The following discussion focuses on two related issues: the risk from impact by NEAs (Near Earth Asteroids) less than 1 km in diameter, especially the risk from impact tsunami, and the atmospheric breakup of NEAs with diameter less than 200 m.
Ever since the original Spaceguard Report in 1992 and the risk assessment published by Chapman & Morrison in Nature in 1994, it has been clear that the greater part of the impact hazard is due to NEAs that can have an adverse effect on the global environment -- roughly those 2 km or larger in diameter. To be safe, the Spaceguard Goal has been set in terms of finding NEAs 1 km or larger. No recent results have substantially changed this conclusion. All of the discussion of sub-km impacts reported here deal with the less than 5% of the risk that is caused by sub-km NEAs. Even when the Spaceguard Goal is reached, the greatest remaining risk will be from the undiscovered 10% of NEAs larger than 1 km, not from the sum of all sub-km impacts on the land or ocean.
The risk from land impacts is fairly well understood, in part drawing upon our knowledge of the effects of nuclear explosions. As hazards go, it is quite small -- much less than the risk of other natural hazards such as earthquakes or severe storms. More uncertainty has been associated with the risk from impact tsunami, since ocean waves are able to propagate energy to large distances from the impact point. Several studies, notably a paper in 2000 by Ward & Asphaug, have attempted to model the formation and propagation of impact tsunami. This work has recently been updated and expanded by Chesley & Ward, who combined tsunami models with information on the distribution of population near coasts to quantify the risk. The first essay below is a summary by Erik Asphaug, Donald Korycansky & Steven Ward of a tsunami workshop held in May at UC Santa Cruz to further examine these issues.
There are still substantial uncertainties about the run-up and run-in of impact tsunami as they reach the shore. However, the main results are clear. Impact tsunami are less dangerous than has been assumed by some in the past, mostly because they have shorter wavelengths than seismic tsunami and therefore can run-in only a kilometer or so at maximum. Further, with even modest warning the loss of human life can be minimized, since it is necessary only to move a few km inland or to an elevation of a few meters to escape the inundation -- thus impact tsunami are primarily a threat to coastal infrastructure, not populations. The risk is rather evenly spread over impactors of different size from 200 m up to 2 km. The lower size limit is imposed by the ability of small NEAs to penetrate the atmosphere, since airbursts do not initiate impact tsunami.
The second set of reports below relate to the atmospheric penetration of small NEAs. All of the recent models allow for energy loss by stony impactors below 200 m diameter, with complete disintegration (an airburst) below a diameter of about 100 m. These assumptions are based on models calculated about a decade ago by Hills, Chyba, Zahnle, and their collaborators. The new result from P.A. Bland & N.A. Artemieva was published in Nature for17 July. (Presented below are the abstract of their paper, a press release, and a press report). This model treats the atmospheric breakup in more detail and concludes that most stony NEAs with diameter less than 200 m will result in airbursts. If they are correct, then ocean impactors smaller than 200 m will not produce tsunami. Partially compensating for the reduced tsunami risk, their models indicate a larger risk from land impacts, since for a given energy an airburst has a larger radius of destruction than a ground impact. The result is little change in the total risk of sub-km impacts, but a slight shift in the risk from tsunami to land impacts.
I find it curious that the publication of various models of NEA populations and (in this case) of atmospheric penetration are often interpreted in the press as changes in the impact hazard. However, few have reported the very real decrease in the hazard that results from the steady progress of the Spaceguard Survey in discovering NEAs larger than 1 km, which is where most of the hazard lies and where we are making the greatest progress in reducing this hazard.
Ocean Waves from Asteroid Impacts
Workshop held May 17-18 2003, University of California, Santa Cruz
An asteroid the size of the Coliseum crashes into the ocean somewhere on Earth, producing a transient cavity miles across. The spectacular collapse launches tsunamis in all directions, casting the asteroid's impact energy far and wide. Everyone living on these shores may be in for a bad day.
Based partly on the assumption that minor asteroid impacts spawn dangerous tsunamis, the National Research Council has recommended that NASA/NSF design and build a "Large Synoptic Survey Telescope" to image the entire sky every 7 days and detect thousands of asteroids to 24th magnitude, especially those down to 300 m diameter on near-Earth orbits. The LSST will have many astronomical uses apart from asteroid hunting; nevertheless, before $100M to $200M gets spent, one must examine the justification. Are sub-km asteroids really a significant threat? If not, then existing survey telescopes are adequate to catalog ~98% of the 1100±100 near-Earth asteroids larger than 1 km diameter in the next few decades (Jedicke et al. 2002). Of these, half have been detected so far, and the rest generally have orbits harder to discover from Earth. Perhaps the LSST money should instead be dedicated to a space-borne observatory for bagging these "holdouts" of a size generally believed (Toon et al. 1997) capable of causing global catastrophe.
In 1908, the remote Tunguska river valley in Siberia was flattened by the airburst of a rocky body ~50 m across. Such impacts happen once per ~1000 years (Harris 2003) but their hazard is negligible, compared with other natural disasters, because it would take a direct hit over a population center to kill many thousands of people, barring a nuclear-armed nation mistaking an airburst for a surprise attack. Because the disaster footprint for sub-km land impacts grows like impactor diameter squared, and the flux of impactors falls like their diameter to the -7/3 power, the time-integrated disaster footprint goes down with impactor size. So we shouldn't worry much about Tunguska airbursts nor any land-impacting asteroids smaller than about 1 km.
Tsunamis enter the picture as the means by which smaller asteroids can have a disproportionately large disaster footprint. The 1960 Chile earthquake killed thousands of locals, but its tsunamis were hugely devastating 10,000 km away in Japan. An asteroid plunking down into an ocean might similarly radiate wave energies throughout the basin, resulting in widespread loss of life and infrastructure - hence the odd linkage of tsunamis with space telescopes.
Our workshop concerned understanding the key distinctions between familiar earthquake-generated tsunamis and never-witnessed impact tsunamis. Because wavelengths shorter than the ocean depth are poorly excited from below, earthquake tsunamis are tens to hundreds of kilometers long. Long wavelength amplitudes decrease with distance as slowly as R-1/2, the nondispersive limit. Impact tsunamis on the other hand produce a busy spectrum of wave energy peaking near the cavity diameter, and include considerable energy at smaller wavelengths. Even in the absence of viscosity or other dissipation, impact tsunamis pull apart rapidly into their components (long waves traveling faster), leading to amplitude attenuation ~R-1 according to linear theory (Ward and Asphaug 2000).
But how linear is the collapse of a ~10 km wide ~3 km deep cavity in 5 km of water? The collapse will overshoot, reaching supersonic speeds, sending a plume of water beyond the tropopause. From here it will collapse again, pumping out waves from a turbulent oscillation. Linear theory captures this rush-in and fall-back of water, but it assumes that waves do not break and ignores vorticity, viscosity, and water-atmosphere interaction. This initial phase of impact tsunami generation is clearly a job for advanced computational fluid dynamics.
The latest hydrocode models differ significantly from linear theory (Gisler et al. 2003). Waves topple and crash and cavitate; kilometer-high waves crest, break and foam; supersonic waves rip through the atmosphere and get blown out. But hydrocodes have the opposite problem of linear theories, in that numerical dissipation is challenging to control or characterize, especially for long model time. Moreover, hydrocodes cannot model wavelengths and amplitudes smaller than the resolution. New techniques, notably adaptive grid methods, allow resolution to increase where needed, but have not been validated on this class of problem.
Wave attenuation in adaptive grid hydrocodes is closer to R-2, even far from the impact. If asteroid impact tsunamis really decay like R-2, then they pose little hazard - certainly nothing comparable to the risk from storm swells. Curiously, the most recent models, presented by Gisler at this workshop, show no sign of transitioning towards linear behavior, even 100 km from the impact where modeled wave heights are <10 m. Moreover, attenuation seems greatest for the smallest cavities. Hydrocode models by Korycansky et al. (2003) show better agreement with linear theory, for the initial phase of wave generation, so the issue is wide open.
Data to resolve this discrepancy may exist in the literature of water explosions (see Le Méhauté and Wang 1996). In 1965, 10,000 kg of TNT was exploded in Mono Lake, California, and the water wave train was described by Van Dorn et al. (1968) in a little-circulated but unclassified report. (To our knowledge, ecological effects were never documented.) In this experiment, the envelope of wave train amplitude agreed very well with linear prediction. Still, these and other explosion waves may be of a different character than the ~1000 Mt class events contemplated here.
As for shoaling, long-period tsunamis don't lose much energy on the continental shelf, whereas Van Dorn et al. (1968) state that shorter period waves from nuclear explosion cavities a few kilometers across will break on the slope, posing a danger to deployed ships and submarines but little hazard to coastal infrastructure and communities behind a shelf. Van Dorn et al. (1968) present no data to support this effect, but related classified reports may have contributed to the decision to deploy the MX missile system on land instead of in small submarines off the coast.
Breaking of ~10 m high waves in ~100 m of water runs counter to traditional criteria. Moreover, it does not address what happens to the relatively unabated momentum after waves break, as spilling (not cresting) breakers are indicated. Also, the wave train from an asteroid impact is long, with many frequencies propagating at different speeds. Following an initial breaker, hundreds of subsequent waves could "set up", potentially leading to even larger run-in at shore.
Perhaps the most relevant observed events, in terms of associated wave periods and amplitudes, are volcanic landslide collapses such as the tsunami generated at Ritter Island in the Bismarck Sea, in 1888. German settlers recorded wave periods and amplitudes around the basin of ~4 minutes duration and ~15 m run-up. The event, which probably killed hundreds of villagers, is well-matched by linear theory (Ward and Day, 2003). Such events, including predicted future collapses on the Hawaiian and Canary Islands, occur with a frequency that may exceed, by a factor of five or ten, the frequency of asteroid impacts of equivalent magnitude. Existing sky surveys have already reduced the earlier impactor rates adopted by Ward and Asphaug (2000) by a factor of 5: thus, even under the assumptions of linear theory, the asteroid impact tsunami hazard is probably small in human terms. A new sky survey to 24th magnitude seems moot, apart from its great science value, unless we happen to be in the gunsights of an undiscovered rock.
A science can be broad and rich without representing a hazard, and a final workshop topic was how waves from asteroid impacts have left their imprint in the marine geophysical record. Documented pelagic impacts range from mega-events preserved in Archean sediments (Hassler et al. 2000) to the Eltanin event which scraped the bottom of a 5 km ocean 2.5 million years ago (Gersonde et al. 2003).
Our workshop identified active topics in this relatively new science (but skirted controversy regarding the plural of tsunami):
- Characterizing the wave train produced by the collapse of deep marine cavities, including those which bottom out.
- Establishing valid regimes for linear tsunami theory.
- Predicting the marine geophysical signature of cavity collapse.
- Understanding interaction with the shelf: is the Van Dorn effect real?
- Knowing how, and how far, the complex wave train from an impact event comes ashore.
Some components are addressed in various reports and volumes; others are nearly unexplored. Until these components are assimilated, we shall have poor understanding of unique geophysical events that have occurred thousands of times in the Tertiary alone. Answers may lie in a more comprehensive assessment of the sedimentological record, which means more astronomers looking at x-rays of drill cores, and more geologists thinking about rocks from space.
This workshop was sponsored by UCSC's Institute for Geophysics and Planetary Physics.
Chesley, S. R. and Ward, S. N., preprint, submitted. A quantitative assessment of the human and economic hazard from impact-generated tsunami. Available at
Gersonde, R., Kyte, F. T., Frederichs, T., Bleil, U. and Kuhn, G. (2003). New data on the late Pliocene Eltanin impact into the deep Southern Ocean. Large Meteorite Impacts (Nördlingen, Germany), Abstract 4094, http://www.lpi.usra.edu/meetings/largeimpacts2003
Gisler, G., Weaver, R., Mader, C., and Gittings, M. L. (2003). Two- and three-dimensional simulations of asteroid ocean impacts. Science of Tsunami Hazards 21, 119-134.
Harris, A. W. (2003). The impact frequency of near-Earth asteroids. American Astronomical Society DDA Meeting 34, 02.04
Hassler, S. W., Robey, H. F., and B. M. Simonson (2000). Bedforms produced by
impact-generated tsunami, similar to 2.6 Ga Hamersley basin, Western Australia.
Sedimentary Geol. 135, 283-294.
Jedicke, R., Morbidelli, A., Spahr, T., Petit, J.-M., and Bottke, W. F. Jr. (2002). Earth and Space-based NEO survey simulations: Prospects for achieving the Spaceguard goal. Icarus 161, 17-33.
Le Méhauté, B. and Wang, S. (1996). Water Waves Generated by Underwater Explosion.
World Scientific, Singapore.
Toon, O. B., Zahnle, K., Morrison, D., Turco, R. P. and Covey, C. (1997). Environmental perturbations caused by the impacts of asteroids and comets. Reviews of Geophysics 35, 41-78.
Van Dorn, W. G., Le Méhauté, B. and Hwang, L.-S. (1968). Handbook of Explosion-Generated Water Waves, Volume I - State of the Art (Tetra Tech, Pasadena CA)
Ward, S. N. and Asphaug, E. (2000). Asteroid impact tsunami: a probabilistic hazard assessment. Icarus 145, 64-78.
Ward, S. N. and Day, S. J. (2003). Ritter Island Volcano - Lateral collapse
and tsunami of 1888. Geophysical Journal International, in press.
Efficient Disruption of Small Asteroids by Earth's Atmosphere
P.A. Bland (Imperial College London) & N.A. Artemieva (Russian Institute for Dynamics of Geospheres)
Nature 424, 288 - 291 (17 July 2003)
Abstract: Accurate modelling of the interaction between the atmosphere and an
incoming bolide is a complex task, but crucial to determining the fraction of small asteroids that actually hit the Earth's surface. Most semi-analytical approaches have simplified the problem by considering the impactor as a strengthless liquid-like object ('pancake' models), but recently a more realistic model has been developed that calculates motion, aerodynamic loading and ablation for each separate particle or fragment in a disrupted impactor. Here we report the results of a large number of simulations in which we use both models to develop a statistical picture of atmosphere-bolide interaction for iron and stony objects with initial diameters up to 1 km. We show that the separated-fragments model predicts the total atmospheric disruption of much larger stony bodies than previously thought. In addition, our data set of >1,000 simulated impacts, combined with the known pre-atmospheric flux of asteroids with diameters less than 1 km, elucidates the flux of small bolides at the Earth's surface. We estimate that bodies >220 m in diameter will impact every 170,000 years.
Fewer Earthbound Asteroids Will Hit Home
Press Release from Imperial College
July 16, 2003
Scientists report in Nature today that significantly fewer asteroids could hit the Earth's surface than previously reckoned. Researchers from Imperial College London and the Russian Academy of Sciences have built a computer simulation that predicts whether asteroids with a diameter up to one kilometre (km) will explode in the atmosphere or hit the surface.
The results indicate that asteroids with a diameter greater than 200 metres (the length of two football pitches) will hit the surface approximately once every 160,000 years - way down on previous estimates of impacts every 2,500 years.
The findings also predict that many more asteroids blow up in the atmosphere than previous estimates, which means the hazard posed by impact-generated tidal waves or tsunamis is lower than previous predictions. The researchers suggest that proposals to extend monitoring of Near Earth Objects (NEO) to include much smaller objects should be reviewed.
Dr Phil Bland of Imperial's Department of Earth Science and Engineering and a Royal Society University Research Fellow, said: "There is overwhelming evidence that impacts from space have caused catastrophes for life on Earth in the past, and will do so again. "On the Moon it's easier to track the number, frequency and size of collisions because there is no atmosphere, so everything hits the surface. On Earth the atmosphere acts like a screen and geological activity erodes many craters too.
"Massive impacts of the type thought to have wiped out the dinosaurs leave an indelible print on the Earth but we have not been able to accurately document the effect of smaller impacts. Now, we have a handle on the size of 'rock' we really need to worry about and how well the Earth's atmosphere protects us."
When small asteroids hit the atmosphere the two forces collide like two objects smashing together, which often breaks the asteroid into fragments. Until now, scientists have relied on the 'pancake' model of asteroid impact to calculate whether the asteroid will explode in the atmosphere. This treats the cascade of fragments as a single continuous liquid that spreads out over a larger area - to form a 'pancake'. But a new model known as the 'separate fragment' (SF) model, which was developed by co-author of the study, Dr Natalya Artemieva of the Russian Academy of Science, has challenged this approach.
"While the pancake model can accurately predict the height from the Earth's surface at which the asteroid will break up, it doesn't give an accurate picture of how the asteroid will impact," explains Dr Bland. "The SF model tracks the individual forces acting on each fragment as it descends through the atmosphere."
To create a more accurate model of how asteroids interact with the atmosphere the researchers ran more than 1,000 simulations using both models. Objects made of either iron or stone, known as 'impactors', were used to reflect the composition of asteroids and experiments were run with varying diameters up to 1 km.
The researchers found the number of impacts for iron impactors were comparable using both models. For stone the pancake model significantly overestimated the survivability rate across the range used. The SF simulations also allowed the researchers to define the different styles of fragmentation and impact rates for iron and stone, which correspond closely with crater records and meteorite data.
"Our data show that over most of the size range we investigated stony asteroids need to be 1,000 times bigger than the iron ones to make a similar sized crater. Much larger objects are disrupted in the atmosphere than previously thought. "But we are not out of the woods yet," added Dr Bland "asteroids that fragment in the atmosphere still pose a significant threat to human life."
Small Stony Asteroids Will Explode and Not Hit Earth, Study Shows
Robert Roy Britt
Space.com, 16 July 2003
When asteroids fall through Earth's atmosphere, a variety of things can happen. Large iron-heavy space rocks are almost sure to slam into the planet. Their stony cousins, however, can't take the pressure and are more likely to explode above the surface.
Either outcome can be dismal. But the consequences vary. So scientists who study the potential threat of asteroids would like to know more about which types and sizes of asteroids break apart and which hold together. A new computer model helps to quantify whether an asteroid composed mostly of stone will survive to create a crater or not.
A stony space rock must be about the size of two football fields, or 720 feet (220 meters) in diameter, to endure the thickening atmosphere and slam into the planet, according to the study, led by Philip Bland of the Department of Earth Science and Engineering at Imperial College London.
"Stones of that size are just at the border where they're going to reach the surface -- a bit lower density and strength and it'll be a low-level air burst, a bit higher and it'll hit as a load of fragments and you'll get a crater," said Bland, who is also a Royal Society Research Fellow. The distinction would mean little to a person on the ground.
Two Ways to Destroy a City
"An airburst would be a blast somewhere in the region of 500-600 megatons," Bland said in an e-mail interview. "As a comparison, the biggest-ever nuclear test was about 50 megatons."
A presumed airburst in 1908, over a remote region of Siberia called Tunguska, flattened some 800 square miles (2,000 square kilometers) of forest. The object is estimated to have been just 260 feet wide (80 meters). Bland said the event was probably equal to about 10 megatons.
"If most of it made it to the ground you might actually be a bit better off, because the damage would be a little more localized," he said. "A lot of energy would still get dumped in the atmosphere, but you'd probably also have a ragged crater, or crater field, extending over several kilometers, with the surrounding region flattened by the blast."
Smaller stony asteroids, say those the size of the car, enter the atmosphere more frequently but typically disintegrate higher up and cause no damage. In fact, as many as two or three dozen objects ranging from the size of a television to a studio apartment explode in the atmosphere every year, according to data from U.S. military satellites.
Separate research in recent years has shown that stony asteroids are often mere rubble piles, somewhat loose agglomerations of material that may have been shattered in previous collisions but remain gravitationally bound.
Pieces and Parts
The new computer model is detailed in the July 17 issue of the journal Nature. It was created with the help of Natalia Artemieva at the Russian Academy of Sciences. Previous models treated the cascade of fragments from a disintegrating asteroid as a continuous liquid "pancake." The new model tracks individual forces acting on each fragment as the bunch descends.
The researchers can plug in asteroid size, density, strength, speed and entry angle at the top of the atmosphere. With "reasonable confidence" a computer program then details how that rock should behave in the air and what will happen at the surface.
The model has implication not just for land-based impacts, but also splashdowns in the ocean that can trigger devastating tsunamis. An airburst is not likely to generate much of a tsunami, possibly lowering that risk compared to what scientists had figured. The results suggest rocks about 720 feet across (220 meters) are likely to actually hit the surface every 170,000 years or so. Some previous research has suggested a frequency of every 4,000 years or less.
The model can also "hindcast" what sort of rock might have generated a certain known crater. "You see a crater field on Mars, we can tell you what sort of object caused it," Bland said. In fact, he and Artemieva have done just that. In their most recent tests, which are not discussed in the Nature paper, they plugged in the atmospheric details of Mars, as well as Venus, and hurled some hypothetical space rocks at those planets. "The simulated crater fields that the model produces look almost exactly like the real thing," Bland said.
For now, the model does not handle very large asteroids, those that could cause widespread regional or even global damage, though Bland said the flaw may be fixable. He is careful to point out that computer models do not provide solid proof for what might happen. "There are still a lot of unknowns in this," he said.