Beginning with this chapter, we turn our attention to the present-day danger posed by near-Earth object (NEO) impacts. Our own contribution to the impact hazard debate has been largely associated with efforts to evaluate the risk in quantitative form. We have also played a role in publicizing the impact hazard and stimulating discussion of public policy issues. In this and the following chapters, we include many personal impressions of the key personalities and critical events that have characterized the recent (and continuing) debates on this issue. This chapter focuses on recognition and early evaluation of the hazard, beginning with a NASA workshop in 1981, the first time this subject was addressed by scientists and policy-makers. One result of this meeting was funding for Project Spacewatch in Arizona, the prototype of modern asteroid searches, pioneered by astronomer Tom Gehrels. It was Spacewatch that reported in 1989 the widely-publicized near-miss by asteroid Ascelpas. In the same year we published our book Cosmic Catastrophes, comparing the risk of death from impact to that of a round-trip commercial air flight. These two events, plus lobbying by the American Institute of Aeronautics and Astronautics, spurred the US Congress to request that NASA study the hazard, propose ways to accelerate the discovery of NEOs, and convene a workshop on the technologies that might be used to protect against a cosmic impact.
CHAPTER 11: SPACEGUARD SURVEY
"It is an intolerable uncertainty that right now there are some 1700 or so Earth-crossing asteroids larger then one kilometer in diameter and we don't know whether any one of these will be hitting the Earth with the force of hundreds of thousands of Hiroshima bombs. That requires the greatest urgency." -- Spacewatch astronomer Tom Gehrels (on Three Minutes to Impact, 1997) The United States Congress took a bold initiative in asking NASA to study the impact hazard. At that time, most NASA managers were skeptical of becoming involved in the issue. There was good reason to be cautious. In 1991, the space agency was suffering from an apparently endless series of technical and public-relations disasters. The Hubble Space Telescope, pride of the NASA space science program, was found to have been launched with an optical flaw that seriously degraded its resolution, giving it fuzzy vision. Erratic performance of the Space Shuttle main engines led to repeated launch delays. The nation's editorial writers and political cartoonists were savaging NASA for both technical and management failures. The request to study the dangers of "space rocks" and "killer asteroids" seemed to NASA officials like an opportunity best declined.
The example of Vice President Dan Quayle reinforced NASA officials' sense of caution. A fan of space exploration, Quayle had established the President's National Space Council as an independent voice in U.S. space policy. One of the Space Council staffers, on detail from the Air Force, had included a reference to the asteroid danger in one of the Vice President's speeches. Always happy to find fault with Quayle as a mental lightweight, the press ridiculed his "Chicken Little" concerns. The Vice President, stung by the criticism, banned all reference to asteroids from his future speeches, and NASA took note. The "giggle factor" had become an ingredient of public discussion of the impact threat, and officials at NASA Headquarters did not want to risk further ridicule in the press.
Although NASA managers may not have relished becoming entangled in the asteroid impact business, a study had been formally requested by the Congress, so they had to comply. In the spring of 1991 two separate committees (or "workshops") were set up to deal with the Congressional request. We participated in both. The first part of the Congressional charge, the search for potentially threatening asteroids, seemed to be a straightforward technical issue, and NASA assembled an international team of astronomers chaired by one of us (Morrison) to prepare a response. The second element of the problem, how to organize defense technologies to deal with an incoming object, lay outside NASA's usual mission, especially if the politically sensitive issue of nuclear weapons were raised. This is exactly the concern from George Wetherill that contributed to the failure of NASA to publish the results of the 1981 Snowmass conference. To chair the second workshop, NASA looked beyond the space science community and appointed John Rather, who had recently joined NASA after a career in defense-related projects. In this chapter we will tell the story of the Morrison team from our perspective as active participants, returning to the space defense workshop in the next chapter.
The members of the "Morrison Committee" were mostly astronomers, including asteroid hunters Gene Shoemaker, Eleanor Helin, Ted Bowell, and Tom Gehrels, and orbit experts Brian Marsden, Don Yeomans, and Steve Ostro. In addition to U.S. scientists, the team included members from Russia (then the USSR), Finland, France, India, and Australia.
The official name of Morrison's team was the NASA International Near Earth Object Detection Working Group. Clearly a better term was needed to characterize both the group and the comprehensive asteroid survey we were charged with designing. The inspiration for a new name came from the opening pages of the novel Rendezvous with Rama by the famous science fiction writer Arthur C. Clarke. In his prologue, Clarke had described a fictional 2077 asteroid impact of catastrophic dimensions that had wiped out Venice and adjacent parts of Northern Italy. According to the novel, humanity responded with an international effort to ensure that the Earth would never again be struck by such a cosmic missile, and this program was named "Project Spaceguard". Spaceguard seemed like an excellent name for what we were proposing. With Clarke's enthusiastic permission, we called our proposal the Spaceguard Survey and named our team the Spaceguard Working Group.
Arthur Clarke followed the progress of the Spaceguard Working Group closely from his home in Sri Lanka, and when asked to write an essay by Time magazine in the summer of 1992, Clarke chose asteroid impacts as his subject. The essay, printed in a special issue of Time, was later expanded into a novel, The Hammer of God, which refers directly to the NASA Spaceguard study. This novel, in turn, is the basis for the 1997 Spielberg film Deep Impact.
The Spaceguard group examined both the nature of the impact hazard and how best to search for potentially threatening asteroids. We needed to look at the hazard issue in order to answer the questions of what objects needed to be searched for: asteroids or comets; large ones or small ones? The first task, therefore, was to subject the risk from impact to a rigorous statistical analysis and place it in context with other hazards. How big is the risk, and what is its nature? In particular, what do impacts by comets and asteroids of various sizes actually do to the environment? In terms of what they do and how often such impacts happen, which are the most hazardous?
Analysis of a problem often begins with concrete examples. Most of our direct experience with the impact hazard was limited to just two events. At one extreme was the gigantic Chicxulub impact of 65 million years ago that had led to the extinction of most species on Earth, including the dinosaurs. Clearly such a global catastrophe would kill most (if not all) human beings, but it is equally clear that such enormous events are very rare, happening perhaps only once in a hundred million years. The other example was the 1908 Tunguska blast, which would have wiped out a city had it hit one, yet it didn't -- such events were calculated to take place on the land area of the Earth about once per millennium. The range of projectile mass from Tunguska to the Chicxulub impactor is enormous, and the effects are entirely different depending on the size of the projectile. To make any sense of the hazard, we had to analyze the effects and risks of different sized projectiles.
At smallest end are the meteorites -- rocks from space that strike the atmosphere yet fail to burn up entirely during their fiery plunge to the ground. Astronomers estimate that several tons of meteorites reach the surface every year, but this is only a minute fraction of the total mass of impacting cosmic debris. Clearly, the great majority of these fragments burn up in the atmosphere and do not reach the ground. Those that do are slowed by atmospheric braking and strike the ground like a freely falling rock, at less than 10 m/s. This means that you have to almost literally be hit on the head to risk injury or death. Since such direct hits are extremely rare, the hazard is negligible. We can walk freely in our neighborhoods with no need to fear being struck dead by a space rock.
An impact by a larger projectile is much more serious. That is not only because it is larger, but such objects can reach the ground with most of their original speed, and hence energy, intact. When such a hypervelocity projectile strikes the ground or disintegrates in the lower atmosphere, it explodes causing widespread destruction. The question is: at what size (energy) does an incoming projectile penetrate the atmosphere without being slowed by atmospheric friction? Only objects larger than this threshold size will cause substantial damage.
At the time the Spaceguard Working Group began its task, conventional wisdom, at least in some quarters, was that the threshold for atmospheric penetration was at a mass of a few tons, corresponding to a diameter of a meter or so. A recent report from Livermore National Laboratory had estimated this threshold at 4 m diameter and concluded that hypervelocity impacts releasing tens of kilotons of energy take place annually on the Earth, an amazing prediction to which we will return in the next chapter. But common sense tells us that natural blasts the size of the Hiroshima bomb are not annual events on Earth. Something must be wrong with such calculations. Fortunately, Chris Chyba and Kevin Zahnle of NASA Ames Research Center had an explanation, as we described in Chapter 7. In 1991 they were just developing computer simulations that realistically accounted for the Tunguska event.
The new analysis showed that the smallest diameter rocky object that can penetrate the lower atmosphere is about 50 m (or 100 m for an icy, cometary projectile), with a corresponding impact energy of about 10 megatons. Smaller impacts detonate harmlessly in the upper atmosphere, except for projectiles made of iron, which fortunately are rare. Thus we need not worry about natural Hiroshima-scale events. Only those much rarer impact explosions with energies greater than the largest thermonuclear bombs penetrate low enough to do any harm -- and they happen somewhere on Earth only about once per century.
An impact above the 10-megaton threshold would be destructive in a populated area, but it would do little harm if the projectile hit in the ocean or in a wilderness like Siberia. Most impacts in the 10 to 100 megaton range would cause brilliant, devastating airbursts. Projectiles with energies exceeding 100 megatons would crash into the ground and form explosion craters -- the more energetic the impact, the larger the resulting crater. At 10,000 megatons, for example -- equivalent to the world's total nuclear arsenals -- the crater would be several kilometers across and the zone of devastation would approach a hundred thousand square kilometers. Such events occur somewhere on Earth roughly once in 100,000 years.
It is easy to estimate the risk we each run of being killed by such an impact, which could, after all, happen anytime -- we do not necessarily have a grace period of 100 millennia! Since such an impact would be lethal only to those near ground-zero, we term such an event a "locally destructive impact" to distinguish it from the larger, rarer events that could harm the whole global ecosphere. We would be safe from such a 10,000 megaton event unless we were less than 200 miles away from ground-zero. The hazardous zone, within 200 miles radius of ground-zero, has an area of about 100,000 square miles. For comparison, the total surface area of the Earth is approximately 200,000,000 square miles, larger by a factor of 2000. Thus, by random chance, only one in 2000 of these locally destructive impacts will be close enough to us to do us harm.
Since the frequency of such impacts on our whole planet is only about one per 100,000 years, the average interval between impacts that endanger us is about 100,000 x 2,000, or 200 million years. Expressed as a probability, this means the chance each year of any one of us being killed by such an impact is only 1 in 200 million. Most people consider such a low probability -- much less than the chances of an average American being killed by a poisonous insect or by a lightning bolt -- to be negligible. (Note our phrase average American. A golfer on the greens during a thunderstorm, or a person with known allergies to bees trying to remove a swarm of killer bees, would run a much higher than average risk. Locally destructive impacts, however, strike at random, and we are all at roughly the same, minimal level of risk.)
There is a major, qualitative difference between locally destructive impacts and impacts large enough to produce a global environmental disaster. The Chicxulub impact was a global holocaust, of course, but even far short of events that cause mass extinctions, an impact can cause short-term changes in global weather with catastrophic consequences for human civilization.
In Chapter 6 we discussed recent research on the environmental effects of impacts by objects of various sizes. To make an estimate of the actual hazard, however, we must first decide exactly what we mean by a global catastrophe. There were many, roughly equivalent concepts: the collapse of modern civilization . . . a new Dark Ages . . . something like the effect of the plague on 14th century Europe . . . the sudden death of a large fraction of the world's population . . . something analogous to the aftermath of global nuclear war. In order to define the onset of global catastrophe, we also needed to understand just which globally destructive environmental consequence can be produced by a blast of any given size. While an impact of 10 million megatons or more was required to ignite global wildfires, a smaller impact might be sufficient to inject dust into the upper atmosphere worldwide, block the sunlight, and lead to crop failures in all nations. In 1991, as the Spaceguard team groped for a definition for the onset of global catastrophe, it seemed that mass starvation due to collapse of world agriculture was the critical factor.
We decided to define a global catastrophe as an event that leads to the death of 25% of the world's human population, primarily due to starvation and attendant disease and instability. As it happens, collapse of world agriculture had already been studied as an environmental consequence of a large-scale nuclear war. We had available to us the results on nuclear winter from the original TTAPS team, and one of the TTAPS scientists, Brian Toon, became interested in what might happen if the stratosphere were loaded with impact-produced dust instead of smoke. Applying the computer simulations that had been developed to study nuclear winter, he and his colleagues concluded that the injection of about a billion tons of fine dust into the stratosphere would lead to a global cold snap and crop destruction. An impact with an energy of a few hundred thousand megatons would produce that much dust. That is the impact energy of a mile-wide asteroid. Toon provided this information to the Spaceguard Working Group for our evaluation of the hazard associated with a global impact winter.
Even as we learn more about the impacts themselves and their potential environmental consequences, we may never understand how human civilization would react to such a global disaster, which would vastly exceed the severity of World War 2. Would such massive stress lead to international cooperation or to war? Would social, political, and economic structures collapse, too, along with agricultural production? Would this be a short-term crisis, or the end of civilization? Europe survived the death of a quarter of its population during the plague, and Germany was prospering less than two decades after losing World War 2. But since all nations would suffer simultaneously from impact winter, there would be no provider of a Marshall Plan. Would our social and religious systems and our individual wills to live muster the hidden strengths that would enable us to re-build? Or would destructive chaos imperil us, as depicted by novelists William Golding in The Lord of the Flies and Larry Niven and Jerry Pournelle in Lucifer's Hammer? No one really knows, and we surely don't want to do an experiment to find out.
Adopting Toon's estimate of about 1 mile for the threshold diameter of an asteroid capable of delivering a globally catastrophic blow to the environment, we can calculate (from the known frequencies of impact by asteroids and comets of various sizes) the relative hazard of locally destructive impacts and globally catastrophic events, and we can compare the probabilities of death. (These risks, for each range of size, are derived in the appendix.) An important conclusion is that the larger the impact, the greater is the total risk, confirming the work we did earlier for the book Cosmic Catastrophes. Of course, the damage increases with the size of the impact, but one might expect that the overall hazard would be more than compensated by the increasing rarity of such events. Not true. Despite the relative infrequency of larger impacts, we are more at risk from the very rare global catastrophe than from all of the smaller, more frequent impacts of the Tunguska class. Nearly 90 percent of the total hazard is associated with impacts at, or a little above, the global threshold of about 1 mile diameter. If we could protect against these impacts, we would not only ensure that there are no global, civilization-threatening disasters, but we would also remove most of the total risk.
Our hazard analysis led to another important conclusion: the probability of death from impact (chiefly the globally catastrophic ones) is actually similar to that from more familiar natural hazards such as earthquakes, volcanic eruptions, and severe storms. Worldwide, the lifetime chances of an individual dying from the aftermath of a large cosmic impact is about 1 in 20,000. We will compare this number with other hazards and discuss people's responses to such levels of risk in a later chapter.
For purposes of writing the Spaceguard Survey Report, settling on the level of the threshold was an important issue. The specific value for the threshold diameter supplied by Toon was 1.7 km, but he realized this value was uncertain by a factor of two either way. We wouldn't want to design and construct Spaceguard to find only those asteroids larger than 1.7 km, and then have later research show that 1.0 km asteroids, or even smaller ones, might be capable of triggering a global catastrophe. So some members of the Spaceguard Working Group felt that we should be conservative, and adopt 0.5 km diameter as the threshold size, to be sure of protecting against anything that might conceivably have such unparalleled consequences as the end of civilization. Other members took an opposite philosophical view of scientific "conservatism". They rejected adoption of the smaller threshold size as being alarmist. To them the "conservative" decision would be to adopt a larger value, perhaps 3 km, a size of projectile that we all could agree would have undoubted catastrophic effects.
Ultimately we compromised on a value of 1 km for the threshold -- at least for the practical purposes of the survey. Even though the real threshold is probably a bit larger, Spaceguard would provide Earth with a margin of safety by searching to somewhat smaller diameters. Although almost no one was satisfied, the Spaceguard Working Group nevertheless had taken its first step toward protecting the Earth from incoming asteroids: We had decided to design the search to locate all, or at least most, potentially threatening asteroids 1 km or larger.
Now we had to figure out a way to locate the objects themselves long before any of them struck the Earth. If we could determine the orbits for each of the Earth-crossing objects larger than 1 km and project them forward in time by computer calculations, we could establish once and for all whether any of them pose a near-term hazard of impact with our planet. Some members of the team calculated that it should be possible to predict most of these orbits accurately for at least a century ahead.
How many asteroids are there with diameters of 1 km or larger? As we discussed in Chapter 2, Gene Shoemaker and others have estimated this number as being approximately 2,000. Of these, only about 100 were known in 1991. Our objective was to find the other 1900 in a timely way.
Asteroid surveys, such as those done with the 18-inch telescope on Palomar Mountain and the 36-inch Spacewatch instrument on Kitt Peak, were finding new Earth-approaching asteroids larger than 1 km in diameter at a rate of about one a month. At this rate, it would require 1900 months, or about 160 years, to discover them all. The challenge was to increase this discovery rate to at least 10 per month and to carry out a comprehensive all-sky survey.
Although these are called near-Earth asteroids because they can come close to Earth, they do so only occasionally in the normal course of their loops around the Sun, when Earth and asteroid both happen to be near the crossing points of their orbits. If we are to complete the survey rapidly, we can't wait for these chance close encounters. The way to increase the discovery rate is to extend our reach to a greater distance (where the objects are fainter) in order to find NEOs even at the far points of their orbits.
Some people misinterpret the Spaceguard Survey as a search for asteroids that are heading straight for Earth -- part of a last-minute defense system analogous to military radars that detect incoming bombers or ballistic missiles. But this is not the case. Spaceguard has never been designed to detect asteroids on their final fatal approach, since long lead time is required for an effective defense. The point is to survey a large volume of space and pick up each asteroid on one of the thousands or millions of times it crosses the Earth's orbit before actually striking. Ideally, the warning of a future impact would come decades, perhaps centuries, before the actual collision with our planet.
Spaceguard team members used computers to model various survey approaches in order to zero in on an optimum search strategy. These simulations showed that for a complete survey to be accomplished in a practical length of time, it would be necessary to increase both the size of the telescopes and the area of sky to be scanned beyond that of Tom Gehrels' Spacewatch system. Basically, the telescope size is determined by the requirement to see dark, 1-km asteroids out to a distance of about 200 million km; anything much less lets too many objects slip past undetected. This requirement in turn demands telescopes of diameter 2 m or larger -- that is, with about 5 times the light-gathering power of the 36-inch Spacewatch telescope. To provide sufficient sky coverage, we concluded that six such telescopes were needed, three in the northern hemisphere and three in the south.
The survey would be impossible to conduct without modern detectors and computers to analyze the data. Visual or photographic techniques are woefully inadequate, and even the electronic approach pioneered by Tom Gehrels would need considerable refinement. We concluded that identification of faint, slowly moving asteroids against a heavenly background of millions of stars is a task ideally suited to a computer, but perhaps slightly beyond the capabilities of state-of-the-art computers in 1991. We were confident, however, that within a few years even relatively inexpensive work stations would be up to the task.
Congress had asked that we estimate the cost of this survey system. While the astronomers on the Spaceguard Working Group lacked expertise in such budgeting, we enquired into the purchase price of a 2-m telescope and concluded that each of the Spaceguard telescopes, complete with the best detectors and computers, could be bought for about $8 million -- a modest sum compared with the $80 million price tag of the largest ground-based telescopes under construction in Hawaii and elsewhere. The comparison with Earth-orbiting telescopes is even more dramatic; each Spaceguard facility would cost less than 1 percent as much as the Hubble Space Telescope. The whole Spaceguard system would cost about $50 million for construction and perhaps $10 million per year for operations. By NASA standards, this sounded like a real bargain!
The Spaceguard Survey Report was completed in January 1992, just in time for the second Congressionally mandated study, which was to deal with defense technologies. If Spaceguard discovered a comet or asteroid on a collision source, did the means exist to intercept it and deflect it? We were about to find out.