Документ взят из кэша поисковой машины. Адрес оригинального документа : http://star.arm.ac.uk/preprints/2009/539.pdf
Дата изменения: Wed Feb 4 13:18:25 2009
Дата индексирования: Tue Oct 2 02:51:38 2012
Кодировка:

NAPIER, ASHER: TUNGUSK

A

The Tunguska impact event and beyond
Bill Napier and David Asher discuss the Tunguska event of 1908. Assessing the impact hazard requires an understanding of the effect of the solar system's galactic environment on the Oort comet cloud. ABSTRACT
Current strategies for dealing with the impact hazard are geared towards the detection and deflection of near-Earth asteroids, which typically have approach speeds ~20 km s1 and involve decades of warning. However, galactic signals in the age distribution of well-dated impact craters suggest that the globally destructive impactors (diameters between 1.5 and 2 km and upwards) ultimately derive from the Oort cloud. Warning times are then measured in months or days, and characteristic approach speeds are ~55 km s1. Concentrations of sub-kilometre debris in meteor streams may also be a significant regional hazard. Intersection with the debris of a large short-period comet may account for the widespread biological and cultural dislocation in North America around 12 900 BP.

Table 1: Possible impact effects based on energy
10 000 Mt impactor 500 m scope land regional fires, blast, and earthquake over 2501000 km uncertain 1 million Mt 12 km civilization-destroying destructive blast, quake and possibly fire over continental dimensions (mega)tsunamis around ocean rims 100 million Mt 10 km species-destroying global conflagration and destructive earthquake ocean rim devastation; cities replaced by mudflats; oceans acidified

sea

air

Sun obscured

land and sea ecologies agriculture collapses; ozone depletion; acid rain; collapse; ozone depletion; acid rain; sky black for sky darkened for years years global warming followed by sharp cooling global warming followed by cosmic winter

climate

possible brief cooling

Uncertainties attend all these thumbnail descriptions. Timescales are also open to debate, as discussed in the text.

T

he Tunguska impact of 30 June 1908, which destroyed 2000 square kilometres of conifer forest in a sparsely populated region, the Central Siberian Plateau, had the energy of a large hydrogen bomb (figure 1). No m e t e o r i t e c r at e r a s s o c i at e d w i t h i t h a s b e e n secu rely identi fied. A nu mber of conferences in Moscow, held around the centenar y of the event, brought home that even this fairly recent, much explored impact holds many mysteries. It is occasionally pointed out that if the 1908 i mpac t had taken place over a met ropolitan area, huge damage would have been in fl icted. I n the case of an impact on London, a bolide brighter than the Sun, and leaving a thick trail of smoke, would have been seen approaching from half way across France. The gun fi re-like bangs of the impact would have been heard across Britain to I reland, north to Orkney and Denmark, and over Europe as far as Switzerland. People would have had their hats knocked off in Glasgow and Edinburgh, topsoil would have been stripped from fields in Cheshire, trains would have been derailed throughout central
1.18

England, and people in Oxford would have been thrown through the air and severely burned. A n incandescent column of matter would have been thrown 20 km into the air over London, and the cit y itself would have been destroyed about as far out as the present-day M 25 ring road. The political ramifications that would have followed the destruction of Edwardian London are a matter for speculation; one may question whether the British Empire would have survived. T h e i m p a c t e n e r g y o f t h e e ve n t r e m a i n s uncer tain. It might have been as low as 3 Mt (megatons T N T equivalent) some simulations suggest this, with fierce vortex winds responsible for destroying an already weak Siberian forest. On the other hand, seismic and barometric data have consistently been interpreted as pointing to a higher impact energy, typically 10 15 Mt (BenMenahem 1975). The occurrence of 20 impacts from a fragmented comet, D/1993 F2 (Shoemaker-L ev y 9), on Jupiter as recently as July 1994 demonstrates that planetary impacts are common at high energies too the characteristic energy of the fragments was about 100 000 Mt

(Asphaug and Benz 1996), enough to cause devastation on continental scales on Earth. Estimates, all of them uncertain, have been made of the damage expected from impactors of various sizes (see table 1). The threshold for a civilization-destroying impact, killing over a billion people, comes in at a 1 or 2 km diameter bolide. Above a certain energy (~106 Mt), vaporized material thrown out from the impact punches out through the atmosphere and spreads globally. Everywhere on the surface of the Earth, the sky is red hot and a global con flagration results. Once the initial heat pulse has passed, micron-sized dust particles and vapour condensates in the atmosphere may take characteristically between 1 and 10 years to settle down, collapsing food chains in the meantime. A nd, because about 10 0 million people live within 2 km of a shoreline, large ocean impacts have the potential to cause severe tsunami devastation.

Nature of the Tunguska bolide
If we rule out crashed spacecraf t, black holes, a n t i m at t e r p a r t i c l e s , n at u r a l H - b o m b s a n d
A&G · February 2009 · Vol. 50

NAPIER, ASHER: TUNGUSK

A

1: The explosion due to an incoming cosmic body over the Tunguska region of Siberia in 1908 flattened trees over 2000 square km.

geophysical explosions as causes of the Tunguska impact, we are down to a comet or asteroid. In 1930, the British meteorologist F J W W hipple suggested that the Tunguska body was a small comet, and this view has generally been s up p o r t e d b y R u s s i a n a s t r o n o m e r s . H u g h e s (1976) likewise considered that the object was probably a small comet. The A merican astronomer Fred W hipple (1975) not the meteorologist thought it more likely that the bolide was an inactive, low-densit y, friable body. The suggestion by K resАk (1978) that the body was a fragment of the short-period comet 2P/ Encke, and therefore part of the Taurid Complex, was s up p o r t e d b y t h e c o i n c i d e n c e i n t h e d at e o f impact (30 June) with the Earth's annual passage through the day time Taurids. T he trajectory of the bolide moreover lies within 20° of that of the comet, a difference explicable by planetary perturbations (Asher and Steel 1998). Sekanina (1983), however, argued that a body c o m p o s e d o f w e a k c o m e t a r y m at e r i a l c o u l d not have survived intact on a journey into the lower atmosphere, and proposed instead that
A&G · February 2009 · Vol. 50

the object was dense and rocky, probably from the asteroid belt. Recent hydrocode simulations by both Russian and A merican groups revealed that, from the perspective of impact mechanics, the object could have been either cometar y or asteroidal. T h e f o r wa rd m o m e n t u m o f a l a r g e fi r e b a l l breaking up even at high altitude can bring it close to the ground. T he severe dear th of cosmic material on the ground may be due to the updraught of the fi reball, which forms a rap idly rising, near-spherical plume. Debris from the plume may be spread over about 150 0 km, and scat tered sunlight from this debris could a c c o u nt f o r t h e w i d e ly r e p o r t e d r e ad i n g of newspapers, overnight cricket, midnight photog raphy and the like du ring the nig ht of 30 June 1908 in England. T he simulations show t hat t he objec t cou ld have been a 50 60 m d i a me ter s tony a s teroid , or a n 8 0 10 0 m comet: either would produce similar effects at the Tunguska site. A 70 m comet falling ver tically could reach the ground, whereas one up to 1 km across, coming in at 5° to the horizontal,

would unload nearly all its kinetic energ y into the atmosphere. A n interesting debate has been stimulated as to whether there is in fact an impact crater at the site. A group in Bologna (Gasperini et al . 20 07) has suggested that Lake Cheko, a 30 0 m lake 8 km downrange from the epicentre, may be such. It is steep-sided and bowl-shaped, and cone-shaped at depth. It does not seem to be a meander lake or volcanic depression, and it is not shown on an 1883 map of the area. A seismic anomaly exists just below the bed of the lake. On the other hand, there is no evidence around the lake of high shock pressure or temperature and no sign of ejected material. Further, it seems that no trees were affected by the postulated impact even at the edge of the lake. Numerical simulations by Collins et al. (20 08) have failed to reconcile these con flicting factors. A nalyses of peat columns in the catastrophe layers have revealed isotopic composition shif ts for carbon-13 and deuteriu m i n add ition to enhanced iridium and enriched siderophile elements. These have been interpreted as evidence
1.19

NAPIER, ASHER: TUNGUSK

A

of a cometary origin for the Tunguska cosmic body (Kolesnikov and Rasmussen 20 08).

Table 2: Impact frequency estimates
author Sekanina and Yeomans Bailey and Emel'yanenko Nurmi et al. Rickman et al. Hughes Stokes et al. Stuart and Binzel Asher et al. year 1984 1998 entity collisions method historic comet encounters comet dynamics comet dynamics comet dynamics Spaceguard close encounters with NEAs close encounters with comets Spaceguard lunar meteorites close encounters with NEAs active comets 43 Myr (mostly 1 km) Halley-type comets "comparable with NEAs"

Impact frequencies
Estimates of Tunguska-like impact frequencies have varied enormously over the years. K resАk (1978), extrapolating from fireball data, thought that such events might happen at about 50 -year intervals. Hughes (1976) considered them to be once in 20 0 0 -year events on the assu mption that the impactor was a small active comet, a figure which has also been derived on the basis of Spaceguard survey observations (table 2). W hat appear to be records of similar impacts are to be found throughout early my thological literature. Hellenistic my th, for example, includes the story of Phaethon, who borrowed the chariot of his father Helios, but was unable to control its horses as they crossed the sky, with the result that the chariot crashed to the ground in a blinding light, flattening and burning forests, poisoning rivers and darkening the sun. Various commentators over the years have regarded these and similar myths as referring to one or more real events (Plato, Goethe, Kugler, Engelhardt). Similar tales are to be found in the earliest Sanskrit literature, throughout the near East and as far away as China. The earliest recorded literature containing such material is to be found in Babylonian cosmology, going back to 20 0 0 BC but probably based on preliterate oral traditions. Megaton- class impacts may therefore have impressed themselves from time to time on early cultures. B eg i n n i ng i n t he 1970 s , helped by satellite o b s e r va t i o n s o f t h e E a r t h a n d S c h m i d t t e l e scope sur veys, quantitative assessments of the i mpac t rate became possible. One approach, then, is to ex trapolate from the known impact craters on Ear th, using some energ y diameter scaling relation. Many impact craters are so large that they are not easily recognized from the ground, the crater diameters ex tending far beyond the visible horizon (as in the case of Lake Manicouagan in Quebec , which is about 10 0 km across). About 170 i mpac t st r uc t u res a re k now n on E a r t h , with another dozen or so candidates. T hey are ver y unevenly distributed, being concentrated around the Baltic and Canadian Shields as well as deser t areas. T h roug hout I ndia, Pakistan, T i b e t , C h i n a a n d t h e Fa r E a s t , t h e r e i s o n l y one recorded impact crater: the L onar Crater i n c e n t r a l I n d i a , ab o u t 1 k m a c r o s s , 50 0 0 0 years old and already heavily eroded it will d i s ap p e a r i n t h e b l i n k o f a g e o l o g i c a l e y e . T hus the data set of terrestrial impact craters is ex tremely incomplete. Moreover, sub -kilo m e t r e b o d i e s w i l l t e n d t o d i s i n t e g r at e i n t h e atmosphere and will be underrepresented in the impact cratering record. T hey may nevertheless generate damaging airbursts, and on timescales of im mediate human interest may be the most
1.20

2001 captured Oort <200 000 yr (>1 km) cloud comets 2001 Jupiter family 2003 1000 Mt 2003 LP/NEA 2004 10 Mt impacts 1000 Mt 2005 10 Mt impacts 1000 Mt 63 000 ± 8000 yr 3000 yr
~1% ~1 Myr (>1 km)

Morbidelli et al. 2002 1000 Mt

20003000 yr 56 000 ± 6000 yr
<300 yr

5005000 yr

Estimates of impact frequencies at various energy levels, from various approaches. NEA = near-Earth asteroids; LP = long-period comets.

dangerous class of impact. E stimates of the collision hazard posed by i nt e r pl a n e t a r y b o d i e s o n s hor t t i m e s c a l e s may be arrived at in other ways. One proce du re is to ex trapolate the know n population of near-Earth objects (N EOs) currently being revealed by several comet and asteroid search programmes (such as the Spacewatch, LI N E A R and Catalina surveys). From these surveys, it is generally considered that completeness of discover y has now approached 10 0% for bodies >3 km in diameter, ~80% for >1 km diameter, declining rapidly to a fraction of a percent for objects around 10 0 m or less. A movie of the known N EOs (an animation by Scot t Manley is at ht tp: //star.arm.ac.uk /neos/ anim.html) would reveal that there is a brisk movement of material between the main asteroid belt and the inner planetary system. It is known furthermore that asteroids deflected from the mai n belt on ti mescales ~106 yr (Morbidelli 1999) are a prime source of hazardous bodies. Generally missing from such movies are comets; and yet the number of N E As (near-Earth asteroids) more than about 5 km in diameter capable of striking the Earth is tiny at present the only bodies in Earth- crossing orbits with diameters of this order are comets. It may be that below a certain threshold, asteroids diverted from the main belt are the prime hazard, while above this, comets are dominant. B y s o m e a c c o u n t s , l o n g- p e r i o d c o m e t s a r e

responsible for only ~1% of terrestrial impacts (Stokes et al. 20 03). Considering that the end of civilization, and perhaps even the human species, might be at stake, even this background hazard may be seen as a mat ter for concern. Moreover, others set the "global catastrophe" comet impact rate at a level comparable with that of the N E A s (Bailey and E mel'yanen ko 1998), and perhaps even dominant (R ickman et al. 20 01). To understand the hazard bet ter, we need to know the relative contribution of asteroids and comets at different energy levels. This mat ters because, mass for mass, comets have an order of magnitude more impact energ y (the mean impact speed of a comet in a Halley-t ype orbit is ~55 km s 1 as against ~20 km s 1 for an N E A); because t he wa r n i ng t i me for a n i ncom i ng comet may be measu red in months or weeks rather than centuries or decades as we would expec t from a well-mapped out N E A population; and because dormant comets may be extremely hard to detect. A nother consideration is that, although we could in principle map out hazardous bodies in asteroid-like orbits within a decade or t wo, mapping of a population of dark bodies in, say, high- eccentricit y Halleyt y pe orbits to 90% complet ion wou ld t a ke bet ween 10 0 0 and 20 0 0 years with present-day technology. They thus constitute an essentially unpredictable hazard. Table 2 lists various estimates of impact rates,
A&G · February 2009 · Vol. 50

NAPIER, ASHER: TUNGUSK
o
7*

A

o o *

o

oo *

o *

oo *

o * *

o

o *

6

5

4

and the fact that the passage occurred in 1983 indicates that we are faced with either another statistical fluke or a hazard that is somehow being underestimated. I A A was a peculiar comet with very low activit y, only about 1% of its surface being active. It was discovered only t wo weeks before its closest approach to Earth. T he suggestion here is that we may be dealing with a population of dark objects, carr ying a lot of kinetic energy, which are not being properly picked up in the Spaceguard surveys.

3

A problem of mass balance
We know that about one bright comet (of absolute magnitude as bright as 7, comparable to Halley's Comet) arrives in the visibilit y zone (perihelion q < 5 AU, say) each year from the Oort cloud. It seems to be securely established that ~12% of these are captured into Halleyt y pe (H T) orbits (E mel'yanen ko and B ailey 1998). T he dy nam ical lifeti me of a body in s u c h a n o rb i t c a n b e e s t i m at e d , f r o m w h i c h the expected number of H T comets is perhaps ~30 0 0. The actual number of active H T comets is ~25. This discrepancy of at least t wo powers of 10 in the expected impact rate from comets as deduced from this theoretical argument on the one hand, and observations on the other, is an aspect of the well-known fading problem of cometar y dynamics. A similar problem holds with regard to Jupiter family comets (orbital periods <20 years): many more dormant comets should exist in such orbits than are obser ved (R ickman et al. 20 01). T hree ways in which comets might fade out have been discussed in the literature. Firstly, the comets may disintegrate to dust (Levison et al. 20 02). To avoid conflict with observation, however, the disintegration needs to proceed with ~99% disruption ef ficiency within one or t wo perihelion passages, and this is not observed. Comets on the way out look much as they did on the way in: the archet ype, Halley's Comet, has been reliably observed for almost 30 revolutions, and all the major meteor streams have an active o r d o r m a nt s o u r c e c o m e t e m b e d d e d w it h i n them. A nother dif ficult y with the hypothesis is that the dust from the disinteg rated comet would be observed as a glowing disc in the sky af ter sunset or before dawn. A third problem is that the greatest dear th of comets is found at larger perihelia, whereas one would expect disinteg ration to proceed most ef ficiently for comets that reach small perihelion distances (R ickman 20 05). S e cond ly, t he come t s m ay b e come dorm a n t , d e ve l o p i n g d a r k m a n t l e s ( B a i l e y a n d Emel'yanenko 1998). The problem here is that even for albedos p 0.04, characteristic of the inactive surfaces on comets, the Spaceguard surveys should by now have detected ~ 40 0 dark comets >2 . 5 km across. However, only ~25, the
1.21

N

2

1

0

0

50

100

time (Myr)

150

200

250

2: The age distribution of 40 impact craters 3km or more in diameter, with ages less than 250Myr known to precision better than 10Myr (data smoothed in this plot by a window of width 8Myr). Impacts occur in discrete episodes of bombardment. The circles represent the formation date for 12 craters over 40km across with ages measured to precision 2.6Myr or better. The asterisks mark out a best-fitting periodicity of ~35Myr for those 12. (See figure 5.)

their method of derivation and the nat u re of i mp a c t o r o r e n e r g y y i e l d r e f e r r e d t o . I f we c o n s i d e r h i s t o r i c e n c o u nt e r s w i t h c o m e t s a n d close approaches with active comets , we fi nd comfor tably long inter vals bet ween collisions. H igher impact rates tend to be estimated from considerations of t he m ig ration or evolution b e t w e e n va r i o u s s u b - p o p u l a t i o n s o f c o m e t s that inhabit the planetar y system. T he number of bodies in a reser voir, and the rate at which they are calculated to transfer bet ween reservoirs, yield theoretical estimates of population based on the assumption of equilibrium. T hese mass balance calculations lead to much higher impact estimates, t ypically by t wo powers of ten, than those obtained from direct obser vations of the arrival of active comets from the various reser voirs. Table 2 shows that this discrepancy holds even down to 10 Mt (Tunguskasized) impactors. If we consider close encounters of small interplanetary bodies with the Earth, which may be less subject to various modelling uncertainties, then the estimates of impact rate become substantially higher. A Tunguska-like impact then becomes something like a 30 0 -year event, as against a 20 0 0 or 30 0 0 -year one as deduced from Spaceguard sur veys (the lat ter implying that such an impact a mere cent u r y ago was something of a statistical fluke). A more direct e s t i m at e c o m e s f r o m t h e t h r e e d o z e n o r s o meteorites ejected from the Moon and found in
A&G · February 2009 · Vol. 50

desert and A ntarctic regions. Assuming that it takes a 10 Mt impact on the Moon to dislodge material at escape velocity with the potential to land on Earth, and from the likeliest survival times on Earth, Asher et al. (20 05) deduced a lunar impact rate which translated to Tunguska impacts at <30 0 -year intervals on Earth. I n summary, there seems to be a discrepancy bet ween what is inferred mainly from population dynamics of comets, and what is observed in the Spaceguard surveys.

A case study: IRAS-Araki-Alcock
I n the N EO Science Defi nition Report of 20 03, it was pointed out that only t wo comets passed within 0.1 AU of the Earth during the 20th century, as compared with 155 N EAs over the same period. Hence the impact ratio bet ween comets and asteroids was taken to be 2 /155, 1%. However, one of the t wo comets, C /1983 H1 (I R AS -A raki-Alcock), passed within 0.03 AU of the Earth (730 Earth radii), and so in terms of numbers the ratio is about 1/17. Further, comet I A A had d i mensions of 11 7 7 k m , close to that of 1P/ Halley, and an encounter speed ~ 4 4 km s 1. T he impact energy of such a body would be ~20 0 million Mt. Clearly, in making the comparison, like is not being compared with like. Based on, for example, the Sekanina and Yeomans (1984) encounter rate with comets, such a passage involving an ac tive comet of this size is expected about once in 560 0 years,

NAPIER, ASHER: TUNGUSK
20 15 10 5 0 5 10 15 20 20

A

Table 3: Impacts resulting from main belt asteroids
family % episode no. of hitting duration >1 km Earth (Myr) impacts 1.5 0.2 0.2 0.02 0.02 0.02 0.007 0.0026 30 10 10 5 5 5 140 90 4 0 12 2 2 0 2 3

phase (Myr)

Flora Vesta Eunomia Gefion Dora Koronis
25 30 35 period (Myr) 40 45 50

Eos Themis

3: Retrieving the periodicity. The motion of the Sun around the galaxy has been simulated using galactic plane data derived from Hipparcos. The flux of comets from the Oort cloud varies pro rata with the local galactic tide and also has random components due to encounters with molecular clouds which exist preferentially in the galactic plane. Synthetic impact craters are extracted from the dynamical model, assuming impact probability to be proportional to the flux of comets from the Oort cloud, and taking account of the disappearance of terrestrial impact craters with time. The synthetic data are then analysed for periodicity (de-trending and applying power spectrum analysis to bootstrapped data). The inbuilt periodicity in the model is well retrieved (P 36 Myr, 2 Myr). Phase is defined as the time elapsed since the most recent episode. Other sets of harmonic solutions sometimes arise, depending on the vagaries of the randomly selected data. In the example above a second group of solutions appears strongly, P 26 Myr, 11 Myr. Interestingly, this is a close match to the periodicity which Raup and Sepkoski (1984) claimed to exist in the extinction record of marine families.

Terrestrial impacts expected from main belt asteroid disintegrations over 108109 yr (ZappalА et al. 1998). These generally cannot reproduce the sharpness and amplitude of the observed bombardment episodes.

20 15 10 5 0 5 10 15 20 20

25

30

35 period (Myr)

40

45

50

4: Bootstrap analysis applied to large (D> 40 km) terrestrial impact craters. Within the range of uncertainty the most probable period and phase (P, ) (35,0) Myr are as expected from the dynamical model, but weaker harmonic solutions are also present.

so- called Damocloids, have been found so far. T hirdly, the comets may develop super- dark mantles , wit h albedos p < 0.01 (Napier et al . 20 0 4). T h is is possible i f t he comet nucleus becomes covered with organic grains ~10 5 cm comprising a bird's nest structure with porosit y ~ 0.7 or more, consistent with that obser ved in Brownlee particles of probable cometary origin. I f sublimation of ices leaves such a structure, t hen van ish i ng ly small albedos become possible. T he nucleus of Comet 19P/ B orrelly has developed patches of albedo ~ 0.0 08, blacker t h a n a ny t h i n g o n E a r t h o u t s i d e o f n a n o engineered su rfaces, and if the entire nucleus became this dark we would probably not know that the comet existed. However, nearly all H T comets would have to become this dark for the problem to be solved. We do not yet k now whether this happens. T he exact nat u re of the hazard due to this c o m e t a r y m a t e r i a l c ap t u r e d t o H T o r b i t s f r o m t h e O o r t c lo u d d e p e n d s o n t h e s i z e of the individual bodies or fragments where the mass is predominantly hidden. For a randomly distributed population of N high- eccentricit y bodies with orbital periods P years, the mean inter val bet ween collisions with the E ar th is t 330 P/N Myr. For a population of N = 30 0 0 dark bodies in H T orbits with P = 60 yr, the c u r rent i mpac t i nter va l is t hen t 7 My r. T he long-term inter val bet ween such speciesdestroying impacts may be 30 60 Myr, but as we are currently immersed in an impact episode (figure 2), this temporarily high rate appears to
A&G · February 2009 · Vol. 50

1.22

phase (Myr)

NAPIER, ASHER: TUNGUSK
2 1.8 1.6 1.4 1.2

A

power

1 0.8 0.6 0.4 0.2 0 20

D > 40 km

D < 30 km

25

30

35 period (Myr)

40

45

50

5: Power spectrum analysis of well-dated impact craters. There is a clear galactic signal among the 12 impact craters >40km in diameter, but little or none for smaller craters. The interpretation is that impacting bodies 1.52km across are comets active or dormant but that smaller craters are predominantly formed by impacting asteroids.

be compatible with the cratering record. Hierarchic disintegration is a common mode of comet decay, and fragments that would make Tunguska-sized projectiles are a common product of these break-ups (such as in the K reutz sungrazing family, just one example of a split comet). Such bodies, if dormant, would largely avoid telescopic detec tion. Could the fading problem be resolved by assuming that comets disintegrate into unseen Tunguska-sized objects with physical lifetimes in excess of their dynamical ones? W hile it is plausible that such bodies are produced, wholesale conversion of cometary mass to such bodies would yield Tunguska-like impacts at ~10 yr intervals. The era of wide-area automated surveys has been under way for only about 10 years. If we are to extrapolate from such a short time base to impact probabilities at the 10 3 10 6 per annum level, then statistical completeness becomes an issue: that is, we have to ask whether all significant t ypes of hazard have manifested themselves over this period. The sporadic nature of comet disintegration, for example, is a potential source of failure of the "statistical completeness" assu mpt ion. Such d isi nteg rat ions a re not uncommon and may yield scores and in ex treme cases perhaps thousands of sub-km frag ments. T hese frag ments may or may not be short-lived, but it is clearly necessary to see w h e t h e r s u c h e ve n t s c o m p r i s e a s i g n i fi c a n t , and perhaps even dominant, impact hazard on timescales of interest to civilization. Otherwise we might be monitoring a swarm of bees while
A&G · February 2009 · Vol. 50

standing on a railway line with an express train d u e ! T h e s u r ve y s t h e m s e l ve s c a n n o t a n s w e r questions about their own statistical completeness: we have to take a broader perspective. It seems fai r to say t hat , u nt i l we u nderstand the fading problem, it would be unsafe to assume that the population of dark objects constitutes a negligible impact hazard at any energy level.

Signatures in cratering record
A quite different approach to the dark comet problem, which avoids the uncertainties involved in the fading f unction, is to look for galactic signatures in the impact cratering record. Since the main asteroid belt is impervious to galactic per tu rbations, any such signatu res would be diagnostic of cometary impacts. T he Oor t Cloud comprises ~5 1011 comets (dow n to 1 k m diameter) with aphelia in the ra nge 3 0 0 0 10 0 000 AU. T he long-p er io d comets i n t he cloud a re on ly just bou nd to the solar system and so are sensitive to galactic dist u rbances. T hese arise primarily from encounters with nebulae, and from a variable, periodic galactic tide coming from the vertical motion of the solar system as it bobs up and down in its orbit around the galax y. About a third of the mass of the galax y is in the form of nebulae, with cold, dense molecular clouds; although encounters with these nebulae have a strong sporadic component, they do tend to concentrate near the galactic plane (Z 50 60 pc). T he in flu x of comets to the planetar y system

varies pro rata with the galactic tide T, which is in turn proportional to the ambient densit y (z) of disc material at the ver tical distance z f rom t he galac tic plane: t hus T 4 G (z). The flux of comets from the Oort cloud may be modelled by adopting a combination of variable tide and sporadic encounters with nebulae. Long-period comets feed into other comet reservoirs (Biryukov 20 07, Emel'yanenko et al. 20 07) and so the overall comet impact rate on E a r t h w i l l r e fl e c t t h e a m b i e nt g a l a c t i c f o r c e s acting on the Oort cloud. Here we adopt a mean in-plane densit y 0.15 M pc3 and a molecular cloud density which declines exponentially with scale height 60 pc (Wickramasinghe and Napier 20 08). This yields a predicted cometary flux of amplitude a few with periodicit y P ~ 36 Myr. Since we passed through the galactic plane only one or t wo million years ago, we should be in or just past the peak of an impact episode now. Extracting crater ages randomly from this flux, we fi nd that the inbuilt periodicity of the model may be recovered from these synthetic datasets using standard procedures of power spectrum analysis (figure 3). E x a m i n at i o n o f t h e 4 0 w e l l - d at e d i mp a c t structures ( < 10 Myr) of the past 250 million years reveals that the larger craters (say 40 km in diameter) in particular were formed in sharp, discrete, statistically significant episodes (figure 2 and Napier 20 06) interspersed by long, quiet intervals. These episodes are too frequent and too strong to have come from the breakup of main belt asteroids (table 3). I f we now apply the periodicit y-hunting procedure to those 40 craters, a periodicit y around 35 Myr emerges, close to that predicted from the model (figure 4). The phase is also close to zero, consistently w i t h o u r r e c e nt p a s s a g e t h r o u g h t h e g a l a c t i c plane and implying that we are currently in a higher than average period of risk. I f we now divide up the craters by size, we find that the periodicity is strongly concentrated in craters more than about 30 40 km in diameter, whereas smaller ones show lit tle sign of galactic modulation (figure 5). The break- even point of 30 40 km corresponds to impactors of bet ween 1. 5 and 2 km diameter. This is around the threshold for global catastrophe, in which one contemplates the destruction of a quarter of mankind by the impact. It seems that below this size, the main impactors are probably asteroids, whereas above it, comets dominate the record. Hence comets, active or dormant, seem to be a major global hazard.

Major airbursts of the 20th century
Recent estimates based on Spaceguard discoveries have suggested a deficiency in relation to downward ex trapolation of the larger objects. T his can be understood at a qualitative level, since the Yarkovsky effect (a thermal effect due to solar heating) will tend to hinder the effects
1.23

NAPIER, ASHER: TUNGUSK

A

6: Chance alignment? (Left): Representative orbits of meteor sub-streams in the Taurid Complex; orbital elements taken from Kronk (1988). (Right): Large NEAs (orbits from the IAU Minor Planet Center) selected only on the basis of orbital size, shape and inclination being similar to Taurid meteors. Orbits of Earth and Jupiter also shown. Do the NEA orbits have any tendency to cluster in longitude around the TC meteors?

of resonances which drive asteroids out of the m a i n b e l t . T h e s u b - k i l o m e t r e p o p u l at i o n i s almost completely unknown, and yet airbursts from such bodies may be the most dangerous of impact phenomena for civilization. We can, however, glean something from three significant airbursts known to have occurred in the 20th century (Steel 1995). One such took place around eight o'clock in the morning of 13 August 1930, in the neighbourhood of the R iver CuruГa in the Brazilian A mazon. The associated energy is uncertain but may have been in the range 0. 2 2 Mt. A nother took place on the evening of 11 December 1935, this time in British Guiana (now Guyana). Even less is known of the energy of this impact, but a local pilot reported seeing an elongated area of destroyed forest more than 20 miles across. A nd of course there is the Tunguska impact itself. All three occurred when the Earth passed through or close to a major meteor stream (table 4). T he occurrence of three airbursts that we know about in the 20th century, each of which had the potential to cause huge damage, does imply that our increasingly crowded planet faces a signi ficant level of risk from sub -kilometre bodies imbedded in meteor streams (table 5).

Table 4: Airbursts and meteor streams
airburst Tunguska River CuruГa River British Guiana date 30 June 1908 13 August 1930 11 December 1935 meteor stream Taurids Perseids Geminids peak 30 June 12 August 13 December

Coincidence? The three greatest known airbursts of the 20th century all occurred when the Earth was passing through major meteor streams.

A large Holocene short-period comet
The evidence for a large comet in a short-period, low-inclination orbit, and continuously disintegrating over the last ~104 105 yr, includes the fact that current replenishing sources are about t wo powers of ten inadequate to yield the mass of the zodiacal cloud (W hipple 1967, Hughes 1996). Since the lifetime of the zodiacal cloud against collisional grinding and radiative effects is ~104 yr, an injection of ~1019 g of dust has been
1.24

required, more if the injection took place more remotely in time. A comet of mass, say, 5 1019 g with the densit y of water has diameter ~50 km. Chiron-sized ( 100 km across) cometary bodies, having mass a hund red times that of the entire cu rrent N E A system, may be injec ted into shor t-period, Ear th- crossing orbits with recurrence time of around 10 0 000 yr (Bailey et al. 1994). Thus the capture of such large comets to the inner planetary system does happen. But how high a lower bound can we put on the diameter of this Holocene comet, based on what is now observed to remain of it? T he Tau rid Complex (TC) a low-inclination, broad meteor st ream spread over 120 ° of sky fits the pict u re of this large comet's remnant debris stream. Not only the Northern and Southern Taurids, but several other meteor showers such as the Nor thern and Southern Orionids (Babadzhanov and Obrubov 1992), are genetically related components of this complex which derives from a single parent object. The existence of northern and southern shower branches, which need 10 4 yr to develop separately, con fi rms the TC 's age.

T he present- day active comet know n in the TC is 2P/ Encke. T his comet or its progenitor (there are no records of Encke itself before 1786) has been feeding meteoroidal material into the TC for at least a few 104 yr. Over this timescale this material both undergoes collisions (Steel and Elford 1986) as well as being dispersed by gravitational and radiative effects, eventually reaching the zodiacal cloud via a broad, sporadic stream (Stohl 1986) surrounding the Taurids. The sporadic stream produces the well known helion (H E) and antihelion (A H) sources seen in radar meteor data (Taylor and Elford 1998, Campbell-Brown 20 08). Wieger t et al . (20 08) found Encke to be by far the dominant contributor to the H E and A H flux. The diameter of Encke is ~5 km (FernАndez et al. 2000), while the total mass in the TC in 1 m and smaller meteoroids alone can be shown from meteor observations combined with stream modelling to be somewhat more than this. These reasons imply a minimum 10 km diameter progenitor, while to satisf y the mass balance of the zodiacal cloud, and noting the absence of other obvious candidates in the past 105 yr, at least,
A&G · February 2009 · Vol. 50

NAPIER, ASHER: TUNGUSK
3 2.5 700 600 500 400 300 200 0.5 0 100 0 40

A

2 count

1.5

1

count

0

50

100

150 200 250 longitude of perihelion (deg)

300

350

400

60

80 100 longitude of perihelion (deg)

120

140

7 (left): Among brighter NEAs (H < 16.5) we extract the 15 objects with semi-major axis a from 1.852.7 AU, eccentricity e from 0.651.0 and inclination i from 014°. These are distributed as shown, with a mean absolute difference of 48° from Comet Encke's longitude of perihelion which is 161°. (Right): If is uniformly random around 360°, the mean difference from Encke should be 90°. This Monte Carlo simulation (20000 trials) indicates that a difference as small as 48° arises by chance only once in 2000 trials.

say, 30 km is required. A similarly substantial original si ze may also follow depending on the amount of mass in larger bodies, through Tunguskas up to possible N E As comparable to Encke. A re these macroscopic bodies present?

Taurid NEAs
W hen the set of discovered N E A s star ted to grow as a result of modern surveys, at Palomar and later elsewhere, it became possible to verif y the expectation that inert bodies of significant size should exist in orbits similar to the comet Encke meteor stream (Clube and Napier 1984). S u b s e q u e n t p ap e r s , a s t h e N E A c a t a l o g u e e x p a n d e d , i d e n t i fi e d f u r t h e r ob j e c t s a l i g n e d with Encke and the TC . Following the "reduced D - criterion" method (a for ma l way to de fi ne orbit a l si m i la r it y) described by A sher et al . (1994), but using a new, up to date N E A dataset, we illustrate the TC alignment in figure 6. Valsecchi (1999) demonstrated an interesting observational selection effect that could favour the discovery of higheccentricit y, low-inclination N E As at cer tain longitudes. I n figure 6 we therefore restrict the N E As to absolute magnitude H <16. 5, corres p o n d i n g t o a m i n i mu m d i a m e t e r o f 1 3 k m depend i ng on albedo, so t hat obser vational incompleteness is less of a problem: H < 16. 5 objects are now being discovered at an annual rate of less than 10, compared to well over 20 a few years ago soon af ter the L I N E A R asteroid sur vey began. I n figure 7 we adopt a slightly simpler defi nition of orbital similarit y than the D - criterion (the alignment, if real, should not be sensitive to the exact technique used) and illustrate the statistical signi ficance. Figure 7 allows us to reject the null hypothesis that lonA&G · February 2009 · Vol. 50

gitude of perihelion is randomly distributed around 360 ° in favour of the hypothesis that bright N EOs whose (a , e, i) are close to that of Comet Encke also tend to have close to that of Encke. So there does seem to be an N EA stream, and it does seem to lie rather close to the TC . The perihelion distances of all these objects are a lit tle greater than those of obser ved Taurid meteors and none of those for which the taxonomic t ype has been determined are among the small fraction of N E As whose types correspond to ex tinct comet nuclei (further discussion on p464 of Jenniskens 20 06), but in fact the internal constitution of the large TC progenitor is quite unknown. A mong fainter N E As obser vational incompleteness increases; when one reaches Tunguska size the vast majorit y are still undiscovered. It turns out moreover that a statistically signi ficant alignment is presently hard to fi nd within the current dataset. Nevertheless, Porubcan et al . (20 06) have demonstrated the association of identi fiable fi laments in the Taurid meteor orbit database w it h several speci fic , k now n N E A s in the hund reds of metre to 2 k m size range. Babadzhanov et al. (20 08) found further examples of meteor shower associations with asteroidal Taurid objects.

Bombardment epochs
The 15 sub-streams or filaments recognized in meteor orbit data by Porubc an et al . (20 06), who also identified times over the past several 103 yr when they may have originated, are direct observational evidence of fi ne structure within the TC , at least in the component of the complex that is Earth-intersecting and can produce meteors. Furthermore, structure in the TC as a

whole is an inevitable consequence of meteor stream dynamics. If a large comet was captured to cis-Jovian space a few 10 4 yr ago, and if at least some products of its continuous disintegration are still present as a coherent, Ear thcrossing meteor stream, what structure should this stream have? Spec tacu la r L eon id meteor d isplays a few years ago helped to reinforce our understanding of the fi ne struct u re in streams. Narrow trails exist within the overall stream; they are essentially the least dispersed components of the stream and have an ex tremely high spatial densit y of meteoroidal material. T h e r e i s a n i mp o r t a nt d i f f e r e n c e b e t we e n the L eonid and Tau rid streams. T he orbit of the L eonid parent, comet 55P/ Tempel-Tut tle, is quite close to Earth orbit intersection at the present epoch. Small d isplacements (due to planetary perturbations) of narrow trails, relative to the comet orbit, can bring the trails to precise Earth intersection and allow the planet to encounter dense concentrations of meteoroids. Encke in contrast misses the Earth's orbit by quite some distance, at the present epoch. Assuming the densest concentrations of Taurid material to lie close to Encke's orbital plane (and in the absence of evidence against, this is the most reasonable assumption), we simply do not encounter these trails at present. T h i s h a s c h a n g e d i n t h e p a s t , a nd w i l l d o so again. Jupiter's gravit y causes the orbits of Encke and of Tau rid par ticles to precess , or twist around in 3-D space. W hen the orbits have t u r n e d a rou nd e nou g h , t h e y c u t t h ro u g h t h e Earth's orbit. So orbital precession makes the Earth intersect dense trails of Taurid material ever y few millennia. Dynamical calculations
1.25

NAPIER, ASHER: TUNGUSK

A

set the spacing of these intersection epochs at three millennia or so, with the nex t such epoch due around A D 30 0 0. Meteors we are seeing now come from outlying parts of a very broad complex. A fair number of Taurid meteors occur at present, even from low- densit y regions of the stream, because the whole complex really is massive. Hierarchic disintegration and fragmentation const it ute a n i mpor t a nt evolut iona r y route for comets ( Jenniskens 20 08). I f large comets disinteg rate hierarchically (cf. progenitors of the TC and of the K reutz sungrazing comets), then sub-kilometre objects may concentrate in comet trails, either from recent breakup or trapping in resonances. A dramatic example of a comet split ting in recent years was Comet 73P/ Schwassmann-Wachmann 3 (period ~5.4 yr). At its 1995 return at least three additional nuclei were identified, and t wo revolutions later at the 20 06 apparition the disintegration had yielded around 60 pieces. T he Taurid bombardments every few millennia are likely to involve multiple Tunguska impacts.

Table 5: Guilty by association?
meteor stream Taurids Perseids Geminids associated body Comet Encke Comet Swift-Tuttle Phaethon period (years) 3.3 120 1.6 encounter speed (km/s) 30 60 35

Parents, or possibly siblings, of the 20th-century impactors. All three are in cometary orbits with high encounter speeds and short warning times.

A possible Holocene cosmic impact
A carbon-rich black layer, ~12 90 0 years old, has been identified at many sites across North A merica (Haynes 20 08). It is closely coincident in age with the abrupt cooling know n as the Younger Dryas, as well as with large-scale mammoth extinctions and "rapid human behavioural shif ts", the lat ter taking place over decades or less. Evidence for an ex traterrestrial cause has been given by Firestone et al. (20 07) in the form of a contemporaneous thin layer at numerous North A merican sites, containing sharp peaks of iridium-bearing magnetic grains, magnetic m i c ro sphe r u le s , n a no d i a mo nd s , f u l le re n e s containing ex traterrestrial helium, and other indicators. They consider these to be evidence for a shower of cometary airbursts (Tunguskal i ke a nd la rger) produci ng t he w ide spread ex tinctions, the abrupt climate downturn and ex tensive biomass burning, along with abrupt cultu ral changes and a decline in the hu man population. T he evidence for an ex traterrest rial cause has more recently expanded into Greenland and Europe (A llen, personal communication), implying a disturbance on a global rather than continental scale. We are currently running simulations to determine whether the Taurid Complex can be convincingly proposed as the cause of this event.

deflection and mitigation strategies have not yet been developed for this class of hazard. Rare, large comets are occasionally thrown into the inner planetary system. I n terms of mass, they dominate the interplanetary environment in the course of their disintegration. T here seems to be a smoking gun, in the form of the Tau rid Complex and a zodiacal cloud which is substantially overmassive in relation to known sources. Disintegrating dormant comets could provide a major fraction of the dust, but the fact that over half the mass of the sporadic meteors are in a broad stream encompassing the Taurid Complex implies that a single large comet was the major contributor. On the evidence of the three major airbursts known to have taken place in the 20th centur y, it seems likely that the most dangerous regional hazards sub-kilometre impactors tend to concentrate in meteor showers associated with this erst while comet and with other active comets. Clai ms have been made by a small ad hoc g roup of geoscientists , the Holocene I mpac t Working Group (http://tsun.sscc.ru / hiwg / hiwg. htm), that impacts have been much more frequent throughout the Holocene than expected from Spaceguard surveys. It remains to be seen whether these claims will continue to hold up; but on the basis of the astronomical evidence described here, they cannot yet be excluded.
Bill Napie r is Honorar y Professor at the C e ntre for A strobiolog y, C ardif f Unive rsit y. David A she r is Research Fellow at Ar m agh Obse r vator y.
References Asher D J and Steel D I 1998 Planet. Space Sci. 46 205211. Asher D J et al. 1994 Vistas Astron. 38 127. Asher D J et al. 2005 Observatory 125 319. Asphaug E and Benz W 1996 Icarus 121 225248. Babadzhanov P B and Obrubov Yu V 1992 Celest. Mech. Dyn. Astron. 54 111127. Babadzhanov P B et al. 2008 MNRAS 386 14361442. Bailey M E and Emel'yanenko V V 1998 in eds M M Grady et al. Geol. Soc. Lon. Special Pub. 140 1117. Bailey M E et al. 1994 in Gehrels T ed. Hazards Due to Comets and Asteroids (University of Arizona, Tucson) 479. Ben-Menahem A 1975 Phys. Earth Planet. Inter. 11 61. Biryukov E E 2007 Solar System Res. 41 211. Campbell-Brown M D 2008 Icarus 196 144163. Clube S V M and Napier W M 1984 MNRAS 211 953968. Collins G S et al. 2008 Terra Nova 20 165.

Conclusions
On the evidence of the galactic signatures in the impact cratering record of the past 250 Myr, comets down to ~2 km diameter seem to be the major contributors to the global impact hazard. Ac tive comets may be too rare to f ul fi l that role, and so it seems likely that dormant bodies are the major contributors. Detection,
1.26

Emel'yanenko V V and Bailey M E 1998 MNRAS 298 212. Emel'yanenko V V et al. 2007 MNRAS 381 779. FernАndez Y R et al. 2000 Icarus 147 145160. Firestone R B et al. 2007 Proc. Nat. Acad. Sci. 104 16016. Gasperini L et al. 2007 Terra Nova 19 245251. Haynes C V 2008 Proc. Nat. Acad. Sci. 105 65206525. Hughes D W 1976 Nature 259 626627. Hughes D W 1996 QJRAS 37 593604. Hughes D W 2003 MNRAS 338 9991003. Jenniskens P 2006 Meteor Showers and their Parent Comets (CUP). Jenniskens P 2008 Earth Moon Planets 102 505520. Kolesnikov E M and Rasmussen K L 2008 100 years since Tunguska phenomenon: past, present and future June 2628, Moscow. KresАk L 1978 Bull. Astron. Inst. Czechosl. 29 129134. Kronk G W 1988 Meteor Showers: A Descriptive Catalog (Enslow). Levison H F et al. 2002 Science 296 22122215. Morbidelli A 1999 Celest. Mech. Dyn. Astron. 73 3950. Morbidelli A et al. 2002 Icarus 158 329. Napier W M 2006 MNRAS 366 977. Napier W M et al. 2004 MNRAS 355 191195. Nurmi P et al. 2001 MNRAS 327 13671376. Porubcan V et al. 2006 Contrib. Astron. Obs. SkalnatИ Pleso 36 103117. Raup D M and Sepkoski J J 1984 Proc. Nat. Acad. Sci. 81 801805. Rickman H 2005 in Kneeviґ Z and Milani A eds IAU zc Colloq. 197 277288. Rickman H et al. 2001 in Marov M Ya and Rickman H eds Astrophys. Space Sci. Libr. (Kluwer, Dordrecht) 261 131142. Sekanina Z 1983 Astron. J. 88 13821413. Sekanina Z and Yeomans D K 1984 Astron. J. 89 154161. Steel D 1995 WGN 23 207209. Steel D I and Elford W G 1986 MNRAS 218 185199. Stohl J 1986 in Lagerkvist C-I et al. eds Asteroids, Comets, Meteors II (Uppsala University, Uppsala, Sweden) 565574. Stokes G H et al. 2003 Report of the Near-Earth Object Science Definition Team (NASA Off. Space Sci. Solar System Exploration Division, Maryland, USA) http:// neo.jpl.nasa.gov/neo/neoreport030825.pdf sb04. Stuart J S and Binzel R P 2004 Icarus 170 295. Taylor A D and Elford W G 1998 Earth Planets Space 50 569575. Valsecchi G B 1999 in Svoren J et al. eds IAU Colloq. 173 353364. Whipple F J W 1930 Q. J. R. Meteorol. Soc. 56 287. Whipple F L 1967 in Weinberg J L ed. The Zodiacal Light and the Interplanetary Medium (NASA, Washington DC) 150 409426. Whipple F L 1975 Astron. J. 80 525. Wickramasinghe J T and Napier W M 2008 MNRAS 387 153157. Wiegert P A et al. 2008 Asteroids, Comets, Meteors July 1418, Baltimore, paper 8166. ZappalА V et al. 1998 Icarus 134 176179.
A&G · February 2009 · Vol. 50