Dear friends and students of NEOs:
This edition of NEO news describes a workshop held on March 16 to discuss the hazard due to tsunami from deep-water impacts by sub-kilometer asteroids. The workshop developed a consensus on the order of magnitude of this hazard, which is substantially less than was estimated years ago. However, the group did not support the position taken by Jay Melosh in a U Arizona Press release today that tsunami from such small impacts do not pose any hazard whatever.
David Morrison
SUMMARY OF TSUNAMI HAZARD WORKSHOP HOUSTON 03/16/03
by David Morrison
The tsunami hazard workshop was held to bring together impact tsunami experts in an informal setting to resolve apparent differences in their assessments of the magnitude of this hazard. In particular, it provided the first opportunity for Jay Melosh (U Arizona) and Steven Ward (UC Santa Cruz), the two main proponents of these different estimates, to meet. A dozen others attended also, with Bill Bottke (SW Research Institute) as convener.
Background: Since the time of the original NASA Spaceguard Survey Report (Morrison, 1992) is has been clear that there is a major qualitative and quantitative difference in the hazard posed by impacts that have global environmental consequences, as compared with smaller impacts that produce only local or regional effects (see also Chapman & Morrison, 1994: Impacts on the Earth by Asteroids and Comets, Assessing the Hazard, Nature 367: 33-39). The total risk (measured by equivalent annual fatalities) for all impacts below the threshold for global disaster is at least a factor of 100 below that associated with global disasters, assuming that the threshold for global disaster is between 1 and 2 km diameter (roughly 1 million megatons; Toon et al., 1997: Environmental Perturbations Caused by the Impacts of Asteroids and Comets, Rev Geophys 35, 41-78). Land impacts in particular are a minor hazard for impactors less than 1 km diameter. However, there has been a general concern that tsunami from deep-water marine impacts could contribute substantially to the hazard for people living near coasts — perhaps amounting to tens of percent of the global population. If the hazard from ocean impacts of asteroids between 200 and 1000 m is substantial, then these ocean impacts are the dominant hazard from “small” asteroids, those below the global threshold. The tsunami hazard is then the primary motive for extending the Spaceguard Survey to smaller asteroids. It was partly in response to concerns about ocean impacts in the sub-kilometer range that two studies by the US National Research Council recommended that a survey be carried out for NEAs smaller than 1 km.
Two science working groups are now studying the hazard associated with sub-kilometer impacts: a NASA Science Definition Team that is preparing recommendations to the NASA Office of Space Science (due in June) and a Science Working Group organized by National Optical Astronomy Observatory that is studying many of the same issues for the NSF. Both groups have agreed that the risk from land impacts of sub-kilometer NEAs is very low. Tunguska-class impacts occur on the land only once every 2-3 millennia, and land impacts that are expected to kill thousands are rarer still. The more important challenge is to evaluate the hazard from tsunami generated by sub-kilometer NEAs.
It has become apparent in recent years, and especially from the work of Steven Ward and Erik Asphaug (2000: Asteroid Impact Tsunami, A Probabilistic Hazard Assessment, Icarus 145, 64-78), that impact-induced tsunami are significantly different from earthquake tsunami. The impact tsunami have shorter wavelengths, comparable to the size of the impact cavity (a few km in diameter). This places them intermediate in other respects as well between the more familiar long-wavelength earthquake tsunami and ordinary storm waves. Their shorter wavelength influences both the likelihood that they will break before reaching shore and the degree of run-in, which is comparable to the wavelength as the wave approaches the shore (typically about 1 km). Ward and Asphaug have modeled all these effects using standard linear seismic theory adjusted for the particular properties of impact tsunami. Others, particularly Galen Gisler and Don Korycansky, have modeled the impact process and initial wave generation using more advanced hydrocodes. All of this work is in general agreement, at the factor-of-two level.
Alan Harris (Space Science Institute) and Steven Chesley (JPL) have used these results to calculate the hazard associated with tsunami from sub-kilometer impacts. Harris has recently recalibrated the NEA impact frequency, finding a reduction by a factor of 5 in the frequency at all sizes relative to the assumptions made by Ward and Asphaug. Thus the Ward and Asphaug models can be transformed readily to agree with current impact frequency estimates. Chesley has combined the Ward and Asphaug results, so modified, with a new global population distribution database. Using this known distribution of population as a function of height above sea level and distance from the shoreline, he can estimate the fraction of the population that is at risk from impact tsunami and can calculate the frequency with which coastal populations are subject to tsunami dangers. In round numbers about 1% of the world population is at risk, considerably less than had been estimated a decade ago. The drop follows primarily from the reduction in the run-in of impact tsunami as dictated by their shorter wavelength. Chesley’s initial results (presented at this workshop) do not take into account shielding of coastal populations by reefs, barrier islands, seawalls, and harbor constructions. Harris and Chesley suggest that at least half of all coastal population is protected by these barriers from short-wavelength tsunami, thus further reducing the at-risk population (and associated annual equivalent fatality rate) by another factor of two (or perhaps even more). The resulting hazard, while still greater than the land-impact risk from sub-kilometer NEAs, is less than has been previously estimated. These models are not yet complete, but Harris and Chesley seem to be converging on a reasonable answer.
The prime reason for this workshop was not to review the progress of this modeling, interesting though that was. It was to deal with a challenge from Jay Melosh suggesting that impact tsunami approaching the shore from deep ocean break near the continental shelf and dissipate their energy offshore, yielding no coastal inundation. Melosh is also presenting a paper at this Lunar and Planetary Science Conference (on 17 March), and the University of Arizona is issuing a press release to this effect, stating that the removal of the alleged hazard due to tsunami “will save taxpayers the cost of financing searches for small NEAs, at a savings of billions of dollars”. Melosh had first presented these results to the NASA team, but most of us attending this workshop had not yet had the opportunity to hear his analysis, and he and Ward had not previously met.
Melosh began by saying that he was “only the messenger”, and that his purpose was to call attention to the 1968 report by W.G. Van Dorn of the Scripps Institution of Oceanography, who had studied explosion-generated waves for the U.S. Navy (TTR Report TC-130, Handbook of Explosion-Generated Water Waves, Volume 1 – State of the Art). While never formally classified, this report has been generally unavailable, and as recently at 1996 Van Dorn himself had asserted that it did not exist. However, a handful of copies had been distributed to academic libraries long ago, and these were eventually located and distributed to the attendees at this workshop. Van Dorn carried out an extensive analysis of the entire subject of “small” tsunami based on both theory and experimental results from nuclear explosions (both on and under the ocean, and up to 10 megatons yield), and also on a series of smaller-scale chemical-explosion tests carried out in Mono Lake. Most important for our purposes is the so-called “Van Dorn Effect”, which asserts that small (short-wave) tsunami break when they cross the continental shelf, generating large-scale turbulence there but relieving the coast of any wave run-in. The Van Dorn Effect apparently has had important implications in nuclear strategy, for example in the basing of ballistic missile submarines.
Much of the workshop was devoted to discussing the Van Dorn Effect and comparing his report with the modern work of Ward, Gisler, and Korycansky. While the report was generally well received by the workshop attendees, parts of it eluded our comprehension. Especially cryptic is the genesis of the Van Dorn Effect, which is not derived or explained in any detail in this report. It is simply asserted in the text and summary. Thus the most critical part of this argument for our purposes is not justified.
Aside from this major mystery, there was a consensus (shared by Melosh) in support of the kind of analysis presented by Ward and Korycansky, and used by Chesley, recognizing the substantial uncertainties embedded in the computations. This consensus extended to each element: (1) estimating the size and shape of the original explosion cavity (as long as it did not extend to the ocean floor; e.g., for NEAs up to about 500 m); (2) modeling the expanding wave front (which can be approximately dealt with by linear theory and which indicates that the maximum wave amplitude varies as about 1/r; (3) shoaling, which results in reduced wavelength and modestly increasing wave amplitude as the wave approaches the shore; and (4) the most uncertain part, the run-in on the shore. It was agreed that a run-in equal to the near-shore wavelength (and associated run-up no higher than about 2 times the height of the deep ocean wave) provides a reasonable and probably conservative approach to estimating the population at risk.
Other effects not yet modeled (such as sheltering of coasts or the mysterious Van Dorn Effect itself, if real) will tend to reduce the estimated hazard. Thus the current Harris-Chesley-Ward approach probably yields an upper limit for the impact tsunami risk. Undoubtedly they will refine their results for the NASA team report to be written in the next two months. Meanwhile, it is reasonable to take their upper limits to estimate the tsunami risk.
In this upper-limit case, the tsunami hazard still dominates over that of land impacts by factors of several for the sub-kilometer NEAs, but both are quite small relative to the risk from NEAs larger than 1 or 2 km. Another way to look at the situation is that the population at risk from tsunami is small, only a few tens of millions, consisting of those who live very close to unprotected coasts (for example in Los Angeles and the coast of Bangladesh, as well as many low islands without barrier reefs.) For those people, the risk from tsunami is comparable to that from a global-scale catastrophe. That is, such a person (a resident of Venice Beach, California, for example) is roughly equally at risk (at a level of about one in a million per year) from impact tsunami as from a global ecological catastrophe. (Of course, this person is even more at risk from earthquakes or seismic tsunami).
The workshop concluded with general agreement, as well as some frustration concerning our inability to fathom the Van Dorn Effect. We will all look forward to the completion of the Chesley et al. hazard analysis, which will add some quantitative meat to this unoficial and rather skeletal summary from the workshop.
NEO News is an informal compilation of news and opinion dealing with Near Earth Objects (NEOs) and their impacts. These opinions are the responsibility of the individual authors and do not represent the positions of NASA, the International Astronomical Union, or any other organization. To subscribe (or unsubscribe) contact dmorrison@arc.nasa.gov. For additional information, please see the website: http://impact.arc.nasa.gov. If anyone wishes to copy or redistribute original material from these notes, fully or in part, please include this disclaimer.
