6.1 Introduction

In the previous chapter we described a search strategy for the NEO survey in which we define the search operations to include both initial observations and verification on a second night. However, the uncertainty in the determination of the NEO orbit, and hence our ability to predict the object's future position, generally increase away from the period spanned by the observational data. If the positional data obtained during the discovery apparition are inadequate, then the uncertainty in the NEO's sky position during the next predicted apparition may be so large that the NEO cannot be recovered. The problem can be alleviated if the object is found in the existing file of observations of unidentified asteroids, but the object must otherwise be designated as lost, and it will remain lost until it is accidentally rediscovered. Clearly, we need to acquire sufficient data to minimize this loss of newly discovered objects.

An important part of the proposed survey involves the precise definition of NEO orbits, for this is a prerequisite to the identification of potentially hazardous objects. The critical first step in this process is to follow up each NEO discovery astrometrically, i.e., by tracking the object optically and/or with radar. Every NEO discovered should be followed astrometrically at least until recovery at the next apparition is assured. Further, we must develop explicit criteria for possibly hazardous ECAs, and any object that appears to fall into the "possibly hazardous" category on the basis of initial observations must be carefully tracked until an improved orbit determination allows a rigorous judgement as to its hazard potential.

In the case of an LPC, which cannot be tracked over several orbital periods, some uncertainty as to where (or even whether) it will strike the Earth may remain almost up to the time of impact. Smaller (Tungunska-class) ECAs may also require extensive tracking to determine their point of impact with sufficient accuracy (say 25 km) to permit rational judgements concerning countermeasures, such as the need to evacuate areas near the target. Finally, some uncertainty in the impact point will always remain due to lack of predictability of aerodynamic forces on the object in the Earth's atmosphere, especially if it breaks up during entry.

Apart from the astrometric follow-up observations, additional physical observations should be made to estimate the size and gross characteristics of the NEO. The rest of this chapter discusses various aspects of the follow-up process in detail.

6.2 Recognition and Confirmation

Immediately after the discovery and verification of an NEO, the principal need is to secure enough astrometric data (observations of position and velocity) that the orbit can be determined with some reasonable reliability. Modern asteroid-hunting practice is to measure carefully the positions of the objects in relation to the stars, and to do so on two nights in quick succession. Although the above procedure is mainly designed for main belt asteroids, its general features apply equally well to NEOs. The principal difference is that, because of its rapid motion, an NEO can generally be recognized as such on the night of its discovery, permitting the discoverer to plan for additional observations. In the case of an object moderately close to the Earth, the difference in perspective (parallax) arising from viewing points that are rotated about the center of the Earth (for example, at the same observatory but at times several hours apart) permits a rather accurate triangulation on the object's distance and hence contributes to the rapid determination of its orbit. In order not to interrupt the actual search process, it may be better to secure the additional initial-night observations with a different instrument or at a different site, although it is generally appropriate for the discoverer to take the responsibility for seeing that these observations are secured.

If an NEO is very close to the Earth, it is possible that enough information to compute a meaningful orbit can be obtained on a single night. Asteroid 1991 BA, which was observed eight times over only a five-hour interval, is an excellent example of this. If an initially computed orbit bears a resemblance to that of the Earth, however, it is quite probable that the object is an artificial satellite. There do exist artificial satellites in highly eccentric orbits with apogees at and even beyond the orbit of the Moon. In the recent case of tiny NEO 1991 VG, the earthlike orbit was verified as more observations became available, thereby introducing the troublesome possibility that this was an uncatalogued artificial object that had completely escaped from the Earth's gravity long ago but that was now returning to the Earth's vicinity. As the quantity of "space junk" increases, similar problems are likely to recur.

The majority of the ECAs discovered will be visible only for relatively short time intervals because, being small, they must be close to Earth to be detectable. Indeed, the simulations discussed in Chapter 5 show that in a 25-yr survey covering the standard 6,000 square degree region to V = 22, the distance of closest approach of ECAs larger than 0.5 km diameter peaks at only about 50 lunar distances. The number of monthly observing runs during which ECAs larger than 0.5 km diameter can be detected in the standard survey region is shown as a function of V in Figure 6.1. At V= 18, 20, 22, and 24, the percentages of ECAs detected in only one run are 59, 41, 20, and 4 percent, respectively. The median numbers of runs in which ECAs are detectable are 1, 2, 4, and 9, respectively, although a few are reobservable almost 30 times. At a diameter threshold of 1 km and for faint magnitudes, the percentages of ECAs observed in only one run are a factor of two smaller, and the median numbers of runs are increased by about 50 percent.

In the strategy described in Chapter 5, we did not directly address the use of the survey telescopes to obtain follow-up astrometric positions near the time of discovery. If follow-up observations were made out to, say, 60 deg longitude from opposition, the percentage of ECAs larger than 0.5 km seen only once to V = 22 would be reduced from 20 to 12 percent. Even greater protection against loss would be afforded by a follow-up strategy in which ECAs discovered were reobserved as long as possible in any accessible region of the dark sky. The question of strategy for this follow-up work needs further study, with the results depending on the availability of other supporting telescopes for astrometric observations.

Since losses after observation in one monthly run can be reduced to small numbers, it is possible that, for deep ECA surveys, follow-up can largely be ignored in favor of the linkage of detections from one run or one apparition to another. In general, such linkage can be achieved unambiguously provided observations are not too sparse. However, care must be taken not to lose the very fast-moving ECAs that may be most hazardous to Earth. Also, because of the large numbers of small ECAs that will be discovered, selection must be made, at least in part, on the basis of the diameter threshold. Both considerations call for a rapid estimate of the diameters of all ECAs discovered near the magnitude limit. To achieve this, the observed brightness can be combined with the distance gauged by means of diurnal parallax. Preference in such work should be given the those objects that appear to be true ECAs, especially those that might pose some threat based on initial orbit calculation.

6.3 Optical Astrometry

For a typical NEO, astrometric follow-up is essential. Much of the follow-up astrometry is most conveniently and efficiently accomplished using conventional reflecting telescopes fitted with CCDs. If conventional reflectors are used, they should generally be in the 1-m aperture range, although larger telescopes should certainly be considered for following up very faint discoveries. A set of semi-dedicated observatories is preferable to a single dedicated observatory (or one in each hemisphere), if only for reasons of weather and availability of observers, and there are certainly times when the more-or-less continuous coverage that may thereby be possible can be very useful.

Existing facilities currently involved with astrometric follow-up of NEOs are listed below in order westward from the principal U.S. discovery sites (the 0.46-m Schmidt at Palomar and the Spacewatch 0.91-m reflector at Kitt Peak), separately for each hemisphere:

Northern hemisphere:

Victoria, B.C., Canada (0.5-m reflector with CCD); Mauna Kea, Hawaii (2.2-m U Hawaii reflector + 3-m NASA IRTF with encoders); Japan (no professional but much amateur activity); Kavalur, India (fledgling Spacewatch program); Kitab, Uzbekistan, and Crimean Astrophysical Observatory, Ukraine (0.4-m astrographs; coordinated by the Institute for Theoretical Astronomy, St. Petersburg, Russia); Klet, Czechoslovakia (0.6-m Maksutov; currently no e-mail communication but should become possible via Prague); Western Europe (not much professional activity, but possibilities at Caussols, France, 0.9-m Schmidt, and La Palma, Canaries, 2.2-m reflector with CCD); Oak Ridge, Massachusetts (1.5-m reflector with CCD); Lowell Observatory, Arizona (1.1-m and 1.8-m reflectors with CCD). Other possibilities include the 1.3-m Schmidt at Tautenburg, Germany, and telescopes at the Bulgarian National Observatory, but these are not currently involved with NEOs, and rapid communication is a problem.

Southern hemisphere:

Mount John Observatory, New Zealand (0.6-m reflector, conversion to CCD in progress); Siding Spring, N.S.W., Australia (U.K. 1.2-m Schmidt, 0.5-m Uppsala Southern Schmidt, 1.0-m reflector with CCD); Perth, W.A., Australia (occasional use of 0.3-m astrograph or 0.6-m reflector); European Southern Observatory, Chile (occasional use of 1.0-m Schmidt, 0.4-m astrograph or 1.5-m reflector). Also there would seem to be a need for participation in southern Africa and eastern South America.

6.4 Radar Astrometry

Radar is also an essential astrometric tool, yielding both a direct range to an NEO and the radial velocity (with respect to the observer) from the doppler-shifted echo. Since most NEOs are discovered as a result of their rapid motion on the sky, these objects are then generally close to the Earth; radar observations are therefore often immediately possible and appropriate. However, radar observations do not become feasible until the object's expected position can be refined (from optical astrometry) to better than about 1 arcmin, and an accuracy of 10 arcsec or better is preferable. A single radar detection has a fractional precision that is two or three orders of magnitude beyond that of optical astrometry, so the inclusion of radar data with the optical data in the orbit solution can quickly and dramatically reduce the future ephemeris uncertainty.

The principal radar instruments are currently those at Arecibo, Puerto Rico, and Goldstone, California. There may also be possibilities at Effelsberg, Germany, Parkes, N.S.W., Australia, and Yevpatoriya, Ukraine Since radars are range limited, radar-detectability windows are narrow, but both Arecibo and Goldstone are being upgraded to enlarge their current windows. There is a clear need for a comparable facility in the southern hemisphere, and some preliminary planning has been done for an "Arecibo-class" radio telescope in Brazil which could also be used as a radar.

The inclusion of radar data in the orbital solutions would allow an NEO's motion to be accurately integrated forward for a few decades (or centuries) to assess the likelihood of future Earth impacts. With optical data alone, such an assessment requires an observational span of several years, which may or may not be possible from the inspection of old photographic plates. The addition of radar data to the orbital solution may allow reliable extrapolations of the object's motion to be made within only days of discovery.

There has hitherto always been a time interval, at least several days long, between discovery and the initial radar work. If the first radar ephemeris is found to have very large delay or doppler errors, the initial radar astrometry is used to generate a second-generation radar ephemeris to enable finer-precision delay or doppler astrometry (by at least a factor of ten) than would have been possible with the first radar ephemeris. This bootstrapping process would be much more efficient than it currently is if a capability to do the computations existed at the radio telescope itself. Ideally, one could input the first measurements of doppler and delay into a program on a computer at the site, generate an improved ephemeris within an hour of initial detection, and proceed immediately to high-resolution ranging. The existence of on-site ephemeris generating capability would be essential if the astrometry that does the critical shrinking of the pointing uncertainty becomes available at the same time as the object enters the radar window, or with an NEO that comes so close that it traverses the telescope's declination-distance window in one day (like comet IRAS-Araki-Alcock at Arecibo in 1983).

6.5 Physical Observations

The impact energy of an NEO that actually hits the Earth depends on both its velocity and its mass. Knowledge of the orbit provides only the velocity, not the mass. The latter quantity can be determined only from physical observations. If astrometric observations are made with a photometric device, such as a CCD, they can also provide information about the most basic of physical parameters, namely, the brightness of the object. In the case of a bright comet, measurements of the brightness will almost certainly include a strong contribution from the atmosphere, whereas what is needed is isolation of the solid nucleus, something that can be satisfactorily attempted only when the comet is faint.

Although an asteroid's brightness is correlated with its size, the known range of asteroid surface reflectivities spans a factor of 20, which leads to a large uncertainty in the volume. The range of densities of asteroids can be inferred from their bulk compositions, which may in turn be suggested by measurements of surface composition.. If only a brightness measurement is available, the deduced mass of the object, and therefore the potential impact energy, can be uncertain by a factor of a hundred. Additional uncertainty arises from the fact that asteroid brightnesses vary as they rotate, sometimes by more than a factor of five.

Measurements of the relative reflectivity of an asteroid at a variety of wavelengths (its spectral reflectivity) can place the object in one of several known taxonomic classes and therefore reduce the uncertainty in the surface reflectivity. At the same time, the composition of the object is constrained, leading to an improved estimate of the bulk density. In a minimal effort, the use of three filters, appropriately chosen to sample spectral features in the ultraviolet and infrared regions, should be employed. With additional filters, greater diagnosticity can be achieved, with a corresponding improvement in reflectivity and composition estimates. With a minimal filter set, however, the range of potential impact energies can be reduced to a factor of about ten.

Radar observations are the only source of spatially-resolved measurements from the ground and hence provide the only source of direct information about an NEO's shape. Moreover, radar can also supply constraints on size that are highly reliable if the echoes are strong enough. Radar also provides some information about the composition and roughness of an NEO's surface.

Even single-color photometry permits a rotation period to be determined, and radar can then provide the spin-pole direction. The angular momentum of a potential hazard can therefore be calculated, and this may be an important consideration in deciding on the technique to be used for dealing with the hazard. In the case of a comet, the detection of persistent cyclic variations in the brightness of the condensation about a stable mean is probably an indication that the bare nucleus has been detected.

That NEOs differ greatly in composition is also evident from a comparison of the effects of encounters. Although the bodies that produced Meteor Crater in Arizona 50,000 years ago and the Tunguska event in Siberia 84 years ago are both thought to have been in the rough size range 50-100 m, one produced a crater that is still very obvious while the other apparently exploded high above the ground, produced no crater, but levelled trees over a much larger area. Knowledge of the likely composition can also play a prominent role in establishing the ameliorative action that might to be taken in the case of a predicted impact.

One could argue that it is not necessary to make physical observations until an object on a collision course has actually been detected. This may not be a prudent course of action, however, for the following reasons. (1) The possibility exists that there will be no further opportunity to study the object in question sufficiently in advance of a collision to provide the necessary information on the potential impact energy and on how to deal with the object. (2) Discoveries of NEOs are often made when they are unusually close to the Earth, and physical observations can be performed more efficiently and with higher precision at these times. (3) We need to learn more about the full range of NEO compositions and structural properties, which are poorly known at present, to plan possible strategies for deflection of these objects in case of a predicted impact. (4) There are significant scientific and possible future commercial benefits that can result from the study of a sizable portion of the NEO population, including the identification of objects with resource potential (substantial sources of water or of nickel-iron and other heavy metals), the providing of selection criteria for possible future spacecraft missions to such objects, the understanding of the link between terrestrial meteorites and the asteroid belt, and important information regarding the origin (cometary versus asteroidal, for example) of these objects.

6.6 Survey Clearinghouse and Coordination Center

Much of the discussion in this chapter has been in the context of current practice.for NEO discoveries. However, the proposed new search strategy described in Chapter 5 means that future NEO discoveries may take place up to 5 magnitudes, or 100 times, fainter than at present. When searches routinely reach magnitude 22 there should be a thousand new NEO candidates each month. With careful organization of the discovery searches, however, the astrometric follow-up data could all be obtained with the same telescopes involved in the discovery. In particular, thought should be given to ensuring that the relevant fields are automatically recorded with a large time separation on either the first or the second night in order to make a parallactic determination of a crude orbit. Month-by-month opposition scanning should also allow, at least in principle, the correct identification of subsequent images of each NEO, but in order to ensure success it would probably be desirable to perform the discovery and confirmation regimen twice during each monthly run.

Bright time (that is, time when the Moon is up) on the discovery telescopes could also be used for physical observations, but radar observations would presumably have to be restricted to close passages by the Earth. Sampling of the physical properties of the smaller NEOs would be important in case they are systematically different from those of the larger NEOs and the main belt asteroids. However, their faintness makes certain observations difficult, so that a large dedicated follow-up telescope with special instrumentation would prove more effective for some physical observations than the survey telescopes themselves.

The dramatic increase in the rate of discovery of NEOs will require considerable extension of the current system for keeping track of these objects and disseminating information about them. Hitherto these functions have principally been carried out by the International Astronomical Union's Central Bureau for Astronomical Telegrams and Minor Planet Center, which since 1978 have been operating together at the Smithsonian Astrophysical Observatory in Cambridge, Massachusetts, under the direction of B. G. Marsden. The Minor Planet Center currently deals with asteroid discoveries (primarily main belt objects) at an annual rate of a few thousand. With the prospect of discovering a thousand NEOs alone in a month, rather than a year, augmentation of the Minor Planet Center's capabilities will be necessary. Procedures for rapidly checking, identifying, computing orbits and providing appropriate ephemerides for new discoveries are already in place, but future enhancement will require acquisition of faster computers and the employment of additional personnel. The future NEO survey clearinghouse would also be undertake the task of actually planning the observations at the various sites, collecting the observations from the sites, and coordinating further observations to cover fields missed by bad weather and to ensure proper follow-up in specific cases.

Further development of procedures and construction and maintenance of software must also be an important component of the work of the survey clearinghouse. For comets and asteroids, the computation of an orbit and ephemeris should include an estimate of the uncertainty in the NEO's location as a function of time, that is, the "positional error ellipsoid." (This is less easily done in the case of comets because of the existence of nongravitational effects that can at best be modelled in a semi-empirical manner.) By projecting the error ellipsoid into the future, one can quantify the likelihood that an NEO will be recoverable, and one can also assess the uncertainty in an Earth-asteroid distance for any future close approaches. Such software will also (1) help to expedite verification of newly discovered objects as NEOs, (2) provide the basis for prioritizing NEOs for follow-up astrometry, both to avoid losing objects and to optimize the use of telescope time and personnel, and (3) permit the reliable identification of NEOs on very close-approach trajectories and the appropriate hazard assessment.

For each newly discovered NEO, data files will have to be established to catalog discovery data and follow-up observations, both astrometric and physical. Orbits and associated error analyses will be required for each object to identify close Earth approaches in the immediate future and to establish optimum observation times for securing the object's orbit and ensuring its recovery at subsequent observation opportunities. Once the need for follow up observations has been established and the optimal observation times determined, the clearinghouse would notify the appropriate people capable of making the required observations and provide them with all the information required efficiently to utilize the limited amount of available telescope time.

Contents

Next Chapter