2
Risk Analysis

Impacts are one of the most fundamental processes shaping planetary surfaces throughout the solar system. Images of many solar system objects are dominated by craters formed throughout the past 4.5 billion years. Smaller airless bodies in particular retain a significant history of collisions. Earth’s Moon has been used to determine variation in the rate of impacts since the earliest days of the solar system. Imagery, coupled with the dating of lunar materials, has allowed scientists to demonstrate that the rate of impacts has gradually diminished since these early times.

Although the frequency of impacts due to bodies of all sizes is considerably less than during the first 700 million years of solar system history, as the planetary orbits have stabilized and a significant proportion of the smaller objects has been accreted, the most significant risk remains from collisions with bodies on oval-shaped orbits (such as comets) and objects with orbits that pass near Earth’s orbit.

The average amount of material accreted daily to Earth is estimated to be in the range of 50 to 150 tons of very small objects (Love and Brownlee, 1993). This material is mostly dust, although there are abundant small objects that burn up quickly in the atmosphere and are evidenced by meteor trails. More rarely, larger objects impact Earth. It is now widely believed that the impact of an approximately 10-kilometer-diameter object formed the Chicxulub Crater near the Yucatan Peninsula about 65 million years ago, very likely resulting in the extinction of the dinosaurs. Its mass is similar to that of the total amount of dust and other small objects accreted to Earth during the time since that impact.

Substantial atmospheres around planetary bodies act as significant filters to incoming objects. Smaller objects, particularly those that are lower in density and more fragile, vaporize in the upper reaches of the atmosphere, while more intact, larger bodies may survive to impact the surface. Thus, small craters are much less common on bodies with dense atmospheres, such as Earth, Venus, and Titan, than they are on Mercury and the Moon, with Mars somewhere in between. Of course there are still substantial numbers of large impact craters even on Venus, with its dense carbon dioxide atmosphere; the lack of weathering and erosion, coupled with low rates of volcanic and tectonic activity over the past 0.5 billion years, has allowed the retention there of a significant number of craters, most largely unaltered since emplacement. By contrast, the movement of water on Earth and the action of plate tectonics have both resulted in the loss of much of the cratering record on this planet. There are more than 170 established impact craters on Earth, including the approximately 1.2-kilometer Meteor Crater in Arizona (Figure 2.1). The largest known terrestrial crater is the 300-kilometer-diameter Vredefort Crater in South Africa, dated at around 2 billion years old.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 12
2 Risk Analysis Impacts are one of the most fundamental processes shaping planetary surfaces throughout the solar system. Images of many solar system objects are dominated by craters formed throughout the past 4.5 billion years. Smaller airless bodies in particular retain a significant history of collisions. Earth’s Moon has been used to deter- mine variation in the rate of impacts since the earliest days of the solar system. Imagery, coupled with the dating of lunar materials, has allowed scientists to demonstrate that the rate of impacts has gradually diminished since these early times. Although the frequency of impacts due to bodies of all sizes is considerably less than during the first 700 million years of solar system history, as the planetary orbits have stabilized and a significant proportion of the smaller objects has been accreted, the most significant risk remains from collisions with bodies on oval-shaped orbits (such as comets) and objects with orbits that pass near Earth’s orbit. The average amount of material accreted daily to Earth is estimated to be in the range of 50 to 150 tons of very small objects (Love and Brownlee, 1993). This material is mostly dust, although there are abundant small objects that burn up quickly in the atmosphere and are evidenced by meteor trails. More rarely, larger objects impact Earth. It is now widely believed that the impact of an approximately 10-kilometer-diameter object formed the Chicxulub Crater near the Yucatan Peninsula about 65 million years ago, very likely resulting in the extinction of the dinosaurs. Its mass is similar to that of the total amount of dust and other small objects accreted to Earth during the time since that impact. Substantial atmospheres around planetary bodies act as significant filters to incoming objects. Smaller objects, particularly those that are lower in density and more fragile, vaporize in the upper reaches of the atmosphere, while more intact, larger bodies may survive to impact the surface. Thus, small craters are much less common on bodies with dense atmospheres, such as Earth, Venus, and Titan, than they are on Mercury and the Moon, with Mars somewhere in between. Of course there are still substantial numbers of large impact craters even on Venus, with its dense carbon dioxide atmosphere; the lack of weathering and erosion, coupled with low rates of volcanic and tectonic activity over the past 0.5 billion years, has allowed the retention there of a significant number of craters, most largely unaltered since emplacement. By contrast, the movement of water on Earth and the action of plate tectonics have both resulted in the loss of much of the cratering record on this planet. There are more than 170 established impact craters on Earth, including the approximately 1.2-kilometer Meteor Crater in Arizona (Figure 2.1). The largest known terrestrial crater is the 300-kilometer-diameter Vredefort Crater in South Africa, dated at around 2 billion years old. 

OCR for page 12
 RISK ANALYSIS FIGURE 2.1 Meteor Crater (also known as Barringer Crater) in Arizona, with the Great Pyramids of Giza and the Sphinx inserted for size comparison. One of the most familiar impact features on the planet, this crater is about 1,200 meters in diameter and 170 meters deep; the interior of the crater contains about 220 meters of rubble overlying bedrock. The crater was formed about 50,000 years ago through the impact of an approximately 40-meter iron-nickel meteorite moving at about 13 kilometers per second (Melosh and Collins, 2005). SOURCE: Crater image courtesy of U.S. Geological Survey; composite created by Tim Warchocki. Over the past several decades, research has clearly demonstrated that major impact events have occurred throughout Earth’s history, often with catastrophic consequences. The Chicxulub impact apparently caused a mass extinction of species, possibly resulting from a global firestorm due to debris from the impact raining down around the planet. It may also have caused dramatic cooling for a year or more and global climatic effects that may have lasted a long time (e.g., O’Keefe and Ahrens, 1989). Many species became extinct at this time (includ - ing perhaps 30 percent of marine animal genera), but many survived and ultimately thrived in the post-dinosaur world. It may be that impacts throughout the history of this planet have strongly helped shape the development and evolution of life forms. Several recent events and new analyses have highlighted the impact threat to Earth: 1. As Comet Shoemaker-Levy 9 came close to Jupiter in 1992, tidal forces caused it to separate into many smaller fragments that then may have regrouped by means of self-gravity into at least 21 distinct pieces (e.g., Asphaug and Benz, 1994). These pieces impacted Jupiter in July 1994, creating a sequence of visible impacts into the gaseous Jovian atmosphere. The resultant scars in Jupiter’s atmosphere could be readily seen through Earth- based telescopes for several months. In July 2009, a second object, though much smaller than Shoemaker-Levy 9, impacted Jupiter, also causing a visible dark scar in the Jovian atmosphere. Such clear evidence of major collisions in the contemporary solar system does raise concern about the risk to humanity. 2. In December 2004, astronomers determined that there was a non-negligible probability that near-Earth asteroid Apophis (see Chapter 4 for more details) would strike Earth in 2029. As Apophis is an almost 300-meter- diameter object, a collision anywhere on Earth would have serious regional consequences and possibly produce transient global climate effects. Subsequent observations of Apophis ruled out an impact in 2029 and also deter - mined that it is quite unlikely that this object could strike during its next close approach to Earth in 2036. How - ever, there likely remain many Apophis-sized NEOs that have yet to be detected. The threat from Apophis was discovered only in 2004, raising concerns about whether the threat of such an object could be mitigated should a collision with Earth be determined to have a high probability of occurrence in the relatively near future. 3. In June 1908, a powerful explosion blew down trees over an area spanning at least 2,000 square kilometers of forest near the Podkamennaya Tunguska River in Central Siberia. As no crater associated with this explosion

OCR for page 12
 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES FIGURE 2.2 The long-lasting airburst trail over Sudan after the impact of 2008 TC3 on October 7, 2008. SOURCE: Courtesy of M. Elhassan, M.H. Shaddad, and P. Jenniskens. was located, scientists initially argued against an asteroid or comet origin. However, subsequent analysis and more recent modeling (see, e.g., Chyba, 1993; Boslough and Crawford, 1997, 2008) have indicated that modest-sized objects (the Tunguska object may have been only 30 to 50 meters in diameter) moving at high supersonic speeds through the atmosphere can disintegrate spontaneously, creating an airburst that causes substantial damage without cratering. Such airbursts are potentially more destructive than are ground impacts of similar-size objects. 4. A stony meteorite 1 to 2 meters in diameter traveling at high supersonic speeds created an impact crater in Peru in September 2007. According to current models with standard assumptions, such a small object should not have impacted the surface at such a high velocity. This case demonstrates that specific instances can vary widely from the norm and is a reminder that small NEOs can also be dangerous. 5. On October 6, 2008, asteroid 2008 TC3 was observed by the Catalina Sky Survey (see Chapter 3) on a collision course with Earth. Although the object was deemed too small to pose much of a threat, the Spaceguard Survey1 and the Minor Planet Center (see Chapter 3) acted rapidly to coordinate an observation campaign over the following 19 hours, with both professionals and amateurs to observe the object and determine its trajectory. The 2- to 5-meter-diameter object entered the atmosphere on October 7, 2008, and the consequent fireball was observed over northern Sudan (Figure 2.2) (Jenniskens et al., 2009). Subsequent ground searches in the Nubian Desert in Sudan located 3.9 kilograms (in 280 fragments) of material from the meteorite. These recent events, as well as the current understanding of impact processes and the population of small bodies across the solar system but especially in the near-Earth environment, raise significant concerns about the current state of knowledge of potentially hazardous objects and the ability to respond to the threats that they might pose to humanity. 1The Spaceguard Survey was mandated by Congress to detect 90 percent of NEOs 1 kilometer in diameter or greater by 2008.

OCR for page 12
 RISK ANALYSIS INVENTORY OF NEAR-EARTH OBJECTS (NEOS) AND POTENTIALLY HAZARDOUS NEOS Introduction Scientists’ ability to detect NEOs is dependent on how bright each individual object appears in the sky—which depends primarily on its distance from Earth, its size, its albedo (how well light reflects from its surface), and its location relative to the Sun. The observation of NEOs that appear very close to the Sun when viewed from Earth is difficult or even impossible. The brightness of each NEO also changes as it moves through its orbit, coming closer to and going farther away from Earth. As a result, it is very difficult to detect all NEOs, particularly smaller (fainter) asteroids, in the entire population. Figure 2.3 shows the distribution (in January 2010) of known asteroids in the inner solar system. (Note that the asteroids represented in Figure 2.3 are not all in the same orbital plane, and so it is more accurate to envision some of the objects above the page and some below it. The image is also very mislead - ing in the sense that on this scale, the asteroids would be invisible. The vast majority of the solar system is empty space, but there are nonetheless many objects present.) Of course, while many NEOs have been located, there are many yet to be discovered, some of which may represent a significant threat of impact on Earth. Using estimates of the distribution and orbits of these undiscovered NEOs, the committee can statistically address the hazard posed by NEOs, particularly those that are large enough to cause significant damage should they impact Earth. To determine what fraction of the entire NEO population has been detected, it is necessary to compute the total expected number of objects from knowledge of the properties of known NEOs and how objects are expected to get brighter and fainter as they and Earth move around their orbits. Using computer models one can determine the fraction of all NEOs of different sizes that will be detected for a particular survey strategy. As surveys approach completion and the knowledge of the NEO population increases, refinements are possible to the computer simula - tions that allow greater confidence in the predicted numbers of NEOs in each size range. Current estimates (Harris, FIGURE 2.3 The distribution of cur- rently known asteroids (in January 2010). The green dots represent asteroids that do not currently approach Earth. The yellow dots are Earth-approaching asteroids, ones having orbits that come close to Earth but that do not cross Earth’s orbit. The red boxes mark the locations of asteroids that cross Earth’s orbit, although they may not necessarily closely approach Earth. Contrary to the impression given by this illustration, the space represented by this figure is pre- dominantly empty. SOURCE: Courtesy of Scott Manley, Armagh Observatory.

OCR for page 12
 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES 2009)2 indicate that there should be a total of about 940 NEOs larger than 1 kilometer in diameter. This includes near-Earth asteroids but does not include long-period comets (orbital periods in excess of 200 years), which are believed to present less than on the order of 1 percent of the total NEO impact threat (Stokes et al., 2003). Based on this estimate and current NEO detections, the committee concluded that nearly 85 percent of all objects 1 kilo - meter in diameter or larger in the near-Earth environment have been detected. The committee has also shown that none of these objects presents a threat of impact on Earth within the next century. Although impacts of objects smaller than 1 kilometer in diameter do less damage than larger ones, it is this smaller class of objects that, owing to their far greater numbers, presents the most frequent threat to humanity. Estimates of the “risk” posed by the portion of the NEO population that has yet to be discovered require the following components: 1. The orbital distribution of undiscovered asteroids and comets capable of producing damage to human life or property. This information is used to compute the collision probabilities and impact velocities of the possible impactors on Earth. 2. The mass distribution of potential Earth impactors. Given the uncertainties about the properties of comets and asteroids, previous works have concentrated on the distribution of brightness of these objects at a standard dis - tance from both Earth and Sun. This distribution is then converted into an “uncalibrated” size distribution by making assumptions based on the present (incomplete) understanding of the average properties of these objects. Thus the committee can estimate equivalent diameters, D, from measurements of brightness, H, where the term “diameter” used here and in the subsequent text refers to the equivalent diameter of a sphere of the same volume. 3. The amount of “damage” produced by impactors when they strike different locations on Earth. Damage is usually calculated from components of the impact. One component is the impact energy distribution, which is computed from points 1 and 2, above. A second component, the worth of things of value on Earth (e.g., human life, infrastructure, and property), can be set in a manner similar to that used by insurance actuarial assessors. As property damage or loss of life will vary significantly with the geographical point of impact, realistic assessments of “damage” must allow for the stochastic nature of impacts and usually involve the use of Monte Carlo computer simulations. The previous reports by Stokes et al. (2003) and NASA PA&E (2006) reviewed available data on NEOs and made extensive calculations of the potential hazard to humankind from various populations of NEOs. The next sections briefly review the computations in Stokes et al. (2003) and NASA PA&E (2006). Both of these documents were fairly extensive in their descriptions and are still close to state of the art. Thus the committee only updates the calculations based on more recent scientific analysis, points out uncertainties and sensitivities of the results to assumptions, and comments where new work is needed. The Distribution of NEO Orbits The basis for the distribution of NEO orbits in both Stokes et al. (2003) and NASA PA&E (2006) comes from the work of Bottke et al. (2002). The method is fairly detailed, but, in brief, they used dynamical modeling to determine the primary source regions of NEOs (e.g., portions of the main asteroid belt, and the trans-Neptunian region that acts as a source of “Jupiter family” comets) and to create probability distributions of the destinations of the NEOs (e.g., into the Sun, interactions with planets, return to the asteroid belt). The probability distributions were then compared to models of observations of known NEOs detected by surveys (e.g., by Spacewatch and the Lincoln Near-Earth Asteroid Research [LINEAR] program; see Chapter 3). 3 These surveys found that most 2Some of the data presented to the Survey/Detection Panel by Harris (2009) will also be published in the upcoming European Space Agency conference proceedings of the April 27-30, 2009, 1st International Academy of Astronautics Planetary Defense Conference: Protecting Earth from Asteroids. 3Spacewatch was one of the first NEO discovery systems, established in 1981 and run by the University of Arizona. The LINEAR program at the Massachusetts Institute of Technology Lincoln Laboratory is funded by the United States Air Force and NASA and was the most suc - cessful NEO search program from 1997 until 2004.

OCR for page 12
 RISK ANALYSIS kilometer-sized NEOs come from the inner and central parts of the asteroid belt. Only a small percentage (<20) comes from the outer main belt or are comets delivered from the trans-Neptunian region. Based on this distribution of orbits, these surveys find about 20 percent of the NEOs have orbits that pass within 0.05 astronomical units (AU) of Earth. NEOs in this class are called “potentially hazardous NEOs.” Stokes et al. (2003) and NASA PA&E (2006) used potentially hazardous NEOs to determine survey strategies. The Bottke et al. (2002) model has held up fairly well over the past several years as scientists have neared 85 percent completion of the survey for objects greater than 1 kilometer in diameter. Some limitations of this model exist for dimmer (or smaller) NEOs. For example, the NEO data used to calibrate the Bottke et al. (2002) NEO model were mainly kilometer-sized objects; few subkilometer-sized objects were known when the model was developed. If the population of kilometer-sized objects has the same distribution of orbits as the subkilometer-sized objects, the Bottke et al. (2002) model should work for the latter group. There are indications, however, that this equivalence may not hold. In particular: • Studies of fireballs (i.e., objects burning in Earth’s atmosphere) indicate that these submeter- to meter-sized objects mainly come from the central part of the asteroid belt (Morbidelli and Gladman, 1998), whereas studies of large NEOs indicate that the primary source of these objects is the inner part. It remains unclear whether these differences in source regions have meaningful consequences for the probabilities of collision with Earth and for the impact velocities for NEOs with diameters between 100 meters and 1 kilometer. • The population of smaller NEOs is more likely to be more affected than are larger objects by collisions or nongravitational force effects (Rubincam, 2000; Bottke et al., 2006; Walsh et al., 2008). The effects of such mechanisms could modify or even disrupt certain NEOs and thereby modify the overall orbit and size distributions of the population. Additional survey and numerical work will be needed to settle these questions. In addition, although the popula- tion models based on Bottke et al. (2002) have predicted rather well the discoveries to date, their model may need to be recalibrated as the survey is extended to smaller objects. Furthermore, as scientists pass 90 percent survey completion, we are approaching the tails of the distribution of orbits where the model is far less robust. The Size Distribution of NEOs and Potentially Hazardous NEOs Most NEOs with diameters under half a kilometer remain undiscovered, although many of the larger objects in this size range have been identified in past surveys. Although the size distribution of these objects can be esti - mated by modeling NEO survey data (e.g., NEO discoveries plus accidental rediscoveries), scientists’ incomplete knowledge of these objects limits our ability to assess the nature of this impact hazard. To this end, Stokes et al. (2003) and NASA PA&E (2006) decided that it was reasonable to attempt a conservative, upper-limit-type estimate of the NEO population over all sizes. They concluded that the cumulative number, N, of NEOs with diameters greater than D could be described by: N = 942D−2.354, where D is in units of kilometers. The exact number of NEOs greater than 1 kilometer in diameter is uncertain, but reasonable estimates as noted above suggest that it is somewhere around 940, in agreement with the above formula. Another calibration point for this function comes from detections of small (1- to 20-meter-diameter) objects entering Earth’s atmosphere (e.g., Brown et al., 2002; Silber et al., 2009; also, Chapter 4). The number of potentially hazardous NEOs has been estimated to be 21 percent of the above function (Stokes et al. 2003; see also Bottke et al., 2002). More recent estimates of the distribution of sizes under 1 kilometer in diameter come from Harris (2009), who “debiased” the existing database of NEO discoveries and accidental rediscoveries using the methods described in NASA PA&E (2006). His work indicates that somewhere between the calibration points described above, the NEO size distribution deviates from the above formula by factors of a few, suggesting that the curve is steeper for very

OCR for page 12
 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES -1 2 5 8 10 10 10 10 10 10 Current population estimate Tunguska 0 Assumed (2003) population 10 Discovered to 1/19/2009 8 10 2 10 Impact Interval, years 6 10 N( OCR for page 12
 RISK ANALYSIS ways before crossing the street: Observation does not affect the distribution of either cars or NEOs, but it is indis - pensable for determining what actions should be taken to remain safe in both situations. Having determined the sizes and distribution of orbits for NEOs, one wants to understand the risk to human life and property that is presented by various sizes of NEOs. Although the impact of a large NEO (diameter greater than 1 kilometer) anywhere on Earth would have major consequences in terms of loss of life and damage to property, the frequency of such impacts is very low (Figure 2.4, Table 2.1), and thanks to Spaceguard, nearly 85 percent of such objects have already been detected. None of those detected objects has a significant chance of impacting Earth in the next century. Damage Produced by the Impact of NEOs To evaluate the risk posed by NEOs, one must estimate the distribution in time of impact energy on Earth. This distribution can be computed from three components: (1) the collision probability of potentially hazardous NEOs with Earth, which is a function of the distribution of orbits; Stokes et al. (2003) estimated that the average collision rate with Earth per single NEO is 1.6 × 10−9 yr−1 and per single potentially hazardous NEO is 8.4 × 10−9 yr−1; (2) the impact velocity distribution of potentially hazardous NEOs with Earth, which again is a function of the distribution of orbits; Stokes et al. (2003) used an impact velocity for potentially hazardous NEOs striking Earth of 20 kilometers per second in their computations; (3) the mass distribution of potentially hazardous NEOs striking Earth; this component is obtained by calculating the masses of the objects on the assumption that they have densities of 2.5 g cm–3. Table 2.1 is based on such information to provide an approximate indication of the average impact interval and impact energy for objects of various sizes. Even if these data were accurate, the determination of impact hazard would remain challenging for the fol - lowing reasons: • The direct and indirect effects produced when an asteroid or comet strikes the land or ocean are only poorly understood at present. • The population of Earth is not uniformly distributed. For example, there is a higher population density near coastlines, where people may be susceptible to impact-driven tsunamis (whose damage potential is very uncertain). • Until the population of small NEOs is understood, the impact effects of undiscovered objects can only be characterized statistically. As noted above, most impact simulations indicate that the likelihood is low that human TABLE 2.1 Approximate Average Impact Interval and Impact Energy for Near-Earth Objects Characteristic Diameter Approximate Impact Approximate Average Impact Type of Event of Impacting Object Energy (MT) Interval (yrs) Airburst 25 m 1 200 Local scale 50 m 10 2,000 Regional scale 140 m 300 30,000 Continental scale 300 m 2,000 100,000 Below global catastrophe threshold 600 m 20,000 200,000 Possible global catastrophe 1 km 100,000 700,000 Above global catastrophe threshold 5 km 10 million 30 million Mass extinction 10 km 100 million 100 million NOTE: This table provides only very approximate long-term average data for impact energy (also known as kinetic yield) and impact interval. The correlation of impact diameter with scale of damage and impact energy is based on assumptions delineated in Stokes et al. (2003). It must be kept in mind that there may be significant variability in the scale of damage and impact energy depending on the velocity and the physical and chemical characteristics of the impacting NEO. MT stands for megatons, which refers to the chemical energy release of a million tons of TNT. SOURCE: NASA PA&E (2006), updated by Harris (2009).

OCR for page 12
0 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES life will be significantly affected by impacts over short time scales (i.e., of less than 1,000 years). However, as all NEOs have not yet been detected and characterized, it is possible (but very unlikely) that an NEO will “beat the odds” and devastate a city or a coastline in the near future. • Actuarial studies provide an assessment of property values and may be used to place a value on a human life, but it is very challenging to measure, for example, the value of religious, historical, ecological, cultural, and political sites, as well as the value of entire societal entities (such as ethnic groups, cities, and nations). These values may vary greatly across communities, regions, and nations. • Beyond very crude estimates, it is not known what the size threshold is for impacts that would lead to a global catastrophe and kill a significant fraction of Earth’s population as a result of firestorms or climate change and the associated collapse of ecosystems, agriculture, and infrastructure. There may not even be a well-defined threshold, because global effects probably depend critically on impact location and surface material properties (e.g., land, sea, ice sheet), season, and so on. As Stokes et al. (2003) provide an in-depth discussion of these issues, there is no need to reproduce it in detail here. Land Impacts That Are Incapable of Producing Global Effects Land impacts correspond to the damage produced by asteroids or comets that either strike the ground or explode low enough in the atmosphere to produce damage on the ground. Stokes et al. (2003) based their damage assessments on the modeling work of Hills and Goda (1993). According to the estimates of Hills and Goda (1993), hard, stone objects between 40 and 150 meters in diameter explode upon entry into Earth’s atmosphere and generate airbursts capable of producing surface damage. In this manner, they are similar to the Tunguska airburst. Larger, more energetic impacts naturally produce destruction over a wider area. As the size of the damage zone increases, more cells within the gridded map in the Monte Carlo code are affected, although damage decreases as a function of distance from the impact site. To account for a range of outcomes, error estimates were included that accounted for minimum, nominal, and maximum numbers of fatalities per event. The results from the Stokes et al. (2003) Monte Carlo analysis indicate that 75 percent of all impacts do not produce any fatalities because they impact the oceans or uninhabited land areas. The most common impact events that produce highly lethal results are the smallest ones (less than 200 meters). Though their blasts are smaller in scope, their larger numbers give them more chances to affect a highly populated region. Scientists’ understanding of the immediate damage caused by land impacts capable of producing craters is reasonably mature because their effects are constrained by nuclear weapons tests as well as by craters on planetary surfaces. For airbursts, however, much work is needed to improve the understanding of their consequences. For example, many groups have studied the 1908 Tunguska blast. Using insights from nuclear blast data as well as seismograms and barograms of the Tunguska event, scientists estimated that the height of the explosion was about 10 kilometers and that the energy yield was 10 to 20 megatons (MT) (Chyba et al., 1993). According to the new estimate of size distribution made by Harris (2009), the average interval between such events on Earth would be on the order of one every 2,000 years. Work by Boslough and Crawford (1997, 2008), however, indicates that a much lower yield could produce the same effects. They found that asteroid airbursts do not act like point explosions in the sky (e.g., like a nuclear bomb explosion) but instead are more analogous to explosions along the line of descent. In an airburst, kinetic energy (see Appendix E) is deposited along the entry path, with significant downward momentum transferred to the ground. Accordingly, these researchers suggest that smaller explosions, with net yields of 3 to 5 MT, may be sufficient to produce Tunguska-like impact events. If true, the average interval between Tunguska-like events using the Harris (2009) size distribution (see Figure 2.4) would be on the order of a few hundred years. These results would increase the calculated hazard from smaller objects, perhaps those as small as 30 meters or so in diameter. Further research is needed to better characterize this threat.

OCR for page 12
 RISK ANALYSIS Tsunamis Produced by Ocean Impacts Ocean impacts from asteroids or comets affect their immediate surroundings but also have the potential to launch tsunamis that inundate coastlines and affect populated areas. Because tsunamis can potentially affect a wide area and because people like to live along coastlines, impact-driven tsunamis may present a disproportionate contribution to the total hazard from small NEOs. In the Stokes et al. (2003) model, impact-driven tsunamis were simulated using the results of Chesley and Ward (2003) (see also Chesley and Ward, 2006; Ward and Asphaug, 2000). Ocean run-up and damage to infrastructure along coastlines were computed as functions of impactor size and distance to the coastline. The population residing along various coastlines was taken from the work of Small et al. (2000). Given the many uncertainties in the model (e.g., the precise shape of the coastline, the depth of the seafloor adjacent to the coastline, harbor obstructions, the distance of people and property from the coastline, and so on), Stokes et al. (2003) assigned large lower and upper bounds to the assignment of damage within each geographic cell in the Monte Carlo analysis. There is considerable uncertainty about the nature and damage produced by impact-driven tsunamis, in large part because (1) direct experiments cannot easily be done; (2) impact-driven tsunamis present a difficult nonlinear modeling problem: computer simulations need extremely high resolution and fidelity to treat important factors such as breaking waves and run-up along a specific coastline; (3) the precise nature of the coast and seafloor near population centers strongly affects the results (e.g., consider the Pacific coast versus the shallow Gulf coast); and (4) a loss of life may be avoided by early warnings of an incoming tsunami. The classic work in this field is from Van Dorn et al. (1968), who used nuclear detonation data to show that the waves produced by a large blast would likely break on the continental shelf. Their motivation for this study was to show that tsunamis produced by nuclear blasts make poor tactical weapons if the goal is to knock out enemy submarines lying along the coast of the United States. The idea that large waves break at considerable distances offshore is now referred to as the Van Dorn effect (e.g., Korycansky and Lynett, 2007). Using the original Van Dorn report as a guide, Melosh (2003) argued that impact-driven tsunamis would have similar wavelengths and thus would also break along continental shelves. He predicted the damage from these events would be minimal. Korycansky and Lynett (2005) numerically confirmed the existence of the Van Dorn effect, but Korycansky and Lynett (2007) pointed out that some ocean run-up is still expected from waves that break. They suggested that their work should be incorporated into next-generation Chesley and Ward (2006)-type models to better determine damage from these events. (Note that the Van Dorn effect could only apply where there are continental shelves. Small amounts of bottom friction may nullify the effectit remains hypothetical.) At present, the assessment of the impact hazard is limited by the understanding of impact-driven tsunamis. The uncertainties of NEO-impact tsunamis therefore suggest three research areas: (1) the coupling of impact energy into ocean wave energy, both through water impacts and through airbursts; (2) the propagation of impact-induced waves across oceans; and (3) the effect on the world’s coastlines. Impacts Capable of Producing Global Effects The motivation for the original Spaceguard Survey was to find all of the NEOs larger than 1 kilometer in diameter capable of striking Earth. According to Toon et al. (1997), impacts by 2- to 3-kilometer-diameter aster - oids may be capable of causing global damage owing to the firestorm generated by the infall of impact debris or indirectly by affecting the climate and producing a so-called asteroid winter. Given the uncertainties in these calculations, Stokes et al. (2003), like other groups before them, decided to be conservative and assumed that all objects greater than 1.5 kilometers rather than 2 to 3 kilometers in diameter would cause a global catastrophe. Nonetheless, the true hazard represented by multi-kilometer impactors is only modestly understood at pres - ent. Other than by Toon et al. (1997) and a few other groups, little modeling has been done on the worldwide environmental effects produced by such impactors other than the one associated with the now-famous impact of an approximately 10-kilometer object 65 million years ago that apparently resulted in the extinction of the dinosaurs. More work in this area is clearly needed.

OCR for page 12
 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES Long-Period Comet Impacts Stokes et al. (2003) provide considerable description of the threat represented by long-period comets, and there is no need to repeat all of their arguments here. In brief, they find that the comet hazard constitutes only a tiny fraction (on the order of <1 percent) of the total risk to life on Earth by impacting NEOs (prior to the Spaceguard Survey) and that producing a complete catalog of hazardous long-period comets is far beyond the abilities of any proposed survey. For these reasons, they suggested that limited resources would be better utilized in finding and cataloging Earth-threatening near-Earth asteroids and short-period comets. With the completion of the Spaceguard Survey (that is, the detection of 90 percent of NEOs greater than 1 kilometer in diameter), long-period comets will no longer be a negligible fraction of the remaining statistical risk, and with the completion of the George E. Brown, Jr. Near-Earth Object Survey (for the detection of 90 percent of NEOs greater than 140 meters in diam - eter), long-period comets may dominate the remaining unknown impact threat. Furthermore, these comets present a difficult challenge, as they are large objects, and there will be only a short warning time (months to a very few years maximum) before impact. Thus mitigation options are very limited, as noted in Chapter 5. Assessing the Hazard From their Monte Carlo analyses, described above, Stokes et al. (2003) estimated the hazard from all potential impactors in terms of fatalities per year. However, since 2003, new information has been presented that affects the shape of the hazard curve (Figure 2.5). For example: • The NEO and potentially hazardous NEO size distributions may not follow the simple law as shown by the dashed line in Figure 2.4 but instead may have a dip, as illustrated by the open circles. If so, the frequency of impacts of objects with diameters in the 50- to 500-meter range might decrease by a factor of two to three below the Stokes et al. (2003) estimates; Land impact/airburst 9 10 Ocean impact/tsunami Global effects Sum of all Fatalities per event 7 10 5 10 3 10 0.1 1 10 Diameter of impactor, km FIGURE 2.5 Model of fatalities per event for impacts of various size NEOs. The solid curve represents the total fatalities associated with both ocean and land impacts, including those with global effects. The sharp increase in the solid (red) curve Figure 2.5.eps reflects the assumption of a large increase in fatalities for an impact that crosses the global-effect threshold. SOURCE: Courtesy of Alan W. Harris, Space Science Institute.

OCR for page 12
 RISK ANALYSIS • The number of fatalities from impact-driven tsunamis in the Stokes et al. (2003) analysis was treated inconsistently, with different numbers of fatalities used in separate parts of the calculation; • The ground damage produced by airbursts from Tunguska-like events may have been underestimated. Increasing the area of damage in the Monte Carlo analysis by such events from impactors in the size range of about 50 to 150 meters in diameter and lowering the size threshold for surface damage increases the risk from such objects. These revised factors, illustrated in Figure 2.5, yield the fatalities per impact event versus the size of impactor (Harris, 2009). There is a tail on the fatalities curve at small diameters, which reflects the increase in statistical risk associated with airburst events, and revision downward in the deaths associated with tsunamis resulting from ocean impacts. However, this latter revision may not be warranted. Above the conservatively assumed global catastrophe threshold from a 1.5-kilometer-diameter impactor, the number of fatalities ramps up from 10 percent of the world’s population to the entire population for impactors above 10 kilometers in diameter. Clearly, there are many assumptions in developing such models that result in difficult-to-determine uncertainties in the calculated fatalities. Nonetheless, Figure 2.5 provides a useful illustration of the significant increase in potential destruction and death with impactor size. To assess the effectiveness in the reduction of statistical risk from the various survey activities, consider the predicted average annual fatalities derived by multiplying the expected deaths per event by the frequency of events of a certain size. This risk is “actuarial” and is an average of many thousands of years with few fatalities and a single low-probability, high-fatality impact year. Nevertheless, it is an objective method that can be used for order-of-magnitude comparisons with other risks that take place on radically different time scales. Figure 2.6 shows such a figure from Harris (2009) for the NEO population as it was known before the Spaceguard Survey. The plot shows estimated average fatalities per year and clearly indicates that most of the threat comes from the larger objects that exceed the global catastrophe threshold, even though the probability of an impact by these objects is very low relative to that for smaller objects. The two sets of histograms are based on (1) the poten - 250 Stokes et al. (2003) 1,254 total Harris (2009) 1,168 total 200 Fatalities per year 150 100 50 0 0.1 1 10 Diameter of impactor, km FIGURE 2.6 Estimated average fatalities per year for impacts by asteroids of various sizes calculated for the circumstances prior to the Spaceguard Survey. One histogram references the data used in the Stokess al. (2003) study. The new revised data Figure 2.6 - riskhist1.ep et include corrections resulting from the understanding of the threat due to tsunamis and airbursts and from recent revisions to the size distribution of NEOs (see Figure 2.4). SOURCE: Courtesy of Alan W. Harris, Space Science Institute.

OCR for page 12
 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES tially hazardous NEO population of Stokes et al. (2003) and their assumed hazard due to airbursts and tsunamis, and (2) the recalculation based on the revised population curves shown in Figure 2.4 and the reassessed impact hazard for airbursts and tsunamis from Harris (2009), as illustrated in Figure 2.5, which, as noted earlier, may not be warranted. Assuming that 85 percent of the NEOs with diameters larger than 1 kilometer have been discovered, which is close to the present state of affairs, Harris (2009) calculated the hazard statistics shown in Figure 2.7. Here the reassessed risk presented by the remaining 15 percent of the NEOs with diameters greater than 1 kilometer is comparable to that from all smaller objects. Figure 2.7 predicts that, in an actuarial sense, there is a long- term statistical average of about 91 fatalities worldwide per year due to impacts. Because the assessed statistical hazard from mid-range objects has dropped, the overall hazard has decreased as well. The drop from >1,000 to 91 expected fatalities per year clearly demonstrates the results of the Spaceguard Survey to date, which has “retired” the statistical risk from most objects above the assumed global catastrophe threshold. Using the Stokes et al. (2003) data for asteroids smaller than 1 kilometer in diameter yields a “humped” distribution with a peak near 300 to 400 meters. This hump may be significantly reduced when more realistic assessments of the effects of impact-driven tsunamis are available. The residual hazard was used to establish the Stokes et al. (2003) goal that a future survey should try to identify 90 percent of the NEOs with diameters of 140 meters or greater. This limiting value, according to survey simulation of potentially hazardous NEOs, could remove a significant proportion of the remaining statistical hazard that still exists after the conclusion of the Spaceguard Survey. The completion of this survey does not change the probability of Earth impact for any undetected NEO. However, if none of the objects detected in the survey is on a collision course with Earth, the total statistical risk of impact is decreased as a result of the reduction in the total number of unknown potentially hazardous NEOs. Nonetheless, this survey may detect one or more NEOs on a collision course with Earth. (Carrying out a survey per se does not remove whatever risk there is; one just learns more about that risk.) In carrying out this survey, a substantial fraction of NEOs with diameters 50 meters and greater will also be discovered and catalogued. Although not specifically designed for the purpose, such surveys may also detect as many as half of the NEO “imminent impactors” larger than 10 meters in diameter in the hours to months prior 25 Stokes et al. (2003) 100 total Harris (2009) 91 total 20 Fatalities per year 15 10 5 0 0.1 1 10 Diameter of impactor, km FIGURE 2.7 Estimated average fatalities per year for impacts by asteroids of various sizes calculated for the circumstances after 85 percent completion of the Spaceguard Survey. One histogram references the data used in the Stokes et al. (2003) study. Figure 2.7 - riskhist2.eps The new data include changes resulting from newer estimates of the threat due to tsunamis and airbursts and from recent revi - sions to the size distribution of NEOs (Figure 2.4). SOURCE: Courtesy of Alan W. Harris, Space Science Institute.

OCR for page 12
 RISK ANALYSIS to their impact with Earth. The discovery of such objects shortly before impact provides an opportunity to save lives by evacuation or by suitable sheltering rather than by human changing of their orbits. Based on these results, one could argue that a change is needed in the minimum diameter of the object to be included in the search, say, from 140 meters down to 50 meters. Nevertheless, the committee concluded that work on detecting these smaller objects should not be at the expense of detecting objects 140 meters and greater in diameter (see the recommendation at the end of Chapter 3). Additional information could change the relative statistical hazard associated with the various size ranges of NEOs as the following data are obtained: • Orbital distributions and collision probabilities for subkilometer-sized impactors; • More reliable estimates of the effects of Tunguska-like and larger impacts, including tsunami damage; and • Maps that more realistically account for human population distribution and growth. As was clearly stated in the Stokes et al. (2003) and NASA PA&E (2006) studies, the completion of the survey as currently conceived will result in a significant amount of the residual statistical risk residing with the long-period comet population. WARNING TIME FOR MITIGATION A key issue associated with the hazard from NEOs is that the length of time needed to execute a mitigation strategy involving orbit change is likely to require acting before the knowledge of the trajectory is sufficiently accurate to know with high confidence that an impact would occur without mitigation. It is possible, therefore, that action to mitigate could be deferred until it is too late if plans are not already in place to act when the probability of impact reaches some level that is well below unity. As addressed in Chapter 5, the time required to mitigate optimally (other than only by means of civil defense) is in the range of years to decades, but this long period may require acting before it is known with certainty that an NEO will impact Earth. Chodas and Chesley (2009) have simulated the discovery of objects that would impact within the 50 years starting at the beginning of the next generation of surveys (see Chapter 3), using estimates of the (decreasing) orbital uncertainty as observations are accumulated. Although there are many assumptions in this approach, the most important is whether or not the surveys and the follow-up programs to determine the orbits will be funded and will operate as assumed. Chodas and Chesley (2009) assume that an NEO is declared “truly hazardous” and worthy of mitigation preparations when the probability of hitting Earth reaches 0.5 (any other assumption regard - ing the decision point is also easily simulated). In this simulation, about 90 percent of impacting NEOs larger than about 140 meters in diameter are discovered in a 10-year survey. The temporal distribution of discoveries in this simulation showed that several percent of the 140-meter-sized objects that impact do so before discovery, but the total number of impactors per century is not large, so that a few percent represents an exceptionally unlikely event. Most of the impactors in this size range are discovered to be truly hazardous within several years of discovery, typically at the next time that the object is in a location in which it is viewable, thus providing warning times of a decade to several decades. By contrast, more than 10 percent of the objects larger than 50 meters in diameter that would impact within 50 years do impact before discovery, and there are many more of these than there are of the larger objects. Such smaller objects would generally be found to be truly hazardous within weeks to months before impact. Objects in the size range of 10 to 50 meters in diameter make up the majority of all potentially hazardous NEOs larger than 10 meters. The damage that could be caused by one of these smaller objects is less than for a larger object, but those smaller ones that are detected are likely to be found, at most, hours to months prior to their final plunge, with civil defense then being the only plausible mitigation strategy. Currently, by far the most probable scenario is that of a small impactor, likely to cause at most only local destruction. However, the assessed probability of any particular scenario is changing with time as the next- generation surveys discover most of the larger objects and the understanding of impact processes, such as airbursts and tsunami generation, improves. Thus, planning for mitigation must continue to evolve over time. Furthermore, when working with the statistics of small samples, and particularly when less likely scenarios have outcomes that

OCR for page 12
26 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES are so much more catastrophic than the most likely scenario, one should not assume that the next event will be the most likely one. SOCIETAL ELEMENTS OF NEO RISKS Unlike most other known natural hazards to humanity, such as earthquakes, volcanic eruptions, tsunamis, hurricanes, and tornadoes, NEO impacts present a very large spread of disaster scales ranging from small property damage to global extinction events. Larger impacts may result in global climatic changes that can result in famine and disease, infrastructure failure and, potentially, societal breakdown. Smaller impacts could be misinterpreted and thereby could conceivably even trigger wars. Numerous small incidents present little risk to people and prop- erty, but major impact events occur very infrequently. Impacts represent the extreme example of “low-probability, high-consequence” events. Although the probability of such a major impact within the next century may be small, a statistical risk of such an impact remains. Because of the nature of the impact threat, the expected fatality rate from impacts is an “actuarial” estimate based on calculations with attempted conservative assumptions. All the other estimates in Table 2.2 are based on the attribution of causes of actual fatalities from ongoing threats that may change in the future. In contrast to other known natural hazards, there has been no significant loss of human life to impacts in historical times, due to the low frequency of major impacts and the higher probability of impact in unpopulated TABLE 2.2 Expected Fatalities per Year, Worldwide, from a Variety of Causes Cause Expected Deaths per Year Shark attacksa 3-7 Asteroidsb 91 Earthquakesc 36,000 Malariad 1,000,000 Traffic accidentse 1,200,000 Air pollutionf 2,000,000 HIV/AIDSg 2,100,000 Tobaccoh 5,000,000 NOTE: The entries in this table are of various types. For example, the fatality rates given for shark attacks, earthquakes, traffic accidents, and HIV/AIDS entries are extrapolations, based on past reported individual deaths due to these causes, estimates of the completeness of these reports, and the assumption that future such deaths will continue at the same average rates (or straightforward extrapolations from them). The asteroid impact entry has been treated in this chapter and is based on models for impact and tsunami effects, an assumption of ecological collapse above some global catastrophe threshold, and a statistical calculation of risk based on the known near-Earth-object size distribution, with the temporal rate expected to vary enormously from the rate given, that is, to be zero most years, sizable in a relatively few years, and enormous in only an extremely few years over a time span of a billion years. The entries for malaria and tobacco fatalities are inferences based on plausible assignments of causes of deaths; such assignments are, individually, far less reliable than, for example, is the case for shark attack fatalities. Mitigation Panel member Mark Boslough wanted an additional entry in this table for fatalities due to climate change. The Steering Committee disagreed with including this entry because it did not think that a reliable estimate is available, among other reasons. Dr. Boslough’s minority opinion is provided in Appendix D. aData from International Shark Attack File, http://www.flmnh.ufl.edu/fish/sharks/ statistics/statsw.htm. bData from Harris (2009) and Figure 2.7. cWorldwide, 1970-2009; data from U.S. Geological Survey, cited in http://earthquake.usgs.gov/earthquakes/eqarchives/year/. dData from http://apps.who.int/malaria/wmr200 � /malaria200 � .pdf. http://apps.who.int/malaria/wmr200 eData from http://whqlibdoc.who.int/publications/2004/9241562609.pdf. fData from http://www.who.int/mediacentre/factsheets/fs313/en/index.html and http://whqlibdoc.who.int/ hq/2006/WHO_SDE_ PHE_OEH_06.02_eng.pdf. gData from http://data.unaids.org/pub/EPISlides/2007/2007_epiupdate_en.pdf. hData from http://www.emro.who.int/TFI/PDF/TobaccoHealthToll.pdf. SOURCES: Data for this table were derived from the sources listed above, as well as the World Health Organization.

OCR for page 12
 RISK ANALYSIS areas (notably the oceans) rather than in populated regions. Unlike the other hazards listed in Table 2.2, the hazard statistics for NEOs are dominated by single events with potentially high fatalities separated by long time inter- vals. Should scientists identify a large life-threatening object on a collision course with Earth, tremendous public resources to mitigate the risk would almost certainly be brought to bear. However, options for effective mitigation become much more limited when threatening objects are identified with only months to years, rather than decades or centuries, before impact. Thus, one of the greatest elements of risk associated with NEOs is the publicís expec - tation that governments will provide protection against any threat from NEOs, even as governments and agencies have been unwilling so far to expend public funds in a concerted effort to identify, catalog, and characterize as many potentially dangerous NEOs as possible, as far in advance of a damaging impact event as feasible. Given these issues, there are a number of concerns that can be addressed by an NEO detection, characteriza - tion, and mitigation program: 1. The statistical risk to human life and property associated with impacts of NEOs is real, but it falls outside the everyday experience of most of humanity. This risk must therefore be communicated effectively to the community at large in the context of other natural disasters, particularly those that the local community is likely to encounter. Scientists must carefully assess and explain the hazard so that appropriate public policy measures, commensurate with the level of risk, can be put into action. 2. There must be an assessment of the statistical risk from NEOs that is reasonable and acceptable to the gen - eral public. The mandate of discovery of 90 percent of objects 140 meters in diameter or greater in the George E. Brown, Jr. Near-Earth Object Survey Act of 2005 was based on many assumptions about impact hazards. However, periodic reassessment of the impact threat needs to be performed as the knowledge base on NEO populations, their physical characteristics, and impact-associated processes increases. 3. It is important to assess the length of time that the public is prepared to wait for scientific surveys to reach target goals of detection and characterization and for mitigation technologies to reach the desired maturity. Whereas surveys will never be 100 percent complete given the diversity of the objects, their origins, and their orbits, surveys should be as close as feasible to 100 percent complete in order to assure the public that all reasonable precautions are being taken. 4. An assessment is needed of the levels of expenditure that the public is prepared to accept in order to reach such goals for detection, and similarly for characterization, and mitigation. Although the costs (other than for advanced mitigation strategies) are almost vanishingly small relative to other elements of the federal budget, public support for such activities may be absent lacking demonstration of a clear and present threat. Undoubtedly issues 2, 3, and 4 above are strongly interrelated, as higher mandated percentage detections of increasingly smaller objects over shorter time periods would drastically increase cost. Equally, a comprehensive near-term mitigation strategy to address the full spectrum of possible NEO threats would be more expensive than a phased program of technology development. In the following chapters, various scales of NEO detection, characterization, and mitigation programs are presented that seek to identify a greater percentage of potentially threatening objects and to expeditiously develop the knowledge and capability to mitigate the risk associated with those objects. In addition, a program of research activities is presented to provide better constraints on the threat by various classes of NEOs impacting in diverse environments. REFERENCES Asphaug, E., and W. Benz. 1994. Density of Comet Shoemaker-Levy 9 deduced by modelling breakup of the parent “rubble pile.” Nature 370:120-124. Baldwin, R.B. 1985. Relative and absolute ages of individual craters and the rate of infalls on the Moon in the post-Imbrium period. Icarus 61:63-91. Boslough, M., and D. Crawford. 2008. Low-altitude airbursts and the impact threat. International Journal of Impact Engineering 35:1441-1448. Boslough, M.B.E., and D.A. Crawford. 1997. Shoemaker-Levy 9 and plume-forming collisions on Earth. Near-Earth Objects, the United Nations International Conference: Proceedings of the International Conference held April 24-26, 1995, in New York, N.Y. (J.L. Remo, ed.). Annals of the New York Academy of Sciences 822:236-282.

OCR for page 12
 DEFENDING PLANET EARTH: NEAR-EARTH-OBJECT SURVEYS AND HAZARD MITIGATION STRATEGIES Bottke, W.F., A. Morbidelli, R. Jedicke, J. Petit, H.F. Levison, P. Michel, and T.S. Metcalfe. 2002. Debiased orbital and absolute magnitude distribution of the near-Earth objects. Icarus 156:399-433. Bottke, W.F., D. Vokrouhlicky, D.P. Rubincam, and D. Nesvorny. 2006. The Yarkovsky and YORP effects: Implications for asteroid dynamics. Annual Review of Earth and Planetary Sciences 34:157-191. Brown, P., R.E. Spalding, D.O. ReVelle, E. Tagliaferri, and S.P. Worden. 2002. The flux of small near-Earth object colliding with the Earth. Nature 420:294-296. Chesley, S.R., and S.N. Ward. 2003. Assessing the human hazard from impact-generated tsunami. American Astronomical Society Division for Planetary Sciences meeting, September 1-6, Monterey, Calif. Abstract. Chesley, S.R., and S.N. Ward. 2006. A quantitative assessment of the human and economic hazard from impact-generated tsunami. Natural Hazards 38:355-374. Chodas, P., and S. Chesley. 2009. Impact warning times for the next generation of asteroid surveys. Proceedings of IAA Planetary Defense Conference. Chyba, C. 1993. Death from the sky. Astronomy 21:38-45. Errata in C. Chyba. 1993. Tunguska corrections. Astronomy 22:12. Chyba, C.F., P.J. Thomas, and K.J. Zahnle. 1993. The 1908 Tunguska explosionAtmospheric disruption of a stony asteroid. Nature 361:40-44. Harris, A.W., Space Science Institute. 2009. The NEO population, impact risk, progress of current surveys, and prospects for future surveys. Presentation to the Survey/Detection Panel of the NRC Committee to Review Near-Earth Object Surveys and Hazard Mitigation Strategies, January 28-30, 2009. Hills, J.G., and M.P. Goda. 1993. The fragmentation of small asteroids in the atmosphere. Astronomical Journal 105:1114-1144. Ivanov, B.A., G. Neukum, W.F. Bottke, and W.K. Hartmann. 2002. The comparison of size-frequency distributions of impact craters and asteroids and the planetary cratering rate. Pp. 89-101 in Asteroids III (W.F. Bottke, A. Cellino, P. Paolicchi, and R. Binzel, eds.). University of Arizona Press, Tucson, Ariz. Jenniskens, P., and 34 colleagues. 2009. The impact and recovery of asteroid 2008 TC 3. Nature 458:485-488. Korycansky, D.G., and P.J. Lynett. 2005. Offshore breaking of impact tsunami: The Van Dorn effect revisited. Geophysical Research Letters 32:10608. Korycansky, D.G., and P.J. Lynett. 2007. Run-up from impact tsunami. Geophysical Journal International 170:1076-1088. Love, S.G., and D.E. Brownlee. 1993. A direct measurement of the terrestrial mass accretion rate of cosmic dust. Science 262:550-553. Melosh, H.J. 2003. Impact-generated tsunamis: An over-rated hazard. Lunar and Planetary Science XXXIV:2013. Melosh, H.J., and G.S. Collins. 2005. Planetary science: Meteor crater formed by low-velocity impact. Nature 434:157. Morbidelli, A., and B. Gladman. 1998. Orbital and temporal distributions of meteorites originating in the asteroid belt. Meteoritics and Planetary Science 33:999-1016. NASA PA&E (NASA Program Analysis and Evaluation). 2006. 2006 Near-Earth Object Survey and Deflection Study. NASA Headquarters, Washington, D.C. O’Keefe, J.D., and T.J. Ahrens. 1989. Impact production of CO2 by the Cretaceous/Tertiary extinction bolide and the resultant heating of the Earth. Nature 338:247-249. Rabinowitz, D.L., E. Helin, K. Lawrence, and S. Pravdo. 2000. A reduced estimate of the number of kilometre-sized near-Earth asteroids. Nature 403:165-166. Rubincam, D.P. 2000. Radiative spin-up and spin-down of small asteroids. Icarus 148:2-11. Silber, E.A., D.O. ReVelle, P.G. Brown, and W.N. Edwards. 2009. An estimate of the terrestrial influx of large meteoroids from infrasonic measurements. Journal of Geophysical Research (Planets) 114:8006. Small, C., V. Gornitz, and J. Cohen. 2000. Coastal hazards and the global distribution of human population. Environmental Geosciences 7:3-12. Stokes, G., D. Yeomans, W.F. Bottke, S. Chesley, J.B. Evans, R.E. Gold, A.W. Harris, D. Jewitt, T.S. Kelso, R. McMillian, T. Spahr, and S.P. Worden. 2003. A Study to Determine the Feasibility of Extending the Search for Near Earth Objects to Smaller Limiting Magnitudes. Report Prepared at the Request of NASA Headquarters Office of Space Science’s Solar System Exploration Division. Toon, O.B., K. Zahnle, D. Morrison, R.P. Turco, and C. Covey. 1997. Environmental perturbations caused by the impacts of asteroids and comets. Reviews of Geophysics 35:41-78. Van Dorn, W.G., B. LeMehaute, and L.-S. Hwant. 1968. Handbook of Explosion-Generated Water Waves. Vol. 1. State of the Art. Tetra Tech, Pasadena, Calif. Walsh, K.J., D.C. Richardson, and P. Michel. 2008. Rotational breakup as the origin of small binary asteroids. Nature 454:188-191. Ward, S.N., and E. Asphaug. 2000. Asteroid impact tsunami: A probabilistic hazard assessment. Icarus 145:64-78.