments

Instead Halliday (1958a) argued that any mix-ture of meteoric atoms and air will tend to slowdown gradually relative to the ambient environ-ment in the wake of the meteor and considera-tion of conservation of momentum shows the relative speed to be on the order of 5 kms after001 s In this scenario the afterglow phenome-non would be caused by the collisional excitationof meteoric metal atoms and air plasma untilthese compounds have slowed down

Indeed BorovicIuml ka and Jenniskens (2000) mea-sured at least one trail segment in a meteor af-terglow that descended at a rate of 2 kms at02 s However that same trail fragment also hadBo 0 indicative of secondary ablation Othertrail segments did not move after formation orshow secondary ablation Of course meteoroidfragments can survive longer than envisioned byOumlpik (1958) if they are in the wake of the mete-oroid or its vapor cloud or if they are rapidly de-celerated by induced backward motion duringfragmentation

There is other evidence that links the afterglowphenomena to meteoroid fragmentation Therapid increase in brightness at the beginning of

the meteorrsquos trajectory may signify a fragmenta-tion that increased the effective surface area of themeteor vapor cloud The increase was much morerapid than predicted for a classical light curve ofa single solid body (dashed lines in Fig 4) Afterthat the vapor cloud seemed to remain constantand the meteor had a light curve expected for asingle body with a peak emission at 1004 km Themeteorrsquos O I wake remained faint until a secondevent that was associated with a rapid exponen-tial increase in brightness Subsequently the O Iwake intensity followed the meteorrsquos light curveclosely and was fainter by only 28 magnitudesor a factor of 13 in intensity (Fig 4 lower curve)In this regime the meteor behaved as if due to asingle solid body of a dimension correspondingto a peak intensity at an altitude of 1060 kmWhen its mass was spent the meteor intensity fellback to the original classical light curve but withlower fraction of mass

We interpret the light curve as follows Duringthe second fragmentation event (about frame235) a fraction of the meteoroid mass may havebroken into smaller fragments that were spent atabout 100 km altitude Part of those fragmentsmay have been slowed down enough to supportcontinuous ablation in the meteor afterglowSmall meteoroid fragments tend to be slowed rel-ative to the larger fragments by collisions in thevapor cloud because of their larger surface-to-mass ratio It is therefore significant that the af-terglow was seen to form shortly after the secondfragmentation event

While fragmentation appears to cause meteorafterglow even in relatively faint meteors it is notnecessarily the fragments that caused the gradualdecay in light intensity Secondary ablation is notalways detected It could be that fragmentation ef-fectively contributed to thermalizing a larger vol-ume of air plasma and that it was the hot plasmathat continued to excite the meteoric atoms

The lack of atmospheric lines and bands in thebright fireball afterglow could mean that the airplasma is enriched in meteoric metals in the re-gion contributing to the afterglow There couldalso be a narrower zone near the center of the me-teoroidrsquos (or meteoroid fragmentrsquos) path than thatfrom which the meteor emissions are observed

The O I wake

In addition to the afterglow there is also theoxygen green-line emission a sensitive indicator

JENNISKENS AND STENBAEK-NIELSEN104

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

of atom density The green line in natural airglowis caused by the catalytic recombination of oxy-gen atoms by meteoric metal atoms

O 1 O 1 M R O2 1 M (3)

O2 1 O R O (1S) 1 O2

The green-line intensity due to the 1S R 1D tran-sition and hence is proportional to [O]3 makingthis a sensitive indicator of O I density Once inan excited state the lifetime is 076 s (Baggaley1976) Jenniskens et al (2000c) calculated that theinitial dissociation of about 15 of the oxygenmolecules explain the intensity of the O I greenline emission in the persistent train of the Chip-penham Leonid fireball For a small meteoroidthe percentage of dissociated oxygen moleculeswould be much less because the oxygen atomswould be spread over the same volume that theyare in larger meteoroids resulting in very lowvolume densities The green-line emission wasobserved to peak at 04 s after the meteoroidand then quickly decayed There was no O I emis-sion in the meteor spectrum Kinetic processesthat control Reaction 3 may be responsible for theinitial increase in O I green-line intensity whilediffusion of the oxygen atoms would have low-ered the O I density and caused the subsequentdecay

IMPLICATIONS

Temperature decay in faint meteors

If meteoric vapor or debris was shielded fromthe impinging air flow or slowed down by fa-vorable induced speed during fragmentationthen the material would have been deceleratedless violently This may enhance the survival oforganic compounds Whether this process is im-portant for all small 150-mm meteoroids at thepeak of their mass influx curve depends onwhether fragmentation occurs in a similar man-ner or whether shielding can occur efficiently

Small grains tend to have a short track and ab-late at an altitude independent of a mass (but in-creasing with speed) of about 89 km Their meanspeed is 25 kms (Taylor and Elford 1998) Whilethe gas cools from 4100 K to 250 K the collisionfrequency at that altitude increases from 2000sto 8000s The ambient temperature is about 220

K Cometary grains are thought to have verysmall 01-mm subunits (Greenberg 2000) sothat abundant fragmentation is expected to occureven in grains 150 mm in diameter Shieldinghowever is less efficient for small grains becausethe vapor cloud is expected to be of similar sizebut less dense We propose that even in smallgrains fragmentation and the relatively larger va-por cloud will increase the effective volume of airbeing processed leading to a more gradual cool-ing in the wake

In an empirical approach the expected dura-tion of the afterglow (B) for smaller meteoroidsis mostly a function of how rapidly the parame-ter D decreases with decreasing mass D is not afunction of Bo (that is the amount of secondaryablation) or of the speed with which the mete-oric plasma can be slowed down (nearly instan-taneous) Rather D is likely a function of theamount of air plasma that is being created If soD may be proportional to the mass loss for thegiven mass altitude (air density) and speeddMdt m23 ra V 3 (McKinley 1961) The massloss rate of a typical 150-mm grain at the peakof the mass influx curve is 8 3 1024 times that ofthe 23 magnitude Leonid studied here while thendash13 magnitude Leonid (at 84 km) had a mass lossrate 4 3 104 times larger If D behaves as a powerlaw D (dMdt)2024 and D 5 100 s21 eV21 fora 150-mm grain If D behaves as a logarithmicfunction then D is approximately 238 log(dMdt) and D 5 30 s21 eV21 If D is a linear orexponential function of dMdt then D 5 19 s21

eV21 as measured for our Leonid For an assumedD 5 100 s21 eV21 the temperature decay is asshown by the dashed line in Fig 5 marked by aquestion mark With this choice there remains anenhancement over the calculated trend of T(t) t212 by Boyd (2000)

Organic chemistry in the meteor wake

In a CO2-rich early Earth atmosphere all of theCO2 would have dissociated at 4300 K to formoxygen atoms and CO (Jenniskens et al 2000a)These products along with electrons would havebeen the main reactive species in the air plasmaO2 and O3 would have formed in very low abun-dance Hydrogen from the organic matter in themeteoroid may have contributed some amount ofH and OH radicals

In a combustion process the efficiency ofchemical processes depends on the generation

METEOR WAKE IN HIGH FRAME-RATE IMAGES 105

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

and propagation of free radicals Indeed O atomsare used to reduce hydrocarbon pollutants ingaseous industrial waste and on cleaning sur-faces In the case of hydrocarbons complete mol-ecular conversion will result in the formation ofcarbon dioxide and water In that process manydifferent reactions occur some involving molec-ular oxygen and other compounds In the pres-ence of only O and CO the possible chemistry islimited to C and H extraction reactions and oxy-gen insertion

In the meteor phase there are about 6 3 104

oxygen atoms relative to all meteoric metal atomsin a 23 magnitude Leonid (Jenniskens et al2004a) or an O I density of about 5 3 1017 atomscm3 at 100 km For a 150-mm meteoroid at 25kms the kinetic energy decreases by a factor of4 3 1027 With O atoms spread out over the samevolume the density of O atoms generated by a150-mm meteoroid is only 2 3 1011 atomscm3Since the standard atmospheric background O Idensity at 100 km is about 2 or 15 3 1012

atomscm3 the meteor-induced contribution ofoxygen atoms is quite small

In a CO2-rich early Earth atmosphere each or-ganic compound released from the meteoroidwould have been bombarded about 2000s 32 3 002 s 5 08 times before the plasma hadcooled to 1800 K (after about 002 s) A similarnumber of collisions with CO would also be ex-pected as well as about 40 collisions with CO2The oxygen atoms would have reacted with or-ganic compounds to make radicals which couldthen react efficiently with other atmospheric com-pounds

R-CH 1 O R R-C 1 OH (4)

followed by

R-C 1 CO2 R R 1 2COR R-CO 1 COR R-C C 1 O2 (5)

No collisions between organic compounds wouldbe expected at this stage If only a few percent ofthe collisions with air compounds were inelasticand led to chemical reactions (Oumlpik 1958) evensmall organic molecules could survive this me-teor phase albeit with some chemical change

In the wake of the meteor these reactionswould have been inhibited by the rapid declineof radical concentrations to ambient background

levels and by the quenching of existing organicradicals After the air plasma had cooled to 1800K the oxygen atom abundance is expected tohave quickly declined to the atmospheric back-ground level if the plasma remained in LTE In alow radical and high ultraviolet background or-ganic radicals would be removed by losing H andby reactions with atmospheric radicals electronsand ions In todayrsquos Earth atmosphere metalatoms help lower the O I abundance in the me-teor column by facilitating catalytic reactionswith ambient ozone

Although radical chemistry does not normallyinvolve activation energy barriers and is not tem-perature dependent the frequency of collisionsand the abundance of radicals and ions in the me-teor wake over time is This work provides the dataneeded to calibrate the conditions in the coolingplasma behind the meteoroid from which detailedchemical modeling can be attempted an effort out-side the scope of this paper

ACKNOWLEDGMENTS

The 2001 Leonid MAC was sponsored byNASArsquos Astrobiology ASTID and Planetary As-tronomy programs This work meets with objec-tive 31 of NASArsquos Astrobiology Roadmap (DesMarais et al 2003)

ABBREVIATIONS

CCD charge coupled device HSI high-speedimager LTE local thermodynamic equilibriumMAC Multi-Instrument Aircraft Campaign UTuniversal time

REFERENCES

Baggaley WJ (1976) The excitation of the oxygenmetastable OI (lsquos) state in meteors Bull Astron InstCzechoslov 27 173ndash181

Baggaley WJ (1977) The velocity dependence of mete-oric green line emission Bull Astron Inst Czechoslov28 277ndash280

Betlem H Jenniskens P Spurny P Docters vanLeeuwen G Miskotte K Ter Kuile CR Zarubin Pand Angelos C (2000) Precise trajectories and orbits ofmeteoroids from the 1999 Leonid meteor storm EarthMoon Planets 82-82 277ndash284

BorovicIuml ka J and Jenniskens P (2000) Time resolved

JENNISKENS AND STENBAEK-NIELSEN106

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

spectroscopy of a Leonid fireball afterglow Earth MoonPlanets 82-83 399ndash428

Boyd ID (2000) Computation of atmospheric entry flowabout a Leonid meteoroid Earth Moon Planets 82-8393ndash108

Bronshten VA (1983) Physics of Meteoric Phenomena DReidel Publishing Co Dordrecht The Netherlands

Ceplecha Z (1992) Earth influx of interplanetary bodiesAstron Astrophys 263 361ndash366

Ceplecha Z BorovicIuml ka J Elford WG ReVelle DOHawkes RL Porubcan V and Simek M (1998) Me-teor phenomena and bodies Space Sci Rev 84 327ndash471

Chu X Liu AZ Papen G Gardner CS Kelley MDrummond J and Fugate R (2000) Lidar observationsof elevated temperatures in bright chemiluminescentmeteor trails during the 1998 Leonid shower GeophysRes Lett 27 1815ndash1818

Des Marais DJ Allamandola LJ Benner SA DeamerD Falkowski PG Farmer JD Hedges SB JakoskyBM Knoll AH Liskowsky DR Meadows VSMeyer MA Pilcher CB Nealson KH SpormannAM Trent JD Turner WW Woolf NJ and YorkeHW (2003) The NASA Astrobiology Roadmap Astro-biology 3 219ndash235

Greenberg JM (2000) From comets to meteors EarthMoon Planets 82-83 313ndash324

Halliday I (1958a) Meteor wakes and their spectra As-trophys J 127 245ndash252

Halliday I (1958b) Forbidden line of O I observed in me-teor spectra Astrophys J 128 441ndash443

Halliday I Mcintosh BA Babadzhanov BP and Get-man VS (1980) Orbit chemical composition atmos-pheric fragmentation of a meteoroid from instanta-neous photos In Solid Particles in the Solar System editedby I Halliday and BA McIntosh D Reidel PublishingCo Dordrecht The Netherlands pp 111ndash116

Hawkes RL and Jones J (1978) The effect of rotation onthe initial radius of meteor trains Month Not R As-tron Soc 185 727ndash734

Jenniskens P (1999) Activity of the 1998 Leonid showerfrom video records Meteoritics Planet Sci 34 959ndash968

Jenniskens P (2001) Meteors as a vehicle for the deliveryof organic matter to the early Earth In ESA-SP 495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-tions Division Noordwijk The Netherlands pp247ndash254

Jenniskens P (2003) Leonid MAC near-real time fluxmeasurements and the precession of comet 55PTem-pel-Tuttle ISAS-SP 15 73ndash80

Jenniskens P and Russell RW (2003) The 2001 LeonidMulti-Instrument Aircraft Campaignmdashan early reviewISAS Rep SP 15 3ndash15

Jenniskens P de Lignie M Betlem H BorovicIuml ka JLaux CO Packan D and Kruumlger CH (1998) Prepar-ing for the 199899 Leonid storms Earth Moon Planets80 311ndash341

Jenniskens P Wilson MA Packan D Laux COKrueger CH Boyd ID Popova OP and Fonda M(2000a) Meteors A delivery mechanism of organic mat-ter to the early Earth Earth Moon Planets 82-83 57ndash70

Jenniskens P Lacey M Allan BJ Self DE and PlaneJMC (2000b) FeO ldquoOrange Arcrdquo emission detected inoptical spectrum of Leonid persistent train Earth MoonPlanets 82-83 429ndash438

Jenniskens P Nugent D and Plane JMC (2000c) Thedynamical evolution of a turbulent Leonid persistenttrain Earth Moon Planets 82-83 471ndash488

Jenniskens P Schaller EL Laux CO Schmidt G andRairden RL (2004a) Meteors do not break exogenousorganic molecules into high yields of diatomics Astro-biology 4 67ndash79

Jenniskens P Laux CO Wilson MA and SchallerEL (2004b) The mass and speed dependence of meteorair plasma temperatures Astrobiology 4 81ndash94

Kondratrsquoeva ED and Reznikov EA (1985) Comet Tem-pel-Tuttle and the Leonid meteor swarm Solar Syst Res19 96ndash101

Le Blanc AG Murray IS Hawkes RL Worden PCampbell MD Brown P Jenniskens P Correll RRMontague T and Babcock DD (2000) Evidence fortransverse spread in Leonid meteors Month Not R As-tron Soc 313 L9ndashL13

Love SG and Brownlee DE (1993) A direct measure-ment of the terrestrial mass accretion rate of cosmicdust Science 262 550ndash553

McCrosky RE (1955) Fragmentation of faint meteors As-tron J 60 170

McKinley DWR (1961) Meteor Science and EngineeringMcGraw-Hill New York

Murray IS Hawkes RL and Jenniskens P (1999) Air-borne intensified charge-coupled device observationsof the 1998 Leonid shower Meteoritics Planet Sci 34949ndash958

Oumlpik E (1958) Physics of Meteor Flight in the AtmosphereInterscience New York

Oroacute J and Lazcano A (1997) Comets and the origin andevolution of life In Comets and the Origin and Evolutionof Life edited by PJ Thomas CF Chyba and CPMcKay Springer Verlag New York pp 3ndash28

Park C and Menees GP (1978) Odd nitrogen produc-tion by meteoroids J Geophys Res 83 4029ndash4035

Ponnamperuma C ed (1981) Comets and the Origin of LifeProceedings of the 5th College Park Colloquium on Chemi-cal Evolution D Reidel Publishing Co Dordrecht theNetherlands

Popova OP Sidneva SN Shuvalov VV and StrelkovAS (2000) Screening of meteoroids by ablation vapor inhigh-velocity meteors Earth Moon Planets 82-83 109ndash128

Spurny P Betlem H van lsquot Leven J and JenniskensP (2000) Atmospheric behavior and extreme beginningheights of the thirteen brightest photographic Leonidmeteors from the ground-based expedition to ChinaMeteoritics Planet Sci 35 243ndash249

Stenbaek-Nielsen HC Hallinan TJ Wescott EM andFicircppl H (1984) Acceleration of barium ions near 8000km above an aurora J Geophys Res 89 10788ndash10800

Stenbaek-Nielsen HC and Jenniskens P (2003) Leonidat 1000 frames per second ISAS SP 15 207ndash214

Taylor AD and Elford WG (1998) Meteoroid orbital el-ement distributions at 1 AU deduced from the Harvard

METEOR WAKE IN HIGH FRAME-RATE IMAGES 107

Radio Meteor Project observations Earth Planets Space50 569ndash575

Taylor MJ Gardner LC Murray IS and JenniskensP (2000) Jet-like structures and wake in MgI (518 nm)images of 1999 Leonid storm meteors Earth Moon Plan-ets 82-83 379ndash389

Thomas PJ Chyba CF and McKay CP eds (1997)Comets and the Origin and Evolution of Life Springer Ver-lag New York

von Zahn U (2001) Lidar observations of meteor trailsEvidence for fragmentation of meteoroids and theirsubsequent differential ablation In ESA SP-495 Pro-ceedings of the Meteoroids 2001 Conference ESA Publica-

tions Division Noordwijk The Netherlands pp303ndash314

Address reprint requests toDr Peter Jenniskens

Center for the Study of Life in the UniverseSETI Institute

2035 Landings DriveMountain View CA 94043

E-mail pjenniskensmailarcnasagov

JENNISKENS AND STENBAEK-NIELSEN108